The World Nuclear Industry Status Report 2023 (HTML)

Executive Summary and Conclusions

Following the worst COVID-19 pandemic years 2020–2021, 2022 was largely dominated by the effects of a global energy crisis exacerbated by the war in Ukraine. For the first time in history, operating commercial nuclear facilities were directly attacked and then occupied by hostile forces during a full-scale war. As of the end of 2023, while attracting little attention in recent months, the occupation of the Ukrainian nuclear power plant Zaporizhzhia is still ongoing, the threats of cuts of power and water supplies persist. The specific risks to a nuclear plant in a full-scale war have been analyzed in detail in WNISR2022.

The World Nuclear Industry Status Report 2023 (WNISR2023) provides a comprehensive overview of nuclear power plant data, including information on age, operation, production, and construction of reactors. WNISR2023 includes a special focus chapter assessing Nuclear Economics and Finance.

WNISR2023 analyses the status of newbuild programs in 13 of the 32 nuclear countries (as of mid-2023) as well as in Potential Newcomer Countries. WNISR2023 includes sections on 12 Focus Countries representing almost one third of the current nuclear countries—plus Germany that closed its last reactor in April 2023 and Poland that, once again, envisages the construction of its first reactors—72 percent of the global reactor fleet, and the world’s five largest nuclear power producers. The comprehensive special United States Focus provides a detailed analysis of the status of the U.S. nuclear program as well as the multiple federal and state-level support initiatives for the sector. For the first time, the Focus Countries chapter includes a section on South Africa.

The situation of Small Modular Reactor (SMR) development is analyzed in a dedicated chapter. The status of onsite and offsite challenges are summarized in the Fukushima Status Report. The Decommissioning Status Report provides an overview of the current state of nuclear plants that have been permanently closed. The chapter on Nuclear Power vs. Renewable Energy Deployment offers comparative data on investment, capacity, and generation from nuclear, wind, and solar energy, as well as other renewables around the world. Finally, Annex 1 presents overviews of nuclear power programs in the countries not covered in the Focus Countries chapter.

Production and Role of Nuclear Power

Prior to the entry into force of the Treaty on the Non-Proliferation of Nuclear Weapons (NPT) in 1970, 14 countries were operating nuclear power reactors. By 1985, 16 additional countries had reactors on the grid. Over the 30-year period 1991–2020 (none in 2021), only five countries started up their first power reactors—China (1991), Romania (1996), Iran (2011), United Arab Emirates (UAE), and Belarus (both 2020); in 2021–2022, no newcomer country started any reactor. Four countries abandoned their nuclear power programs, Italy (1987), Kazakhstan (1998), Lithuania (2009), and Germany (2023).

Reactor Operation and Capacity. As of 1 July 2023, a total of 407 reactors—excluding Long-Term Outages (LTOs)—were operating in 32 countries, four units less than in WNISR2022,1 eleven less than in 1989, and 31 below the 2002-peak of 438. At the end of 2022, the nominal net nuclear electricity generating capacity had peaked at 368 GW,2 having added 5.3 GW during the year, 1 GW more than the previous 2006-record of 367 GW, but it dropped again to 364.9 GW by mid-2023.

IAEA versus WNISR Assessment. Between September 2022 and April 2023, the International Atomic Energy Agency (IAEA) significantly modified its statistics—including retroactively—as displayed in its online-Power Reactor Information System. This in turn impacts the perception of nuclear industry trends. Until September 2022, PRIS showed a historic peak in officially operating reactors, both in terms of number (449) and capacity (396.5 gigawatt), in 2018. In July 2023, PRIS shows the peak in the number of units occurring as early as 2005 at a maximum of 440 and the maximum capacity still in 2018 at 374 GW. Both indicators have declined since, with PRIS showing 410 units as operating with 368.3 GW of capacity as of mid-2023.

Until September 2022, the IAEA had included 33 units in Japan in its total number of reactors “in operation” in the world while only 10 of these units had effectively restarted and 23 have not produced electricity at least since 2010–2013 (of which, three since 2007). As of mid-2023, the IAEA had pulled those 23 units, together with four reactors in India, from the list of operating reactors retroactively since shutdown and added them to a new category labelled “Suspended Operation”.

WNISR had called on the IAEA to adapt its statistics to industrial reality since 2014 when it created its own Long-Term Outage (LTO) category. As of mid-2023, WNISR classified 31 units as LTO, of which 23 in Japan, three in India, two in Canada, and one each in China, France, and South Korea—the number increased by two compared to WNISR2022.3

Nuclear Electricity Production. In 2022, the world nuclear fleet generated 2,546 net terawatt-hours (TWh or billion kilowatt-hours) of electricity. Production dropped by 4 percent compared to 2021 to the level of pandemic-year 2020. China continued to generate more nuclear electricity than France for the third year in a row and remains second—behind the United States (U.S.)—in countries operating nuclear power plants. Outside of China, nuclear production dropped by 5 percent in 2022 to a level last seen in the mid-1990s.

Share in Electricity/Energy Mix. Nuclear energy’s share of global commercial gross electricity generation in 2022 dropped by 0.6 percentage points—the largest drop since post-Fukushima year 2012—to 9.2 percent, 47 percent below the peak of 17.5 percent in 1996.

Reactor Startups and Closures4

Startups. In 2022, seven reactors were connected to the grid, of which three were in China and one each in Finland, Pakistan, South Korea, and the UAE. In the first half of 2023, four units were connected to the grid, one each in Belarus, China, Slovakia, and the U.S.

Closures.5 In 2022, five reactors were closed, three in the United Kingdom (U.K.), and one each in Belgium and the U.S. In the first half of 2023, another five units were closed, three in Germany and one each in Belgium and Taiwan.

Over the two decades 2003–2022, there were 99 startups and 105 closures. Of these, 49 startups were in China which did not close any reactors. As a result, outside China, there has been a drastic net decline by 55 units over the same period, and net capacity declined by over 24 GW.

Construction Data6

As of 1 July 2023, 58 reactors (58.6 GW) were under construction, that is five more than in last year’s WNISR, but 11 fewer than in 2013 (five of those units have subsequently been abandoned).

Four in five reactors are being built in Asia or Eastern Europe. 16 countries are building nuclear plants, one more than in WNISR2022. The list includes Egypt and the construction restart in Brazil but leaves out Belarus since it completed its second unit. Only four countries—China, India, Russia, and South Korea—have construction ongoing at more than one site. Construction started on ten reactors worldwide in 2022: five are in China while the other five are implemented by Russia in Egypt (2), in Turkey (1), and domestically (2). The building of three reactors got underway in the first half of 2023, two of them in China, and one by Russia in Egypt. Chinese and Russian government-owned or -controlled companies were responsible for all 28 reactor construction-starts in the world over the 42-month period from the beginning of 2020 to mid-2023.

Building vs. Vendor Countries

• As of mid-2023, China had by far the most reactors under construction with 23 units or 40 percent of the total. However, China is currently not building anywhere outside the country.
• Russia is the dominant supplier the international market with 24 units under construction in the world as of mid-2023. Five of these are being built domestically. The remaining 19 units are being constructed in seven countries, including four each in China, India, and Turkey, as well as three in Egypt.7 It remains uncertain to what extent these projects have or will be impacted by sanctions imposed on Russia and other consequential geopolitical developments following the invasion of Ukraine.
• Besides Russia’s Rosatom, only French and South Korean companies are acting as leading contractors building nuclear power plants abroad; France in the U.K. and South Korea in the UAE.8

Construction Times

• For the 58 reactors being built, an average of six years has passed since construction start—lower than the mid-2022 average of 6.8 years—but many remain far from completion.
• All reactors under construction in at least 10 of the 16 countries have experienced often year-long delays.
• Of the 24 reactors clearly documented as behind schedule, at least nine have reported increased delays and one has reported a delay for the first time over the past year.
• WNISR2021 noted a total of 12 reactors scheduled for startup in 2022. At the beginning of 2022, 16 were planned to be connected to the grid within the year (including four pushed back from 2021 to 2022) but only seven of these generated first power; the other nine were delayed at least into 2023.
• Initial construction of the Mochovce-4 reactor in Slovakia started 38 years ago and its grid connection has been further delayed, currently to 2024. Bushehr-2 in Iran originally started construction in 1976, over 47 years ago, and resumed construction in 2019 after a 40-year-long suspension. Grid connection is currently scheduled for 2024.
• Seven additional reactors have been listed as “under construction” for a decade or more: Angra-3 in Brazil, the Prototype Fast Breeder Reactor (PFBR), Kakrapar-4, and Rajasthan-7 & -8 in India, Shimane-3 in Japan, and Flamanville-3 (FL3) in France. The French and Indian projects have been further delayed this year, and the Japanese reactor does not even have a provisional startup date.

Construction Starts

• Construction started on ten reactors in 2022, including five in China. Russia began work on reactors in Egypt (2), in Turkey (1) and in Russia (2), and on a barge in China which is to be equipped with two reactors in Russia.9 In other words, of the global total of ten, seven reactors were designed by the Russian and three by the Chinese industry.
• Construction of three reactors started in the first half of 2023, two of them in China, and one of Russian design in Egypt.
• Chinese and Russian government-owned or -controlled companies launched all of the 28 reactor constructions in the world over the 42-month period from the beginning of 2020 to mid-2023.

Operating Age

• The average age (from grid connection) of operating nuclear power plants has been increasing since 1984 and stands at 31.4 years as of mid-2023, up from 31 years in mid-2022.
• A total of 265 reactors—five less than mid-2022—two-thirds of the world’s operating fleet, have operated for 31 or more years, including 111—more than one in four—for at least 41 years.
• If all currently licensed lifetime extensions and license renewals were maintained, all construction sites completed, and all other units operated for a 40-year lifetime (unless a firm earlier or later closure date has been announced), in the years to 2030, the net balance of operating reactors would turn negative as soon as 2024, and slightly positive for the years 2026–2027; but overall, an additional 88 new reactors (66.5 GW)—almost one unit or 0.7 GW per month—would have to start up or restart to replace closures. This would necessitate almost doubling the annual startup rate of the past decade from six to eleven over the remaining period to 2030 just to maintain the current number of reactors in the world. Considering the long lead times, this appears to be a highly unrealistic scenario.

Focus Countries

The following 11 Focus Countries include those home to almost one third of the current nuclear countries as well as Germany that closed its last reactors in April 2023 and Poland which plans to build its first reactors. Some key developments in 2022 and the first half of 2023:

Belgium. Nuclear generation dropped by 13 percent in 2022. Under the framework of the phaseout policy, one reactor was closed in September 2022, and another one in January 2023. Five reactors remain operational. The current plan is to close three by 2025 and extend operation by 10 years for the two most recent ones to 2035. A legally binding agreement is expected to be closed in early 2024.

China. Nuclear power generation increased by 3.2 percent—a modest development compared to the 11-percent boost in 2021—and provided a stable 5 percent of total electricity generation. Meanwhile, wind energy output grew by 16 percent and solar by 31 percent. Non-hydro renewables produced 15.5 percent of national gross power generation, more than three times the nuclear contribution.

France. After experiencing a declining performance since 2015, the year 2022 represented an “annus horribilis”, in the words of an EDF director. Due to a cumulation of generic technical failures, issues related to ageing, climate impact, and social movements, nuclear generation dropped below the level of 1990 or about 120 TWh below the 2005–2015 level of around 400 TWh. That drop in output is larger than the 100 TWh Germany lost in nuclear generation since 2010 due to its phaseout policy. On average, French reactors generated zero power on 152 days in 2022. For the first time since 1980, France turned into a net importer of electricity, with Germany playing a key role as an exporter. Utility Électricité de France (EDF), facing potential bankruptcy over record losses and unprecedented net debt levels (€202364.8 billion or US$70 billion as of mid-2023), was renationalized.

Germany. The country’s nuclear fleet generated 32.8 TWh net in 2022, a decline by half over the previous year after three reactors were closed at the end of 2021, and only a fraction of the peak generation of 162.4 TWh in 2001. Nuclear plants provided 6 percent of Germany’s gross electricity generation, compared to the historic maximum of 35.6 percent in 1999. The three last operating reactors were closed on 15 April 2023, 62 years after nuclear electricity was first generated in the country.

Japan. No additional reactor has been restarted since WNISR2022 (none was slated for closure). A mere 10 units are considered operational with 23 in LTO as of mid-2023. After a major increase in 2021, nuclear generation dropped again (-15.3 percent) to provide 6.1 percent (-1.1 percentage points) of the country’s electricity. A special investigation committee found hundreds of falsification cases at Japan Steel Works, one of the most important manufacturers of large forgings which have been supplied to nuclear power plants around the world.

Poland. In October 2020, the government adopted a long-term Polish Nuclear Power Program aiming to commission 6–9 GW of nuclear capacity by 2043. The country abandoned construction of two Russian-designed VVER reactors in the 1980s, and several subsequent relaunch attempts were aborted. Meanwhile, Poland has one of the fastest growing solar programs in the E.U. increasing capacity by 61 percent in 2022 to reach 12.4 GW.

Russia. Nuclear power generation increased slightly to reach a new record of almost 210 TWh. The country operates 37 reactors and has a further five under construction. Abroad, Russia maintained its role as the leading nuclear power plant builder in the world with 19 units under construction in seven countries as of mid-2023. Eight European countries, including four in the E.U., remain highly dependent on Russian fuel assemblies for 38 operating reactors.

South Africa. Nuclear generation dropped by 17 percent to just over 10 TWh providing 4.9 percent of electricity. The drop was the result of lengthy outages for extensive refurbishment in view of a 20-year lifetime extension, coupled with unplanned outages due to technical incidents. As the country’s large fleet of coal plants also experienced severe technical problems, the country was faced with severe power shortages.

South Korea. Nuclear power production increased by 11.3 percent to 167.5 TWh providing just over 30 percent of the electricity in the country. The increase is due to the startup of one new reactor (Shin-Hanul-1) and the better performance of some units. However, state-owned utility KEPCO filed a record loss of US$202225 billion with its net debt jumping by 32 percent to an unprecedented US$2022149 billion. KEPCO stock lost 70 percent of their value over the past seven years.

United Kingdom. The nuclear program is shrinking rapidly. Two additional reactors have closed since WNISR2022, leaving only nine units operating. In total, there are now 36 closed units awaiting decommissioning, the second largest number after the U.S. The nuclear share in the electricity mix has almost halved since 1997 when it made up 28 percent. However, due to the dramatic production plunge in France, in 2022, the U.K. turned into a net power exporter in 2022 for the first time in four decades. Meanwhile, following repeated delays, the cost estimate for the two reactors under construction at Hinkley Point C has continued to rise and reached US$202144 billion in February 2023, with grid connection of the first unit planned for June 2027 at the very earliest.

United States. Nuclear output declined slightly (-0.9 percent) to 771.5 TWh, the lowest in a decade. The nuclear share of commercial electricity generation declined to 18.2 percent, its lowest level in 25 years. The U.S. nuclear fleet is still the largest in the world, with 93 units, and one of the oldest with a mean age exceeding 42 years. After 10 years of construction, the first of two new reactors at Plant Vogtle was connected to the grid in April 2023. All-in cost estimates for the two Vogtle units now exceed US$35 billion. Substantial new subsidy programs for uneconomic operating reactors and for new projects have been further expanded at the federal and state levels and are impacting previous retirement planning. Over the past five years, the seven closed reactors averaged an operational age of just over 47 years, far below their licensed lifetimes of 60-years. Emerging business models include the coupling of nuclear output with projected consumption by data centers, crypto mining, or hydrogen production. Various criminal investigations continue to plague the nuclear sector. In March 2023, the former CEO of the utility in charge of the later abandoned V.C. Summer newbuild project was sentenced to 15 months in prison, the payback of US$1 million in “ill-gotten income”, and a US$200,000 fine for lying on the real construction status of the project.

Fukushima Status Report

Eleven years have passed since the Fukushima Daiichi nuclear power plant disaster began, triggered by the East Japan Great Earthquake on 11 March 2011 (referred to as 3/11 throughout the report). The situation is still far from having been stabilized.

Overview of Onsite and Offsite Challenges

Onsite Challenges

Spent Fuel Removal. All spent fuel from the pool of Unit 3 had been removed by February 2021. Preparatory work is still underway on Units 1 and 2, with removal further delayed, now to begin in FY 2027–2028 and to be completed by the end of 2031, more than 20 years after the disaster began.

Fuel Debris Removal. Due to technical challenges, operations have been postponed several times. An investigation into the state of the structure supporting the reactor pressure vessel of Unit 1 raises concerns about its potential collapse, as much of the concrete around the rebars has apparently melted.

Contaminated Water Management. As water injection continues to cool the fuel debris, highly contaminated water has continued to run out of the cracked containments into the basements mix with water from an underground river that has penetrated the basements. The combination of various measures have reduced the influx of water from up to 540 m3/day to about 90 m3/day. Every day, an equivalent amount of water is partially decontaminated and stored in 1,000-m3 tanks. Thus, a new tank is still needed almost every 10 days.

As of 24 August 2023, about 1.3 million m3 of treated water were stored in 1,046 tanks.

The safety authority agreed to operator TEPCO’s plan to release the contaminated water into the ocean. As of the end of March 2023, about two thirds of the water must be treated again, and the water diluted by a factor of 100 (or more) before it is released into the ocean through a one-kilometer-long sub-seabed tunnel. Release of the first batch of partially decontaminated water began on 24 August 2023. The operation will take at least three decades. The plan remains widely contested, including overseas.

Offsite Challenges

Offsite, the future of tens of thousands of evacuees, food contamination, and the management of decontamination wastes, all remain major challenges.

Evacuees. As of 1 May 2023, about 27,000 residents of Fukushima Prefecture were still living as evacuees, down from a peak of nearly 165,000 in May 2012. In 2022, evacuation orders for some parts of the so-called “difficult to return areas” were lifted for the first time; these areas continue to have significant exposure levels and are designated as “reconstruction and revitalization areas”. The rate of return varies greatly from 1 percent to 90 percent.

Food Contamination. According to official statistics, of a total of 36,309 samples that were analyzed in financial year 2022, 135 from ten prefectures exceeded the radionuclide concentration limit. Whether the testing program provides an adequate picture of the situation remains open. As of 1 July 2023, 12 countries and regions—down from a peak of 54—still had import restrictions for Japanese food items in place. In July 2023, the European Commission lifted its remaining import restrictions for the E.U.

Decontamination and Contaminated Soil Management. The contaminated soil in the temporary storage area in Fukushima Prefecture is currently being transferred to intermediate storage facilities in eight areas. As of the end of March 2023, four out of a total of ten storage facilities were filled to maximum capacity, and about 88 percent of total storage capacity was filled with contaminated soil. The government is legally responsible for the final disposal of the contaminated soil.

Decommissioning Status Report

As more and more nuclear facilities either reach the end of their pre-determined operational lifetime, or close due to deteriorating economic conditions, timely decommissioning is becoming a key challenge (note that the status of radioactive waste management is not part of this analysis).

• As of mid-2023, the number of closed power reactors reached 212 units—eight more than one year earlier. Thus, almost one third of the reactors connected to the grid in the past 70 years have been closed. These had a total operating capacity of 105 GW, exceeding 100 GW for the first time.
• 190 units are awaiting or are in various stages of decommissioning, eight more than one year earlier.
• Only 22 units, or 10 percent of the closed reactors, have been fully decommissioned, no change over the past year: 17 in the U.S., four in Germany, and one in Japan. Of these, only 11—one more than in WNISR2022—or 5 percent of all closed reactors have been released from regulatory oversight.
• The average duration of the decommissioning process is about 21 years, with a large range of 6–45 years (both extremes are for reactors with very low power ratings of respectively 22 MW and 63 MW).
• The analysis of 11 major nuclear countries hosting 84 percent of all closed reactors shows that progress in decommissioning remains slow: of 159 units in various stages of advancement, six are in the post-operational phase, 75 are in the “warm-up stage”, 27 are in the “hot-zone stage”, 12 are in the “ease-off stage”, while 39 are in “long-term enclosure”.
• To date, none of these early nuclear states—U.K., France, Russia, and Canada—has fully decommissioned a single reactor.

Potential Newcomer Countries

Three potential newcomer countries had nuclear reactors under construction as of mid-2023: Bangladesh, Egypt, and Turkey. All these projects are implemented by the Russian nuclear industry. The impact of sanctions and potential other geopolitical developments on the future of these projects remains uncertain albeit some effects have already been documented.

Other countries like Kazakhstan, Nigeria, Saudi Arabia, and Uzbekistan have more or less advanced plans, but so far none of them has selected a design nor raised necessary financing. Several countries, including Indonesia, Jordan, Thailand, and Vietnam have suspended or cancelled earlier plans. Some key developments:

Bangladesh. Two reactors of Russian design have been under construction since 2017–2018. They were scheduled to start up in 2023 and 2024. Reportedly, sanctions have led to delays in the delivery of some equipment and the commissioning of Unit 1 has been pushed back to late 2024 at least.

Egypt. Construction of the first, Russian-designed nuclear power plant was launched at the El-Dabaa site on 20 July 2022, even as the war in Ukraine was ongoing. Building of Units 2 and 3 began in November 2022 and May 2023 respectively.

Kazakhstan. Several potential suppliers had been considered for the construction of small or large reactors, but no technology has been chosen, no site selected, and no financing package announced.

Nigeria. The country signed nuclear cooperation agreements with several countries and considered the option of developing up to 4 GW of nuclear capacity. However, when in early 2023 Nigeria launched its Energy Transition Plan (ETP) with the goal of carbon neutrality by 2060, nuclear power did not feature amongst the options outlined for electricity generation.

Saudi Arabia. In early 2023, the government confirmed it had received bids from China, France, Russia, and South Korea for the construction of two large reactors.

Turkey. Construction of four units started between 2018 and 2022 at the Akkuyu site. Construction on Unit 4 started in July 2022. Turkish authorities had hoped to connect Unit 1 to the grid in 2023, to coincide with the 100th anniversary of the foundation of the Republic of Turkey. That target was missed, and startup of the first unit is now expected in 2024, and commercial operation in 2025.

Uzbekistan. In May 2022, officials announced that a site for the construction of two Russian-designed VVER-1200 reactors had been chosen in the Farish district of the Jizzakh region, near Lake Tuzkan. The financing package had been under negotiations then and no further information was released.

Small Modular Reactors (SMRs)

Just as in previous WNISR editions, this year’s update on the development status and prospects of Small Modular Reactors (SMRs) does not reveal any major advances despite increasing media attention and additional public funding commitments. The country-by-country status:

Argentina. The CAREM-25 project has been under construction since 2014. Following numerous delays, the current estimated date for startup remains 2027. An updated cost estimate has not been released, but the last released one—predating the latest delays—suggests that on a per kilowatt basis CAREM-25 will cost roughly twice as much as the most expensive Generation-III reactors.

Canada. Strong federal and provincial government support for the promotion of SMRs continues. The largest commitment, of over US$2022745 million, came from the Federal Infrastructure Bank for an SMR project at the Darlington site. Several designs have gone through a “pre-licensing vendor design review” none has yet been certified by the safety authority.

China. It took ten years between construction start and first full power in December 2022 for two high-temperature reactor modules, twice as long as anticipated. Since then, the operational record has been apparently disappointing. Construction started on a second design, the ACP100 or Linglong One, in July 2021. This is six years later than planned, with scheduled startup now by February 2026.

France. In February 2022, President Macron announced a US$20221.1 billion contribution to finance the development of the Nuward SMR design and other “innovative reactors”. Currently, “basic design” studies are to be completed by 2026 and construction is scheduled to start in 2030.

India. An Advanced Heavy Water Reactor (AHWR) design has been under development since the 1990s, but its construction has been continuously delayed. There have been no signs that construction could start any time soon. There have been reports about plans for “a roadmap for studying the feasibility and effectiveness” of SMRs.

Russia. Russia operates two SMRs on a barge called the Akademik Lomonosov. Both reactors were connected to the grid in December 2019, nine years later than planned. Since then, their performance has been mediocre. Construction on a second SMR project, a lead-cooled fast reactor design called BREST-300, started in June 2021. The project has been discussed for a decade and was originally to be deployed by 2018.

South Korea. In 2012, the System-Integrated Modular Advanced Reactor (SMART) design received approval by the safety authority, but there have been no orders since. Several other designs are reportedly in very early stages of development. Foreign SMR developers have started proposing their competing designs in the country, but without tangible success beyond symbolic Memoranda of Understanding.

United Kingdom. Since 2014, Rolls-Royce has been developing the “UK SMR”, a (now) 470 MW reactor (exceeding the size-limit of 300 MW for the generally adopted SMR definition). The regulator is currently carrying out a Generic Design Assessment (GDA) that is scheduled to be completed by August 2026. Six other SMR designs are under review. The U.K. government is aiming for a Final Investment Decision by 2029.

United States. The Department of Energy (DOE) has already spent more than US$1.2 billion on SMRs and has announced further awards over the next decade that could amount to an additional US$5.5 billion. However, there is still not a single reactor under construction. Only one design, NuScale, has received a (conditional) final safety evaluation report. However, since then, the design capacity has been increased from 50 MW to 77 MW per module, and many issues remain unsolved. In October 2021, eight municipalities withdrew from the only investment project, in the Western states, leaving the 6-module 462 MW project with subscriptions amounting to just 101 MW. By January 2023, cost estimates had ballooned to US$9.3 billion, and in early November 2023, the entire project was terminated, officially because “it appears unlikely that the project will have enough subscription to continue toward deployment”.

Nuclear Economics and Finance

Overview

Nuclear power plant projects are amongst the most expensive construction projects of any kind. Some of the main selling points of nuclear—a firm rather than variable power source (although that is questionable in light of recent performances e.g. in Belgium, France, and Japan), low-carbon, dispatchable, and generating heat that can be used for other purposes—are all attributes that are under pressure from a wide range of other, increasingly innovative options throughout the system. These innovative pressures are not limited to generation but extend to all attributes affecting the cost and reliability of the service as well—for example, efficient use or demand response, electric-vehicle-to-grid integration, or power storage to address the variable nature of wind and solar generation. Already some models show that solar photovoltaics (PV) plus storage can have load factors of 50–70 percent. Long-term contracts pairing solar, wind, and storage are already being struck.

In Key Markets, Nuclear Finance Driven by Geopolitics, Not Economics

While a reliable comprehensive, global overview of credit data is not available, partial data indicates strong credit support especially from Russia and China for overseas projects. “Lavish financing” conditions are key to the relative success of both countries. According to a former Nuclear Energy Agency (OECD-NEA) official, “privately-owned equity companies in the nuclear sector are no longer competitive in international markets” and “China and Russia are in the process of putting the West’s nuclear industry out of business”. China’s investments beyond Hinkley Point C in the U.K. are slated to ramp up quickly, with 30 reactor projects abroad by 2030 and an associated investment of more than US$145 billion. How many of these projects will come to fruition is highly uncertain, especially considering U.S. government blacklisting of the main Chinese nuclear companies. However, there seems to be a trend towards an increasing role for Export-Import Banks and various international development banks to finance nuclear projects. The U.S. EXIM Bank has issued letters of interest in multi-billion-dollar financing of newbuild projects in Poland, Romania, and Ukraine. State intervention has been increasing in many countries for some time. WNISR estimates that already roughly 45 percent of global nuclear capacity is fully state-owned.

Operating Reactors Face Continued Competitive Pressure, Receive State Support

In recent years, operating reactors have been facing financial challenges in many countries. Unplanned outages have cut into output, and aging reactors or unexpected problems have sharply driven up plant repair and reinvestment costs, particularly in France and Japan. Plant performance has also suffered from climate-related impacts, such as cooling water availability, heat sink capacity, and storm events. While the effect on overall output remains limited until now, climate-related disruptions of nuclear generation have increased eight-fold over the past 30 years and can have significant impact on available capacity for limited periods of time. Competition by low-cost natural gas, and increasingly wind and solar, represents serious competitive risks for nuclear, especially during certain periods of the year or times of day. For example, in Finland, surging renewables production and negative wholesale power prices forced curtailment of generation at the much-delayed Olkiluoto-3 plant, a month after it commenced commercial operation. Similar cuts have been made at Spanish reactors. In the U.S., 13 reactors officially closed between 2013 and 2022 (including three reactors that had ceased electricity production in 2009 and 2012). Cost pressures are most evident in competitive power markets.

Arguing that plant closures would drive up carbon emissions and that their product, labelled “low-carbon, reliable power”, was not being properly valued by the market, the industry has tagged the closures as premature, and has lobbied for—and increasingly often successfully obtained—large subsidies to support operating uneconomical plants. In the U.S., state-level taxpayer-funded subsidies were granted to 19 reactors; these last from five to 12 years and are estimated to exceed US$15 billion by 2030. Federal subsidies called Zero-Emission Nuclear Production Credits offer a maximum of US$15/MWh for plants operating from 2024 to 2032. They can likely be combined with other subsidies, e.g. for hydrogen production. In addition, the Civil Nuclear Credit (CNC) program funded a national pool of US$6 billion in subsidies to keep economically distressed reactors from closing.

The largest nuclear operator in the world, the French utility EDF, has been fully renationalized. The French government is also lobbying to allow the possibility of accessing various E.U. financing mechanisms to subsidize its existing nuclear fleet. In Belgium, the government has agreed in principle to share the economic risk of a planned 10-year lifetime extension of two reactors beyond the previously agreed closure date of 2025 by setting up a joint company with utility Engie-Electrabel. Japan has de facto nationalized the Fukushima operator TEPCO injecting unlimited funding for compensation of victims and disaster remediation. In order to expedite the restart of reactors shuttered since 3/11, the Japanese government is also considering subsidies that would guarantee income to winning bidders for the subsequent 20 years. This would be an extension of the “long-term decarbonized power supply auction,” slated to begin in early 2024.

Economics of New Reactors in the Context of Government Support

The OECD-Nuclear Energy Agency’s overnight cost (excl. financing and other costs) estimates for Light Water Reactors (LWR) vary by a factor of two from US$20182,157 per installed kilowatt (South Korea) to US$20184,250/kW (U.S.). An independent assessment from the Workgroup for Infrastructure Policy (WIP, at Technical University Berlin) and the German Institute for Economic Research (DIW) based on an 88-reactor database found much higher values, including about US$6,000/kW for mean overnight costs for LWRs.

Overnight cost analyses have some significant limitations for the assessment of nuclear competitiveness: the exclusion of financing and other costs, although financing is frequently recognized as a key component; the very limited number of real cases to serve as reference; the frequent assumption for nth of a kind implementation supposing learning effects through the building of a series of units, but without clearly defining the number n, which can range from five to hundreds (in the case of SMRs). However, the production scales of nuclear’s main competitors are in entirely different orders of magnitude. The installed base of wind turbines is more than 300,000 globally, with more than 25,000 installed in 2022 alone. Solar PV module (each panel has multiple modules) production translates to a unit count in the hundreds of millions per year, with well-documented associated learning effects and cost reductions.

An academic analysis of delays and cost overruns of “megaprojects”, leady by Bent Flyvbjerg, found that nuclear waste projects top of the list with mean cost overruns of 238 percent, and nuclear power plants rank third with mean cost overruns of 120 percent.

The most advanced SMR design in the U.S., NuScale, terminated a six-module project to be implemented for a conglomerate of Utah municipalities, in early November 2023. Cost estimates had spiked to US$20,000/kW. Despite massive federal subsidies estimated to exceed US$4 billion, the projected cost of electricity appeared too high for most candidate municipalities.

Trends in Nuclear LCOE Estimates

Levelized Cost of Energy (LCOE) assessments incorporate not only construction expenses (so-called overnight costs) but also operating and maintenance costs, build times, load factors, and discount rates to generate an average cost per unit energy produced over the plant’s lifetime. Values here have been scaled to 2018 US$. Analysis by the OECD’s International Energy Agency (IEA) and their Nuclear Energy Agency (NEA) highlights the sensitivity of applied discount rates to nuclear LCOEs. While there are no pure market-based benchmarks for nuclear cost-of-capital, historical cost overruns and delays suggest rates should be higher than for energy pathways with more predictable build costs. As discount rates rise, nuclear becomes less and less competitive with other energy policy options.

Further, nuclear LCOE estimates span a wide range even when the same discount rate is assumed. Analysis of IEA’s Electricity Survey estimates mean LCOEs range from US$51/MWh in non-OECD countries to US$62/MWh at a 5 percent discount rate. This is far below the mean nuclear LCOE of U$100/MWh in an independent meta-analysis (including the IEA datasets) of 88 planned and completed nuclear projects (WIP/DIW). IEA’s Net Zero assessments indicate a range from US$102/MWh in the U.S. to US$145/MWh in the E.U. at 8 percent discount, with the World Energy Outlook indicating a range from US$87/MWh in the U.S. to US$129/MWh in the E.U. at the same discount rate.

Asset-management firm Lazard concluded from similar analysis that aside from natural gas peaking plants at discount rates of less than 5.4 percent, nuclear turned out always the most expensive resource on an LCOE basis. At a 7.7 percent discount rate, nuclear came out at US$158/MWh. At a discount rate of 10 percent, and excluding firming costs, nuclear is nearly four times the LCOE of onshore wind.

LCOE estimates for non-OECD countries tend to be lower than that within OECD countries, though based on more limited data. Given the large role of these countries in newbuild projects, improved data access would be very helpful.

Missing and Underestimated Costs

Beyond the nuclear generating station, there are ancillary requirements of the nuclear fuel chain that are more expensive and more complex than for most other forms of energy generation. These other elements are not always well-captured in the economic evaluations of the resource, and explicitly excluded in some assessments. Key questions are whether decommissioning—not only of the power plant but also of the fuel chain facilities—as well as waste management costs are included in the cost assessments; and, if so, whether those assessments are comprehensive. Earmarked funds need to be of appropriate scale and prudently invested to meet needed targets when needed. Unfortunately, adequate funds are often not collected during the full operation of the facility. In other cases, collected funds have been misappropriated due to structural weaknesses in controls.

Decommissioning cost estimates vary widely, and empirical data area limited. In the U.S., reactor decommissioning estimates span a range of US$478–1,435/kWe for publicly-owned reactors and US$615–2,148/kWe for investor-owned reactors.10 The reasons behind the much higher cost projections for investor-owned utilities are not clear.

A detailed reactor-level WNISR analysis estimated decommissioning costs for the three nuclear phaseout countries Germany, Italy, and Lithuania at approximately US$20206.8, US$202016 and US$202015.7 per MWh, respectively, for high-capacity commercial reactors—at least an order of magnitude larger than most international estimates and at a level that could affect the competitiveness of nuclear, especially on wholesale markets. These cases are particularly significant as the total generation of nuclear electricity is known and thus allow to allocate costs to a fix number of kWh.

An IAEA analysis on funding mechanisms of decommissioning and waste management costs found that about 30 percent of the countries rely on government funding or that of a state-owned enterprise. For the other 70 percent, coverage security is uncertain. All countries rely on taxpayer money to make up for shortfalls.

Detailed European case studies highlighted large aggregate shortfalls between provisioned funds for decommissioning and the expected costs. This gap amounted to estimated US$202310.9 billion in France, US$20236.6 billion in Germany and US$20232.7 billion in Sweden. In the U.S., the transfer of ownership of closed reactors—together with their access to decommissioning funds—to private companies carries specific risks in the case of cost overruns, bankruptcy, or a major accident that could rapidly drain available funding.

Cost estimates for nuclear waste management from the operation of reactors and fuel chain facilities as well as from their decommissioning have reached astronomical levels. For spent fuel disposal alone, estimates for the U.S. reach up to US$2018168 billion and for Canada over US$202019 billion; for the French high-level waste repository construction, there is a “target cost” of US$201628 billion; and if including all radioactive waste streams, estimated disposal costs reach US$2023163 billion for Germany; and US$202121 billion for Switzerland.

Nuclear waste management costs per kWh for SMRs are likely to be higher still than in the case of large reactors.

Insufficient Liability Coverage for Nuclear Accidents

Inadequate or subsidized insurance to cover offsite damages from accidents at nuclear power plants or fuel chain facilities, or during transportation, is common worldwide. Focusing on reactor accidents as an example, liability requirements for offsite damages are set by domestic statute. Additional tiers may be provided by national governments once the operator liability limit is reached; and then by a third tier of coverage provided by series of international treaty agreements (which include the Paris Convention, Vienna Convention, various Joint Protocols and Supplementary Conventions). However, even the total coverage in the U.S., which is the largest liability pool in the world for nuclear accidents, is well below expected damages from even a moderate accident. For example, the Japanese Government estimated the cost of the 2011 Fukushima accidents at US$2021223 billion, more than sixteen times the total U.S. insurance pool of US$13.6 billion. The size of the pool declines as older reactors close. Smaller reactors such as SMRs have much lower primary limit requirements via the mandated purchase of a reactor-specific insurance policy. These depend on the size of the reactor but cover a damage range of only US$4.5–74 million. Further, if the reactors are smaller than 100 MWe, they need not participate in the retrospective premium pool at all.

Industry Claims Regarding Uncompensated Benefits, Future New Markets

Industry proponents sometimes claim there is inadequate compensation for nuclear’s role as a provider of firm and high-capacity low-carbon electricity that is also dispatchable. Capacity payments already compensate providers for firm, high-capacity generation in many U.S. power markets and increasingly in Europe as well. Carbon pricing in the E.U.’s emissions trading system, and to a lesser extent some parts of the U.S., already benefit nuclear providers relative to their fossil competitors. The case for dispatch remains unproven, as the sector’s ability to ramp power production to boost supply flexibility remains limited technically, and associated reductions in load factors needed to spread high fixed costs counter incentives to curtail the resource.

Emerging market services that are supposed to help make the economics of nuclear work include hydrogen production, water desalination, supplying industries in need of high temperature process heat, and behind-the-perimeter uses such as data centers and crypto mining. Because most of these uses involve capital-intensive customers relying on nearly 24/7 production to be economic, a nuclear supplier would need to allocate a fixed percentage of production to that user rather than selling intermittent power surpluses. Thus, the alternative markets would compete with existing power customers, not supplement them. Should some use configurations (for example for low-carbon nuclear used to produce hydrogen) be able to stack multiple subsidies on top of each other, nuclear diversions from power markets could rise, as potentially would carbon emissions. Expansion of the reactor base through newbuild would address concerns regarding diversion of existing low-carbon power supply. However, the costs of power are widely viewed as too expensive relative to alternatives to support these ancillary markets.

Chapter Conclusion

Overall, the economic headwinds for nuclear will remain challenging. Research and deployments will rely primarily on government money, absorption of risks, and direct ownership. Even “private” reactor projects will operate in heavily government-supported environments. In the broader energy marketplace, it is likely that by the time cost improvements could occur, technological developments in competing generating technologies, energy storage, demand side management, and energy efficiency will have moved the economic costs down still further and the reactors will remain too costly. No-regrets policies such as putting an appropriate price on carbon would help nuclear economics as well as other decarbonization pathways, though in a more market-neutral way than most of the current “policy support”.

Nuclear Power vs. Renewable Energy Deployment

Events in Ukraine, which roiled energy markets in 2022, continue to have significant effects on energy-policy decisions for the short and medium term. Some countries have clearly boosted their investments in renewable energies but nuclear power has remained high on the political agendas even though little has followed on the ground so far.

Investment. In 2022, total investment in non-hydro renewable electricity capacity reached a new record of US$495 billion, up 35 percent compared to the previous year, and 14 times the reported global investment decisions for the construction of nuclear power plants of about US$35 billion for 9.4 GW. Investment in solar surged by 50 percent to reach US$307 billion following a 37 percent increase in 2021. Investments in wind power plants increased by 19 percent to US$174 billion. Investments in renewables constitute an estimated 74 percent of all power generation investments in 2022: in contrast, investment in nuclear energy accounted to only 8 percent, the same level as for new coal plants. China’s renewables investment was more than a factor of two larger than the combined European and U.S. investments and larger by than the total global investment in nuclear power over the past decade.

Installed Capacity. A record 348 GW of new renewable energy capacity (including hydro) was installed in 2022, with wind adding around 75 GW of new capacity. The estimates of new solar PV capacity vary widely from 191 GW (IRENA) to 243 GW (REN21) taking total installed capacity beyond 1 terawatt for the first time (in both estimates). These numbers compare with a net addition of 4.3 GW in operating nuclear power capacity.

Electricity Generation. In 2021, the combined output of solar and wind plants surpassed nuclear power generation for the first time. In 2022, wind and solar facilities generated 28 percent more electricity than nuclear plants. Load factors have improved significantly and, as of 2020, stood at 16 percent for utility scale PV, 36 percent for onshore wind, and 44 percent for offshore wind. A floating offshore Scottish wind farm has achieved an average load factor of 54 percent over its first five years of operation, higher than the 52 percent for the French nuclear fleet in 2022.

Share in Power Mix. In 2022, wind (7.2 percent) and solar (4.5 percent) together reached 11.7 percent share of electricity generation, with all non-hydro renewables increasing to 14.4 percent, while the contribution of nuclear energy declined to 9.2 percent.

China. Solar PV produced a total of 423 TWh of electricity in 2022, for the first time overtaking nuclear power that generated 397 TWh. Wind outpaced nuclear in 2012 and has stayed ahead every year since. Wind power plants produced 755 TWh, nearing the double of nuclear power generation. Adding other non-hydro renewables like biomass to solar and wind, total generation of 1,346 TWh net represents 3.4 times the nuclear output, or more than twice the total consumption (577 TWh gross) of Germany, the world’s third largest economy.

European Union. In 2022, renewable electricity generation (including hydropower) reached a new record of 1,080 TWh (gross), with solar energy contributing 203 TWh, up 24 percent from the previous year. Solar and wind plants together produced 624 TWh—more than nuclear energy with 613 TWh, natural gas with 557 TWh, and coal with 447 TWh. All renewable sources combined accounted for over 38 percent of the E.U.’s electricity production.

India. During 2022, 13 GW of solar power capacity was added to reach a total of 62.8 GW. Solar PV generated 94.2 TWh during the year. Since 2021, solar plants have generated more power than wind turbines, which contributed 69 TWh in 2022. Wind has outpaced nuclear in power generation since 2016. Solar passed nuclear generation in 2019. Wind and solar together produced 3.7 times more power than nuclear plants in 2022.

United States. In 2022, nuclear generation declined by 4.7 percent to 772 TWh or 18.2 percent of the electricity mix while wind and solar energies together contributed 14 percent. Including other power sources like biomass and geothermal, non-hydro renewables generated 709.4 TWh (net). If hydropower plants are included contributing 256 TWh, for the first time, with 965.4 TWh, renewables generated more power than coal with 904 TWh (gross).

Introduction

2023 is not over yet, but it is obvious that the war in Ukraine will not have ceased at year-end. Another brutal war has started in the Middle East with protagonists already warning that it will be a long one. Countless scenarios for a regional escalation are possible. And, if Hamas missiles went all the way to Tel Aviv, the Israeli military nuclear complex Dimona in the Negev desert is clearly within reach.

Ukrainian nuclear power plants remain in the middle of an active war zone with one site, the Zaporizhizhia nuclear plant, still occupied by Russian military forces, assisted by engineers of Russian state-owned company Rosatom. As long as the war goes on in Ukraine, there remains a significantly heightened risk of a major nuclear disaster. WNISR2022 detailed why a nuclear reactor needs a functioning cooling system at all times, meaning it also needs reliable electricity supply at all times—during operation and after shutdown.

Repeated calls by various stakeholders, including the Ukrainian Government and the European Parliament, to extend sanctions against Russia to the nuclear sector have remained largely unheeded, aside of U.S. sanctions against Rosatom subsidiary Rusatom Overseas that used to implement Rosatom projects in various countries. Interestingly, that April 2023 decision did not trigger any mainstream media coverage at all, apart from a piece in the French satirical journal Le Canard Enchaîné in August 2023.11

Dependencies of many countries on Russia as nuclear service and hardware provider remain deep. In the European Union (E.U.), Bulgaria, the Czech Republic, Finland, Hungary, and Slovakia operate Russian designed VVERs and are depending on Russian fuel to a great extent. Westinghouse, besides Rosatom the only manufacturer able to manufacture fuel assemblies for the Soviet designed reactors, has so far supplied VVER fuel mainly to Ukraine. These fuel supplies were so far limited to VVER-1000 reactors and have had technical difficulties, but Westinghouse reported in September 2023 to have delivered the first batch of VVER-440 fuel, fabricated in its Swedish plant in Västerås. This will be used in Ukraine for the two-unit Rivne (Rovno) nuclear power plant.12 Ukraine’s Minister of Energy German Galushchenko commented:

The greatness of this day is the end of the Russian monopoly in this segment of the nuclear fuel market. This will pave the way for not only Ukraine, but the whole region, to achieve true nuclear energy independence.

This development is indeed of great significance also to four E.U. countries that operate VVER-440 reactors in the E.U.13 VVER operators have shown interest in Westinghouse fuel in the past and that interest has obviously significantly grown since February 2022. However, following the signature of a “Strategic Cooperation Agreement” with Rosatom in December 2021, French manufacturer Framatome continues to count on its Russian partner and wishes to manufacture VVER fuel in its manufacturing plant in Lingen, Germany, and market the fuel through a Rosatom/Framatome Joint Venture. Why Framatome did not seek cooperation with Westinghouse—whose President and CEO is a French national—rather than cooperate with Rosatom remains unclear. Framatome and Westinghouse cooperate in other areas of nuclear power (e.g. emergency diesel generators, maintenance, Cobal-60 production). Obviously though, the region remains far from “true nuclear energy independence”.

Despite the war in Ukraine, Russia continues to enjoy the top spot in the niche sellers’ market of nuclear reactor building around the world. Since the official construction start of the second Hinkley Point C unit in the U.K. in December 2019 and until mid-2023, work began on 28 reactors in the world, of which 17 in China and all 11 others implemented by Rosatom in various countries. Since Russia’s full-scale invasion of Ukraine in February 2022 and up to mid-2023, Rosatom started building three reactors in Egypt, and one each in China and Turkey.

The question about the role of the International Atomic Energy Agency (IAEA) had been raised in the Introduction to WNISR2022. The Agency’s Director General Rafael Mariano Grossi repeatedly visited the Ukrainian nuclear sites and confirmed Rosatom’s presence in Zaporizhizha. Meanwhile, Mikhail Chudakov, appointed by President Putin and former longtime official of Rosatom companies, remains Grossi’s Deputy Director General and Head of the IAEA’s Department of Nuclear Energy.

Two IAEA General Assemblies passed since the beginning of the all-out war in Ukraine, and not a word has come out of the meetings on potential discussions about basic conditions for technical assistance now and in the future. Russia remains the country that implements by far the most newbuild projects around the world, of which many, if not all, with the assistance of the IAEA. It remains unclear under what conditions Russia, state-owned Rosatom, and its many subsidiaries can be seen as responsible nuclear partners now and in the future—or rather, how the general, applicable, non-negotiable IAEA conditions for nuclear assistance and cooperation would be defined. Neither political decision-makers nor the international media have addressed the issue.

The international media continues to provide large-scale coverage of early, often vague developments of Small Modular Reactor (SMR) designs, despite no significant progress on the ground to report—at least not outside China and Russia—with no startups, no construction starts, not even a design certification. On the contrary, the most advanced project in the western world, the U.S.-based NuScale project with a conglomerate of Utah municipalities was terminated in early November 2023. The company NuScale lost more than 80 percent of its stock market value in little more than a year. Unmoved by the foreseeable NuScale project meltdown, the European Commission launched precisely at the same time a “European Industrial Alliance on SMRs”.

The key element for the NuScale debacle was the dramatically increased cost estimate of the project to US$9.3 billion, which brought the estimated cost per kilowatt to US$20,000 for the six-module 462 MW plant, about twice the cost estimate of the most expensive European Pressurized Water Reactor (EPR).

In the chapter Nuclear Economics and Finance, WNISR2023 assesses in great detail the various cost elements of nuclear power and why the economic headwinds for nuclear will remain challenging. Competing generating technologies, energy storage, demand side management, and energy efficiency will continue to move the economic costs down still further and the reactors will remain too costly to compete. The latest Goldman-Sachs analysis provides only the latest example of many. It forecasts that costs for batteries used in electric vehicles will fall by 40 percent between 2022 and 2025 to US$99/kWh and an average of 11 percent per year until 2030.14 Already dozens of natural gas plant projects are being shelved around the world in favor of large grid-connected batteries.

The 2023-United Nations Environment Program (UNEP) Emissions Gap Report15 demonstrates the extent to which current Greenhouse gas emissions trajectories will overshoot the temperature guidelines of the 2015 Paris Agreement—avoid global temperatures rising 1.5 degrees Celsius above pre-industrial levels. This, coupled with the extreme weather events that are occurring at an alarming and ever-increasing rate, and 2023 expected to be hottest year on record, are once again leading to calls for urgent international action to reduce emissions.

While action is needed across all sectors and societies, one of the highest profile initiatives is the call for a trebling of the current use of renewables and the doubling in energy efficiency by 2030. These targets, already embraced by 70 countries as of mid-November 2023, are expected to be endorsed by the global community at COP 28 taking place in December 2023. If fully implemented, they are expected to lead to three times the current level of wind power and five times the installed capacity of solar PV. This would need to be accompanied by the transformation of the power sector with priority given to measures that increase system flexibility, such as dynamic demand, energy storage, and transformed power grids moving further away from a system using centralized generators, like coal and nuclear power.

Considering the data presented in WNISR2023, a similar pledge to triple nuclear power generation by 2050—considering the long lead-times involved in nuclear construction—seems highly unrealistic and, so far, attracted relatively little support with 10 countries signing up by mid-November 2023.

There is an ever-widening gap between media attention, political announcements, public perception on one side and the industrial reality on the other side. The comprehensive documentation and analysis that WNISR2023 provides on the status and trends of the nuclear industry is a description of an economic sector that struggles to maintain ageing operating fleets, accumulates significant delays and cost overruns at construction projects, and fails to timely develop competitive new designs.

General Overview Worldwide

Role of Nuclear Power

In 1970, the Treaty on the Non-Proliferation of Nuclear Weapons (commonly known as the nuclear Non-Proliferation Treaty, or NPT) entered into force. It was seen as a key tool to limit nuclear weapons programs to the five “official” nuclear weapon states China, France, Russia (then the Soviet Union), the United Kingdom, and the United States.16 In return for not acquiring nuclear weapons capabilities, countries were guaranteed access to technology for nuclear power. Article IV of the NPT stipulates that “nothing in this Treaty shall be interpreted as affecting the inalienable right of all the Parties to the Treaty to develop research, production and use of nuclear energy for peaceful purposes without discrimination.”17

Russia is currently the dominating global reactor builder outside China and works closely with the International Atomic Energy Agency (IAEA), especially in potential newcomer countries. The Russian Ministry of Foreign Affairs in its introductory statement to the First Session of the Preparatory Committee for the 11th Review Conference of the Parties to the NPT in August 2023 stressed:

Russia considers the efforts to promote the nuclear energy development central to the IAEA work. We cooperate with the Agency in implementing the initiative launched in 2017 to develop the nuclear energy infrastructure of newcomer countries. Russia is the initiator and leading donor of the IAEA International Project on Innovative Reactors and Fuel Cycles, in which 43 countries and the European Commission participate. (…)

We note that all NPT-compliant countries should have access to peaceful nuclear energy without any additional conditions.18

As of mid-2023, 32 countries operated nuclear power programs in the world, one less (Germany) than a year earlier. Figure 1 illustrates how the spread of nuclear power throughout the world took place at a significantly slower pace and smaller scope than anticipated in the early 1970s:

• Fourteen countries had operating nuclear power reactors (grid connected) when the NPT entered into force in 1970.
• Sixteen additional countries were operating power plants by 1985, the year when reactor startups peaked.
• Four countries (Romania, Iran, the United Arab Emirates and Belarus) started up power reactors for the first time over the past 30 years, of which two in 2020.
• The number of countries operating power reactors in 1996–1997 reached 32. It took another 23 years to reach a new peak at 33 countries.
• Four countries (Germany, Italy, Kazakhstan and Lithuania) abandoned their nuclear programs.
• Thirteen of the 32 nuclear countries have active reactor construction programs.
• Nineteen countries are not constructing any reactors currently; of these, seven countries have either nuclear phase-out, no-new-build, or no-program-extension policies in place. Some of these policies, such as in the Netherlands and Sweden, are currently being revised. However, while policy changes in some countries reopen the door for nuclear newbuild, actual work on the ground would be many years away.

1. Figure 1 | National Nuclear Power Programs Development, 1954–2022

Sources: compiled by WNISR, with IAEA-PRIS, 2023

Notes: This figure only displays countries with operating or once operating reactors.

* Japan is counted here among countries with “active construction”; it is however possible that the only project under active construction (Shimane-3) will be abandoned.

In 2022, the world nuclear fleet generated 2,546 net terawatt-hours (TWh or billion kilowatt-hours) of electricity19, (see Figure 2). After a decline in 2020, nuclear production increased by 3.9 percent in 2021, but stayed just below the 2019 level, and dropped by 4 percent in 2022. China, with a 3-percent increase (compared to 11 percent in 2021), produced more nuclear electricity than France for the third year in a row, and remains in second place—behind the U.S.—of the top nuclear power generators. Outside of China, nuclear production decreased 5 percent to its lowest level since the mid-1990s.

Nuclear energy’s share of global commercial gross electricity generation in 2022 dropped to 9.2 percent—the lowest value in four decades—and over 45 percent below the peak of 17.5 percent in 1996.20

Nuclear’s main competitors, non-hydro renewables, grew their gross output by 14.7 percent and their share in global gross power generation increased by 1.6 percentage points to 14.4 percent.

In 2020, in a global economic environment depressed by the COVID-19 pandemic, fossil fuel consumption in the power sector slumped: oil by 9.7 percent, coal by 4.2 percent, and natural gas by 2.3 percent. In 2021, the trend was reversed with significant increases in oil +8.9 percent and coal +8.5 percent, while natural gas-based electricity increased by only 2.3 percent. In 2022, oil consumption for power generation remained rather stable (-0.7 percent) while coal and gas slightly increased by 1 percent.

In 2022, nuclear commercial primary energy consumption decreased by 4.7 percent while its share in global consumption slightly decreased to 4 percent; it has been around this level since 2014. In the European Union (E.U.) nuclear primary energy consumption decreased by 17 percent.

Non-hydro renewables, including mainly solar, wind and biofuels, continued their growth, with a 13 percent increase, to reach a share of 7.5 percent in primary energy. While the share of non-hydro renewables is now 1.9 times larger than the nuclear share, both figures illustrate how modest the current contribution of both technologies remains in the global context.

In 2022, there were eight countries that increased the share of nuclear in their respective electricity mix, including one newcomer country, the United Arab Emirates (UAE)—versus six in 2021—while thirteen decreased, and 12 remained at a constant level (change of less than 1 percentage point). Besides the UAE, seven countries (China, Czech Republic, Finland, India, South Korea, Pakistan, Russia) achieved their largest ever nuclear production. China, Finland, Pakistan, South Korea, and the UAE started up new reactors during the year, while the Czech Republic and Russia recorded only marginal increases (below 1 percent) and India slowly increased performance of Kakrapar-3, connected to the grid in January 2021 but in commercial operation only in June 2023.

1. Figure 2 | Nuclear Electricity Generation in the World... and China

Sources: WNISR, with IAEA-PRIS and Energy Institute, 2023

Note: IAEA-PRIS production data for the year 2022 does not include Ukraine (data unavailable). Net nuclear production for Ukraine for the year 2022 represented 59 TWh according to the Energy Institute’s “Statistical Review of World Energy” dataset.21 The total number is thus based on IAEA-PRIS plus the production figure for Ukraine from the Energy Institute.

The following noteworthy developments for the year 2022 illustrate the volatile operational situation of the individual national reactor fleets (see country-specific sections for details):

• Argentina’s nuclear production dropped by 26.5 percent, primarily due to months-long—planned and unplanned—maintenance and reparation outages at one of its three reactors.
• Belgium had an exceptional 2021 after years of struggling with technical issues greatly varying nuclear power generation, only to experience a drop of 13 percent in 2022.
• China started up three units in 2022, just as in 2021, with nuclear generation increasing a modest 3.2 percent following an 11.2 percent in 2021.
• France’s nuclear generation dropped by a record 22.7 percent to below 300 GW for the first time since 1990 and remained below 400 TWh for the seventh year in a row. The outlook for 2023 remains dire with forecasted 300–330 TWh generation.
• Germany, subject to intense political pressure in the middle of a severe energy crisis, stretched operation of its remaining three units beyond the previously planned closure date of the end of 2022 to mid-April 2023 when the nuclear phaseout was completed.
• Japan has restarted ten reactors after all of them were down in 2014. In the past few years, nuclear reactors have generated greatly varying amounts of electricity. After a significant increase in 2021, production dropped again by 15.3 percent in 2022.
• South Africa still has a highly volatile nuclear generation pattern. In 2022, output dropped again by 17 percent contributing less than 5 percent to total power generation.
• In the U.K., after decreasing steadily between 2016 and 2021, nuclear generation increased by 4.3 percent in 2022. However, the previous decreasing trend will continue as three more reactors have been closed in 2022. Consequently, output dropped 21.5 percent in the first half-year 2023 compared to the same period in 2022.

1. Figure 3 | Nuclear Electricity Generation and Share in National Power Generation

Sources: IAEA-PRIS, with national sources for France and Switzerland, and Energy Institute data for Ukraine, compiled by WNISR, 2023

Note: For comparison purposes, data used in this graphic are IAEA-PRIS data, except for France, Switzerland, and Ukraine, and may differ from data used in the country sections.

Similar to previous years, in 2022, the “big five” nuclear generating countries—the U.S., China, France, Russia, and South Korea, in that order—generated 72 percent of all nuclear electricity in the world (see Figure 3, left side).

In 2002, China was 15th in terms of global production levels; in 2007, it was tenth, and reached third place in 2016. In 2020—earlier than anticipated due to the mediocre performance of the French fleet—China became the second largest nuclear generator in the world, a position that France held since the early 1980s.

In 2022, the top three countries, the U.S., China, and France, remained at around 57 percent of global nuclear output, underscoring the concentration of nuclear power generation in a very small number of countries.

In many cases, even where nuclear power generation increased, the addition is not keeping pace with overall increases in electricity production, leading to a nuclear share below the respective historic maximum (see Figure 3, right side). Eight countries achieved their historically largest nuclear share in the 1980s and seven in the 1990s, in other words, almost half of the nuclear countries had seen the peak before the turn of the century.

Besides the United Arab Emirates, which started its second reactor in September 2021 and the third one in October 2022, three countries, Pakistan, Slovakia, and Slovenia, in 2022 reached new historic peak shares of nuclear in their respective power mix. Pakistan’s nuclear share advanced by 4.7 percentage points to 16.4 percent, Slovakia’s almost 7 percentage points to 59.2 percent, and Slovenia’s 6.1 percentage points to 42.8 percent. China remained stable at 5 percent, its highest share.

Operation, Power Generation

Since the first nuclear power reactor was connected to the Soviet power grid at Obninsk in 1954, there have been two major waves of startups. The first peaked in 1974, with 26 grid connections. The second reached a historic maximum in 1984 and 1985, just before the Chernobyl accident in 1986, reaching 33 grid connections in each year. By the end of the 1980s, the uninterrupted net increase of operating units had ceased, and in 1990 for the first time the number of reactor closures22 outweighed the number of startups.

The 1993–2002 decade globally produced almost twice as many startups than closures (51/27), while in the decade 2003–2012, startups hardly exceeded half of the closures (33/63). Furthermore, it took the whole decade to connect as many units—33—as in a single year in the middle of the 1980s (see Figure 4).

In the past decade 2013–2022, 66 reactors were started-up—of which 39 (60 percent) in China—and 42 were closed.

1. Figure 4 | Nuclear Power Reactor Grid Connections and Closures in the World

Sources: WNISR, with IAEA-PRIS, 2023

Notes: WNISR considers reactors closed as of the date of their last electricity production, and not as of their closure announcement (which can be made years after the reactor ceased production).

Over the two decades 2003–2022, there were 99 startups and 105 closures. Of these, 49 startups were in China which did not close any reactors. As a result, outside China, there has been a drastic net decline by 55 units over the same period (see Figure 5). As larger units were started up (totaling 90.7 GW) than closed (totaling 68.5 GW) the net nuclear capacity added worldwide over the 20-year period was 22.2 GW. However, since China alone added 46.8 GW, the net capacity outside China declined by almost 25 GW.

In 2021, six units were connected to the grid, of which three were in China, one each in India, Pakistan and the UAE, and eight were closed.

In 2022, seven reactors were connected to the grid, of which three in China and one each in Finland, Pakistan, South Korea, and the UAE, and five were closed, three in the U.K, and one each in Belgium and the U.S.

In the first half of 2023, four units were connected to the grid, one each in Belarus, China, Slovakia, and the U.S., and five were closed, three in Germany and one each in Belgium and Taiwan. (See Figure 5).

1. Figure 5 | Nuclear Power Reactor Grid Connections and Closures – The Continuing China Effect

Sources: WNISR, with IAEA-PRIS, 2023

As of 1 July 2023, a total of 407 nuclear reactors were operating in 32 countries, down four from the situation in mid-2022.23 The current world fleet has a total electric net capacity of 365 GW, after it peaked at 368 GW at the end of 2022. As the annual statistics always reflect the status at year-end, the situation might change again by the end of 2023.

The number of operating reactors remains by eleven below the figure reached in 1989 and by 31 below the 2002 peak (see Figure 6).

For many years, the net installed capacity has continued to increase more than the net number of operating reactors. This is a result of the combined effects of larger units replacing smaller ones and “uprating”. In 1989, the average size of an operational nuclear reactor was about 740 MW, in 2022 it was almost 900 MW. Technical alterations raised capacity at existing plants resulting in larger electricity output, a process known as uprating.24 In the U.S. alone, the Nuclear Regulatory Commission (U.S. NRC) has approved 172 uprates since 1977. The cumulative approved uprates in the U.S. total 8 GW, the equivalent of eight large reactors..These include seven minor uprates (<2 percent of reactor capacity) approved since mid-2020, of which only one since mid-2021.25

A similar trend of uprates and major overhauls in view of lifetime extensions of existing reactors has been seen in Europe. The main incentive for lifetime extensions is economic but this argument is being increasingly challenged as refurbishment costs soar and alternatives become cheaper.

1. Figure 6 | World Nuclear Reactor Fleet, 1954–mid-2023

Sources: WNISR, with IAEA-PRIS, 2023

IAEA Unexpectedly and Quietly Revises Operating Reactor Data

Until September 2022, the IAEA’s online Power Reactor Information System (PRIS) database counted 33 reactors as operational/operating in Japan, whereas 20 of these had not produced power since 2010–2012, and an additional three units had been shut down even since the Niigata Earthquake in 2007.

For almost a decade WNISR has been calling for an appropriate reflection in world nuclear statistics of the unique situation in Japan. The approach taken by the IAEA, the Japanese government, utilities, industry and many research bodies as well as other governments and organizations to continue classifying the entire stranded reactor fleet in the country as “in operation” or “operational” was clearly misleading.

Faced with this dilemma, the WNISR team in 2014 decided to create a new category with a simple definition, based on empirical fact, without room for speculation: “Long-Term Outage” or LTO. Its definition:

A nuclear reactor is considered in Long-Term Outage or LTO if it has not generated any electricity in the previous calendar year and in the first half of the current calendar year. It is withdrawn from operational status retroactively from the day it has been disconnected from the grid.

When subsequently the decision is taken to close a reactor, the closure status starts with the day of the last electricity generation, and the WNISR statistics are retroactively modified accordingly.

Applying this definition to the world nuclear reactor fleet, as of 1 July 2023, leads to classifying 31 units in LTO, of which 23 in Japan, three in India (Madras-1, Tarapur-1 & -2), two in Canada (Bruce-6 and Darlington-3, restarted after refurbishment in the second half of 2023, after WNISR’s statistical deadline), one in China (CEFR, which has been retrieved altogether from the IAEA-PRIS database in May 2023), one in France (Penly-1, restarted in July 2023 after statistical deadline), and Kori-2 in South Korea, whose license expired in April 2023, and is in the process of seeking a license renewal.

IAEA: Change is Coming – New Category “Suspended Operation”

Ten years ago, on 16 January 2013, the IAEA moved 47 reactors in Japan, most of them shut down in the aftermath of the Fukushima events in 2011, from the category “In Operation” into “Long-term Shutdown”26 that existed in the IAEA statistical system until October 2022. Only two days later, the move was labelled a “clerical error” and the action was reversed at the request of the Japanese government.27

It is only in September 2022, that in the IAEA-PRIS database, twelve Japanese reactors28 were gradually withdrawn from the list of “operating” or “operational” reactors, and their status changed to “Long-term Shutdown” (LTS). By mid-October 2022, the category title was changed to “Suspended Operation” on the PRIS website29, and in November 2022, four more Japanese units30 joined the new category as well as one Indian reactor (Rajastan-1) that has not generated any power since 2004 and is considered closed by WNISR.

As of the end of 2022, the PRIS database still counted 17 Japanese reactors as “in Operation”. Whereas ten have effectively restarted since the beginning of the Fukushima disaster (also referred to as 3/11), the remaining seven have not produced any electricity since 2010–2012. Then, in April 2023, those seven units also joined the “Suspended Operation” category, followed in May 2023 by three additional Indian reactors, that have not produced power since 2018 (Madras-1) and 2020 (Tarapur-1 & -2).

The definition of the new category is as follows:

A reactor is considered in the suspended operations status, if it has been shut down for an extended period (usually more than one year) and there is the intention to re-start the unit but:

1. restart is not being aggressively pursued (there is no vigorous onsite activity to restart the unit) or

2. no firm restart date or recovery schedule has been established when unit was shutdown [shut down].

Suspended operations may be due to [due to] technical, economical, strategic or political reasons. This status does not apply to long-term maintenance outages, including unit refurbishment, if the outage schedule is consistently followed, or to long-term outages due to regulatory restrictions (licence suspension), if restart (licence recovery) term and conditions have been established. Such units are still considered “operational” (in a long-term outage). If an intention not to restart the shutdown unit has been officially announced by the owner, the unit is considered “permanently shutdown [shut down]”.31

It is important to understand that the application of this new rule modifies retroactively all of the IAEA’s statistics on operating reactors—in most cases as of day of last production—back to 2007. This dramatically modifies the IAEA’s representation of the Japanese nuclear reactor fleet’s evolution (see Figure 7). The changes obviously also impact the IAEA’s representation of the long-term evolution of the entire global nuclear power-reactor fleet (see Figure 8).

1. Figure 7 | Evolution of the Japanese Nuclear Reactor Fleet, 1963 to mid-2022

Sources: IAEA-PRIS, July 2022 and July 2023

Note: All reactors listed in the “Suspended Operation” category as of July 2023, were considered Operational in July 2022 statistics.

While now reflected on the PRIS Homepage32 and the PRIS Japan Country Details page33, all of those changes happened without any public announcement or online explanation. The IAEA has argued in the past that they only serve as the “database manager”, the IAEA being only in a position to provide suggestions, with all changes ultimately being decided by Member States officials and implemented in the PRIS database by the respective Government appointed data providers, the “correspondents”.

Apparently, there have been lengthy discussions for several years between the IAEA and the Japanese correspondents on how to address the obvious mislabeling of stranded reactors as “in operation”. In view of public perception, the Japanese government was eager to avoid the term “shutdown” as many of the reactors were officially planned to be restarted (see Japan Focus).

The differences with WNISR statistics are greatly reduced, and the remaining ones mostly relate to official closure dates, as WNISR statistics consider the end of electricity production as reference for dating closures, and not the “announcement” or “political decision” to permanently withdraw a reactor from the grid (see also IAEA Unexpectedly and Quietly Revises Operating Reactor Data above).

1. Figure 8 | World Nuclear Reactor Fleet – IAEA-PRIS Statistics Evolving Over Time

Sources: IAEA-PRIS statistics as of July 2022, January 2023 and July 2023

Notes: The IAEA data used for this graph includes at least three reactors that have been later withdrawn from the PRIS statistics for operating reactors (Niederaichbach, VAK-Kahl and HDR Großwelzheim, in Germany, now only appearing as “Decommissioning Completed”). On the other hand, the Swiss research reactor in Lucens is not included.

The Chinese CEFR was retrieved from the IAEA-PRIS statistics in May 2023 and is therefore only included in July 2022 and January 2023 datasets. Reactors classified as in “Suspended Operation” by the IAEA are not represented here.

As of July 2022, according to the IAEA-PRIS statistics, the evolution of the world nuclear fleet showed a peak of officially operating reactors in 2018, both in terms of number and capacity, with 449 reactors and a maximum capacity of 396.4 GW, declining since. The corresponding data for the end of 2021 showed 437 reactors in operation with a capacity of 389.5 GW.34

The July 2023 data obviously offers a different picture: If the operating capacity still peaked in 2018 in those revised statistics, it only reached 374.1 GW, 22.3 GW less than the 396.4 GW previously indicated, whereas the number of operating reactors never exceeded the number of 440, reached already in 2005.

IAEA vs. WNISR Assessment

WNISR’s assessment of “operating” reactors has shown significant differences with IAEA statistics since the beginning of the Fukushima disaster in 2011. However, after major changes in the PRIS statistics (see IAEA Unexpectedly and Quietly Revises Operating Reactor Data and IAEA: Change is Coming – New Category “Suspended Operation” above), those differences were reduced to minor disparities during the period September 2022 to May 2023, compared to WNISR2022.

The following section provides a detailed explanation and justification of the differences.

Figure 9 presents the evolution of the number and capacity of the world reactor fleet “in operation” as reported by the IAEA vs. WNISR.

As of July 2023, the evolution of the world nuclear fleet in the PRIS statistics shows a peak of 440 reactors operating in 2005, while the operating capacity reached a maximum of 374 GW in 2018; as of end 2022, the operating capacity was 371 GW. In the WNISR statistics, which consider reactors closed from the day they stop producing electricity, and systematically apply the LTO status to reactors not operating for a certain period, a maximum number of 438 reactors was reached as soon as 2002, and again in 2005. The operating capacity slightly increased in 2022 beyond the previous peak of 2006, to reach a maximum of 368 GW.

Although not the only case, the Japanese fleet still provides the main and more visible differences, especially over the past decade. This applies both to reactors that did not produce electricity for many years before they returned to service (designated as “LTO later restarted” or “Restarted from LTO”), or which were declared permanently closed years after they stopped producing electricity (“Closed at a later date”).

Applying this definition to the world nuclear reactor fleet, as of 1 July 2023, leads to classifying four units considered “in operation” by the IAEA as LTO:

• Bruce-6 and Darlington-3, under refurbishment since January and July 2020 respectively. They came back online in the second half of 202335 (see section on Canada in Annex 1).
• Penly-1, shut down on 2 October 2021 for its third decennial inspection, reconnected to the grid on 13 July 2023 (see France Focus).36
• Kori-2 in South Korea, shut down in April 2023, after 40 years of operation, is expected to be restarted at an unknown date, and is therefore considered in LTO.

1. Figure 9 | World Nuclear Reactor Fleet – IAEA vs. WNISR, 1954–July 2023

Sources: IAEA-PRIS and WNISR, 2023

Notes: The IAEA data used for this graph includes at least three reactors that have been later withdrawn from the PRIS statistics for operating reactors (Niederaichbach, VAK-Kahl and HDR Großwelzheim, in Germany, now only appearing as “Decommissioning Completed”). On the other hand, the Swiss research reactor in Lucens is not included. Reactors classified as in “Suspended Operation” by the IAEA are not represented here.
Although the total number of reactors in operation according to WNISR statistics has always remained, albeit slightly, inferior to IAEA-PRIS data, it contains Chinese reactors not accounted for in PRIS (see below).

But on the other hand, WNISR statistics do include additional reactors in China:

• Shidao-Bay-1: The IAEA considers the two 100-MW modules as one reactor as they drive a single 200-MW turbine. WNISR considers that each module is a separate reactor.
• CEFR: The IAEA has simply deleted the file for the reactor without any indication of reasons. Chinese sources have argued it should have never been in the IAEA’s PRIS database in the first place as it is to be considered an experimental reactor. However, as this is a nuclear power reactor, it is considered as such by WNISR. Its current operational status is uncertain. In the absence of operational data, WNISR considers it in LTO as of May 2023 (but still operating as of December 2022).37

The biggest difference between IAEA-PRIS and WNISR is found as of the end of 2012, with 29 units less operating according to WNISR criteria: the IAEA-PRIS counts 30 reactors (detailed in Table 1) that are not considered operating according to WNISR, but on the other hand has retrieved the Chinese CEFR it previously considered operational at this date.

1. Table 1 | WNISR Rationale for the Classification of 30 Reactors as Non-Operational as of end 2012

   Countries	Officially Closed at a Later Date 21 Reactors		Restarted from LTO 9 Reactors
   	Reactors that last produced electricity in (or prior to) 2012, officially closed after 2012 (either considered closed by WNISR as early as 2012, or after a certain period in LTO). Most of those reactors were considered “in operation” for many years before their official closure date.		Reactors in LTO as of December 2012 Restarted prior to 1 July 2023
   	Reactors considered closed in 2012	Reactors in LTO prior to closure
   Japan	6 Reactors Fukushima Daiichi 5–6 Fukushima Daini 1–4 Officially Closed in 2013 and 2019	11 Reactors Last production in 2010–2012 Officially closed 2015–2019	8 Reactors Restarted 2015–2021
   South Korea			1 Reactor Wolsong-1, Restarted in 2015
   Spain	1 Reactor Santa Maria de Garoña Last production in 2012 Officially Closed in 2017*
   U.S.	3 Reactors San Onofre-2 & -3 Last production in 2012 Officially closed in 2013 Crystal River-3 Last production in 2009 Officially closed in 2013

Sources: IAEA-PRIS and WNISR, 2023

Note: *Garoña was subsequently considered in “Suspended Operation” during 2013–2016 by the IAEA until its official closure.

The differences between the IAEA and WNISR are not limited to the effects of the Fukushima disaster. Even prior to 3/11, WNISR and IAEA-PRIS data had differences, reaching up to 10 units at the end of some years. These differences were mainly due to the definition of the closure date that the IAEA sometimes sets at last production and sometimes as closure-decision date while WNISR systematically applies the day of last electricity generation (when available).

Overview of Current New-Build

As of 1 July 2023, 58 reactors were considered as under construction, five more than the WNISR reported a year ago, but 11 fewer than in 2013 (of the 69 reactors under construction at the end of 2013, four units have subsequently been abandoned). The number includes 23 units (40 percent) being built in China.

Four in five reactors are built in Asia or Eastern Europe (see Building vs. Vendor Countries). In total, 16 countries are building nuclear plants, with a (provisional?) construction restart in Brazil, new construction in Egypt, and Belarus having started up its second and only reactor under construction, that is one more country than in WNISR2022.

However, only four countries—China, India, Russia, and South Korea—have construction ongoing at more than one site, and eight countries only have a single reactor under construction (see Table 2 and Annex 3 for details). Since mid-2022, construction of ten new units was launched worldwide, including four in China and three in Egypt.

The 58 reactors listed as under construction by mid-2023 compared with 234 units—totaling more than 200 GW—in 1979. However, many (48) of those projects listed then were never finished (see Figure 10). The year 2005, with 26 units listed as under construction, was the lowest since the early nuclear age in the 1950s.

1. Figure 10 | Nuclear Reactors “Under Construction” in the World

Sources: WNISR, with IAEA-PRIS, 2023

Notes: This figure includes construction of two CAP1400 reactors at Rongcheng/Shidaowan, although their construction has not been officially announced (see China Focus). At Shidao Bay, the HTR plant, where construction started in 2012, has two reactor modules on the site and is therefore counted as two units as of WNISR2020. Grid connection of the first unit of the twin reactors officially took place on 20 December 2021. No date was provided for startup of the second reactor, which is considered as operating in WNISR2023 as of end-2022 (see China Focus for details).

Compared to the year before, the total capacity of the 58 units under construction in the world in mid-2023 increased by 5.3 GW to 58.6 GW, with an average unit size of 1,010 MW.

1. Figure 11 | Nuclear Reactors “Under Construction” – China and the World (as of 1 July 2023)

Sources: WNISR, with IAEA-PRIS, 2023

Building vs. Vendor Countries

As of mid-2023, China has by far the most reactors (23 units) under construction in the world. However, it is currently not building anywhere outside the country and has only exported to Pakistan. Russia is in fact largely dominating the international market as a technology supplier with 24 units under construction in the world, as of mid-2023, of which only five are domestic and 19 in seven different countries, including four each in China, India, and Turkey, three in Egypt and two in Bangladesh. It is uncertain to what extent these projects will be impacted by the various layers of sanctions imposed on Russia following the invasion of Ukraine.

Besides Russia’s Rosatom, there are only French and South Korean companies building abroad (see Table 2 and Figure 12).

1. Tble 2 | Nuclear Reactors “Under Construction” (as of 1 July 2023)38

Sources: Various, compiled by WNISR, 2023

Notes:

(a) - Of the eight reactor projects under construction, all are delayed or likely to be delayed, with all Kudankulam reactors under construction “likely to be impacted” by the war in Ukraine. Six is the number of reactors “formally” delayed. See section on India in Annex 1, and Annex 3.

(b) - Angra-3 in Brazil is a Konvoi design originally developed by Siemens/KWU now owned by EDF/Framatome. The construction completion is managed by the Brazilian state-controlled ENBpar. It remains unclear who will be carrying out the work.

(c) - Mochovce -4 is a Russian VVER design being completed by a Czech-led consortium.

This table does not contain suspended or abandoned constructions. It does include construction of two CAP1400 reactors at Rongcheng/Shidaowan, although that has not been officially announced (see China Focus) as well as two floating reactors of Russian design to be deployed in Russia—thus counted under Country-Russia, but with the barges built in China.

1. Figure 12 | Nuclear Reactors “Under Construction” by Technology-Supplier Country

Sources: WNISR, with IAEA-PRIS, 2023

Construction Times

Construction Times of Reactors Currently Under Construction

A closer look at projects listed as “under construction” as of 1 July 2023 illustrates the level of uncertainty and problems associated with many of these projects, especially given that most builders still assume a five-year construction period:

• For the 58 reactors being built, an average of 6 years has passed since construction start—slightly lower than the mid-2022 average of 6.8 years—and many remain far from completion.
• All reactors under construction in at least 10 of the 16 countries have experienced often year-long delays. Almost half (28) of the building projects are delayed or likely to be delayed. Most of the units which are nominally being built on-time (yet) were begun within the past three years or have not yet reached projected startup dates, making it difficult to assess whether they are on schedule. Significant uncertainty remains over construction in China because of lack of access to information. Five of six units that started building prior to 2020 and are not yet documented as delayed are located in China and one in Bangladesh. The latter, Rooppur-2 is likely to be late, but it is not yet documented. It remains also unclear what will happen with Russian designed and/or implemented projects in six other countries, as sanctions have or will likely have an impact on supply chains.
• Of the 24 reactors clearly documented as behind schedule, at least nine have reported increased delays and one has reported a delay for the first time over the past year.
• WNISR2021 noted a total of 12 reactors scheduled for startup in 2022. At the beginning of 2022, 16 were still planned to be connected to the grid (including four pushed back from 2021 to 2022) but only seven of these made it, while the other 9 were delayed at least into 2023.
• Initial construction start of the Mochovce-4 reactor in Slovakia dates back 38 years and its grid connection has been further delayed, currently to 2024. Bushehr-2 in Iran originally started construction in 1976, over 47 years ago, and resumed construction in 2019 after a 40-year-long suspension. Grid connection is currently scheduled for 2024.
• Seven additional reactors have been listed as “under construction” for a decade or more: Angra-3 in Brazil, the Prototype Fast Breeder Reactor (PFBR), Kakrapar-4 and Rajasthan-7 & -8 in India, Shimane-3 in Japan, and Flamanville-3 (FL3) in France. The French and Indian projects have been further delayed this year, and the Japanese reactor does not even have a provisional startup date. Angra-3 construction, which initially started in 2010, was halted in 2015, apparently resumed in 2022, with an expected startup date of 2028. However, construction activities have been interrupted repeatedly.

The actual lead time for nuclear plant projects includes not only the construction itself but also lengthy licensing procedures in most countries, complex financing negotiations, site preparation, and other infrastructure development.

Construction Times of Past and Currently Operating Reactors

Since the beginning of the nuclear power age, there has been a clear global trend towards increasing construction times. National building programs were faster in the early years of nuclear power, when units were smaller, and safety and environmental regulations were less stringent. As Figure 13 illustrates, average times between construction start and grid connection of reactors completed in the 1970s and 1980s were quite homogenous, while in the past two decades they have varied widely.

The eight units completed in 2020–2022 in China took on average 6.4 years to build, while it took 10.5 years to finalize one project in Russia (compared to an average 15 years for the period 2018–2020).

As Figure 14 shows for the period 2020–2022, the longest construction time was for the Olkiluoto-3 (OL3) reactor (16.6 years), a Franco-German project, the first European Pressurized Water Reactor (EPR) to start up in Europe, twelve years later than planned. The longest construction times in Russia and China were seen for the EPR at Taishan-2 (9.2 years), the first reactor of the two HTR module at Shidao Bay-1 (9.1 years) and Leningrad 2-2 (10.5 years).

1. Figure 13 | Average Annual Construction Times in the World

Sources: WNISR, with IAEA-PRIS, 2023

The mean time from construction start to grid connection for the seven reactors started up in 2022 was nine years, 1.7 years more on average than construction times of units started up in 2021 (7.3 years). In the case of the four units connected in the first half of 2023 to power grids in Belarus, China, Slovakia, and the U.S., the average time from first basemat concreting to first power generation was 16 years. This includes Mochovce-3 in Slovakia, with construction starting first in 1985.

Over the three years 2020–2022, only two of 18 units connected to the grid in seven countries started up on-time. Those are Tianwan-4 and -5 in China, Russian-designed but mainly Chinese-built VVER-1000s (model V-428M), that the designers claim to belong to Generation III (Gen III) classification, but few details are known.

The longer-term perspective confirms that short construction times remain the exceptions. Ten countries completed 66 reactors over the decade 2013–2022—of which 39 in China alone—with an average construction time of 9.4 years (see Table 3), slightly higher than the 9.2 years of mean construction time in the decade 2012–2021.

1. Figure 14 | Delays for Units Started Up 2020–2022

Sources: Various, compiled by WNISR, 2023

Note: Expected construction time is based on grid connection data provided at construction start when available; alternatively, best estimates are used, based on commercial operation, completion, or commissioning information.

At Shidao Bay, the HTR plant, where construction started in 2012, has two reactor modules on the site and is therefore counted as two units as of WNISR2020. Grid connection of the first unit of the twin reactors officially took place on 20 December 2021. No date was provided for startup of the second reactor, which is considered as operating in WNISR2023 as of end-2022, and total construction time set at 10 years.

1. Table 3 | Duration from Construction Start to Grid Connection, 2013–2022

Sources: Various, compiled by WNISR, 2023

Construction Starts and Cancellations

The number of annual construction starts39 in the world peaked in 1976 at 44, of which 11 projects were later abandoned. In 2010, there were 15 construction starts—including 10 in China—the highest level since 1985 (see Figure 15 and Figure 16). That number dropped to five in 2020 (including four in China, while building started on ten units in 2021 (including 6 in China), as well as in 2022 (including five in China). The other five units are implemented by the Russian nuclear industry in Egypt (2), in Turkey (1) and domestically (2), while two of the construction starts in China were also carried out by the Russian industry. In other words, of the global total of ten, seven reactors were by Russian builders and three by the Chinese industry.

Three reactors got underway in the world in the first half of 2023, two of them in China, and one of Russian design in Egypt. Chinese and Russian government-owned or -controlled companies launched all 28 reactor constructions in the world over the 42-month period from the beginning of 2020 to mid-2023.

1. Figure 15 | Construction Starts in the World

Sources: WNISR, with IAEA-PRIS, 2023

Notes: Construction of Bushehr-2 in Iran started in 1976, was considered abandoned in earlier versions of this figure. As construction was restarted in 2019, it now appears as “Under Construction”. Albeit of uncertain future, construction of Angra-3 in Brazil is considered restarted.

Over the decade 2013–2022, construction began on 65 reactors in the world, of which almost half (31) in China. Two of these building sites have been abandoned over the period (V.C. Summer-2 and -3 in the U.S.). As of mid-2023, 17 of the remaining 63 units had started up, while 46 remain under construction.

Seriously affected by the Fukushima events, China did not start any construction in 2011 and 2014 and began work only on seven units in total in 2012 and 2013. While Chinese utilities started building six more units in 2015, the number shrank to two in 2016, only a demonstration fast reactor in 2017, none in 2018, but four each in 2019 and 2020, six in 2021, five in 2022 and two in the first half of 2023 (see Figure 16). While this increase represents a sign of the restart of commercial reactor building in China, the level continues to remain far below expectations. The five-year plan 2016–2020 had fixed a target of 58 GW operating and 30 GW under construction by 2020. As of the end of 2020, China had 49 units with 47.5 GW operating, one reactor in LTO (CEFR), and 17 units (16 GW) under construction, much lower than the original target. At the end of 2022, 56 reactors with a total capacity of 52.2 GW were operating and 22 units (23.1 GW) were under construction (for details, see China Focus).

1. Figure 16 | Construction Starts in the World/China

Sources: WNISR, with IAEA-PRIS, 2023

Experience shows that having an order for a reactor, or even having a nuclear plant at an advanced stage of construction, is no guarantee of ultimate grid connection and power production. The two V.C. Summer units, abandoned in July 2017 after four years of construction and following multi-billion-dollar investment, are only the latest in a long list of failed significantly advanced nuclear power plant projects.

French Alternative Energies & Atomic Energy Commission (CEA) statistics through 2002 indicate 253 “cancelled orders” in 31 countries, many of them at an advanced construction stage (see also Figure 17). The United States alone accounted for 138 of these order cancellations.40

1. Figure 17 | Cancelled or Suspended Reactor Constructions

Sources: WNISR, with IAEA-PRIS, 2023

Note: This graph only includes constructions that had officially started with the concreting of the base slab of the reactor building. Many more projects have been cancelled at earlier stages of construction/site preparation.

Of the 800 reactor constructions launched since 1951, at least 92 units in 18 countries had been abandoned or suspended, as of 1 July 2023. This means that 11.5 percent—or one in nine—of nuclear constructions have been abandoned.

Close to three-quarters (66 units) of all cancelled projects were in four countries alone—the U.S. (42), Russia (12), Germany and Ukraine (six each). Some units were 100-percent completed—including Kalkar in Germany and Zwentendorf in Austria—before it was decided not to operate them.

Operating Age

In the absence of significant, successful newbuild over many years, the average age (from grid connection) of operating nuclear power plants has been increasing since 1984, and as of mid-2023 is 31.4 years, up from 31 years in mid-2022 (see Figure 18).41

A total of 265 reactors—five less than mid-2022—two-thirds of the world’s operating fleet, have operated for 31 or more years, including 111—more than one in four—for at least 41 years.

1. Figure 18 | Age Distribution of Operating Reactors in the World

Sources: WNISR, with IAEA-PRIS, 2023

In 1990, the average age of the operating reactors in the world was 11.3 years; in 2000, it was 18.8 years and it stood at 26.3 years in 2010. The leading nuclear nation also has the oldest reactor fleet of the top-five nuclear generators. The average age of reactors in the U.S. passed 40-years in 2020 and reached 42.1 years as of the end of 2022. France’s fleet exceeded 37 years. Russia’s fleet age peaked in 2017 and declined for a few years before increasing again starting in 2020 and its average fleet age of 29.4 years, as of the end of 2022, caught up with that of 2018. South Korea’s reactors at 22.6 years remain almost half as old as the U.S. fleet, and China has an average fleet age of just 9.3 years. (See Figure 19).

Many nuclear utilities envisage average reactor lifetimes of beyond 40 years up to 60 and even 80 years. In the U.S., reactors are initially licensed to operate for 40 years, but nuclear operators can request a license renewal from the Nuclear Regulatory Commission (NRC) for an additional 20 years. An initiative to allow for 40-year license extensions in one step was terminated in June 2021 after NRC staff recommended that the Commission “discontinue the activity to consider regulatory and other changes to enable license renewal for 40 years.”42

As of mid-2023, 84 of the 93 operating U.S. units had received a 20-year license extension, applications for three further reactors were under NRC review. The owners of three other reactors (Diablo Canyon-1 and -2, Clinton-1) plan to submit applications in late 2023 and early 2024. The Diablo Canyon units, scheduled to close when their current licenses expire in 2024–2025, might defer closure until 2029 and 2030.43

As of July 2023, the NRC had granted Subsequent Renewed Operating Licenses to six reactors, which permit operation from 60 to 80 years. However, the NRC effectively suspended four of these licenses in February 2022, while it develops a new environmental assessment for subsequent license renewals. A further ten reactors have their applications still under review. See “Extended Reactor Licenses” in United States Focus for details and references.

1. Figure 19 | Reactor-Fleet Age of Top 5 Nuclear Generators

Sources: WNISR, with IAEA-PRIS, 2023

Only nine of the 41 units that have been closed in the U.S. had reached 40 years on the grid. All nine had obtained licenses to operate up to 60 years but were closed long before mainly for economic reasons. In other words, at least one quarter of the 134 reactors connected to the grid in the U.S. never reached their initial design lifetime of 40 years. Only one of those already closed had just reached 50 years of operation (Palisades, closed after 50.4 years). The mean age at closure of those 41 units was 22.8 years.

On the other hand, of the 93 currently operating plants, 49 units have already operated for 41 years, of which ten have been on the grid for 51 years or more; thus, over half of the units with license renewals have entered the lifetime extension period, and that share is growing rapidly with the mid-2023 mean age of the U.S. operational fleet exceeding 42.1 years (see United States Focus).

Many countries have no specific time limits on operating licenses. In France, for example, reactors must undergo in-depth inspection and testing every decade against reinforced safety requirements. The French reactors have operated for 38 years on average. The Nuclear Safety Authority (ASN) has evaluated each reactor, and most have been permitted to operate for up to 40 years, which is considered the limit of their initial design. However, the ASN assessments are years behind schedule. For economic reasons, the French state-controlled utility Électricité de France (EDF) prioritizes lifetime extension to 50 years over large-scale new-build.

EDF’s approach to lifetime extension has been reviewed by ASN and its Technical Support Organization. In February 2021, ASN granted a conditional generic agreement to lifetime extensions of the 32 reactors of the 900 MW series. However, lifetime extensions beyond 40 years require reactor-specific licensing procedures involving public inquiries in France. For an assessment of the status of fourth decennial inspections see “Lifetime Extension – Fact Before License” in France Focus.

Recently commissioned reactors and the ones under construction in South Korea do or will have a 60-year operating license from the start. EDF will certainly also aim for 60-year operating licenses for its Flamanville-3 project and the Hinkley Point C units in the U.K.

1. Figure 20 | Age of World Nuclear Fleets

Sources: WNISR, with IAEA-PRIS, 2023

Note: This figure only takes into account reactors operating as of 1 July 2023, thus excluding reactors in LTO, in particular Tarapur-1 & -2 in India, that have passed 50 years.

Figure 20 shows that the average fleet age in 23 of the 32 countries that operate nuclear reactors as of mid-2023, is over 30 years, and in eight countries over 40. Over half, that is 19 of the countries have been operating one or more reactors for more than 40 years, but only five countries operate reactors that are over 50 years, while some others are approaching the milestone.

In assessing the likelihood of reactor fleets being able to operate for 50 or 60 years, it is useful to compare the age distribution of reactors that are currently operating with the 212 units that have already closed (see Figure 18 and Figure 21). In total, 97 of these units operated for 31 years or more, and, of those 97, 41 reactors operated for 41 years or more. Many units of the first-generation designs only operated for a few years. The mean age of the closed units is about 28 years.

1. Figure 21 | Age Distribution of Closed Nuclear Power Reactors

Sources: WNISR, with IAEA-PRIS, 2023

While the operating time prior to closure has clearly increased continuously, the mean age at closure of the 29 units taken off the grids in the five-year period between 2018 and 2022 was 43.5 years (see Figure 22).

As a result of the Fukushima nuclear disaster (elsewhere referred to as 3/11), many analysts have questioned the wisdom of operating older reactors. The Fukushima Daiichi units (1 to 4) were connected to the grid between 1971 and 1974. The license for Unit 1 had been extended for another 10 years in February 2011, just one month before the catastrophe began. Four days after the initial events in Japan, the German government ordered the closure of eight reactors that had started up before 1981, two of which were already closed at the time and never restarted. The sole selection criterion was operational age. Other countries did not adopt the same approach, but clearly the 3/11 events in Japan had an impact on previously assumed extended lifetimes in other countries. Some of the main nuclear countries closed their oldest units, at the time, before or long before age 50, including Germany at age 37, South Korea at 40, Sweden at 46, and the U.S. at 49. France closed its two oldest units in spring 2020 at age 43.

1. Figure 22 | Nuclear Reactor Closure Age

Sources: WNISR, with IAEA-PRIS, 2023

Lifetime Projections

Nuclear operators in many countries continue to implement or prepare for lifetime extensions. As in previous years, WNISR has created two lifetime projections. A first scenario (40-Year Lifetime Projection, see Figure 23), assumes a general lifetime of 40 years for worldwide operating reactors—not including reactors in Long-Term Outage (LTO).

Forty years, explicitly or implicitly, corresponds to the design lifetimes of most operating reactors. Some countries have legislation or policy in place—including Belgium (even if the currently debated lifetime extension for two units was implemented), South Korea (in the course of being amended by the incoming administration) or Taiwan—that limit operating lifetime, for all or part of the fleet, to 40 or 50 years. Recent designs, mostly reactors under construction, have a design lifetime of 60 years (e.g. APR-1400, EPR). For the 122 reactors that have passed the 40-year lifetime as of mid-2023, we assume they will operate to the end of their licensed, extended operating time.

A second scenario (Plant Life Extension or PLEX Projection, see Figure 24) takes into account all already-authorized lifetime extensions as of mid-2023 and assumes that the respective reactors will operate until the expiration of their license—a very conservative assumption considering empirical evidence from the past.

The lifetime projections allow for an evaluation of the number of reactors and respective power generating capacity that would have to come online over the next decades to offset closures and simply maintain the same number of operating plants and level of capacity, if all units were closed after a lifetime of 40 years or after their licensed lifetime extension.

Considering all units under construction scheduled to have started up, 13 additional reactors (compared to the end of 2022 status) would have to be commissioned or restarted prior to the end of 2023 in order to maintain the status quo of operating units. Without additional startups, installed nuclear capacity would decrease by 12 GW by the end of 2023.

1. Figure 23 | The 40-Year Lifetime Projection

Sources: Various, compiled by WNISR, 2023

Notes pertaining to Figure 23, Figure 24 and Figure 25:

Those figures include one Japanese reactor (Shimane), two Chinese 1400 MW-units at Shidao Bay and two Russian 55 MW RITM reactors, for which the startup dates were arbitrarily set to 2025, 2024 and 2027, as there are no official dates.

Restarts or closures amongst the 31 reactors in LTO as of 1 July 2023 are not represented in Figure 23 and Figure 24, although at least two Canadian, two Japanese and one French reactors that were in LTO have restarted since, and will thus be later closed as well. Those are counted as “operating” in Figure 25 (under the criteria of the PLEX projection).

The figures take into account current political decisions or legally binding obligations as of end of July 2023 to close reactors prior to 40 years (South-Korea). These decisions are under discussions and might be reversed after the editorial deadline of WNISR2023, as is the case in Belgium, with discussions on a ten-year lifetime extension for two reactors beyond the current license expiration in 2025.

In the case of reactors that have reached 40 years of operation prior to 2023, the 40-year projection also uses the end of their licensed lifetime (including 6 reactors licensed for 80 years in the U.S., even though the licenses of four of these units have been suspended).

In the case of French reactors that have reached 40 years of operation prior to 2023 (startup before 1983), we use the deadline for their 4th periodic safety review (visite décennale) as closing date in the 40-year projection. In case this deadline is or will be passed by the end of 2023 (10 reactors), we use a 10-year extension, although no licensing procedure has been completed for this extension besides Tricastin-1. For all those that have already passed their 3rd periodic safety review, the scheduled date of their 4th periodic safety review (or 10-year extension for the cases previously mentioned) is used in the PLEX projection, regardless of their startup date.

In the remaining years to 2030, in addition to the units currently under construction, 141 new reactors (121 GW)—over 17 units or 15 GW per year—would have to be connected to the grid to maintain the status quo, almost three times the rate achieved over the past decade (66 startups between 2013 and 2022, that is 6.6 units or 6.5 GW per year).

The relative stabilization of the situation by the end of 2023 is only possible because most reactors will likely not close, regardless of their age. The number of reactors in operation will probably continue to stagnate at best, unless—beyond restarts—lifetime extensions become the rule worldwide. Such generalized lifetime extensions—far beyond 40 years—are clearly the objective of the international nuclear power industry, and, especially in the U.S., there are numerous attempts to obtain subsidies for uneconomic nuclear plants in order to keep them on the grid (see Securing Subsidies to Prevent Closures in United States Focus).

Developments in Asia, including in China, do not fundamentally change the global picture. Reported ambitions for China’s targets for installed nuclear capacity have fluctuated in the past. While construction starts have picked up speed again in 2021–2022, Chinese medium-term ambitions appear significantly lower than anticipated in the pre-3/11 era.44

1. Figure 24 | The PLEX Projection (not including LTOs)

Sources: Various, compiled by WNISR, 2023

Notes: see Figure 23.

Every year, WNISR also models a scenario in which all currently licensed lifetime extensions and license renewals are maintained, and all construction sites are completed. For all other units, we have maintained a 40-year lifetime projection, unless a firm earlier or later closure date has been announced. By the end of 2023, the net number of operating reactors and operating capacity would remain almost stable (+ 1 unit / - 0.3 GW).

In the remaining years to 2030, the net balance would turn negative as soon as 2024, and slightly positive for the years 2026–2027 but overall, an additional 88 new reactors (66.5 GW)—almost one unit or 0.7 GW per month—would have to start up or restart to replace closures.

The PLEX-Projection would still mean for the remaining years to 2030, a need to almost double the annual startup rate of the past decade from six to eleven units (see Figure 23, Figure 24 and the cumulated effect in Figure 25).

However, as documented in detail above, construction starts have not been picking up over the past decade. Between 2013 and 2017, a total of 29 constructions were launched around the world, of which 12 in China and two later abandoned in the U.S. Between 2018 and 2022, constructions started at 36 units, of which 19 in China, thus an average of 6.5 units per year were launched and sustained so far, hardly an increase over the past and hardly more than half of the startup rate needed according to the PLEX Projection over the remaining years to 2030 just to maintain the current number of operating reactors in the world.

1. Figure 25 | Forty-Year Lifetime Projection versus PLEX Projection

Sources: Various, compiled by WNISR, 2023

Notes: This figure illustrates the trends, and the projected composition of the current world nuclear fleet, taking into account existing reactors (operating and in LTO) and their closure dates (40-years Lifetime vs authorized Lifetime Extension) as well as the 58 reactors under construction as of 1 July 2023.

The graph does not represent a forecasting of the world nuclear fleet over the next three decades as it does not speculate about future constructions.

This figure takes into account the restarts of Bruce-6, Darlington-3, Penly-1, Takahama-1 &-2 during the second half-year of 2023.

Further detail, see Figure 23.

Focus Countries

Belgium Focus

After a decade of ups and downs due to multiple technical issues and a record nuclear production of 48 TWh in 2021, nuclear generation dropped by 13 percent in 2022 to 41.7 TWh.

In 2022, Belgium operated seven pressurized water reactors (PWRs) on the Tihange and Doel sites that contributed 46.4 percent of Belgium’s electricity, a 4.4 percentage-point drop over 2021. The historic maximum nuclear share was 67.2 percent in 1986.

In the framework of the Belgian nuclear phaseout legislation, the nuclear operator closed Doel-3 on 23 September 2022 and Tihange-2 on 31 January 2023. The average age of the Belgian fleet is 44.2 years.

Belgium remains highly dependent on fossil fuels as contributions to final energy consumption in 2022 represented 47.2 percent for oil, 24.6 percent of natural gas (together 71,8 percent) with nuclear at 8.4 percent and renewables at only 7 percent.45

The gas-price increase in the fall of 2021 and the war in Ukraine have reopened the debate about the possibility of lifetime extension of the two most recent units, Tihange-3 and Doel-4, and the government has introduced corresponding preliminary legislative proposal on 1 April 2022. However, as of mid-October 2023, no new legislation had been approved, there is no final binding contractual agreement between the Government and the operator, while there is no longer a debate about potential lifetime extensions of the remaining three of the seven Belgian reactors beyond the closure schedule specified by current law.

Legally the country remains bound to a nuclear phase-out target of 2025. In January 2003, legislation was passed that requires the closure of all of Belgium’s nuclear plants after 40 years of operation, so based on their startup dates, plants would have been closed progressively between 2015 and 2025 (see Table 4). Practically, however, after lifetime extension to 50 years was granted for three reactors, five of the seven reactors would have gone offline in the single year of 2025. The planned buildup of alternative power generation capacity had not taken into account the energy crisis and following constraints on the natural gas market. The lifetime extension option gained momentum, long and complex negotiations followed.

1. Table 4 | Belgian Nuclear Fleet (as of 1 July 2023)

Sources: Belgian Law of 28 June 201546, WNISR various.

Lifetime Extension of Tihange-3 and Doel-4?

Operator Electrabel, a subsidiary of French energy group Engie, had previously signaled that it was interested in extending the lifetime of two or three units beyond 2025 but warned that it would need legislation to be adapted by the end of the year 2020.47 This did not happen and Engie decided “to stop preparation works that would allow for the 20-year extension of two nuclear units beyond 2025”.48

In July 2022, the Belgian government inquired whether Tihange-2, slated for closure on 1 February 2023, could be kept operating until the end of March 2023. Engie stated that a lifetime extension of Tihange-2 “had never been on the table” and that on such short notice, without any preparatory work having been done, “it is not possible due to both technical and nuclear safety constraints”.49 In another statement Engie explained that any lifetime extension of Tihange-2 was “not an option” and pointed out that “taking into account the concrete situation, considering such a scenario in haste, without the necessary preliminary studies having been carried out, is not possible with regard to the imperatives of nuclear safety (...)”.50 Accordingly, Tihange-2 was closed on 31 January 2023.

In the fall of 2021, pressure increased to reassess the potential lifetime extension of Tihange-3 and Doel-4, and in January 2022, the Federal Agency for Nuclear Control (FANC) issued a report commissioned by the government concluding a lifetime extension “would be possible from a nuclear safety point of view but only if the facilities were updated”.51

On 16 July 2022, Tinne Van der Straeten, Minister for Energy, stated in an interview: “The biggest concern is France, which is experiencing the largest unavailability of its nuclear fleet in its history. (…) We are not sure we will be able to import as much electricity as expected from France.” Belgium has been however a net power exporter over the year since 2019. The minister confirmed that the operation of Doel-3, slated for closure by 1 October 2022, could not be extended due to a lack of fuel.52

On 22 July 2022, the government signed a “non-binding declaration of intent” with Engie to “evaluate the feasibility and the conditions of a [license] renewal of the two most recent reactors”, Tihange-3 and Doel-4, for a 10-year period starting in November 2026. Engie, that had reoriented corporate strategy away from nuclear, is requesting stiff conditions for a deal. While Engie would remain the operator, the Belgian state would enter a joint company and provide half of the capital. In addition, decommissioning and waste management costs—for all seven reactors—should be determined in a study and would then be capped.53 A final agreement was to be negotiated by the end of the year 2022. That did not happen. Instead, on 9 January 2023, the government—represented by the Prime Minister and the Green Party Energy Minister—jointly announced the signature of a “Heads of Terms and Commencement of LTO [Long-Term Operation] Studies Agreement” with Engie, stating that

This agreement in principle constitutes an important step, and paves the way for the conclusion of full agreements in the upcoming months. It also provides for the immediate start of environmental and technical studies prior to obtaining the authorizations related to this extension. (…)

With this agreement, both parties confirm their objective to make reasonable endeavours to restart the Doel 4 and Tihange 3 nuclear units in November 2026.54

Green-Party Co-President Rajae Maouane commented : “I’m part of this new generation of environmentalists for whom nuclear power is no longer a taboo.”55

Between 20 March and 20 June 2023, the Belgian government held a transboundary public consultation on the basis of the “Environmental Impact Assessment in the context of postponing the deactivation of the Doel 4 and Tihange 3 nuclear power plants”.56

According to ENGIE, the intermediate agreement signed with the Belgian government on 29 June 2023, only nine days after the end of the public consultation, contains the following key points:57

• “The commitment from both parties to use their best efforts to restart the nuclear units of Doel 4 and Tihange 3 as early as November 2026, or, subject to the effective implementation of an announced relaxation of regulations, as early as November 2025, with the aim to strengthen the security of supply in Belgium.”
• The Doel-4 and Tihange-3 reactors will be co-owned in a 50-50 percent partnership.
• The remuneration will be based on a Contract for Difference model.
• ENGIE will pay a lump sum of €15 billion (US$16 billion) for “the future costs of nuclear waste management” of all seven of ENGIE’s nuclear reactors in Belgium. The amount is to be paid in two instalments, one at closing in the first semester 2024 for intermediate- and high-level nuclear waste, and a second payment in 2026 for low-level waste.
• Electrabel has already ordered fuel and the nuclear regulator has determined the scope of inspections and work to be carried out for the operation of ten additional years.

That agreement was followed by another “intermediate agreement” signed on 21 July 202358 and to be followed by the final, legally binding agreement by the end of October 2023 (but had not been announced as of 31 October 2023), which then must be approved by the European Commission. Closure of the deal is expected in the first half of 2024.59

On 20 July 2023, the Federal Agency for Nuclear Control (FANC) communicated its expectations to ENGIE Electrabel to allow for the lifetime extensions beyond 2025. The regulator proposes to stagger upgrading work to 2028 to allow for the two reactors to be available during the winters 2025–2026 and 2026–2027. ENGIE Electrabel now has to come up with concrete proposals on how and by when to implement the requested upgrading work.60

Many technical and legal challenges remain to be solved prior to the operation of Doel-4 and Tihange-3 beyond 2025. In February 2023, ENGIE has ruled out the lifetime extension of the three other remaining operating reactors Doel-1 and -2, and Tihange-1 calling the option “unthinkable”.61 In March 2023, FANC ruled out the prolongation option for the three units on safety grounds.62

Previous Lifetime Extensions

In summer 2012, the operator identified an unprecedented number of hydrogen-induced crack indications in the pressure vessels of Doel-3 and Tihange-2, with respectively over 8,000 and 2,000 previously undetected defects, which later increased to over 13,000 and over 3,000. In spite of widespread concerns, and although no failsafe explanation about the negative initial test results was given, on 17 November 2015, FANC authorized the restart of Doel-3 and Tihange-2 (see previous WNISR editions for details).

The Belgian government did not wait for the outcome of the Doel-3/Tihange-2 issue and decided in March 2015 to draft legislation to extend the lifetime of Doel-1 and Doel-2 by ten years to 2025. The law went into effect on 6 July 2015. On 22 December 2015, FANC authorized the lifetime extension and restart of Doel-1 and -2.63

On 6 January 2016, two Belgian NGOs filed a complaint against the 28 June 2015 law with the Belgian Constitutional Court, arguing in particular that the lifetime extension had been authorized without a legally required public enquiry. Following a 22 June 2017 pre-ruling decision, the Court addressed a series of questions to the European Court of Justice (ECJ), in particular concerning the interpretation of the Espoo and Aarhus Conventions, as well as the European legislation. 64

On 29 July 2019, the ECJ stated that the lifetime extension of a reactor

must be regarded as being of a comparable scale, in terms of risks of environmental impact, to the initial commissioning of those power stations. Consequently, it is mandatory for such a project to be the subject of an environmental impact assessment provided for by the EIA directive.65

In addition, as the Doel-1 and -2 reactors are particularly close to the Belgian-Dutch border, “such a project must also be subject to the transboundary assessment procedure”. The judgement permitted to delay the implementation of the order, if a national court considers it is

justified by overriding considerations relating to the need to exclude a genuine and serious threat of interruption to the electricity supply in the Member State concerned, which cannot be addressed by other means or alternatives, inter alia in the context of the internal market. Considers [t]hat maintenance may only last for the amount of time strictly necessary in order to remedy that illegality.66

On 5 March 2020, the Belgian Constitutional Court nullified the lifetime extension legislation in its entirety but gave the government until the end of 2022 “at the latest” to carry out an appropriate Environmental Impact Assessment (EIA) and a transboundary consultation.67

The Belgian government argued that the lifetime extension “plays a vital role in securing its supply of electricity until 2025” and sent a notification for consultation to a number of European governments inviting them to comment on the “project” (that is the well engaged lifetime extension of Doel-1 and -2).68

The Belgian precedent has significant consequences on lifetime extension projects in European Union Member States that now will all have to carry out full-scale EIAs and organize transboundary consultations prior to granting permission for lifetime extensions.

National Energy and Climate Policy

The National Energy and Climate Plan (Plan National Énergie-Climat or PNEC) was passed in late 2019 and defines the strategy of compensation for the 6 GW of nuclear power that would have been closed by the end of 2025. A capacity market shall attract the necessary investments into other generation capacity and flexibility options. The renewable energy target is set at 40 percent by 2030. The interconnection with neighboring countries, already on a high level, will be further improved.69

Part of the nuclear phase-out strategy was the buildup of offshore wind capacities. In 2020, Belgium reached 2.3 GW installed capacity.70 Offshore wind development shall continue with the designation of a second zone in the North Sea that will see the first turbines connected to the grid in 2027–2028 and ultimately add 3.1–3.5 GW to the national fleet.71

In 2022, a year with exceptionally low wind speeds, offshore wind farms generated 6.7 TWh (gross) compared to 5.3 TWh (gross) for onshore turbines, providing together about as much energy as in 2021 and just over half of the renewable contribution to overall electricity production. Solar electricity generation increased by a remarkable 25.7 percent to 7.1 TWh.

Cumulated installed generating capacity of wind and solar reached 11.7 GW or just over 45 percent of total electricity. All renewable energies combined generated 23.7 TWh (gross) or 24.9 percent, more than natural gas plants with 22.2 TWh (gross) or 23.4 percent.72

Brazil Focus

Brazil’s two commercial nuclear reactors—Angra-1 and -2—are operated by state-controlled company Eletronuclear at the Central Nuclear Almirante Alvaro Alberto (CNAAA) site and provided the country with a stable 13.7 TWh or 2.5 percent of its electricity in 2022.73 According to Eletronuclear, Angra-1 achieved the highest monthly electricity output of its operational history in January 2023.74

After being suspended in 2015, construction of a third reactor at CNAAA resumed in November 2022. The works were interrupted again in April 2023 due to a dispute with local government, causing a costly delay and threatening the projects viability. The quarrel was seemingly settled over the summer, but a clear path forward is not yet guaranteed, as a financing model and an Engineering, Procurement and Construction (EPC) contract are yet to be approved. The Ministry of Energy had previously indicated that a decision on the future of Angra-3 is expected by the end of the year. Various estimates of the remaining investment requirements are all around BRL20 billion (~US$4 billion), and the latest disclosed commissioning target is 2029.

Brazil is expanding its uranium enrichment capacities and expects to manufacture the entire fuel supply requirements of its then three reactors by 2037. The deployment of further nuclear capacity has long been on the agenda of successive governments, but no definite newbuild plans have been revealed by the previous or the current administration of President Lula da Silva. Over the years, as Angra-3 sunk in turmoil such ambitions had gradually been relegated further into the future, but lobbying efforts continue.

The first contract for constructing a nuclear power plant, Angra-1, was awarded to Westinghouse in 1970. The 609-MW PWR eventually went critical in 1981 and is licensed to operate until December 2024. In late 2019, Eletronuclear formally applied for a 20-year lifetime extension with the regulator (CNEN),75 and in October 2020, Westinghouse signed a contract to conduct engineering analyses critical to safety, reliability, and long-term operation as part of the program to extend the working life of Angra-1 until 2044.76 As of September 2020, the process was expected to cost BRL1.2 billion (US$2020230 million).77 In September 2022, Eletronuclear indicated it received the first share of a US$22.3-million loan guaranteed by U.S. Export-Import Bank (EXIM), with a further long-term loan of US$430 million under negotiation.78 The remaining share of the US$22 million-loan was released in December 2022.79

A Pre-“Safety Aspects of Long Term Operation (SALTO)” follow-up mission led in June 2022 by the IAEA, reviewed twenty-one issues that had been identified in 2018 during a previous pre-SALTO mission, and assessed that eleven of these issues were “resolved”, eight were subject to “satisfactory progress” and two had seen “insufficient progress”.80 Overall, the experts concluded that preparation work was progressing “in a timely manner”. A full scope SALTO mission was expected to take place in 2023,81 but is now scheduled for early June 202482. In December 2023, Eletronuclear is expected to submit its third Periodic Safety Reassessment (Reavaliação Periódica de Segurança – RPS) to the safety authority.83

Angra-2 is a large German-designed PWR with a capacity of 1275 MW that was connected to the grid in July 2000, 24 years after construction initially started. A 30-year license set to expire in 2041 was issued in 2011 but Eletronuclear has announced in the past that it will likely request a 20-year extension.84 The company indicated in 2022 that studies were already underway to outline a program for the management of “aging of systems, structures and components at the plant, along the same lines as Angra 1.”85

As reported in WNISR2022, after years of uncertainty, successive setbacks and controversy, in 2022, the Bolsonaro Government finalized the privatization of Eletrobras, the biggest power company in Brazil and, until then, parent entity of Eletronuclear. Requirements for the privatization to succeed included some major restructuring designed to maintain nuclear activities under state control.86 Hence, a new state agency taking over Eletrobras’ activities “that cannot be privatized”—Empresa Brasileira de Participações em Energia Nuclear e Binacional S.A. (ENBpar)—was created by presidential decree on 10 September 2021,87 and announced to be “active” by the responsible Ministry of Mines and Energy, on 4 January 2022.88 In June 2022, corporate control over Eletronuclear was transferred to ENBPar, through capital injection of BRL3.5 billion (US$2022677 million).89

Further institutional changes of recent years include the creation of a new agency to improve the independence of the nuclear regulator. A decree signed by then President Jair Bolsonaro in May 2021 provided for a new regulatory framework and the creation of ANSN (Autoridade Nacional de Segurança Nuclear) which has been reassigned CNEN’s (Comissão Nacional de Energia Nuclear) responsibilities to monitor, regulate and inspect nuclear activities and facilities. CNEN will remain in charge of planning, overall policy, and advocacy for nuclear energy.90 The new allocation and organization was signed into law in October 2021,91 the statutory structure and organization were approved by decree in July 2022,92 but, as of July 2023, ANSN has “not yet started to function” as no “Director-President” has yet been appointed. Consequently, in July 2023, the Joint Budget Committee and Parliament approved an Executive Bill, aimed at opening a special credit line of BRL22.9 million (US$4.7 million) in the 2023 Budget to provide CNEN with the resources considered necessary to carry out ANSN’s duties.93

The Angra-3 Saga

Preparatory work for the construction of Angra-3—a 1405-MW PWR designed by Siemens/KWU—started in 1984. It is unclear how much progress was made before a lengthy interruption starting in 1986. In May 2010, Brazil’s Nuclear Energy Commission issued a construction license, and the IAEA in its Power Reactor Information System (PRIS) recorded that construction (re)started on 1 June 2010.

In early 2011, the Brazilian National Development Bank (BNDES) approved a BRL6.1 billion (US$20113.65 billion) loan for work on the project and in November 2013, Eletronuclear signed a €1.25 billion (US$20131.7 billion) contract with French builder AREVA for the completion of the plant.94

However, a corruption probe led to waves of arrests among plant management, contractors, politicians, heads of state, and senior Eletronuclear executives between 2015 and 2020, and derailed the project altogether (see earlier WNISR editions). In 2015, construction was halted, by 2017 funding had collapsed and the contracts for the construction work were declared void.95 In August 2017, an audit by the Federal Court of Accounts (TCU) of Eletronuclear studies which evaluated the necessary investment to resume works at BRL17 billion (US$20175.3 billion), noted that “… the increase will have a significant impact on the sale price of the energy to be produced and, consequently, on the viability of the enterprise.”96

In September 2018, TCU lifted its recommendation to suspend the program due to irregularities97, and shortly after, the reference value for the price of power from Angra-3 was more than doubled compared to the 2016-value. However, no partner was found to invest in the endeavor, so that in June 2020, the Bolsonaro Government approved plans for carrying out the project, “with or without a partner joining Eletronuclear.” That was despite the ongoing corruption investigation, and Eletronuclear’s various statements at the time that an additional BRL14.5–15 billion (US$20202.8–2.9 billion) of investment would be needed to complete the unit.98 Altogether, at that stage, the unit was said to be 62.8 percent complete, while 80 percent of the equipment was reportedly bought and stored, costing about BRL 25 million (US$4.6–4.8 million) per year in “upkeep and insurance”.99

In March 2021, Eletrobras approved a “Critical Path Acceleration Plan” to complete Angra-3 by 2023 and reach commercial operation by the end of 2026.100 At that time, Leonardo Mendes Cabral, director of privatizations at BNDES, said he expected a financing arrangement to be ready by the end of 2022. The Brazilian Government and Eletrobras had hired BNDES to develop the project, with an estimated additional cost of US$3–4 billion.101 In turn, BNDES released a statement in June 2021 indicating that they had hired Angra Eurobras NES—a consortium composed of Belgium’s Tractebel Engineering SA, Spanish engineering firm Empresarios Agrupados Internacional SA, and led by Tractebel Engineering Ltd. (a subsidiary of French energy company Engie)—to structure the project going forward. This includes identifying the remaining work needed and the means to contract construction companies, providing investment estimates, and accordingly outline a schedule to complete construction.102

In October 2021, ahead of the privatization of Eletrobras, the guidelines for pricing of Angra-3 were approved, clarifying that prices of electricity from Angra-3 would be based on BNDES calculations, taking into account “the economic and financial viability of the project” and “its financeability under market conditions”.103

Meanwhile, in February 2021, Eletronuclear had launched a tender with the intention to hire a contractor in the second half of 2022 for civil works and electromechanical assembly with the expectation that the unit—which was now said to be 65 percent complete—would enter commercial operation in November 2026.104 In July 2021, Eletronuclear announced that a consortium, made up of Ferreira Guedes, Matricial and ADtranz, had won the tender with a winning bid of BRL292 million (US$202154.1 million).105 In February 2022, a contract was signed with the consortium.106

An Angra Eurobras NES-presentation dated May 2022 indicated that on-site construction was planned to resume in the third quarter 2022. The document enclosed a provisional schedule which projected commissioning of the unit in December 2026 and commercial operation in February 2028.107 At the time, an additional BRL19.4 billion (US$20223.8 billion) was said to be needed to complete the project.108

In June 2022, the privatization of Eletrobras occurred, bringing construction of Angra-3 one step closer to resumption, as it was said to be crucial to the completion of the project.109

In September 2022, Angra-3’s environmental license was renewed for six years by the Brazilian Institute for Environment and Renewable Natural Resources (Ibama).110 And finally, on 11 November 2022, Eletronuclear announced the “resumption of concrete pouring”, marking the official restart of construction.111

A few days later, Tractebel announced in more cautious terms that Angra Eurobras NES had finalized “the first stage of the project that will enable to resume the construction”. The consortium delivered an “Engineering, Procurement and Construction [EPC] Contract Specification Report” with the promise that it will “enable BNDES to elaborate the modeling and will provide reliable data for the economic and financial assessment, the fund-raising process, and for the elaboration of the final EPC contract. It is crucial as it will mitigate the project’s risks.”112 Modelling by BNDES would then have to gain approval from Eletronuclear and be reviewed by the Ministry of Mining and Energy and TCU, before a final EPC agreement can be contracted.

As of December 2022, the Angra-3 project—with admirable precision—was said to be 66.97 percent complete with an expected operation date of July 2028.113 However, on 19 April 2023, the City Government of Angra dos Reis ordered the halt of work on the grounds that the project as implemented differed from the initially approved plans. The city indicated that they would grant a new construction permit upon review and approval of the changes, and once Eletronuclear honors its 2009-commitment to a socio-environmental compensation equivalent to BRL264 million (US$54.5 million) in 2023-value.114

It appears noteworthy that the dispute builds on recent tensions between the company and local government. In March 2023, a Public Civil Action was filed against Eletronuclear over an incident that occurred on 16 September 2022 during Angra-1’s refueling outage which the operator failed to disclose to regulatory agencies. According to available information, on 29 September 2022, Ibama was alerted of a contaminated water discharge that led to a joint inspection and continued monitoring with CNEN, which concluded that the measured levels “did not pose any risk to the population and the environment.” In February 2023, Ibama fined Eletronuclear over BRL2 million (~US$392,000) for illegal disposal of contaminated water, and BRL101,000 (~US$19,800) for neglecting to promptly alert the regulator of the event. The following month, the Public Prosecutor filed a public civil action against Eletronuclear,115 which prompted a police search on-site in May 2023 and debates and hearings in Parliament.116

Eletronuclear firmly rejected the allegations of non-compliance with the 2009-agreement and tried to lift the suspension117 while establishing legal action as an option should the administrative proceedings and dialogue attempts fail.118 As of early June 2023, the dialogue on compensation funding seemed to reach some progress119, and a month later, Eletronuclear indicated it was reviewing projects submitted by the municipality to assess if these were eligible to receive parts of the funds earmarked towards socio-environmental compensation.120 A few days later, it was announced that an agreement had been outlined under which Eletronuclear would distribute more than BRL300 million (~US$62 million) in five settlements to three municipalities neighboring CNAAA until 2027, including the BRL264 million for Angra dos Reis.121 Early in the negotiations, in May 2023, Eletronuclear CEO Eduardo Grivot had indicated “I signed the commitment of R$264 million, which was presented by the city council, but I won’t have the money to pay it all.” The new accord was to be signed by early August 2023; however, no precise information was disclosed concerning the reissuance of construction permits.122

The delays come at great cost that could “threaten the financial viability” of the project, as it could force Eletronuclear to repay debts and loan obligation of BRL 6.2 billion (US$1.25 billion) prior to commissioning.123 This could also adversely impact stakeholder Eletrobras that states in financial documents filed in April 2023 that “We may incur substantial financial liabilities as well as unexpected expenses until we complete the construction of the Angra 3 nuclear power plant.”124

Governmental support remains crucial. President Luiz Inácio Lula da Silva backed the “relaunch” of Angra-3 during his previous presidency (2003–2010),125 so the support of his administration upon taking office in January 2023 did not only appear guaranteed but was also reaffirmed on several instances. Notably during a parliamentary commission hearing held in early May 2023. On that occasion, Secretary of Electricity Gentil Nogueira de Sá Junior also disclosed that commissioning would not occur before 2029 and that abandoning the project would cost about BRL13.6 billion (US$2.7 billion), while the funding options of the remaining investment required—amounting to BRL20 billion (US$4 billion)—were still under BNDES review. According to the Secretary of Electricity’s presentation before Parliament, the revised cost of the project increased to BRL27.8 billion (US$5.5 billion).126

ENBPar had earlier hinted towards even higher costs in its Annual Report for 2022, when it referred to an “ongoing due diligence report”—seemingly quoting from Angra Eurobras NES’ review—which estimates the remaining investment needed at BRL21 billion (US$4.3 billion).127

Eletronuclear’s Annual Report 2022 noted that a bidding notice for EPC was expected by the end of 2023, and contract signature in the first trimester of 2024.128 In any way, as then-president of Eletronuclear Leonam Guimaraes summarized in May 2020, “It is much easier to attract partners with a project that is under way than with one that is paralyzed.”129

In late June 2023, Energy Minister Alexandre Silveira de Oliveira had stated that the decision on whether to restart this “big challenge” was still pending, with a final ruling expected by year’s end.130

The matter has become increasingly sensitive to the administration, whose indecisiveness is reflected in its new “Growth Acceleration Program” or PAC (Programa de Aceleração do Crescimento) released in August 2023.131 The nationwide program maps BRL1.7 trillion (US$360 billion) of public and private investment towards a wide range of sectors, such as urbanization, health, education, or culture, until 2026. Of the BRL75.7 billion (US$16 billion) allocated to power generation, just BRL1.9 billion (US$402 million) in state funds are allocated to new nuclear capacity. However, the plan only lists the modernization of Angra-1 as explicit recipient. Angra-3 is not considered an ongoing project and is solely referenced regarding its “technical, economic and socio-environmental feasibility study”.132 Reports indicate that it could still be included in an updated version of PAC, once financing and contracting models are approved.133 On the matter, Minister of Energy, Silveira was quoted as saying “We need to have economic security that the energy that Angra 3 will supply… will also be economical for the consumer, because it is the consumer who pays the energy bill.”134

So far, broader political support for the project and further newbuild seems relatively strong. A joint parliamentary group—composed of 217 elected representatives of the Chamber of Deputies and the Senate (of a total of 513 Deputies and 81 Senators)—created earlier in the year to promote new nuclear projects,135 has been “working to show the government how important, necessary and strategic it is to restart the work on Angra 3 with the utmost urgency” according to its initiator and President, Júlio Lopes (Partido Progressistas).136

Expanding Brazil’s nuclear capacity beyond Angra-3 has been a clear aspiration of the previous administration for the longer term. In November 2021, at COP26, then Minister of Mines and Energy Bento Costa Lima said the country would add 10 GW of nuclear power over the next 30 years,137 as envisaged by the “National Energy Plan to 2050” or PNE 2050 (Plano Nacional de Energia 2050) and amended by the Government in December 2020.138

However, short-term projections remain limited. In January 2022, the Ministry of Mines and Energy published its “Ten-Year Energy Expansion Plan” or PDE 2031 (Plano Decenal de Expansão de Energia 2031), which unveiled a plan to commission a new 1 GW unit by 2031, bringing nuclear power’s share in the national electricity production to 4 percent for 33 TWh of generated power.139 In its final months, the Bolsonaro administration issued the PDE 2032, which projects 1.4 GW of new capacity derived from nuclear over the next decade.140

A few known steps were taken in 2022 to further expand nuclear capacity. The Bolsonaro Government released a statement in March 2022 indicating that it has signed a cooperation agreement with the Electric Energy Research Center (Cepel) to identify appropriate sites for new nuclear plants.141 No locations were named, although in May 2023, the Municipality of Angra dos Reis mentioned in a statement that “the federal government has already announced its intention to build a fourth nuclear power plant in the city.”142 In 2022, interest towards Small Modular Reactors (SMRs) translated into various preliminary governmental and industrial cooperation agreements with Russia and France.143 That ambition also has a voice in parliament through Julio Lopes who is championing the examination of building an SMR at Angra.144

As of July 2023, PEN 2050 and PDE 2032 had not been updated, and it is not entirely clear if and how the incoming Government of President Lula da Silva will revise or implement the current targets.145 It is not clear either, which administration—past or present, or both—has expressed the ambition of a fourth unit at Angra, as disclosed by local officials (see above). Historically, during Lula’s second term, his administration intended to build four reactors starting in 2015, and Eletronuclear had the confidence to plan the construction of six reactors adding 8 GW of nuclear capacity by 2030.146 However, these targets have long slipped away, and while there are clear efforts to keep the option on the table, the overall prospects of nuclear newbuild in Brazil is likely bound to the increasingly uncertain fate of Angra-3.

Expansion of Uranium Enrichment Capacities and Nuclear Fuel Diversification

In November 2022, Indústrias Nucleares do Brasil (INB) inaugurated the tenth cascade of ultracentrifuges for uranium enrichment at its fuel manufacturing facility (Fábrica de Combustível Nuclear – FCN) in Resende, Rio de Janeiro. The expansion of its uranium enrichment capacities deems INB capable of covering 70 percent of the yearly fuel supply necessary to operate Angra-1. Brazil expects to provide the entirety of fuel required by Angra-1 and -2 by 2033, and be completely “self-sufficient” by 2037, though this only entails the needs of the two operating Angra units plus Angra-3, not of further potential future units.147

For now, Brazil relies on nuclear fuel imports, and in December 2022, INB and a Rosatom subsidiary signed their first long-term contract for the fuel supply of Angra-1 and -2, from 2023 to 2027.148

The Russian nuclear industry remains a regular supplier to its Brazilian counterpart. On 13 March 2023, Rosatom announced that its subsidiary TVEL had won the tender for the supply of more than 100 kg of lithium-7 hydroxide for the reactor cooling system of Unit 1 and 2, indicating that contract signature and shipment was expected to occur before the end of the year.149 However, Eletronuclear indicated that due to Russia’s invasion of Ukraine, it had encountered difficulties in acquiring the product, prompting the company to seek to diversify its supply.150 A contract was signed in May 2023 with Rosatom’s Tenex, during the “Nuclear Trade & Technology Exchange” conference for the supply of natural uranium hexafluoride (UF6), after the Russian corporation won a tender for the supply of 330 tons of UF6 in 2022.151

Strong Expansion of Renewable Energy Generation

Meanwhile, according to the Energy Institute, the share of fossil fuels in the country’s electricity generation mix dropped by close to half in one year, falling from 20 percent in 2021, to 10.1 percent in 2022. Their production decreased by over half for natural gas (from 87 TWh to 42.1 TWh), by half for oil (from 20.2 TWh to 10.1 TWh) and by 30 percent for coal (24.2 to 16.5 TWh). The output from non-hydro renewable sources grew by 10 percent (from 144.8 TWh to 164.5 TWh) and that of hydro by 17.7 percent (from 362.8 TWh to 427.1 TWh), resulting in a contribution of 24.3 percent (154.6 TWh) from non-hydro renewables and a remarkable 87.3 percent (591.6 TWh) of renewables including hydroelectricity. Nuclear generation remained stable at 14.6 TWh in 2022 (compared to 14.7 TWh in 2021), for a 2.2 percent contribution.

The government indicates that 80 percent of the additional power covered by PAC will be low-carbon, of which 79 percent will originate from renewable sources.152

China Focus

As of mid-2023, China had 56 reactors in operation with a total capacity of around 53 GW. The count of 56 is slightly different from the IAEA’s count of 55 in its PRIS database because WNISR records the Shidao Bay as twin High-Temperature Reactor Pebble-bed Modules (HTR-PM) with two reactors of 100 MW each. For unknown reasons, the China Experimental Fast Reactor (CEFR) is no longer mentioned in the PRIS database since May 2023, and has been placed in LTO as of this date in WNISR statistics. With 23 reactors under construction, China continues to be the global leader in hosting nuclear newbuild projects.

Nuclear plants produced 395.4 TWh in 2022, marginally higher (+3.2 percent) than the 383.2 TWh generated in 2021. The electricity generated was 5 percent of the total electricity produced in 2022, the same as in 2021. In comparison, the 2023 “Statistical Review of World Energy” records nuclear power’s share of total electricity produced (gross) as 4.7 percent, again the same as 2021.

Since the publication of WNISR2022, only two nuclear reactors have started operating: Fangchenggang-3, a 1000-MW Hualong One, became critical on 27 December 2022, was connected to the grid on 10 January 2023, and was declared as operating commercially on 25 March 2023.153 The reactor’s first pour of concrete was on 24 December 2015, which represents a construction period of 84.5 months.

At the Shidao Bay HTR-PM plant, grid connection of the second of the twin reactors has not been announced. While the production of the plant is not reported, WNISR nevertheless considers both modules to be operating since the end of 2022. According to a report in World Nuclear News (WNN), the plant “achieved the initial full-power operation of the dual reactors and ‘tested the operation control capability’ of it in ‘two reactors with one machine’ mode”, which suggests that both reactors were operational, and the “first reactor reached first criticality in September 2021 and the second one that November. The connection of the first of the unit’s twin reactors took place in December 2021.”154

China has imported reactor technologies from Canada, France, Russia, the U.S. and from a U.S.-Japanese consortium (Westinghouse/Mitsubishi Heavy Industries). The first foreign unit, Daya Bay-1 designed by Framatome, started building in July 1987, the latest one, Xudabu-4 a Russian VVER-1200, started construction in May 2022.

It is interesting to assess the construction durations of the 57 units connected to the Chinese grid between 1991 and July 2023. The 41 reactors of Chinese or Sinicized design had an average construction time of 5.7 years with a range from 4.1 to 10 years, while it took on average respectively only 4.5 years for two Canadian CANDUs, but 6.6 years for six French units (4.4-9.2 years), 6.9 years for four Russian reactors (5–11.2 years), 8.6 years for two U.S. AP-1000s, and 9 years for two AP-1000s built by a U.S.-Japanese consortium (see Figure 26).

1. Figure 26 | Construction Times of Reactors Built in China

Sources: WNISR with IAEA-PRIS, 2023

China has a further 23 reactors under construction, with a combined capacity of around 24.5 GW (see also Annex 3 – Table 29 • “Nuclear Reactors in the World Under Construction”):

• The two CAP1400 reactors, Shidao Bay 2-1 and Shidao Bay 2-2, (since 2019) which are not listed in the IAEA’s PRIS database.
• Four units started construction since WNISR2022: Haiyang-3 (7 July 2022), Lufeng-5 (8 September 2022), Sanmen-4 (22 March 2023) and Haiyang-4 (22 April 2023).155
• Other light water reactors being built are Fangchenggang-4 (since 2016); Zhangzhou-1, Taipingling-1; Taipingling-2, Sanaocun-1, and Zhangzhou-2 (since 2020); Changjiang-3 and -4, Sanaocun-2, Tianwan-7, and Xudabu-3 (since 2021); Tianwan-8, Xudabu-4 and Sanmen-3 (since 2022).
• The Xiapu two fast reactor units started being built on 29 December 2017 and 27 December 2021 respectively.156
• The SMR Changjiang (or Linglong-1) is under construction since 2021.
• The only reactor construction that is currently officially past the deadline for starting is Fangchenggang-4, an HPR-1000 or Hualong One which was originally scheduled to start operating in 2022 and is now scheduled to be connected to the grid in the first half of 2024.157

Chinese government authorities have plans for many more. In May 2023, the Ministry of Ecology and Environment “approved in principle” the Environmental Impact Reports for two Hualong One units at the Fangchenggang site and two CAP-1000 units at the new Bailong site, around 30 km away from the Fangchenggang site.158 These have so far not been approved by the State Council. Plans for the CAP1000 units go back to at least 2015 when a report produced in part by the U.S. Department of Commerce listed them as “nearer-term planned”.159

China’s ambitions include exporting nuclear power plants all over the world. In 2016, the president of China National Nuclear Corporation (CNNC) announced that “China aims to build 30 overseas nuclear power units… by 2030”.160 As described in Annex 1 (see section on Pakistan), China has exported several reactors to that country and is continuing to do so. But so far there has been no other country that has imported a nuclear power plant from China, possibly because of the United States blacklisting Chinese nuclear firms in 2019, accusing them of helping acquire U.S. technology for military use.161 Also, in 2019, the U.S. Department of Commerce added China General Nuclear Power Group (CGN) to its “entity list”, as a result of which U.S. companies cannot sell “products and services to the firm without written approval”.162

Therefore, the February 2022 agreement signed by CNNC and Nucleoeléctrica Argentina SA (NA-SA) to build Atucha-3 was seen as an important beginning.163 But, as NA-SA President Jose Luis Antunez clarified in an interview with Nuclear Intelligence Weekly in early 2022, the agreement to execute the project required “precedent conditions” to be met, including CNNC “transferring the technology for fabricating the metallic component of the fuel in Argentina”.164

Argentina’s demand that it be allowed to “manufacture the reactor fuel” is reportedly becoming an obstacle.165 The president of Argentina’s National Atomic Energy Commission has told the press: “We are trying to establish the best conditions to transfer the knowledge for making the fuel”.166 The growing trade deficit between Argentina and China is also becoming a problem, especially given the economic challenges Argentina is going through, and the Atucha-3 project has reportedly “hit a stumbling block over finances”.167 (See Annex 1 – section on Argentina.)

Renewable sources (not including large hydropower) produced 15.4 percent of the total electricity, over three times the contribution from nuclear power plants. Electricity produced by renewable sources increased by 19 percent in 2022168 (see also Case Study on China in Nuclear Power vs. Renewable Energy Deployment).

China’s renewable energy capacity continues to grow very rapidly. In June 2023, the official English-language communication platform of China’s State Council announced that the country’s installed capacity of non-fossil energy power generation now accounts for 50.9 percent of the total capacity.169 The China Electricity Council reports an installed solar capacity of 392.6 GW and installed wind capacity of 365.4 GW as of the end of 2022, an annual increase of 11.2 percent and 28.1 percent respectively.170 The trend is accelerating. The installed capacity of solar projects that came online in the first quarter of 2023 was 155 percent above the same period in the previous year, with related investments going up 178 percent.171

1. Figure 27 | Age Distribution of the Chinese Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2023

France Focus

Overview

WNISR2022 pointed out that “2020 was considered ‘particularly difficult for the French nuclear sector’, but 2022 is likely to be significantly worse”. It did turn out much worse, disastrous in fact, an “annus horribilis”, according to nuclear utility EDF’s Executive Director of Generation and Engineering of the Existing Nuclear and Thermal Fleet.172 Nuclear output dropped below the level of 1990 when the installed nuclear capacity was some 5 GW lower. Nuclear generation actually peaked in 2005 at over 430 TWh and in nine of the following ten years, output exceeded 400 TWh, which was considered the norm until 2015. In 2022, French reactors produced 279 TWh, a drop of over 120 TWh from the 2005–2015 period.

To put this decline into perspective, it significantly exceeds the loss of 106 TWh of annual nuclear generation between the years 2010 and 2022 in Germany (see Germany Focus) due to the progressive decrease following the phaseout decision in 2011. The drop of over 150 TWh between France’s historic peak nuclear generation of 430 TWh and the 2022-output exceeds the annual average of 148 TWh of total nuclear electricity generated in Germany between 2001 and 2010. Germany’s nuclear generation peaked at 162 TWh in 2001.

While nuclear production had increased by 7.5 percent in 2021 compared to 2020, the discovery in December of that same year of cracks in emergency core cooling systems led to the shutdown of the four largest (1500 MW) and most recent French reactors. The event represented an unexpected loss of 6 GW of capacity in the middle of the winter when consumption peaks in France. More than in any other European country, France has close to one third of the buildings using inefficient electric space heating. The four units did not generate a single kilowatt-hour throughout the year 2022.

Subsequently, it turned out that certain 1300-MW reactors—there are 20 such units—were also showing similar symptoms and, as of mid-2022, 12 reactors were shut down due to the problem. One of them, Penly-1, remained off-grid between October 2021 and July 2023.

Inspection techniques providing reliable results were a challenge. Inspections take time and it took until the end of July 2022 for the Nuclear Safety Authority (ASN) to judge EDF’s inspection strategy “appropriate in the light of the knowledge acquired concerning the phenomenon and the corresponding safety issues”.173 Once defaults are detected, it takes time to fabricate replacement parts, and then do the replacement work. High profile, experienced nuclear welders are rare and there are many competing requirements for these specialists on the French nuclear fleet, including the construction site of the EPR at Flamanville, and there are significant radiation doses involved in the work that could quickly lead to regulatory exposure limits. Additional welders were flown in from Canada and the U.S., while replacement pipes were manufactured in Italy.174 EDF intends to inspect the entire fleet of 56 reactors only by 2025.175

Concerns were growing over the year that a cold winter 2022–2023 could lead to power shortages, and even rolling blackouts were envisaged. For the first time since 1980, France turned into a net importer of electricity (16.7 TWh)176 with Germany playing a key role exporting 15.3 TWh net.177

Following the discovery of the corrosion issue, on 13 January 2022, EDF published a downwards revised forecast for nuclear generation, and the French government announced the same day that it would force EDF to provide its competitors 20 percent more power, at fixed price, than expected—120 TWh instead of 100 TWh—to limit the effect of sky-rocketing market prices for the consumer. The move indeed limited the price increase of the regulated tariff to 4 percent instead of over 40 percent but significantly contributed to EDF’s catastrophic 2022-results with a negative impact estimated at €8.34 billion (US$20238.80 billion).178

As early as July 2022, some estimates put EDF’s expected net debt as high as €65 billion (US$202268 billion) at year-end,179 and the government announced it would hit the emergency brake and fully re-nationalize EDF. The estimates proved extraordinarily precise, as net debt grew by 50 percent to reach €64.5 billion (US$202267.9 billion) at year-end, and €64.8 billion (US$202370 billion) at mid-2023, according to EDF’s financial results.180

This chapter does not even cover complex fuel chain issues, climate impact, and social movements. The plutonium-economy part of the industry is experiencing its own—underreported—crisis. The throughput of the equally ageing spent fuel reprocessing plant at La Hague dropped to 925 tons in 2022 (for a licensed capacity of 1,700 tons per year), a level last seen in the early 1990s. Consequently, the spent fuel pools are nearing saturation. The project to build a large new cooling pool is encountering fierce local opposition. The uranium-plutonium mixed-oxide (MOX) fuel fabrication facility MELOX at Marcoule plummeted to below 60 tons per year in 2021–2022, that is below 30 percent of its licensed capacity.181 Consequently, the stocks of unirradiated plutonium have increased to the unprecedented level of 92 tons, an increase of spectacular 24 tons since 2018.182

All of these new challenges for an already strained industry did not prevent the National Assembly from picking up on the French President’s landmark “nuclear renaissance” speech of 10 February 2022 and in June 2023 passing legislation for the “acceleration of procedures for the construction of new nuclear facilities near existing nuclear sites and for the operation of existing facilities”.183 The President had expressed his “wish” that “six EPR2 be built and that we launch the studies for the construction of eight additional EPR2”.184

The new law requires the government, prior to tabling legislation on the next pluriannual energy planning, to transmit to Parliament a report that assesses the consequences of the construction of 14 nuclear power reactors on the nuclear industry, the electricity market, and public finances; on nuclear safety and security; on the nuclear fuel chain; and on the means of the Local Information Commissions (CLI). The law simplifies certain administrative procedures, decrees that a nuclear power reactor automatically “constitutes an imperative reason of major public interest” and dilutes some environmental protection rules. For example, a new nuclear power reactor will not be considered in local limitation targets for soil artificialization or consumption of natural, agricultural, or forest areas.185

Currently, the EPR2 does not even exist on the drawing board; no detailed design is available yet. The government administration estimated in an October 2021 internal note that 19 million engineering hours still had to be deployed to get from “basic design” to the “detailed design” stage and that, if everything goes well, the first EPR2 could start up by 2039–2040. In case unexpected industrial difficulties occur—as they did in the past and do currently—it could take until 2043 to commission the first EPR2, the project review states.186

Largely unreported, the science community in France is far from offering unanimous support of the newbuild initiative. As of the end of October 2023, close to 1,200 scientists, doctors, teachers, engineers, academics, and researchers had signed “Call by scientists against a new nuclear program” claiming:

…with neither a real democratic debate, nor a serious assessment of past choices and the options available today, our leaders are preparing to relaunch a program of construction of new nuclear power stations. Under the pretext of the climate emergency, but on the basis of truncated, simplistic, even grossly erroneous arguments, lobbyists with significant media influence are working to organize amnesia of nuclear disasters and revise history. (…)

In the immediate future, the industrial and financial efforts that this new program would require, would for a long time monopolize the financial and human resources necessary to face the combined challenges of the climate crisis, the collapse of biodiversity, generalized pollution and resource depletion.187

In addition to the national initiatives to relaunch the nuclear sector, the French government has been leading a large group of a dozen E.U. countries to collectively lobby the European institutions to create favorable conditions for the nuclear industry in the process of the restructuring of the European electricity market and of the definition of various legislative tools of European climate policy. Much of these negotiations are still ongoing and the outcome will likely be a compromise with a group of countries led by Germany strongly favoring a strategy based on sufficiency, efficiency, and renewable energies.

Another Worst Performance in Decades

Until the closure of the two oldest French units at Fessenheim in the spring of 2020, the French nuclear fleet had remained stable for 20 years, except for the closure of the 250 MW fast breeder Phénix in 2009, two units in Long-Term Outage (LTO) within the period 2015–2017, and another one within the period 2021–2023 (see Figure 28). Penly-1, subject to the stress-corrosion cracking issue, was offline between 2 October 2021 and 13 July 2023.188 While the four units at Civaux and Chooz-B did not generate power throughout 2022, they did not meet the LTO criteria as they were restarted prior to mid-2023.

1. Figure 28 | Operating Fleet and Capacity in France

Sources: WNISR with IAEA-PRIS, 2023

No new reactor has started up since Civaux-2 was connected to the French grid in 1999. The first and only PWR closed prior to Fessenheim was the 300-MW Chooz-A reactor, which was retired in 1991. The other closures were eight first-generation natural-uranium gas-graphite reactors, two fast breeder reactors and a small prototype heavy water reactor (see Figure 29).

1. Figure 29 | Startups and Closures in France

Sources: WNISR, with IAEA-PRIS, 2023

Notes: PWR: Pressurized Water Reactor; GCR: Gas-Cooled Reactor; HGWGCR: Heavy Water Gas Cooled Reactor; FBR: Fast Breeder Reactor.

In 2022, the 56-reactor fleet189—of which one in LTO and four that did not generate any power but did not meet the LTO criteria—produced 279 TWh190, a drop of 22.7 percent over the previous year; nuclear generation was below 300 GW for the first time since 1990, and the seventh year in a row that it remained below 400 TWh. Grid operator RTE stated:

This was the first time since the construction of the existing nuclear fleet was completed that annual output was this low, falling 30% below the average of the prior 20 years. In absolute terms, it is the lowest level on record since 1988, when installed nuclear capacity in France stood at just 51 GW, or 83% of today’s total capacity (eight fewer reactors).191 [bold emphasis in original]

In 2005, nuclear generation peaked at 431.2 TWh. It took the fleet five years to build up to that maximum generation, and with a quasi-stable installed nuclear capacity between late 1999 and early 2020, performance plunged after 2015 (see Figure 30).

1. Figure 30 | Nuclear Electricity Production vs. Installed Capacity in France, 1990–2023

Sources: RTE, 2000–2023, EDF 2023

Note: In Figure 30, reactors in LTO are counted in the “installed capacity”.

In 2022, nuclear plants provided 62.7 percent (–6.3 percentage points) of the country’s electricity, even less than in 2020. According to RTE, the nuclear share peaked in 2005 at 78.3 percent. As of mid-2023, EDF estimates the production range for the year at 300–330 TWh, for 2024 at 315–345 TWh and for 2025 at 335–365 TWh192 (see Figure 30 and Figure 31).

1. Figure 31 | Nuclear Electricity Production vs. Nuclear Share in France, 1990–2023

Sources: RTE, 2000–2023, EDF 2023

Monthly production has continued to deteriorate in early 2023 with a lower output in every month of the first quarter of the year than in any year over the past decade, and while output significantly improved in the second quarter, it remained below the 2021 level (see Figure 32).

Electricity represented 25 percent of final energy in France in 2022. As nuclear plants provided 62.7 percent of electricity, nuclear plants covered 15.7 percent of final energy. The largest share being covered by fossil fuels at over 60 percent, with oil at 42.9 percent and natural gas at 17.4 percent (coal <1 percent), while renewables contributed only 11.1 percent just as in the previous year.193

1. Figure 32 | Monthly Nuclear Electricity Generation, 2012–mid-2023

Sources: RTE and EDF, 2021–2023194

Nuclear Unavailability Review 2022

In 2022, there were 8,515 reactor-days—an increase of 2,704 reactor-days or +46.5 percent compared to 2021—an average of 152 days with zero-production per reactor. This does not include load following or other operational situations with reduced output but above-zero. The number is 58 percent higher than the average 96 days per reactor in pre-COVID year 2019, and 32 percent higher than in 2020 (see Table 5). All 56 reactors were subject to outages lasting four to 365 days (see Figure 35). Five reactors were offline during the whole year. Over half of the French nuclear reactor fleet (29 units) was not available during at least one third of the year, including one third (18 units) that was not available for more than half of the year.

1. Table 5 | Total Unavailability at French Nuclear Reactors, 2019–2022 (in Reactor-Days)

Sources: RTE and EDF REMIT Data, 2019–2023

1. Figure 33 | Reactor Outages in France in 2022

Sources: compiled by WNISR, with RTE and EDF REMIT Data, 2021–2023

Note: For each day in the year, this graph shows the total number of reactors offline, not necessarily simultaneously as all unavailabilities do not overlap, but on the same day.

The unavailability analysis for the year 2022 on Figure 33 further shows:

• On 357 days (98 percent of the year), at least 10 units and up to 34 were down during the same day.
• On 280 days (77 percent of the year), 19 or more units were shut down for at least part of the day.
• At least nine reactors were down (zero capacity) simultaneously at any day of the year.
• At least 20 reactors were offline simultaneously during the equivalent of 273 days.
• On 22 August 2023, a total of 33 reactors, or 59 percent of the fleet, was offline.

1. Figure 34 | Availability of the French Nuclear Fleet Over the Year, 2015–2022

Sources: RTE, 2023195

RTE provides a monthly availability analysis (see Figure 34) with the following comments:

The availability of France’s nuclear fleet was historically low throughout 2022, with a yearly average availability of 54% compared with an average of 73% between 2015 and 2019. 

An all-time low of 21.7 GW was recorded on 28 August 2022, when nearly 65% of the fleet [capacity] was offline. [bold emphasis in original] (…)

The gap with prior years was particularly pronounced during the summer, which saw a concentration of unscheduled outages following the discovery, in late-2021, of stress corrosion cracking in several reactors. These outages, or outage extensions to carry out maintenance, tests and repairs where needed, primarily involved the newest reactors in the fleet (N4 and P4’ designs), i.e. reactors that were not targeted for investment in the Grand Carénage refit programme. These additional outages added to an already busy operational calendar made even busier by the postponements of maintenance caused by the COVID-19 crisis.196

According to EDF’s classification of “planned” and “forced” unavailabilities, in 2022:

• 24 reactors did not experience any “forced” outage,
• at eight units “forced” outages lasted less than one day,
• at 18 their cumulated duration represented between one and ten days,
• and at five reactors “forced” outage cumulated between 18.8 and 47 days over the year (see Figure 35).

1. Figure 35 | Forced and “Planned” Unavailability of Nuclear Reactors in France in 2022

Sources: compiled by WNISR, with RTE and EDF REMIT Data, 2021–2023

Notes: This graph only compiles outages at zero power, thus excluding all other operational periods with reduced capacity >0 MW. Impact of unavailabilities on power production is therefore significantly larger.

“Planned” and “Forced” unavailabilities as declared by EDF.

However, EDF’s declaration of “planned” vs. “forced” outages is highly misleading. EDF considers an outage as “planned” whatever the number and length of extensions (or, in rare cases, reductions) of its total duration if the outage was first declared as “planned”.

Detailed WNISR analysis for earlier years shows a different picture.

“Unplanned unavailability added up to 1,330 days, an increase of 30 percent beyond the expected outage durations.”

The complete assessment of 240 outages in 2021, shows that 161 were declared “planned” and 79 “forced”. In the case of “forced” outages, a generic duration of one day was first declared in most cases (75 percent) and is then readjusted. The additional duration of “forced” outages represented less than 100 days. For “planned” outages, additional unplanned unavailability represented 1,238 days that EDF nevertheless labeled as “planned”. In fact, almost 25 percent of the full-outage durations were unplanned.

Of the 240 full outages, 86 experienced a prolongation exceeding 1 day and up to 156 days (Chooz-2) in 2021197; the cumulated prolongation over the year was over 1,500 days. On the other side, 18 outages were shorter than planned by at least one day; the cumulated reduction over the year was 171 days. (These cases are likely due to outage re-scheduling rather than net savings of outage days.) As a result, the net additional unplanned unavailability added up to 1,330 days, an increase of 30 percent beyond the expected outage durations.

The cumulated outage analysis over the four years 2019–2022 reveals the following (see Figure 36):

• Four reactors were down half of the time or more (Flamanville-1 and -2, Chooz-1 and -2);
• 26 reactors were generating zero power for 30 percent of the time, that is 109 days and more per year on average.
• 39 reactors were off-grid for at least one quarter of the time, in other words, they did not generate any power for the equivalent of one in four years.

1. Figure 36 | Unavailability of a Selection of French Nuclear Reactors, 2019–2022

Sources: compiled by WNISR, with RTE and EDF REMIT Data, 2019–2023

Note: The categorization follows EDF’s classification. However, it is not reflecting reality as a “planned” outage remains in that category even if it lasts much longer than “planned”.

Status of Stress Corrosion Cracking Issue

Severe stress corrosion cracking had been first identified in late 2021 at the safety injection systems of the four largest and most recent French reactors at Chooz and Civaux.198 Later additional reactors were identified and a program of pre-emptive replacement of particularly sensitive piping sections was decided for the “P’4” reactor series. While apparently so far rare, the phenomenon has also been identified on other 1300-MW and some 900-MW reactors (see Table 6 for details). EDF decided to inspect its entire reactor fleet by the end of 2025.

In February 2023, an additional issue has been identified during destructive examination at Penly-1. Close to a weld of a line of the safety injection system that had been repaired during construction of the plant, a 15.5 cm long—about one quarter of the circumference—and up to 2.3 cm deep crack—for a 2.7 cm thick tube—was identified. The origin has been determined as thermal fatigue rather than stress corrosion cracking. This discovery meant that an extensive inspection program of all repaired welds had to be added to the stress corrosion cracking investigations. According to planning, 90 percent of the repaired welds in the safety injection and shutdown cooling systems of the entire reactor fleet are to be inspected until the end of 2024 with the remaining ones in 2025.199

According to EDF, as of mid-2023, 11 of the 16 reactors identified as most sensitive to stress corrosion—the four 1500-MW units and 12 P’4 1300-MW reactors—had been repaired or preemptively treated while two, Cattenom-1 and Bellevillle-2, were undergoing repairs, and two more, Belleville-1 and Nogent-1, were to be fixed before the end of the year. The remaining unit, Cattenom-4, is to be repaired during its fourth 10-year inspection.200

1. Table 6 | Stress Corrosion Cracking - Inspected and Repaired Reactors (as of 30 June 2023)

Source: EDF, 2023201

Note: CP0, CP1, CP2, P4, P’4, and N4 designate identical (or almost) design series of reactors.

Lifetime Extension – Fact Before License

By mid-2023, the average age of the 56 nuclear power reactors exceeds 38 years (see Figure 37). Lifetime extension beyond 40 years—51 operating units are now over 31 years old of which 20 are over 41 years—requires significant additional upgrading. Also, relicensing is subject to public inquiries reactor by reactor.

EDF will likely seek lifetime extension beyond the 4th Decennial Safety Review (VD4) for most, if not all, of its remaining reactors. President Macron in his February 2022 programmatic speech made it clear that the government has no intention of closing reactors anymore. He stated: “While the first extensions beyond 40 years have been implemented successfully since 2017, I’m asking EDF to examine the conditions of the [lifetime] extensions beyond 50 years, in conjunction with the nuclear safety authority”.202

The first reactor to undergo the VD4 was Tricastin-1 in 2019. Bugey-2 and -4 were scheduled in 2020, and Tricastin-2, Dampierre-1, Bugey-5 and Gravelines-1 started in 2021… until the COVID-19 pandemic further disrupted the safety review schedule.203 Until mid-2023, 11 units had undergone their VD4 and a further five were underway (see Table 7).

While the President of the Nuclear Safety Authority (ASN) judged the VD4-premiere on Tricastin-1 “satisfactory”, he questioned whether EDF’s engineering resources were sufficient to carry out similar extensive reviews simultaneously at several sites.204 Beyond the human resource issue, the experience raises the question of affordability. EDF had scheduled an outage for Tricastin-1 of 180 days in 2019, which was first extended by 25 days to 205 days. Including further, unrelated unavailabilities, the reactor was finally in full outage for two thirds of that year (232 days).

1. Table 7 | Fourth Decennial Visits of French 900-MW Reactors, 2019–2023

Sources: compiled by WNISR, based on EDF REMIT-Data205

Notes: The expected duration is based on outage dates in use as of outage start, or within the few days after the reactor has been disconnected from the grid.

For ongoing decennial visits, end of outage date is the date in use as of 1 November 2023, and can vary from the original date:

* Expected duration as of Outage start

** Revised date, as provided as of 1 November 2023

EDF expects these VD4 outages to last six months, much longer than the average of three to four months experienced through VD2 and VD3 outages. The Chief Technical Officer of EDF Group and CEO of EDF R&D, Bernard Salha, told French Parliament in February 2023 that the work volume of a VD4 was five times larger than that of a VD3. He also said investments into the operating fleet have doubled over the past decade.206

As illustrated, many factors could lead to significantly longer outages. EDF has already started negotiating with ASN for the workload to be split in two packages, with the supposedly smaller second one to be postponed four years after the VD4.207

On 23 February 2021, the ASN issued detailed generic requirements for plant life extension.208 The key aspects of ASN’s decision were not the five short administrative articles but the two annexes setting the technical conditions and the timetable for work to be carried out. The challenge for operator EDF will be high, as ASN outlines:

Over the coming five years, the nuclear sector will have to cope with a significant increase in the volume of work that is absolutely essential to ensuring the safety of the facilities in operation.

Starting in 2021, four to five of EDF’s 900 Megawatts electric (MWe) reactors will undergo major work as a result of their fourth ten-yearly outages. (…)

All of this work will significantly increase the industrial workload of the sector, with particular attention required in certain segments that are under strain, such as mechanical and engineering, at both the licensees and the contractors.209

This was prior to the corrosion issues that struck EDF’s fleet at the end of 2021. ASN has shown remarkable tolerance for extended timescales of refurbishments and upgrades in the past; many of the post-Fukushima measures have not yet been implemented eleven years after the events, for example. As of the end of 2020, none of the 56 French reactors were backfitted entirely according to ASN requests issued in 2012. According to some estimates, the completion of the work program could take until 2039.210

Additionally, the implementation of work to be carried out as part of the lifetime extension beyond 40 years stretches over 15 years until 2036, when the last 900 MW reactor is supposed to be upgraded: Chinon B-4, connected to the grid in 1987, gets the 15-year delay to implement 15 of a total of 37 measures. By then, the unit will have operated for 49 years. This is just one example, and it is the newest of the operating 900 MW reactor. ASN has accepted similar timescales for all 32 of the 900 MW units. The French Nuclear Safety Authorities have proven flexible, and—considering the dire state of the reactor fleet—pressure for even more flexibility might increase in the future, particularly in the winter 2022–2023.

1. Figure 37 | Age Distribution of French Nuclear Fleet (by Decade)

Sources: WNISR, with IAEA-PRIS, 2023

Financial Trouble

Operating costs have increased substantially over the past few years (see also previous WNISR editions). The Court of Accounts calculated the operating costs for the year 2019 at €43.8/MWh (US$201949/MWh) when using an “accounting” methodology and €64.8/MWh (US$201972.6/MWh) when applying an “economic” approach (taking into account past investments) as chosen by the Court. Lifetime extension from 40 to 50 years would cost over €201535 /MWh [€202239/MWh or US$202241/MWh] based on EDF figures”, without considering the effect on post-operational costs.211 Whatever the uncertainties of the respective cost estimates, there is no doubt that the additional costs for refurbishment and upgrades in view of lifetime extensions remain far below any cost estimate for newbuild.

The Energy Regulatory Commission recalculated the electricity generating costs of the French nuclear fleet (incl. the Flamanville-3 EPR) for the years 2026–2030 in the range of €202253.8–60.7/MWh (US$202256.6–63.9/MWh) depending on the definition of the scope.212

Outages that systematically exceed planned timeframes are particularly costly. EDF’s net financial debt increased by about €10 billion (US$202310.6 billion) over the period 2019–2021 to a total of €43 billion (US$202151 billion)—as of the end of 2021.213 In 2022 alone, net debt jumped by €21.5 billion (US$202222.6 billion) to €64.5 billion (US$202267.9 billion) at the end of the year. In the first half of 2023, the debt load rose to €64.8 billion (US$202370 billion).214 Luc Rémont, EDF’s incoming CEO, stated during a hearing at the Finance Commission of the National Assembly:

We are on the eve of an industrial challenge which, in reality, is out of all proportion with the Group’s history for several reasons. The first is that we are beginning this steep path towards greater investment in electrification with the somewhat heavy rucksack of a 65 billion euro debt which is—I’m sure, even for the Finance Commission, 65 billion euros is a significant amount—I can assure you for a company, it is the heaviest amount a company can experience in Europe and so, naturally, it is part of the elements that define our capacities and the ways in which we can envisage this new investment cycle.215

Rémont added that the Group never before had to invest on the order of €25 billion per year (US2023$27 billion/year) of which 80 percent in France while “debt can hardly increase more”.216

EDF had been losing 100,000–200,000 clients per month for several years. However, as the sky-rocketing price increases continued into 2022, some consumers returned to EDF’s regulated tariffs that profited from the government-imposed price control mechanism. EDF claims an increase of about half a million clients between September 2021 and May 2022217, a further half a million until the end of 2022, and 400,000 until mid-2023.218 The drawback was that during low nuclear production and excessively high prices on the market, this forced EDF to “buy volumes [of power] at a price that is higher than we [EDF] resell it to the clients at the regulated tariff”, an EDF executive director stated.219

The Flamanville-3 EPR Saga Continued

“The EPR is an overly complicated, virtually unbuildable machine...”

Henry Proglio, Honorary Chairman, EDF220

The 2005 construction decision of Flamanville-3 (FL3) was mainly motivated by the industry’s attempt to confront the serious problem of maintaining nuclear competence. Fifteen years later, the regulator ASN still drew attention to the “need to reinforce skills, professional rigorousness and quality within the nuclear sector.”221

In December 2007, Electricité de France (EDF) started construction on FL3 with a scheduled startup date of 2012. The project has been plagued with design issues and quality-control problems, including basic concrete and welding difficulties similar to those at the Olkiluoto (OL3) project in Finland, which started construction two-and-a-half years earlier. (See earlier WNISR editions.) These problems never stopped.

In March 2020, EDF had stated that fuel loading would be delayed to “late 2022” and construction costs re-evaluated at €12.42015 billion (US$201513.8 billion), an increase of €1.52015 billion (US$20151.7 billion) over the previous estimate.222 In addition to the overnight construction costs, as of December 2019, EDF indicated more than €4.2 billion (US$20194.7 billion) was needed for various cost items, including €3 billion (US$20193.4 billion) of financial costs.

In January 2022, EDF estimated the overnight costs at €201512.7 billion (US$201514.1 billion).223 In December 2022, the figure was updated to €201513.2 billion (US$201514.6 billion).224 In 2020, the French Court of Audits estimated the total cost, including financing and other associated costs, at €201519.1 billion (US$201521 billion).225 The Court estimated that the cost of electricity from FL-3 would be €2015110–120/MWh (US$2015122–133/MWh). This estimate has not been publicly updated.

The fuel issue that struck the Taishan EPRs and kept Unit 1 off-grid for over one year had consequences for FL3. EDF decided to refabricate 64 of the 241 fuel assemblies that had already been produced. These were approved by ASN and delivered to the site.

As of mid-2023, the latest projected date for fuel loading is the first quarter 2024. Because of a fabrication default (see earlier WNISR editions), the vessel head will have to be replaced at the end of the first refueling cycle scheduled for the second half of 2025.226

Conclusion

The French nuclear industry remains under a high level of stress. The full re-nationalization of EDF, analysts agree, will not solve its structural problems: an ageing nuclear fleet with lowest performance in decades, manpower and competence challenges, unprecedented investment needs at times of unprecedented net debt, and never-ending problems at the only active construction site at Flamanville.

Not covered here, but to this list should be added serious fuel chain issues, climate impact, social movements, and some unexpected opposition. Especially the plutonium-economy part of the industry is experiencing its own crisis with historically low throughput at the spent fuel reprocessing plant at La Hague and at the uranium-plutonium mixed-oxide (MOX) fuel fabrication facility MELOX at Marcoule. Consequently, spent fuel pools are filling up and the stocks of unirradiated plutonium have increased to unprecedented levels.

Confronted with this avalanche of problems, the French government has chosen to insist on the launch of a nuclear newbuild program—supported by a majority in the National Assembly. And EDF follows suit:

On 29 June 2023, EDF announced that it was making the applications for approval to launch construction of the first pair of EPR 2 reactors at Penly, and starting other administrative procedures required for their completion and connection to the electricity transmission network. EDF’s objective is to begin preparatory work in mid-2024.227

The EPR2 does not even exist on paper. It increasingly looks as if the current administration and nuclear establishment have not learned the lessons of the Flamanville EPR1 disaster, as spelled out in the chapter headlines of a 2019-assessment commissioned by EDF’s President: “An unrealistic initial [cost] estimate; (…) An inappropriate project governance; Struggling project teams; (…) Insufficiently advanced studies at launch; (…) Generalized loss of competence.”228

Largely unreported, the science community in France is far from offering unanimous support of the newbuild initiative. As of the end of October 2023, close to 1,200 scientists had signed the aforementioned “Call by scientists against a new nuclear program”.

Germany Focus

Nuclear Power in Germany – The Last 25 Years in a Nutshell

Since the beginnings of commercial nuclear operations, there has always been substantial opposition towards the technology in Germany. Protests in the 1970s (Wyhl, Brokdorf, Gorleben, and others) with up to 100,000 participants led to the formation of a politically strong anti-nuclear movement that culminated in the formation of the Green Party.229 The 1998-election led to the formation of the first “Red-Green” government, a coalition made up of the Social Democrats (SPD) and the Green Party, led by Chancellor Gerhard Schröder. The ensuing first Renewable Energy Act (“Erneuerbare Energien Gesetz” or EEG)230 laid the groundwork for Germany’s renewable energy expansion, and the so-called “consensus agreement” (“Atomkonsens”), that limited operational lifetimes of German nuclear power plants to a maximum of 32 years, and, most notably, involved no financial compensation for utilities.231 It was molded into legislation in 2002. In 2010, a conservative-liberal, pro-business, and pro-nuclear Government, consisting of the Christian Democrats (CDU & CSU) and the Liberal Democratic Party (FDP), led by physicist Chancellor Angela Merkel, passed legislation that extended the lifetimes of nuclear power plants completed before 1981 by eight years and all other plants’ lifetimes by 14 years.232 However, immediately after the 2011-Fukushima disaster, the same Government implemented a three-month operational moratorium for seven reactors built before 1980 and temporarily suspended the above-mentioned lifetime extensions for all other plants.233

The Ethics Commission for a Safe Energy Supply, instated by Chancellor Merkel, came to the conclusion:

The Ethics Commission is strongly convinced that the withdrawal from nuclear energy can be completed within one decade using the measures presented here for the energy transition. Society should commit to this objective and the necessary measures. It is only by having a clear, scheduled objective as a basis that the necessary decisions on planning and investment can be taken. (…)

The withdrawal from nuclear energy is necessary and is recommended to rule out future risks that arise from nuclear in Germany. It is possible because there are less risky alternatives.234

The closure of eight of Germany’s oldest235 reactors and the progressive phaseout of the remaining nine by the end of 2022 was drafted into legislation, effectively reactivating the former “consensus agreement” (see Table 8 for the phaseout schedule). With no political party dissenting, it looked virtually irreversible under any political constellation. On 6 June 2011, only one week after the Ethics Commission submitted its report, the German Bundestag passed a seven-part energy transition legislation almost by consensus that came into force on 6 August 2011 (see earlier WNISR editions for details).236 This renewed phaseout scheme prompted the utilities to sue for compensation that, after ten years of legal battles in German courts of law and international arbitrations courts, led to the payment of a total of €2.4 billion (US$20212.8 billion) in 2021.237

In September 2021, legislative elections saw the SPD become the largest political party in Germany. But even in a coalition with the Green Party they would not have had a parliamentary majority, so after complex negotiations, an unprecedented “traffic light” (“Ampel”) coalition-government was formed by adding the FDP (yellow) to the SPD (red) and Greens.

One year into the legislative period, on 5 September 2022, Green party member Robert Habeck, Minister for the Economy and Climate Protection and Vice-Chancellor of Germany, presented the results of a second stress test of the electricity system’s resilience for the winter 2022–2023. He announced that he would recommend to the Government to transfer two of the three remaining operating nuclear reactors, namely Isar-2, and Neckarwestheim-2 into “reserve status” as of the end of 2022. Emsland would be shut down as planned by 31 December 2022.238 This left the FDP, that had over the course of 2022 taken over a role as nuclear advocacy party, dissatisfied, prompting infighting within the coalition, mainly between Finance Minister and FDP leader Christian Lindner, and Robert Habeck. On 17 October 2022, Chancellor Olaf Scholz ended the dispute by announcing in an executive order that all three nuclear power plants would remain operational until 15 April 2023. The order also determined that no new fuel assemblies would be acquired.239 The reactors would merely operate in “stretch mode”, exhausting the fuel in the core. The required change of the Atomic Energy Act was adopted by the cabinet a few days later, and by the Bundestag in November 2022. On 15 April 2023, all three plants were closed.240

Sky-rocketing energy prices in late 2021, the war in Ukraine, and high German dependency on Russian fossil fuel imports (gas, oil, and coal) provided a further opportunity for some pro-nuclear voices in the country to receive considerable attention. In fact, the discourse of the “German isolated phaseout decision in a world going all nuclear” had entered the main media already in the past few years.

An Unexpected Debate Over Potential Lifetime Extensions

The war in Ukraine triggered a public controversy that hardly assessed options based on factual understanding of their respective implications but often consisted of a fact-free opinion debate. Are you for or against lifetime expansions? Never mind legal aspects, technical feasibility, costs, and potential safety implications. A whole series of opinion polls showed comfortable majorities in favor of stretching the operation of the three remaining reactors by a few months or even up to five years. The public perception linked continued operation of the reactors to the hope for more independence from Russian gas.241 A mirage, as reports commissioned in the spring of 2023 showed after the dreaded winter had been overcome without the severe blackouts that had been predicted by some:242 the lifetime “stretching” had close to no effect on security of supply, and impact on wholesale electricity prices in 2022 and 2023 was limited to under 1 percent.243 Instead, mild temperatures in winter and active reduction of consumption by consumers had reduced German gas demand in 2022 by 14 percent compared to the previous four-year average.244

On 7 March 2022, three days after the Russian army attacked and then occupied the Zaporizhzhia nuclear power plant, the German Government issued a 5-page joint statement of the Ministries of Environment and Economy assessing a potential restart of the three reactors that were closed at the end of 2021 and the potential lifetime extension of the remaining three operating reactors beyond the legal closure date of end of 2022:245

• The restart of the three units closed end of 2021 is “out of the question” notably due to the expired operating license.
• The lifetime extension of the still operating units would not lead to additional power generation in the winter 2022/2023, as there is no new fuel available before fall 2023 at the earliest.246
• A lifetime extension of the currently still operating three units beyond the end of 2022 would require an in-depth safety assessment of each of the reactors last carried out in 2009. The outcome and potential backfitting and upgrading work needed cannot be reliably predicted.
• A lifetime extension could not be economically justified for 2–3 years and would not make sense under 3–5 years considering the safety related issues and the need to re-train staff. The two ministries consider that in that timeframe there are other options.
• From a constitutional rights perspective, a lifetime extension would require a comprehensive, new risk-benefit assessment by the legislator. “Against this background, the expected lawsuits against a possible lifetime extension would definitely have promising chances of success.”
• The operators have signaled that a lifetime extension would essentially mean the takeover of legal and economic risks by the state. As the two ministries consider that compromising on safety is not an option, lifetime extension could mean lengthy backfitting programs in the period 2022–2024.
• In conclusion, the two ministries “cannot recommend a lifetime extension of the three still operating nuclear power plants”.

Four days after the government statement and two weeks after Russia had launched its all-out war against Ukraine, the parliamentary group of the far-right AfD (Alternative für Deutschland/Alternative for Germany) tabled a proposal for a resolution in which the German Bundestag would “call on the Federal Government to implement, together with the Länder Governments a lifetime extension of the nuclear power plants” and “immediately give nuclear power plant operators unambiguous and binding assurances that the nuclear power plants may be operated without restriction until their technically reasonable end of life.”247 The proposal was rejected by all of the parliamentary committees and, on 7 July 2022, received a unanimous rejection by all parliamentary groups from the far left to the Christian Democrats. The vote ended 581 to 67, whereas only AfD members and one independent voted for the proposal.248

In June 2022, all three operators of the remaining plants, EnBW, E.ON, and RWE, opposed lifetime extensions citing technical and regulatory challenges that would have to be overcome.249

Over the summer of 2022, noteworthy developments included the following:

• A legal analysis commissioned by Greenpeace concluded on 22 July 2022 that any form of operation of the remaining reactors beyond the end of the year would violate constitutional law, necessitate significant backfitting, and require cross-border consultations under E.U.-Environmental Impact Assessment legislation and ESPOO Convention.250
• On 26 July 2022, the smallest government coalition partner FDP called for a lifetime extension of all three reactors to 2024, arguing: “This is the period when we face energy shortages. That is why we must be prepared for it.”251
• On 28 July 2022, five key SPD parliamentarians on energy and climate issues, led by the parliamentary group’s Vice-President Matthias Miersch, sent a 4-page letter to party members pointing to a comprehensive list of issues highlighting problems around the potential lifetime extension, like the “challenges in times of gas shortages are in the industry and the provision of heat – not in the power sector”; while less suitable than gas plants, coal plants are more suitable to make up for shortages than nuclear plants, as they were more flexible; under regular circumstances, the three nuclear plants would have had to undergo a comprehensive decennial safety inspection in 2019, which they were exempted from considering the anticipated closure in 2022—that safety review would be “mandatory”, could last several years and entail “significant investment needs”; the operators do not want to bear the legal, economic, and safety risks, that would have to be covered by the state.252
• Early September 2022, a draft motion for the regular Green Party congress scheduled for October 2022 was circulated and called on the federal party executive board, the parliamentary group, and the federal government “to stick to the 31 December 2022 phaseout date for the last three nuclear power plants in Germany.”253

Between mid-July and early September 2022, the four grid operators in Germany carried out a second stress test on security of supply and stability of the grid for the winter 2022/2023 under significantly more stringent assumptions. The hour-by-hour analysis included the potential contributions or needs of neighboring countries. A sensitivity analysis found the greatest potential impact with the performance of the French nuclear fleet and the water levels of rivers in Germany (in particular for the shipment capacity of coal).

The French Government had assured the German Government, “orally and in writing”, so said Minister Habeck on 5 September 2022, that 50 GW of the installed total of 61 GW of French nuclear capacity would be operational in the winter.254 The French assurances for winter 2022/2023 had seemed to be based on highly optimistic assumptions, and the German grid operators consequentially judged it necessary to model scenarios with a French nuclear capacity limited to 45 GW and 40 GW respectively.255 The most challenging scenario combined the limited nuclear capacity with the assumption of unavailability of half of the reserve capacity (mainly coal) and half of the gas plants in southern Germany.256

Minister Habeck concluded from the stress test results that “it remains highly unlikely that we will face a crisis or an extreme scenario”, but due to the cumulation of circumstances, “given all these risks, we cannot rely on our neighboring countries to have enough power stations available to help stabilize our power grid at short notice in the event of grid congestion.”257 Therefore, the ministry decided to propose the creation of a new reserve capacity, limited in time, in the form of the two southern nuclear plants Isar-2 and Neckarwestheim-2. The two reactors should “remain available until mid-April 2023 so that they can, if necessary, make an additional contribution to the power grid in southern Germany this winter.”258

Other countermeasures recommended by the grid operators were implemented, including additional production in biogas plants and the increase of transmission capacity and effectiveness. The ministry clarified that the two nuclear units should be “deployed only when it seems likely that the other instruments will be insufficient to avert a supply crisis.” The extension beyond mid-April 2023 or the reactivation in the winter 2023/2024 “is not possible due to the safety status of the nuclear power plants and the fundamental considerations about the risks of nuclear power.” 259

The idea was to monitor European capacity availability throughout the winter and, should it have appeared in November or early December 2022 that a severe shortage was to emerge in January 2023—e.g. due to lower than expected French nuclear capacity—the two southern reactors would keep operating until their fuel exhausted. Otherwise, the units would have been shut down at year-end as stipulated under the current legislation and restarted only should a crisis situation have occurred later in the winter. This would not have been a stop-and-go kind of operation, but once restarted, the reactors would have kept operating until fuel exhaustion.

Meanwhile, the French government, faced with an unprecedented unavailability level of its own nuclear power fleet, called on Germany, in the name of mutual solidarity, to extend the operation of the three remaining reactors “for a few months”, while assuring to upgrade the gas links to Germany in return.260 In 2022, French nuclear production fell to the lowest levels since 1988 due to extended, unplanned outages that kept up to two thirds of the French fleet-capacity down, resulting in neighboring countries having to export large quantities of power to France which, for the first time since 1980, turned into a net power importer over the year. Germany has been a net power exporter to France for many years, especially in winter. In 2022, annual net export reached 15 TWh. (See France Focus).

Following the publication of the stress test results and the conclusions of the Ministry of Economy and Climate Protection, coalition member FDP reiterated the call for a lifetime extension at least until 2024, making a 180 degree turn from statements of the year before when party leader Lindner had said that nuclear power “may be CO2-free, but certainly not sustainable”.261 The party leader of the Christian Democrats (CDU), Friedrich Merz, called the potential closure of the three reactors at year end “completely absurd”.262 Other conservative politicians even called for nuclear newbuild in Germany. Former Federal Transport Minister Andreas Scheuer of Bavarian CDU-equivalent CSU stated: “My formula is 3+3+3: Three nuclear plants must continue operation, three must be reactivated and three new plants must be built”.263

The political feud between Greens and FDP escalated when Lindner refused to accept the proposition once it was brought before cabinet on 11 October 2022 and advocated for the continued operation of all three reactors instead. Meanwhile, on 14 October 2022, the Green party conference approved Habeck’s plans for stretch-operation until 15 April 2023 but explicitly opposed the procurement of new nuclear fuel, which would be required for continued operation until 2024, as proposed by the FDP and conservatives.264

In an attempt to mediate between Greens and FDP, several talks were held at Chancellor Scholz’s office. As these talks had led to no conclusions, in the late afternoon of 17 October 2022, Chancellor Scholz issued an executive order, ending the dispute between the two junior coalition partners. Thereby, all three plants, Emsland, Isar -2, and Neckarwestheim -2, were to remain on the grid until 15 April 2023, a minor win for the FDP. Supposedly to sweeten the deal for the Greens, the order included plans to draw up “ambitious legislation towards energy efficiency increases” and to politically push for an early coal-phase out in the federal state of North Rhine-Westphalia in 2030.265 Scholz demanded that “the relevant proposed regulations [on the “stretched operations”] be presented to the cabinet as soon as possible as part of the distribution of responsibilities.” Lindner said that “it is in the vital interest of our country and its economy that we maintain all our energy production capacities this winter”, and Green parliamentary leaders Britta Hasselmann and Katharina Droege pointed out that the limited lifetime extension at Emsland was “unfortunate and had no factual or technical reason”. All three nuclear operators positively commented on Scholz’s decision, saying that now that it was clear what would be happening, they could begin planning for continued operation until mid-April 2023.266

On 19 October 2022, the draft bill to extend operations of Emsland, Isar-2, and Neckarwestheim-2 to 15 April 2023 received cabinet approval and was passed on to the German Bundestag. Habeck emphasized that no new fuel rods would be ordered, and that he trusted “that the FDP will stick to the [coalition] agreement and not damage the authority of the chancellor [by calling for further extensions]”.267 In the press conference following the cabinet meeting, Environment Minister Steffi Lemke (Green Party) stated:

The phaseout of nuclear power will remain the same. Germany will finally phase out nuclear power on 15 April 2023. There will be no extension of the service life and no procurement of new fuel assemblies - and therefore no additional highly radioactive waste. The draft law will contribute to the stability of the power grid, which is compatible with nuclear safety because it limits the duration of nuclear power plant operation to a short period this winter. Even in the current energy supply crisis, we must keep an eye on the risks of nuclear power.268

On 11 November 2022, the Bundestag approved the 19th amendment of the Atomic Energy Act and thus stretched the operational lifetime of the three remaining nuclear power plants by three and half months. Legislation was approved with 375 votes in favor, 216 opposing (consisting of conservative parties CDU and CSU, the Left-wing party Die Linke, and several Green members of parliament). The AfD parliamentary group abstained (with one vote against).269 During the same session, CDU and CSU put an amendment to a vote with the aim to extend the operation of the three units to 31 December 2024 to a vote, and the AfD proposed two legislative measures for unlimited operational lifetimes and increased funding into nuclear research. All three propositions failed to gather a majority.270

In the days and weeks leading up to the final closure of the three reactors, the debate continued. Bavarian prime minister Markus Söder (CSU) called for all three plants to operate until 2030, saying that the phaseout was a “mistake” and even a “sin”.271 Leading member of the FDP Wolfang Kubicki continued to usher warnings about hypothetical consequences of the phaseout:

Shutting down the world’s most modern and safest nuclear power plants in Germany is a dramatic mistake that will still have painful economic and ecological consequences for us.272

Parliamentary leader of the FDP, Christian Dürr, suggested that three reactors remain in a “strategic reserve” and delay decommissioning because “one could switch them back on if a difficult [energy supply] situation arises”. This was dismissed as “utter nonsense” by prominent Green party member and former Environment Minister Jürgen Trittin.273

Outside of the political debate, a shift seemed to be emerging in German society from a general acceptance of the phaseout to gradual opposition, leading to, depending on the poll, up to two thirds of Germans surveyed in the Spring of 2023 opposing the planned phaseout, citing fears of energy security and rising prices274, although official Government simulations (see above) and other calculations had come to the conclusion that there would be little to no effect on electricity prices or security of supply.275 Nonetheless, prominent industry representatives issued warnings of rising electricity prices and the subsequent locational disadvantage of Germany.276 Chief of the German Technical Inspection Association (TÜV) Joachim Bühler became a prominent advocate for lifetime extensions and even restart of closed reactors after the South-German association TÜV Süd had issued a note on the technical feasibility of lifetime extensions at Isar-2 and the restart of Gundremmingen-C.277 This 7-page paper however was dismissed as “biased” in a legal opinion commissioned by Greenpeace, mainly due to the neglect of necessary safety inspections and expected ensuing measures that would have needed to be implemented, as the last in depth decennial inspection had been conducted in 2009.278 The debate as a whole was criticized by other experts as a “phantom debate” as technical, organizational, financial and liability-related issues were too high to extend operational lifetimes or even restart reactors.279

On 15 April 2023, Emsland, Neckarwestheim-2, and Isar-2 were finally disconnected from the grid. Since then, decommissioning preparation or actual dismantling has commenced at all three plants (see Decommissioning Status Report).

In the months after the reactor closures, trade data showed that Germany was importing more electricity than it exported—a situation due to price developments on the European power market and not because of capacity shortages—that nevertheless swiftly led German conservative and liberal voices, and the French minister for the energy transition, criticizing Germany's energy policy.280 Some political actors criticized Germany as the “only wrong-way driver” in energy policy and demanded the restart of up to eight closed reactors.281 This number comes from a report issued by pro-nuclear Radiant Energy Group that claims that eight reactors could be restarted in as soon as nine months for costs of €100–200 million each (US$109–218 million).282 Given that most German reactors are well underway with decommissioning (See Decommissioning Status Report), and that the utilities have repeatedly confirmed their decision to move away from nuclear power, these estimations seem unrealistic.

The 65 Years of the German Nuclear Program 1958–2023

In 1955, a ten-year post-World War II moratorium on reactor construction and uranium procurement ended in the Federal Republic of Germany (FRG), and the government swiftly opened the first German nuclear research facility in Karlsruhe in 1956 and began constructing the first research reactor in Garching, Bavaria, only one year later. On 1 January 1960, the first Atomic Energy Act came into force and the first West German demonstration reactor VAK Kahl came online in 1961.

The first commercial power plant however was built in the German Democratic Republic (GDR): Rheinsberg began electricity production in 1966. The first West German commercial plant, Gundremmingen-A, was connected to the grid seven months later, in 1966, but closed in 1977 following a radioactive steam leak.283 Most German nuclear power plants began construction in the late 1960s to early 1980s, and were met by major opposition through the whole of society (see above).284 Most of these reactors were light-water reactors, while some attempts at establishing other technologies were made, e.g., the fast breeder SNR-300 in Kalkar, or the pebble-bed high-temperature reactor (THTR-300) in Hamm-Uentrop, they never properly operated.285 Kalkar never started up and has since been transformed into an amusement park.286

In 1989, the total installed capacity of East and West German nuclear power plants reached its maximum of 22.9 GW. After unification, mostly due to liability concerns, former GDR plants at Greifswald and Rheinsberg were closed (and have been undergoing decommissioning since; see Decommissioning Status Report), and construction at three additional units at Greifswald and two at Stendal was halted.287 Before the “consensus agreement” was negotiated by the SPD-Greens government of 1998-2002, West German reactors Lingen, Mülheim-Kärlich, and Würgassen had been taken off the grid for various reasons.288 Reactors Stade and Obrigheim were the only two that were closed as a consequence of the agreement, but by August 2011, another eight were closed resulting from the reinstated phaseout legislation. Further plants were closed successively between 2015, and mid-2023 (see Figure 38).

1. Figure 38 | Construction and Operational History of the German Nuclear Reactor Fleet

Sources: WNISR, with IAEA-PRIS, 2023

Nuclear Power, Renewables, Fossil Fuels, and Efficiency

Germany’s nuclear fleet generated 32.8 TWh net in 2022, a decline by half over the previous year after three reactors were closed at the end of 2021, and only a fraction of the peak generation of 162.4 TWh in 2001. In 2022, nuclear plants provided 6 percent of Germany’s gross electricity generation, compared to the historic maximum of 35.6 percent in 1999, according to data from AGEB.289

Renewables generated 254 TWh (gross), a significant 8.5 percent-increase over the previous year, Consequently, the share of renewables rose five percentage points from 40.2 percent to 44.5 percent.290 In the first half of 2023, while, due to unfavorable climatic conditions, the renewables output slightly declined (-1 percent) compared to the same period in 2022, their share rose nevertheless from 49 percent to 52 percent as consumption dropped significantly.291

Figure 39 summarizes the main developments of the German power system between 2010—the last year prior to the post-3/11 closure of the eight oldest nuclear reactors—and 2022.

The increase in renewables (+148.8 TWh) and the decline in consumption (-68.3 TWh) still overcompensate the decline in fossil fuel (-95.4 TWh) and nuclear generation (-105.9 TWh), allowing for an increase in net exports (+13.2 TWh).

1. Figure 39 | Main Developments of the German Power System Between 2010 and 2022

Sources: WNISR, based on AGEB, 2023

Developments within the fossil-fuel generating segment:

• Lignite peaked in 2013 and then declined—especially in 2019–2020—before increasing again by 20.2 percent in 2021 and another 5.3 percent in 2022. Lignite generation in 2022 thus exceeded 2019 levels by 2.2 TWh but stayed 20.4 percent below the 2010-level.
• After declining constantly between 2013 and 2019, hard coal electricity generation increased for the second year in a row, by 18 percent year on year, to 64.4 TWh remaining 45 percent below the 2010-level.
• Natural gas consumption for electricity in 2022 declined by 11.6 percent compared to 2021 to 79.8 TWh, the lowest value since 2016 and 10 percent below the 2010-level.

1. Table 8 | Legal Closure Dates for German Nuclear Reactors, 2011–2023

   Reactor Name (Type, Net Capacity)	Owner/Operator	First Grid Connection	End of License (latest closure date)
   Biblis-A (PWR, 1167 MW) Biblis-B (PWR, 1240 MW) Brunsbüttel (BWR, 771 MW) Isar-1 (BWR, 878 MW) Krümmel (BWR, 1346 MW) Neckarwestheim-1 (PWR, 785 MW) Philippsburg-1 (BWR, 890 MW) Unterweser (BWR, 1345 MW)	RWE RWE KKW Brunsbüttel (a) PreussenElektra KKW Krümmel (b) EnBW EnBW PreussenElektra	1974 1976 1976 1977 1983 1976 1979 1978	6 August 2011
   Grafenrheinfeld (PWR, 1275 MW)	PreussenElektra	1981	31 December 2015 (closed 27 June 2015)
   Gundremmingen-B (BWR, 1284 MW)	KKW Gundremmingen(c)	1984	31 December 2017
   Philippsburg-2 (PWR, 1402 MW)	EnBW	1984	31 December 2019
   Brokdorf (PWR, 1410 MW) Grohnde (PWR, 1360 MW) Gundremmingen-C (BWR, 1288 MW)	PreussenElektra/Vattenfall( d) PreussenElektra KKW Gundremmingen	1986 1984 1984	31 December 2021
   Isar-2 (PWR, 1410 MW) Emsland (PWR, 1329 MW) Neckarwestheim-2 (PWR, 1310 MW)	PreussenElektra KKW Lippe-Ems (e) EnBW	1988 1988 1989	15 April 2023

Sources: WNISR with IAEA-PRIS, July 2023

Notes: Krümmel and Brunsbüttel were officially closed in 2011 but had not been providing electricity to the grid since 2009 and 2007 respectively.

PWR: Pressurized Water Reactor; BWR: Boiling Water Reactor; KKW: Nuclear Power Plant (Kernkraftwerk);
RWE: Rheinisch-Westfälisches Elektrizitätswerk Power AG; EnBW: Energie Baden-Württemberg AG.

a - Vattenfall 66.67%, E.ON 33.33%

b - Vattenfall 50%, E.ON 50%.

c - RWE 75%, E.ON 25%.

d - E.ON 80%, Vattenfall 20%.

e - RWE 87.5%, E.ON 12.5%.

Other Nuclear Developments in Germany

The closure of the commercial nuclear power plants has not led to the end of industrial activities in the sector in Germany, in particular considering the nuclear fuel manufacturing facility in Lingen and the uranium enrichment plant in Gronau.

The facility at Lingen is operated by Advanced Nuclear Fuels GmbH (ANF), a subsidiary of French state-owned company Framatome.292 An application to cooperate with Rosatom subsidiary TVEL which would enable ANF to manufacture fuel assemblies for Soviet-designed VVER reactors located mainly in Eastern Europe had been submitted to the German Office for Independent Competition (Bundeskartellamt) in February 2021. The application was withdrawn several days before the Russian attack on Ukraine.293 Instead, Framatome and Rosatom founded a joint venture in France.294 ANF has since reapplied for a license extension to produce hexagonal fuel rods in Lingen. This faces opposition from the responsible Environment Ministry in Lower Saxony, led by Minister Christian Meyer (Green Party) who said that “deals with Putin should be ended, […] especially in the nuclear sector.”295

In the past, depleted uranium hexafluoride had been transported from Gronau to Russia where it had been re-enriched,296 these contracts had expired before the Russian attack on Ukraine.297 Owner Urenco indicated it has since cut all ties with its last remaining (unnamed) Russian supplier.298

Meanwhile, the search for a final repository site for highly active nuclear waste in Germany is underway. Initial plans to select a site by 2031 were questioned in a report from the federal company in charge, the Bundesgesellschaft für Endlagerung (BGE), according to comments made in the media by the Federal Environment Ministry on 10 November 2022.299 The overseeing Federal Office for the Safety of Nuclear Waste Management (BASE) had repeatedly urged BGE to provide plans for the process, and in December 2021, BGE had stated that there were “no signs that the goal of finding a site for the final repository by 2031 would fail.”300 In the aforementioned dedicated report, dated 28 October 2022, that was reportedly passed on by the Federal Ministry of the Environment to BASE on 17 November 2022,301 BGE envisions site selection for 2046–2068, contradicting current legislation requiring site selection by 2031. As of December 2022, discussions between agencies were ongoing in this regard.302 Meanwhile, BGE announced that it was experiencing delays at Schacht Konrad, a former iron ore mine that is being rebuilt as final repository for low and intermediate waste and could therefore not stick to the original plan of completion in 2027.303

Conclusion: From Electricity Generation to Management and Disposal of Nuclear Waste

After 75 years of nuclear power history, the last three operating nuclear power reactors were closed in April 2023. Germany has joined three other countries that have phased out national nuclear power programs, namely Italy, Kazakhstan, and Lithuania. However, some industrial nuclear activities are still ongoing, such as nuclear fuel manufacturing in Lingen and uranium enrichment in Gronau; and Germany will remain active in international organizations like the IAEA, the OECD’s Nuclear Energy Agency, and the various instances of the European Union. At the same time, other topics that have so far gotten little attention are moving to the forefront, particularly nuclear facility decommissioning and nuclear waste disposal.

With hindsight, the socio-technical discussions in Germany were rather similar to those in other countries. The discussion centered around the multiple implications of a “technical controversy”, i.e. safety issues related to commercial and military uses of nuclear power. The German historian Joachim Radkau noted as early as 1983 that the anti-nuclear movement was not like any other socio-political movement but that it was enshrined in a deep technical debate about the feasibility of “sustainable” nuclear power, a debate that reappears today.304 The movement in Germany was similar to those in other countries, but it was particularly “successful”: the general deployment of nuclear power plants had not been stopped in its early days but the movement succeeded in making the economic and technical complexity of the issue widely known, and developed a convincing argumentation on costs, safety, environmental, and societal issues that had not been identified elsewhere or were identified but pushed aside (like in France).305 The initial resistance against early nuclear power applications in Wyhl (1975)—inspired by the opposition movement against the Fessenheim nuclear power plant project in France—rapidly spread to other sites. The latest construction start in Germany of a completed nuclear reactor (Neckarwestheim-2) took place in November 1982 and it was started up in January 1989.

A historical example is the decision of Germany not to pursue the plutonium route with commercial spent fuel reprocessing, as both projects, at Gorleben (Lower Saxony) in the late 1970s and at Wackersdorf (Bavaria) in the late 1980s, were abandoned after fierce opposition. Other projects were implemented despite widespread protests and were perceived by the movement as “failures”, such as Brokdorf, a nuclear reactor debated since the 1970s that was brought online shortly after the 1986 Chernobyl accident.

Controversial debates about nuclear power were also at the origin of the “Energiewende”, the socio-ecological transformation that started in Germany in the 1970s and provided a book title in 1980.306 The first energy transformation scenarios suggested to end nuclear power and oil consumption but still contained significant amounts of coal. It was only after the Rio Conference (1992) and the emergence of climate considerations that the end of coal (“coal exit”) gained a dominant position in the public debate. It is not far-fetched to suggest that without the antinuclear movement the breakthrough of the “Energiewende” and the successful mass introduction of renewables might not have happened.307

As in other market economies, German energy companies were pushed by the government to develop nuclear power, starting with the “Gundremmingen” model of state guarantees and subsidies and ending in captured customers having to pay (high) cost-plus tariffs to their local or regional monopolistic supplier. In that context, the “liberalization” of the electricity and natural gas sectors in the 1980s and 1990s heralded the end of nuclear power investments, showing clearly that nuclear power was not competitive under market-economy conditions (see Chapter on Nuclear Economics and Finance). When the unification of East and West Germany occurred in 1990, the energy industry could have built new nuclear power plants, or it could have at least completed the ongoing projects at Greifswald/Lubmin and Stendal, inherited from the GDR. Instead, the projects were scrapped, and operating plants were closed, whereas an entire new fleet of lignite plants was built in East Germany with substantial government support. Thus, inherently, the decision to end commercial nuclear power had been taken already in 1990, the rest of the process being political struggles about distributing the significant economic rents.308

After the closure of the last three reactors, the discussion is now rapidly moving from commercial nuclear power—besides some marginal requests for “newbuilds” by a handful of opposition politicians in need of public profile, a few research organizations in need of funding,309 and the usual lobby organizations and propagandists in need of attention—to challenges of decommissioning and disposal of nuclear waste. Decommissioning will take much longer and will be more expensive than planned (see Decommissioning Status Report). Disposal of nuclear waste, 62 years after the first generation of nuclear electricity, is at its very beginning, with decisions on a deep geological storage site expected in the 2040s at the earliest, and thus a final date for the disposal of the last nuclear waste container deep in the 22nd century.

Is Germany’s path an exception, a “Sonderweg”, in global nuclear trajectories? Yes and no. “Yes”, because the intensity of public debate was particularly high, and the societal consensus on ending commercial nuclear power generation was broader than in most other countries. Also, few other nuclear countries have set legal target dates for the phaseout of nuclear power use. However, “No” too, because there is not a single market economy that has succeeded the challenge of subsidy-free commercial nuclear deployment. Rather, the diminishing share of nuclear power can be observed globally since 1996 and more reactors have been closed over the past two decades than started up. Germany has merely accelerated a global declining trend that, despite all the newbuild announcement noise, will most likely quietly continue to erode the relevance of the nuclear industry in the energy markets.

Japan Focus

Overview

During Financial Year 2022, which runs from April 2022–March 2023, the number of nuclear reactors considered “operable” remains at 10 with a capacity of 9.6 GW310. The average load factor for the whole Japanese nuclear power plants has worsened from 22.1 percent in 2021 to 18.7 percent in 2022 (calendar years).311 As a result, the total nuclear power generation decreased from 61.3 TWh in 2021 to 51.9 TWh in 2022.312 The share of nuclear power in the total power generation also decreased from 7.2 percent in 2021 to 6.1 percent in 2022.313 (See Figure 40)

1. Figure 40 | Rise and Fall of the Japanese Nuclear Program

Sources: WNISR with IAEA-PRIS, 2023

The current reactor fleet consists of 33 units (33.1 GW, gross) of which 25 units (24.8 GW, gross) have applied for an operating license under the new post-Fukushima regulations.314 So far, new licenses have been granted for 17 units while eight applications remain under review. The national safety authorities have not issued any new operating license during the past year.

As of 1 July 2023, nine reactors out of 10 operable reactors (Ikata-3, Mihama-3, Takahama-3 & -4, Ohi-3 & -4, Genkai-3 & -4, Sendai-1) were operating and one was shut down for a periodic inspection (Sendai-2).

WNISR considers 23 in Long-Term Outage (LTO) and 10 in operation. In the past year, the IAEA has adopted a new category called “Suspended Operation” (see dedicated section in General Overview) and has reclassified 23 reactors that WNISR considers as in LTO. In other words, Japan and the IAEA have adopted an approach similar to the LTO concept that WNISR introduced in 2014. (See Figure 7). While the Japan Atomic Industrial Forum (JAIF) has not changed the definition of the category of “operating reactor”, the report from the government to the IAEA seemed to have changed when the responsible agency as the IAEA-correspondent moved from the Nuclear Regulation Authority (NRA) to the international affairs division of the Ministry of Economy, Trade and Industry (METI)315.

1. Figure 41 | Status of the Japanese Reactor Fleet

Sources: Various, compiled by WNISR, 2023

Twelve years after the Fukushima accident began, the reactors in operation are all PWRs although five BWRs (Kashiwazaki-Kariwa-6 & 7, Tokai-2, Onagawa-2, and Shimane-2) have received confirmation from the NRA to satisfy new regulatory requirements set by NRA in 2013.

As of mid-2023, the Japanese nuclear fleet consisting of 33 units including 23 in LTO had reached a mean age of 32.4 years, with 18 units over 31 years (see Figure 42).

1. Figure 42 | Age distribution of the Japanese Nuclear Fleet

Sources: WNISR with IAEA-PRIS, 2023

Tokyo Electric Power Co.’s (TEPCO’s) Kashiwazaki-Kariwa-6 was the first BWR to receive approval from NRA on 27 December 2017. However, due to lack of approval from Niigata Prefecture as well as due to nuclear security violations in 2021, it is not known when the reactors at this site will restart operating. On 17 May 2023, the NRA decided to maintain its ban on moving fresh nuclear fuel within the plant. TEPCO planned to restart the plant in October 2023, but that is impossible without moving nuclear fuel. The NRA imposed the ban in April 2021 after they found that TEPCO had failed to take adequate security measures against the threat of nuclear terrorism (see detailed explanation in WNISR2022). The NRA inspected 27 places but said that the plant still had problems in four areas.316 On 28 June 2023, Dr. Shinsuke Yamanaka, Chairperson of NRA, stated that “We need to confirm whether the major violations are being used as lessons learned, and how the organizational culture and safety culture has been affected. If there are any additional safety-related changes in TEPCO’s activities, we would like to see them as well” when NRA investigates further regarding TEPCO’s qualification as an operator of Kashiwazaki-Kariwa nuclear power plant. The NRA maintained its ban on loading fresh fuel because of violation of nuclear security regulations in 2022.317

Japan Atomic Power Co’s Tokai-2 was the first BWR to get lifetime-extension approval from NRA in November 2018 but currently the work for installation of a Specialized Safety Facility (SSF) against terrorism is underway. The facility is planned to be completed in September 2024.318 Tohoku Electric Power’s Onagawa-2 received official approval by NRA of conformity to new regulatory requirements on 26 February 2020, and work on remaining safety measures is expected to be completed in November 2023. It is planned to restart operation in February 2024.319 Chugoku Electric Power Co’s Shimane-2 received approval from NRA on 15 September 2021 and received local governor’s approval in June 2022.320 But because of delay in safety related work, Chugoku Electric Power announced that it will delay the restart of operation until 2024.321

Kansai Electric Power Co (KEPCO) has the largest number of reactors (seven in total, all PWRs) of which five (Mihama-3, Takahama-3 and -4, Ohi-3 and -4) are currently operating (as of July 2023). Mihama-3 license extension to 60 years was granted on 16 November 2016.322 For both Takahama-3 and-4, KEPCO applied for license extension beyond 40 years on 25 April 2023. The current 40-year license will expire in 2025 for both reactors.323

Shikoku Electric Power’s Ikata-3 reconnected to the grid on 26 May 2023 following regular inspection which started on 23 February 2023.324

Kyushu Electric Power Co’s Genkai-3 was shut down on 21 January 2022, and operation of SSF started on 5 December 2022 while the set deadline was 24 August 2022. It was reconnected to the grid on 12 December 2022.

Genkai-4 was shut down on 30 April 2022 for regular inspection and resumed operation on 13 July 2022.325 It was shut down again on 12 September 2022, as it could not meet the SSF deadline of 13 September 2022. SSF was finally available on 2 February 2023 and power generation resumed on 9 February 2023. Both Sendai-1 and -2 applied license extensions beyond 40 years on 12 October 2022. Licenses will expire on 3 July 2024 for Sendai-1 and on 27 November 2025 for Sendai-2.

As of July 2023, Takahama-1326 and -2 were scheduled to restart in fall of 2023 after NRA approved a beyond 40-year operating license for both reactors on 20 June 2016. Work on safety measures was completed on the two units on 18 September 2020 and on 31 January 2022 respectively. The deadline for the installation of SSFs for the two units was 9 June 2021. Takahama-1 is scheduled to resume power generation in early August 2023 followed by Takahama-2 in mid-September 2023.327

As no additional reactor has been declared for permanent closure during the past year, the total number of closed reactors remains unchanged at 27 reactors328 (including 21 reactors closed because of the Fukushima accidents, as shown on Table 9).

Legal Cases Against the Restart of Reactors

The legal cases against operation of existing reactors continue. The following are two key decisions made during the past year, both of which rejected “injunction” appeals made by local residents.

On 24 March 2023, Hiroshima High Court rejected local residents’ “injunction” appeal to stop the restart of Ikata-3 nuclear power plant operated by Shikoku Electric Power Co. Ikata-3 was shut down for regular inspection from 23 February 2023 until 19 June 2023.329 The case was brought by seven residents of Hiroshima and Ehime prefectures who live between 60 and 130 km from the reactor. The main focal issue was whether the operator’s estimate of seismic ground motion was adequate or not. As reported in WNISR2022, the Hiroshima district court dismissed similar requests and ruled against the injunction. The Hiroshima Hight Court followed the district court decision and ruled that Shikoku Electric’s seismic estimate was to be considered adequate.330

On 24 May 2023, the Sendai district court rejected the appeal for injunction against the restart of Tohoku Electric Power Co’s Onagawa-2 nuclear power plant in Miyagi prefecture. Tohoku Electric Power Co plans to restart the reactor in February 2024 after a long shutdown period after the Fukushima nuclear accidents in 2011. The main issue was the adequacy of the evacuation plan. The case was brought by 17 residents of the city of Ishinomaki, claiming the evacuation plans prepared by the city and prefectural government are not sufficient. But the ruling was not based on the adequacy of the evacuation plans, but on the dismissal of the notion of “specific danger” of a nuclear accident the plaintiffs claimed. The Court said that “it cannot be assumed that a specific danger of an accident exists”, as the burden of proof is with the plaintiffs. Noboru Hara, 81-year-old spokesperson for the plaintiffs said they will “consult with lawyers with a view to filing an appeal.”331

Reactor Closures and Spent Fuel Management

No additional reactor(s) operating (or in outage) at the time of the Fukushima events, were formally declared for decommissioning in the year to 1 July 2023. The 11 commercial Japanese reactors now confirmed to be decommissioned (not including the Monju Fast Breeder Reactor and the ten Fukushima reactors) had a total generating capacity of 6.4 GW, representing about 15 percent of Japan’s officially operating nuclear capacity as of March 2011. Together with the ten Fukushima units, the 21 units total 15.2 GW or just under 35 percent of nuclear capacity prior to 3/11 (see Figure 41 and Table 9). In total, Japan has 27 closed reactors (17.1 GW) (see Case study on Japan in Decommissioning Status Report).

Regarding spent fuel from demonstration reactors, on 24 June 2022, the Japan Atomic Energy Agency (JAEA) signed a final €250 million (US$2022263 million)-contract with French company Orano for the transport and reprocessing of spent fuel from the Fugen ATR,332 which first reached criticality in 1978 and was closed in 2003. Work was set to start in 2023 and be completed by March 2027,333 but no update on ongoing works has been communicated as of July 2023. Prior to the final agreement, on 20 June 2022, it was reported that JAEA would transfer to France the plutonium extracted from spent fuel from its Fugen reactor.334 (See WNISR2022 – Japan Focus for more detail).

In March 2022, similar reprocessing contracts with Orano were said to being proposed for spent fuel from the Monju FBR335—which first reached criticality in 1994, was connected to the grid for only three and a half months when it had an accident in December 1995, and was officially closed in 2017—but no official agreement or decision was communicated as of mid-2023. Meanwhile, spent fuel removal has been completed and by 22 April 2022, all spent fuel from Monju had been moved to a temporary storage tank filled with liquid sodium and relocated to a water-cooled storage pool by October 2022.336

JAEA, which manages the decommissioning work of Monju, plans to start the extraction of the liquid sodium from the reactor in 2023, and eventually transfer the spent fuel “to domestic and foreign operators with licenses for reprocessing in Japan or in countries with which Japan has signed agreements for cooperation on the peaceful uses of nuclear energy.”337

On 13 June 2023, Kansai Electric Power Co (KEPCO), along with the Federation of Electric Power Company (FEPCO), announced that they will ship 200 tons of spent fuel (10 tons of spent LWR-MOX fuel and 190 tons of usual spent uranium fuel338) to France for “demonstration” of spent MOX fuel reprocessing. KEPCO promised to Fukui Prefecture that they will remove spent fuel from the prefecture and find a candidate site for interim storage of spent fuel outside of Fukui prefecture by the end of 2023. Although 200 tons is only about 5 percent of the spent fuel KEPCO stores in Fukui prefecture. Nozomu Mori, President of KEPCO, said that “it carries an equal weight to temporary storage in that spent nuclear fuel will be transported out of the prefecture. The promise has been fulfilled for now.” It is not clear whether Fukui prefecture will be satisfied with this explanation and the plan for the rest of spent fuel stored in Fukui prefecture is not known yet.339

1. Table 9 | Official Reactor Closures Post-3/11 in Japan (as of 1 July 2023)

Sources: JAIF and JANSI, compiled by WNISR, 2023

Notes: This table only lists the 22 reactors closed after the Fukushima accidents, thus not including the Fugen Advanced Thermal Reactor (ATR), Japan Power Demonstration Reactor (JPDR), as well as Hamaoka-1 & -2 (Chubu Electric Power) and Tokai-1 (JAPCo).

BWR: Boiling Water Reactor; PWR: Pressurized Water Reactor; FBR: Fast Breeder Reactor; LTS: Long-Term Shutdown.

JAPC: Japan Atomic Power Company; JAEA: Japan Atomic Energy Commission

(a) – Unless otherwise specified, all announcement dates from JANSI, “Licensing status for the Japanese nuclear facilities”, Japan Nuclear Safety Institute, 26 February 2020, see 

, accessed 27 July 2020.

(b) – Unless otherwise specified, all closure dates from individual reactors’ page via JAIF, “NPPs in Japan”, Japan Atomic Industrial Forum,
see http://www.jaif.or.jp/en/npps-in-japan/, as of 27 July 2020.

(c) – Note that WNISR considers the age from first grid connection to last production day.

(d) – WNN, “Shikoku decides to retire Ikata 2”, World Nuclear News, 27 April 2018,
see http://www.world-nuclear-news.org/C-Shikoku-decides-to-retire-Ikata-2-2703184.html, accessed 22 July 2018.

(e) – The Mainichi, “Japan decides to scrap trouble-plagued Monju prototype reactor”, 21 December 2016,
see http://mainichi.jp/english/articles/20161221/p2g/00m/0dm/050000c, accessed 21 December 2016.

(f) – The Monju reactor was officially in Long-Term Shutdown or LTS (IAEA-Category Long Term Shutdown) since December 1995. Officially closed in 2017.

(g) – The decision to close the reactor was announced in October 2018.

Japan Steel Works (JSW) Falsification Incident Update

On 9 May 2022, Japan Steel Works (JSW), a global leading manufacturer of key nuclear reactor components, published a report on the discovery of “inappropriate conduct in quality inspections” at its subsidiary, Japan Steel Works M&E and announced that it would establish a special investigating committee.340

On 14 November 2022, the special investigating committee submitted its findings to company management. The report said that a total of 449 inappropriate conducts, including data falsification of inspection data, and 20 incidents involving components related to nuclear power had been identified341 (see Table 10). Out of 20, six cases were related to French EDF orders, including the nozzle support ring of a steam generator of the Cruas-1 reactor. EDF claimed that its own analysis showed that the integrity of the equipment was not jeopardized.342 On 29 November 2022, JSW issued a statement saying that eight senior executives, including the former and current presidents of the company, will receive corporate punishment (salary cuts by 30 percent for three months).343

On 9 May 2023, Mr. Toshio Matsuo, President of JSW issued a statement on this issue, saying that the company “will reform the system so that no single department manages everything from specifications and delivery dates to even customer relations, in order to transform the organizational structure”.344

1. Table 10 | Typology of Falsification Cases at Japan Steel Works

Source: Japan Steel Works, 2023345

Original notes by Japan Steel Works:

(a) Most of the cases were emergency measures that were triggered by sudden events that occurred in the manufacturing process, a finding that was confirmed in the investigation report by the Special Investigation Committee. There were circumstances that would not have otherwise been a problem if they had been reported to or discussed with the customers, but they were covered up without reporting to or discussing about with the customers, which constitutes a deviation from the procedural specifications sought by customers.

(b) There was no deviation from the specifications agreed on with final customers, but instead from the internal control values of M&E, whose customer is our Company (Hiroshima Plant).

New Energy Policy and the Role of Nuclear Energy

As reported in WNISR2022, in July 2022, Prime Minister Kishida’s government expressed its intention to promote nuclear energy, while the detail of new policy was not known at that time.346 On 10 February 2023, the Cabinet of PM Kishida’s government approved the so-called “Green Transformation Basic Policy” which includes various measures to promote nuclear energy.347 The main stated policy objective is to realize the goal of “Carbon neutrality by 2050” with an investment roadmap for ¥150 trillion (more than US$1.1 trillion) of public-private financing over the next 10 years. One of the main new policies is to “maximize the utilization of nuclear power”. This is the major change from current energy policy which says Japan will “reduce dependence on nuclear energy as much as possible”. The new policy also emphasizes the unstable energy situation caused by the war in Ukraine. Securing a stable energy supply is thus mentioned as a major reason to promote nuclear energy.348

On 31 May 2023, Japan’s parliament passed a bill, so-called “GX bundled bill” which includes amendment of Nuclear Reactor Regulation Law, Electricity Utility Industry Law and Atomic Energy Basic Law. Those three laws specify the main features of the new policy as follows:

• Extension of the “licensing period” (generally 40 years and 60 years for exceptional cases) allowing operators to apply for an extension of “certain shutdown period due to ‘non-technical’ or ‘unplanned’ reasons” (through amendment of the Nuclear Regulation Law and Electric Utility Industry Law)

This has become one of the most controversial issues of the GX Basic Policy. The licensing-period limitation was introduced after the Fukushima accidents primarily for two reasons. One is the safety concern over the aging reactors as Fukushima Daiichi-1 was just 40 years old (it started commercial operation in 1971 and had been given a 10-year lifetime extension one month prior to its accidental destruction) and all six Fukushima Daiichi units started commercial operation in the 1970s. The other reason was to facilitate the nuclear phaseout policy.349 It was argued that there is no scientific basis to determine the lifetime of reactors and thus METI and the utility industry would like to extend the operation period beyond 40 and 60 years from the beginning of power generation for the periods during which reactors were shut down for “unplanned” reasons (beyond regular inspection period due to non-technical reasons such as licensing activities or socio-political reasons). In 2020, Japanese utilities filed a similar request with NRA before, but NRA rejected their request saying in July 2020: “It is difficult to determine extension period based on scientific and technical reasons as safety assessment should be made considering conditions of reactor by reactor”.350

However, on 5 October 2022, NRA accepted METI’s proposal to amend the lifetime extension regulation. NRA chairman Shinsuke Yamanaka said at a press conference that “extending operational period is a matter of energy policy and NRA is not in a position to comment” quoting the same July 2020 statement351. On 21 December 2022, NRA decided on possible changes in safety regulation for lifetime extension, preempting the amendment made by METI.352 On 14 February 2023, NRA voted to accept the amendment of the Nuclear Regulation Law to allow METI to give approval for the extension of the operating period. It was unusual for NRA to take a vote as typically decisions are made on a consensus basis. But this time, one of the Commissioners, Akira Ishiwatari opposed the revision, saying NRA has not yet specific regulations for an entire 60-year operational lifetime and it is not logical and very strange that the longer the NRA takes to conduct a rigorous inspection the longer the operating period of a reactor life will be, as the inspection outage would not be included in the lifetime calculation.353 Another commissioner, Tomoyuki Sugiyama said he felt the discussion was “rushed” as a result of government pressure. But NRA chairman Shinsuke Yamanaka denied that NRA yielded to government pressure.354 Then it was revealed that NRA staff and METI officials met privately several times to discuss amendments of Nuclear Regulation Laws without consulting NRA commissioners or keeping any records.355 This is apparently against rules No. 1 (Independence) and No. 3 (Openness and Transparency) of NRA’s Guiding Principles.356 But the law passed on 31 May 2023, and now the METI Minister can determine lifetime extensions based on the condition that the reactors will pass the NRA safety review. NRA will review the conditions of reactors at least every 10 years after 30 years of operation.357

• Clarification of government’s responsibility to support:

a) the utility industry to build and construct innovative advanced reactors;

b) nuclear industry to maintain and strengthen industrial base;

c) smooth operation of decommissioning and disposal of radioactive waste (through amendment of the Atomic Energy Basic Law).

This amendment to the Atomic Energy Basic Law has attracted attention as it is unusual to introduce specific policy measures into basic framework legislation. The Atomic Energy Basic Law was passed in 1955 and was treated like a “Constitution for Atomic Energy” as it stipulates three basic principles (Autonomous, Democratic, Open) as well as guarantees that atomic energy is used only for peaceful purposes. However, this amendment clarifies “Government Responsibility” to support promotion of nuclear energy such as: assist electric utility industry by making “institutional arrangements” for building new reactors when they face difficulties under liberalized-market conditions. Some experts claim the amendment is against the spirit of the Basic Law and may lead to unnecessary tax money spending as well as to the reemergence of the “safety myth”.358

• Institutional enhancement for decommissioning and radioactive waste disposal (through amendment of the Reprocessing Fund Compulsory Contribution Law)

This is similar to the obligation established by the Reprocessing Fund Compulsory Contribution Law which requires nuclear utilities to contribute an annual reprocessing and MOX-fabrication fee for spent fuel generated. Now the nuclear utilities are required to contribute a certain fee to cover future decommissioning costs. The amended law also added decommissioning of commercial nuclear reactors to the missions of the Nuclear Reprocessing Organization (NURO).359

In response to the first passage of the “GX Bundled Bill”, several civil society organizations have raised their voices against the bill. For example, Citizens’ Nuclear Information Center (CNIC), one of the leading anti-nuclear organizations, issued a statement on 28 April 2023, entitled “GX Nuclear Power Plant Bill Passed by Japanese House of Representatives After Diet Deliberations Full of Deceit and Fabrication”.360 Another leading civil society platform, the Citizens’ Commission on Nuclear Energy (CCNE), also initiated a campaign to oppose the GX Bill calling for signatures from researchers and experts on this issue with 21 experts supporting the “emergency appeal” and hundreds of individuals and experts joining the campaign.361

Prospects for Nuclear Power

The new nuclear energy policies introduced under the GX Transformation laws represent a major shift as they allow for the construction of new reactors in Japan for the first time since the Fukushima disaster. It also amends the nuclear regulation laws to allow for lifetime extensions beyond 60 years. These new policies, which aim to maximize the use of nuclear power, are in fact inconsistent with the policy to reduce dependence of nuclear power as much as possible as stated in the current Energy Basic Plan.

A recent public-opinion survey suggests that support for the restart of existing reactors exceeds opposition to restarts for the first time since 3/11.362 However, at least in the short term, it remains unclear how these new policies would change the conditions for utilities to restart reactors, and it is even less certain what the impact on the potential construction of new reactors could be. In addition, many issues associated with the decommissioning of the Fukushima Daiichi reactors remain unresolved (see Fukushima Status Report). Also, legal cases against reactor restarts and in favor of compensation for the impact of the Fukushima disaster continue. In short, the future of nuclear power in Japan is still far from certain.

Poland Focus

Poland planned the development of several nuclear power stations in the 1980s and started construction of two VVER1000/320 reactors in Żarnowiec on the Baltic coast, but both construction and further plans were halted following the Chernobyl accident in 1986.363 Since then, there has been a long, expensive, and time-consuming series of attempts to restart the program.

Once again, in 2008, Poland announced that it was going to re-enter the nuclear arena.364 The Council of Ministers adopted a resolution providing for the development of a nuclear power program in January 2009, and the “Polish Energy Policy until 2030” in November 2009, which set a roadmap for the inclusion of nuclear to the country’s energy infrastructure. The policy assumed that by 2030 three units (4.8 GW) would generate “over 10 percent” of the country’s electricity, with the first unit put into operation “no[t] sooner than in 2020”.365 The following years saw negotiations with potential vendors, successive revisions of the project, various announcements, and delayed decisions (see past WNISR editions).

On 28 January 2014, the Polish Government adopted the “Polish Nuclear Power Programme” outlining the framework of the strategy. The plan included proposals to build 6 GW of nuclear power capacity at an estimated cost of PLN40–60 billion (US$201412.6–19 billion), with the first reactor starting up by 2024 and two units operating by 2035. A first site was to be named by 2016.366 That did not happen.

Prior to the Government’s 2014 strategy publication, state-owned utility Polska Grupa Energetyczn (PGE) had followed earlier attempts by declaring plans to build two nuclear power reactors in 2009. By February 2012, PGE’s supervisory board ratified a strategy plan for 2012 to 2035 that included the construction of two reactors with a total capacity of 3 GW, with the first envisioned to be operational by 2025. Together with two other state-owned utilities Tauron Polska Energia and Enea, in cooperation with copper supply company KGHM Polska Miedz, PGE had agreed in 2013 on the supply of shares of PGE EJ1, a subsidiary of PGE that had been set up for the construction and operation of a potential new plant.367

In March 2017, PGE EJ1 launched site selection studies at Lubiatowo-Kopalino and Zarnowiec, both locations are close to the Baltic coast in the northern province of Pomerania.368 A year later, rumors circulated on PGE corporation’s declining interest in nuclear development as the company had supposedly shifted its attention towards offshore wind farms.369 Nonetheless, the push for a nuclear strategy continued, and in November 2018, the Government published a draft strategic energy development program, which called for the construction of up to four reactors (providing 4–6 GW of capacity) by 2040, with the first in operation by 2033, and up to a total of six units with a combined capacity of 6–9 GW to be put into operation by 2043.370 In May 2019, the Ministry of Energy envisaged the site selection for the first plant in 2020, while the technology would be chosen in 2021.371

In October 2020, the Council of Ministers adopted a revised long-term Polish Nuclear Power Program.372 It maintains the objective to build and commission nuclear power plants in Poland with a total installed capacity of approximately 6–9 GW based on Generation III (+) pressurized water reactors, with the start of operation during the 2030s, while the share of nuclear power in the electricity mix is predicted to reach about 20 percent by 2045. According to the documentation, the timetable was as follows:

• 2021: choice of technology for the first (EJ1) and second plant (EJ2);
• 2022: site license for EJ1;
• 2026: building permit and construction start of EJ1;
• 2028: site license for EJ2;
• 2032: building permit and construction start of EJ2;
• 2033–2037: operating license by the President of the National Atomic Energy Agency (PAA) and commissioning of three units (EJ1);
• 2038–2043: operating license by President of PAA, and commissioning of three units (EJ2).373

In the same month, the U.S. and Polish governments signed an agreement on the “cooperation towards the development of a civil nuclear power program and the civil nuclear power sector in […] Poland”. The agreement includes cooperation plans on the development of financing regulations and schemes, technological knowledge transfer, and the “development, construction, and financing of the first [nuclear power plant] project, intended to be operational during 2033.” The agreement came into force in February 2021.374 In June 2021, a first grant was issued by the U.S. Trade and Development Agency to fund a front-end engineering and design study for Polskie Elektrownie Jądrowe (PEJ).375

PEJ is the direct descendant of PGE EJ1. In March 2021, the four owners PGE (70 percent of shares), Enea, Tauron and KGHM (10 percent each) had sold ownership to the Polish State Treasury “in preparation for reali[z]ation of the Polish nuclear power [program]”. Negotiations had begun in October 2020, and the transaction cost the Treasury around PLN531 million (US$2021137.5 million).376 In June 2021, “PGE EJ1” was renamed “Polskie Elektrownie Jądrowe”, or “PEJ”.377

In late December 2021, PEJ announced it had chosen the village of Choczewo in Pomerania for the first reactor.378 In March 2022, PEJ submitted the Environmental Impact Assessment report for the project.379

Reportedly, the actual offers submitted between October 2021 and September 2022 included the plans of Korea Hydro & Nuclear Power (KHNP) for six APR-1400 (8.4 GW) for US$26.7 billion, Westinghouse’s proposal to build six AP-1000 (6.7 GW) for US$31.3 billion, and EDF’s preliminary offer of four to six EPRs (6.6–9.9 GW) for US$33–48.5 billion.380

In May 2022, KHNP Deputy CEO Lim Seung-yeol told the Polish Press Agency, the company would envisage taking a 20–30-percent equity stake in the newbuild project, which “would be […] KHNP’s direct contribution to the investment. The rest would be covered by financial institutions. On the Korean side, it would be export credit-agencies.”381 It remains unclear whether the offer to inject capital would cover the first three units only or the entire package of up to six APR-1400. In any case, the Korean initiative represented a financing offer that would be difficult to match for EDF or Westinghouse.

Regardless, in November 2022, Westinghouse was formally appointed as the contractor to deliver three reactors to the Pomeranian project at costs of around US$20 billion.382 In January and September 2022, Westinghouse had already signed MoUs with 10 then 22 Polish supply companies, for cooperation on various potential tasks such as steel manufacturing, translation services and machine maintenance.383 Given that KHNP’s initial offer was cheaper by several billion US$, it is understood that the decision is of a more geopolitical nature, i.e. to strengthen ties between the governments of Poland and the U.S.384 However, as discussed below, South Korean actors might come to build nuclear reactors in Poland after all. Opposition to the project was voiced by four East German states (Brandenburg, Saxony, Mecklenburg-Vorpommern, and Berlin) during the consultation period of the environmental impact assessment process.385 Nonetheless, cooperation agreements were signed between Westinghouse and PEJ in December 2022.386 These were further advanced when in February 2023, a contract covering front-end engineering, early procurement work and program development was signed between Westinghouse and PEJ,387 followed in May 2023 by an agreement “defining the principles of the parties’ [Westinghouse, PEJ and Bechtel] cooperation in the design and construction of Poland’s first nuclear power plant.”388 On 13 April 2023, PEJ had applied to the Ministry of Climate for a “decision-in-principle” on the project,389 which was granted in July 2023, allowing for further administrative applications to proceed.390 At this stage, construction work is planned to begin in 2026, with electricity generation to commence in 2033.391

In parallel, in a notable development, in October 2022, Polish utility Zespół Elektrowni Pątnów-Adamów-Konin (ZE PAK) and PGE as well as KHNP signed a letter of intent to develop plans for a second nuclear power plant based on KHNP’s APR-1400 technology in Pątnów, central Poland, at the site of a lignite power plant. On the same day, Poland’s Minister of Assets, the Deputy Prime Minister, and South Korea’s Minister of Trade, Industry and Energy also signed a Memorandum of Understanding (MoU) “to support the nuclear energy project in Patnow [Pątnów] and tighten cooperation in the scope of necessary information exchange”. This nuclear plant would constitute the second phase of the 6–9 GW nuclear capacity envisioned in Poland’s Nuclear Power Program from 2021.392 The project however might come under E.U. investigation due to possible noncompliance with competition regulation that requires multiple equally treated bidders to be allowed to compete for such large infrastructure projects.393 Regardless, ZE PAK and PGE announced in March 2023 they would establish a joint venture to “represent the Polish side at all stages of the [Pątnów] project”, now planned with at least two APR-1400 reactors delivered by KHNP, scheduled to be on the grid by 2035.394 This joint-venture, named PGE PAJ Energia Jądrowa, was established with a 50/50 share by both companies in April, and in August 2023, submitted an application to the Polish Ministry of Climate for a “decision-in-principle” on the construction of a nuclear power plant consisting of two APR-1400 reactors.395 In the meantime, South Korean and Polish firms signed six MoUs relating to nuclear generation at the Korea-Poland Business Forum held in Warsaw in July 2023. Two of those were signed between Doosan Enerbility and Polish companies on the construction of nuclear power plants in Poland.396

In an attempt to block KHNP’s participation in the competition (and possibly hinder KHNP’s further expansion to other Eastern European Countries, e.g., the Czech Republic) Westinghouse filed a lawsuit against KHNP and its owner Korea Electric Power Corp. (KEPCO) before the U.S. Federal Court in October 2022.397 Westinghouse argues that KHNP is infringing on intellectual property rights owned by Westinghouse regarding “System 80 reactor technology” that were originally held by Combustion Engineering, a company that was taken over by Westinghouse in 2000.398 Arguably, KHNP would require permission to export this technology, to which KHNP states that all necessary regulations had been followed. 399 An attempt to settle the decades-old dispute outside of judiciary was made in January 2023 by KHNP and KEPCO by suggesting a split of potential profits of a nuclear project with Westinghouse.400 The parties had until 17 March 2023 to come to some form of agreement which did not happen.401 The Korean Commercial Arbitration Board begun assessing damages claimed by both sides, possibly amounting to several hundred million US$, in August 2023.402

In addition to negotiations around potential orders of large reactors, Poland eyes the possibility of investing in Small Modular Reactors (SMRs). Various cooperation agreements have been signed including between the Polish state-owned company Enea S.A. and U.S. SMR developer Last Energy to cooperate on the deployment of SMRs.403 In April 2023, the U.S. Export-Import Bank and the U.S. International Development Finance Corporation both signed letters of interest to provide loans, up to US$3 billion and US$1 billion respectively, to the Orlen Synthos Green Energy (OSGE) project.404 The project emerged in March 2022, when PKN Orlen, Poland’s largest oil company, joined forces with Synthos Green Energy to “invest in the development of micro and small modular reactor technologies”.405 In March 2023, GE Hitachi (GEH), Tennessee Valley Authority (TVA), Ontario Power Generation (OPG) and OSGE agreed to collaborate on the global development of the GEH BWRX-300 reactor. In June 2023, OPG and OSGE separately signed a letter of intent to cooperate on various SMR-related activities.406 The project envisions the construction of up to 20 BWRX-300 reactors in Poland, with launch of the first one expected in 2029.407

Polish efforts to become a country operating commercial nuclear power plants have intensified over the past several years. The planned parallel implementation of three different technologies (Westinghouse’s AP-1000, KHNP’s APR-1400 and SMRs) in a country that has only little experience in the construction of nuclear power plants (dating back four decades), their operation, and the regulation thereof seems ambitious.408 Whether Poland will be able to pull off these plans, especially given that many details on contracts and financing remain undisclosed, remains uncertain.

The Polish electricity mix is highly dependent on coal, which contributed 69 percent to the electricity mix in 2022, followed by wind (11 percent), natural gas (7 percent), and solar (4.5 percent). The remainder is generated from various other fossil and renewable sources such as bioenergy and hydro.409

The extension of onshore wind capacities ceased in 2016 when restrictive distance laws (“10H legislation”) essentially brought onshore newbuild to a standstill. By 2022, only a total of about 8.3 GW had been installed. A 2022-amendment of the law might foster some project development, while the Government’s target lies at only 14 GW by 2030 and 20 GW by 2040. The first offshore wind farm is expected to come online in 2026, and a total of 12 GW of offshore capacity is planned.410

In comparison, solar energy is rapidly gaining significance. Over the course of 2022, solar capacity grew from 7.7 GW in 2021, to 12.4 GW, a 61 percent increase. For context, in 2019, solar generation accounted for only 0.4 percent of Polish electricity, an 11-fold increase of the solar share in three years. Provisional announcements of updates to the Polish Energy Strategy envision a total of 27 GW to be installed by 2030.411

Russia Focus

In 2022, nuclear energy contributed 20 percent to the country’s electricity mix, with another record production of 209.5 TWh, up from 208.5 TWh in 2021. 2022 did not see the startup or closure of any reactors, and as of mid-2023, 37 reactors were operating, and ten have been permanently closed.

There are five reactors under construction in or for Russia, including two barges built in China but destined to Russia. Two are large units at Kursk II, a significant project, as it involves the first of the latest Russian design, the VVER-TOI (VVER-V-510), officially expected to cost around US$3.5 billion, although this is likely to be a significant underestimate.412 These are 1200 MW, Generation III+ design, and are also earmarked for export. When construction started on Unit 1, project completion was scheduled for late 2023, and in April 2020, the first deputy director for construction claimed that the project was on schedule.413 In November 2022, plant director Alexander Uvakin was quoted as saying “We hope that 2024-2025 will see the physical start-up and commercial operation of the first and then the second unit of the Kursk-II NPP”.414 While no completion date has been confirmed, in July 2023, Rosatom announced that the last structural element was installed, and therefore, completion is likely to be some way off.415 In February 2023, public hearings began on the planned building of Units 3 and 4.416

Construction of an innovative SMR fast reactor design using liquid lead as a coolant and uranium-plutonium nitride for fuel started in June 2021. The objective for the BREST-OD-300 reactor is for it to operate by 2026, and it is said to cost 100 billion rubles (US$20211.4 billion).417 In June 2020, Rosenergoatom announced that preparation work would begin for the construction of four new reactors, Units 3 and 4 at Leningrad II (also referred to as Leningrad-II NPP Units 7 and 8 when including the previous four RMBK reactors), as well as two reactors at Smolensk II.418 In December 2022, concrete was poured for the first buildings for the new units at Leningrad, which are due to be completed at the end of 2023, after which formal construction on Unit 7 could begin in 2024.419 It is unlikely that they will begin generating electricity this decade. The last reactor to start up in Russia, Leningrad 2-2 in 2020, took 10.5 years to build.

In August 2022, Rosatom announced the keel-laying ceremony—considered construction start for floating reactors—in China of the first Arctic-type Nuclear Floating Power Unit (NFPU) to be equipped with two RITM-200C reactors and to be deployed in Russia, in the framework of the Cape Nagloynyn project.420

In March 2021, in its strategic review, Rosatom said that by 2045, nuclear energy should provide 25 percent of the country’s electricity. According to Rosatom CEO Alexei Likhachev, this will require the commissioning of 24 blocks, including at new sites and in new regions.421 Rosatom reiterated its intentions in May 2022. The list of sixteen new reactors in the plan for 2035 includes:

• Kursk-II: Units 1–4; Leningrad-II: Units 3 & 4 (VVER-1200 reactors);
• Smolensk-II: Units 1 & 2 (VVER-TOI reactors);
• Baimsky GOK: four modernized FNPP units (RITM-200 reactors);
• Small reactor in Yakutia: Unit 1 (RITM-200 reactor);
• ODEK in Seversk: BREST-OD-300;
• Kola-II: Unit 1 (VVER-S or VVER-600 reactor);
• and Beloyarsk: Unit 5 (BN-1200M fast reactor),422

the majority of which will be at or close to existing nuclear power plant sites, although these include three new sites in Biamsky and Yakutia (in the far East), and the proposed Seversk facility in the Tomsk oblast, a closed city and site of military nuclear facilities.

Russia has closed ten power-generating reactors: Beloyarsk-1 and -2, Bilibino-1, Leningrad-1 and -2, Kursk-1, Novovoronezh-1–3, and Obninsk-1, with a further ten units to potentially close by 2030 without operating lifetime extensions.423

The average age of the Russian reactor fleet is 29.9 years as of mid-2023, with close to two-thirds being 31 years or more, of which 12 operated for 41 years or more (see Figure 43). Therefore, a vital issue for the industry is managing its aging units.

There are six classes of reactors in operation: the RBMK (a graphite-moderated reactor of the Chernobyl type), the VVER-440, the VVER-1000, the VVER-1200, the KLT-40 and FBRs. Designed for an operational lifetime of 30 years, both the RBMKs and VVER-440 designs have been granted 15-year lifetime extensions to enable them to operate for 45 years. The process of annealing, whereby the reactor pressure vessel is heated for an extended period of time has been undertaken in VVER 440 reactors in Armenia, Bulgaria and Ukraine and is hoped to extend the operating life of the vessels until up to 60 years,424 while the VVER-1000s are expected to work for up to 50 years. Consequently, the closure of Leningrad-1 and -2 after 45 years of operation, in 2018 and 2020 respectively, is potentially a significant event, as it could indicate that a 60-year operational lifetime is beyond the RMBK potential. The current operating licenses for Units 3 and 4 expire in 2025 and 2026 when they are likely to close.425 At the same time, the RMBKs at Kursk are also closing after 45 years, with Unit 1 closed since 2021 and Unit 2 set to close in 2024.426

1. Figure 43 | Age Distribution of the Russian Nuclear Fleet

Sources: WNISR, with IAEA-PRIS, 2023

The country also operates two Fast Breeder Reactors (FBRs) at Beloyarsk (Units 3 and 4). The older and smaller of the two reactors is a 600 MW unit, which was connected to the grid in 1980 with an expected operational lifetime of 30 years. This was extended for the second time in April 2020 for a further five years to enable the unit to operate until 2025, but plans are being developed to enable the unit to operate for 60 years.427. The new VVER-1200 reactors in Novovorenezh II and Leningrad II have a design lifetime of 60 years, with some studies said to enable up to 100 years of operation for their pressure vessels. A new “ultra pure” nickel material would even allow for 120 years of irradiation of future vessels, Rosatom claims.428

Russia is an aggressive exporter of nuclear power, with, according to one report from Rosatom, 33 separate projects in various stages of advancement.429 These claims must be taken with some skepticism, as the same source claims seven reactors are under construction domestically when, by most accounts, there are three, plus the two barges for floating reactors that are being built in China. As the reactors will be added to the barges in Russia and the plant is to operate in Russia, WNISR considers this as a domestic Russian project. As of 1 July 2023, Rosatom is involved as the main contractor of the following projects abroad in various stages of active construction:

• Bangladesh – Construction started on two reactors at Rooppur in 2017 and 2018, which were expected to begin operation in December 2023, but commissioning will not start before 2024 and more likely 2025.430 See section on Bangladesh in Potential Newcomer Countries.
• China – Two reactors each at Tianwan and Xudabu (or Xudabao). Construction started for the respective first units in 2021 for the respective second units in 2022.431 See China Focus.
• Egypt – Three reactors are under construction at El Dabaa, with the fourth expected to start construction in late 2023. The plant is supposed to be fully operational between 2028 and 2031432 and cost US$30 billion433. See section on Egypt in Potential Newcomer Countries.
• India – Four reactors are under construction at Kudankulam. Construction started on the first of the units in June 2017 and on the most recent ones in December 2021. Completion of the first of these units is supposed to be reached in 2025.434 See section on India in Annex 1.
• Turkey – Four reactors are being built at Akkuyu. Construction started on the first unit in 2018 and on Unit 4 in 2022. Unit 1 is now supposed to start in 2024, but commercial operation appears to be delayed to 2025.435 See section on Turkey in Potential Newcomer Countries.
• Iran – Construction of Bushehr-2 (also called Busheer-2) originally started in February 1976 by the German company KWU-Siemens and was suspended in 1978. Work resumed in 1996, with Rosatom subsidiary ASE as the nuclear island provider. In 2022, completion has been delayed to 2026. See section on Iran in Annex 1.
• Slovakia – Mochovce-4, a Russian VVER design that started construction in 1985, is being completed by an international consortium and scheduled to finally be commissioned in 2024.436 See section on Slovakia in Annex 1.

In addition, negotiations continue with Hungary around the construction of Paks II, which has been delayed and is now not expected to be completed until 2032.437 The European Commission gave its approval for the contract changes for Paks II in May 2023, despite the ongoing conflict between Europe and Russia on energy and on the war in Ukraine.438 As of July 2023, Rosatom’s subsidiary JSC ASE was carrying out preparatory work onsite.439 The Rosatom list also includes a nuclear reactor to be built in Finland but, due to the invasion of Ukraine, the consortium in Finland cancelled the project.440

It remains clear that Rosatom is the primary constructor and exporter of reactors with, as of mid-2023, building 24 out of the 58 constructed around the world (see Figure 12 and Table 2).

The relative success of Russia’s export drive in a niche market of state-funded projects is not primarily the technology but the access to cheap financing accompanying the deals. According to Rosatom, it sold US$10 billion of products in 2022, an increase of 15 percent on the previous year and has an overseas order book of US$200 billion over 10 years.441 While the value of its order book is likely to be overinflated, Rosatom is clearly pushing to remain the most influential exporter of nuclear technologies and fuel chain facilities, a ‘full-service’ package—as one commentator described it: “Russian nuclear power is on a roll.”442

Nuclear Interdependencies and Sanctions

In April 2023, the U.S. Government expanded its ‘Russia sanctions’ to Rosatom subsidiary Rusatom Overseas, which is—or at least was—in charge of implementing the construction projects of nuclear power plants in other countries (see section above).

While the E.U. has introduced eleven different rounds of sanctions against Russia, despite many of these addressing the energy industry, these have not included measures against the nuclear sector, despite the ongoing trade in electricity, nuclear fuel, and fuel chain services.

In February 2023, the European Parliament passed a resolution that called for the expansion of the sanctions again to include individuals and entities present on the E.U. market, including Rosatom.443 However, despite initially suggesting it would propose sanctions against the Russian commercial nuclear sector, the European Commission was reported to have abandoned such plans in February, and none have subsequently been applied.444 There is one exception, that is the sanctions decided in February 2023 against Atomflot, a Russian company that maintains Russia’s nuclear icebreaker fleet, also sanctioned by other countries including the U.S., U.K., and Canada.445 The reason given by the European Council read:

The icebreaker fleet managed by Atomflot is designed specifically to meet Russia’s maritime transportation objectives along the Northern Sea Route—the Arctic shortcut between Europe and Asia. The Northern Sea Route has emerged as a new strategic opportunity for unlocking and monetising Russia’s vast oil and gas reserves in the Arctic, thereby providing a substantial source of revenue to the government of the Russian Federation.

There are many economic and political reasons for the European inaction otherwise. According to the World Nuclear Association, Russia supplies about a fifth of all uranium conversion services and 46 percent of enrichment globally, as well as 5 percent of the world’s uranium production, but Kazakhstan provides 43 percent and Uzbekistan 6.6 percent (both of which are significantly influenced by Russia).446 According to an analysis published by the U.K. think tank Royal United Services Institute for Defence and Security Studies, based on customs data, in the year since the start of the war, Russia exported nuclear technologies and fuels worth over US$1 billion. This includes a significant increase to China, with other increases in trade with Hungary, India, and Turkey.447

Rosatom provided 31 percent of uranium enrichment services to E.U. nuclear utilities in 2021448 and represented the largest foreign provider at 24 percent to U.S. nuclear operators in 2022.449 According to the Euratom Supply Agency, the nameplate capacity of uranium conversion and enrichment plants in the E.U. would be “sufficient for the EU to be self-dependent”, but the “Global West” would be missing enrichment capacity of 3,500–8,000 tSWU (thousand Separative Work Units) without Russia. The Agency warns that the construction of “additional conversion and enrichment capacity will take several years.”450 E.U. and U.S. nuclear utilities alone have an annual enrichment service-need of about 24,000 tSWU.451

Furthermore, as there are five E.U.-countries—Bulgaria, Czech Republic, Finland, Hungary, and Slovakia—operating in total 19 Soviet-designed VVER reactors, the diversification of fuel supply is more complex. Westinghouse has become a fuel supplier in Ukraine and Framatome is in the ranks to start fabricating VVER fuel. However, Westinghouse’s experience had been limited to manufacturing VVER-1000 fuel, while 15 of the units in the E.U. are VVER-440 that still use different fuel and no non-Russian company has yet delivered any assemblies of that design (see Table 11). All of the countries dependent on VVER fuel have also a relatively high share of nuclear power in their respective power mix, between one third and almost 60 percent.

Westinghouse is confident to supply a first batch of VVER-440 fuel to Ukraine by the end of 2023.452 Details on agreements about the transfer of design property-rights are unknown, thus the level of ongoing dependence on Rosatom is unclear.

1. Table 11 | Operating Soviet-designed Reactors in Europe (as of mid-2023)

Sources: WNISR, with IAEA-PRIS, 2023

How Dependent Remain Non-Russian VVER-Fuel Manufacturers on Rosatom’s Cooperation/Good Will?—General Legal and Technical Observations

Background–Ever since Russia’s invasion of Ukraine in 2014, efforts have accelerated to extend and develop alternatives to nuclear fuel assembly supplies for Russian designed reactors in the Ukraine (15) and in the European Union (19). Westinghouse was the first to offer VVER-1000 fuel for Ukraine and the Czech Republic. This did not go without technical glitches, and the Czech operator had switched back to Russian fuel while it remains unclear whether price or performance drove the decision. A series of new agreements and contracts has been signed between Westinghouse and Framatome—a newcomer to the VVER-fuel market that has yet to produce its first assembly—and VVER operators in the European Union and Ukraine. Does this mean future independence of the original Russian designer?

Enterprises, particularly technology companies, must continually seek competitive advantage to secure their existence and profitability. This advantage primarily resides in the collective knowledge of their employees, who understand the company’s products and manufacturing processes. Additionally, national laws and privileges may play a role, such as exclusive access to essential raw materials. A company’s knowledge is embodied in its employees, especially engineers and technicians who have both product and manufacturing expertise. Complex products, like nuclear reactor fuel assemblies, require years of development and repeated attempts to accumulate the necessary knowledge. As has been seen in Westinghouse’s attempts to build fuel elements for Russian VVER-1000 reactors, it takes many years and many failed attempts until the necessary knowledge has been obtained.

To safeguard their knowledge, companies keep it secret, invest in research and development, and protect inventions with patents. Patents provide exclusive rights for 20 years, allowing the patent holder to take legal action against potential infringers.

This applies more when a company expands its business through a joint venture with a partner to prevent “leakage” of the company’s knowledge to the partner beyond the joint venture. This is likely to be the case for the reported establishment of a joint venture between TVEL and Framatome in France with the purpose of manufacturing VVER-fuel in Germany. It should be noted that Framatome could have adopted Westinghouse’s technology, but given Westinghouse’s problems to develop this technology, Framatome decided to cooperate with TVEL, despite Russia’s war in Ukraine and potential sanctions against Russian companies.

Framatome’s hopes for complete knowledge transfer and independence are unlikely to be fulfilled, as TVEL has a strong interest in maintaining control over its expertise. This dependency is further reinforced by legal instruments such as patents and contracts.

In conclusion, the competitive advantage and knowledge protection are vital for the survival and success of enterprises, particularly in technology-related fields, such as nuclear fuel management. Joint ventures may offer collaboration opportunities but should be approached with a clear understanding of knowledge protection and the interests of the involved parties.

Russia has also developed into a major hub for nuclear education. In a recent official statement, Russia’s Ministry of Foreign Affairs claimed: “We actively participate in the training and retraining of personnel for the nuclear power industry. Over 2,000 students from 65 countries study at Russian universities specializing in nuclear and related disciplines.”453

As Russia has turned into the dominant supplier of reactor technology in the world—in fact, all 11 construction starts in the world outside China since the construction of Hinkley Point C officially began in the U.K. in 2019 and up to mid-2023, were carried out by the Russian industry (see Overview of Current New-Build)—component suppliers also largely depend on Russian projects. Examples of this co-dependency include the nuclear turbine manufacturer GEAST in France. GEAST produces the Arabelle turbines, is thus highly dependent on the niche virtually entirely controlled by the Russian nuclear industry over the past three and a half years. Reportedly, Rosatom represents about half of the GEAST turnover.454 It was therefore no surprise that, just prior to Russia’s invasion of Ukraine, the French government had offered to sell Rosatom a 20-percent share in the company.455 The project is currently on hold.

EDF subsidiary Framatome originally planned to set up a joint-venture company with Rosatom subsidiary TVEL for the manufacturing and marketing of VVER fuel elements in its Lingen plant which is located in Germany. But when it became clear in spring of 2023 that the German government would likely oppose the deal, the Franco-Russian company was set up in France with a 25-percent participation of TVEL.456 Whether the Lingen plant—that continues to import Russian uranium—will be able to start the manufacturing of VVER fuel elements remains open. Framatome subsidiary Advanced Nuclear Fuels (ANF) that operates the Lingen plant has submitted a licensing application for the extension of the manufacturing plant with a dedicated VVER-fuel production line. While the Lower Saxony government, which acts as the local licensing authority is opposed to the project,457 it can only examine the application based on the Atomic Law but does not have a veto right. It is up to the federal government to greenlight or block the initiative. That decision had not been taken as of mid-2023.

The German electronics giant Siemens in cooperation with Framatome has contracted Instrumentation and Control (I&C) equipment to Rosatom for the four Akkuyu reactors in Turkey (under construction) and for the Paks II project in Hungary (in planning) as well as a range of other Russian reactor projects around the world, including in Russia itself.458 In the case of the Turkish project, apparently, the German authorities have not yet issued any export license for the items in question.

Interdependencies between western and Russian nuclear industry interests cover all the fuel chain elements and reactor-related activities, from uranium mining to the backend services. In 2009, Rosatom acquired 100 percent of German company NUKEM Technologies focusing on Engineering and Consulting especially in Decommissioning and Waste Management services.459

South Africa Focus

South Africa, the only country in Africa currently operating a nuclear plant, is at a crossroads in its energy trajectory. In the past year, the country has been experiencing persistent power cuts and record electricity shortfalls that at times exceeded 6 GW.460 South Africa is also one of the leading global carbon emitters.461

The economically devastating power cuts have turned electricity security into probably the leading item of public discourse and is set to dominate political debates in the lead-up to the national elections in mid-2024. Discussions around the future of the existing almost 40-year-old Koeberg nuclear power plant and the possible construction of further nuclear facilities have therefore also become more widespread.

During the year covered in this report, Koeberg has been partly shut down for major maintenance and upgrading work aimed to secure a 20-year lifetime extension beyond its originally projected 2024 closure date (see Figure 45). The shutdown has exacerbated the national electricity crisis, especially as the work is taking considerably longer than projected. The costs to the financially severely constrained national electricity utility Eskom are also proving much higher than originally announced.

In 2010, a 9.6 GW mega nuclear newbuild program had been touted as a solution to South Africa’s looming electricity shortfall, but this initiative failed, partly due to financial considerations and industrial issues, but also because of the extremely controversial manner in which the project was driven. The process to take the program forward was eventually declared illegal and the entire initiative was effectively terminated. The present power crisis has however led to a revival of intense lobbying for new nuclear power plants, including the promotion of small modular reactors (SMRs). In parallel to the pro-nuclear lobbying, there has also been persistent campaigning against renewable energy, and these activities have succeeded in securing support for nuclear solutions among influential individuals, including many politicians.

Historical Background

The history of nuclear power in South Africa started with the establishment of the South African Atomic Energy Board in 1944. The coming to power of the National Party in 1948 and the subsequent institutionalization of the policy of Apartheid led to increasing international isolation and ostracization of South Africa. Partly in reaction to this, the forty years thereafter saw growing prioritization of the development of the South African nuclear sector, both civil and military. The Apartheid government saw this as both a deterrent to potential military action against the South African state as well as an energy backup option in an environment of growing international sanctions. The building of Koeberg, the first nuclear plant in Africa, started in the late 1970’s, was an outcome of this strategy.462

When the great transition to democracy happened, culminating in the 1994 national elections, and in a global environment favoring nuclear de-escalation, South Africa had already voluntarily relinquished its nuclear military capacity. South Africa remained however a nuclear power producer, and this led to a continued push for the expansion of the nuclear capacity in the early years of the 21st century, culminating in an official plan being adopted in 2010–2011 for a massive nuclear newbuild of 9.6 GW.463 The attempted implementation thereof, especially in an environment where the public became increasingly agitated by perceived corruption and cronyism in government, galvanized into ultimately successful opposition to this nuclear newbuild proposal. Since then, South Africa has however slipped into a worsening electricity crisis, and lobbying for new nuclear plants is again increasing.

Nuclear Capacity in South Africa

South Africa operates one nuclear plant at Koeberg with a capacity of 1854 MW. It also hosts a large facility in Pelindaba with an associated small reactor that is used for research, development, nuclear waste disposal studies, and isotope production. During the Apartheid era up to 1994, the country had a nuclear weapon program that was however terminated in 1989. South Africa has also for many decades now been associated with the mining and export of uranium.

The Koeberg Nuclear Plant

The Koeberg nuclear plant consists of two Pressurized Water Reactors (PWRs) with a nominal net capacity of 924 MW and 930 MW respectively. They were constructed according to the French Framatome CP1 design. The plant construction was started in 1976, with the two units being connected to the grid on 4 April 1984 and 25 July 1985.464 The building work was carried out by Framatome in the face of by then considerable international opposition to any form of cooperation with the white minority government and its Apartheid policies. The commissioning of the plant was delayed by a year due to the damage caused by limpet mine explosions on 18–19 December 1982. The mines had been planted during the construction by an underground operative of the then outlawed African National Congress (ANC).465

At its commissioning it received a 40-year operational license that is set to expire in 2024. While mostly operating without major incidents, there were some events that led to outages and investigations. The most publicized of these was perhaps the ‘bolt incident’ in 2005–2006.466 Overall, in the course of its 39-year operational lifespan both units had, up to the end of 2022, achieved modest cumulative load factors of 72 percent (see Koeberg’s Troubled Operational History).467

The Apartheid Era Nuclear Weapon Program

As an international pariah state in the 1960’s to 1980’s, South Africa has long had the ambition to develop a nuclear weapon arsenal that it envisaged would act as a deterrent to foreign pressure. While the development of nuclear energy plants was common around the world at that time, especially in countries that enjoyed levels of technological know-how similar to South Africa, the construction of a nuclear plant was particularly welcomed in the country, as it enabled justifying the growth of nuclear skills and capacity as being for peaceful purposes. Furthermore, while rich in coal reserves that guaranteed a level of energy security, South Africa has no oil reserves and only limited gas, so additional electricity options in the form of nuclear power assisted in mitigating the growing threat of economic sanctions.

The large-scale investment in the growth of nuclear skills resulted in a large cohort of engineers and scientists being recruited to the nuclear sector. While activities included genuine scientific research, applied technologies for medical diagnosis and treatment, and the production of isotopes, substantial effort focused not only on the development of what would become the Koeberg nuclear plant, but also on the eventual production of nuclear arms.

While it has never been clarified exactly to what extent this latter aim was achieved, on 22 September 1979 the U.S. VELA satellite detected a flash that appeared consistent with a nuclear explosion over the far Southern Ocean. It was suspected to be a nuclear test carried out either by South Africa, or by Israel with full South African cooperation.468

In 1989, on the eve of the South African transition to democracy, South Africa terminated its nuclear weapon program. In 1991 the country signed the Nuclear Non-Proliferation Treaty (NPT). Two years after that, just before South Africa’s first democratic elections, then president, F.W. de Klerk, conceded that the country had constructed six nuclear warheads, and that these had been dismantled in line with its obligations towards the NPT.469

Pelindaba Facility

The Pelindaba facility, approximately 30 km west of South Africa’s capital city, Pretoria (now Tshwane), dates back to 1961, when the National Nuclear Research Institute was set up at the location. It became the site of the 20 MW SAFARI-1 reactor, designated as a research reactor.470 The reactor has been progressively converted from high to low enriched uranium.471 Previously operated uranium conversion, enrichment, and fuel-fabrication plants that have supplied the nuclear weapons program, the SAFARI-1 reactor, and the Koeberg plant (until 1995) have been decommissioned.472

In 1999, the Pelindaba facility and associated entities became a state-owned public company named the South African Nuclear Energy Corporation (Necsa). Necsa’s subsidiaries NTP and Pelchem manufacture radioisotopes and fluorochemicals respectively. Necsa also used to manage the Vaalputs radioactive waste disposal site, later transferred to the National Radioactive Waste Disposal Institute which was created in 2009.

In November 2007, Pelindaba experienced a mysterious break-in by two teams of armed intruders that has been suspected of being a sophisticated attempt with insider cooperation to steal high-enriched uranium for use in nuclear explosive devices. One guard and a firefighter (not site staff, present unexpectedly) were seriously injured. The government has downplayed the event, and the intruders were never arrested.473 More recently, the facility has experienced various breakdowns, safety scares and intense contestation for the leadership of the organization.474

The Vaalputs Radioactive Waste Disposal Site

Low-level radioactive waste from Koeberg and Pelindaba is transported to the waste storage site at Vaalputs, an isolated locality in the semi-desert approximately 400 km north of Cape Town and Koeberg. This site was opened in 1986.475 Its license to operate was suspended for a short period in 2012 due to instances of non-compliance.476 While not involving high-level waste, which remains in local storage at Koeberg and Pelindaba, local politicians have expressed concern about the worsening state of the gravel road leading to the site which could cause an accident followed by radioactive contamination, claiming that this state of affairs is in breach of international regulations.477

Uranium Mining

Uranium has been identified and sometimes extracted in several of South Africa’s many gold mines. In addition to these, there are several smaller designated uranium mines.478 Forty years ago South Africa ranked amongst the world’s top three producers of uranium with 14 percent of the global output, but in recent years the country’s share in global production dropped to below 1 percent and even 0.1 percent in 2020.479

The significance of owning local uranium supply in the event of a nuclear power boom came into the spotlight a decade ago when a politically connected family controversially purchased the Shiva mine at a time when government was actively pushing for a massive nuclear newbuild.480

The Pebble Bed Modular Reactor Initiative

In the early 2000’s, considerable efforts were invested in South Africa to develop and commercialize an SMR design originally developed in Germany in the 1980s.481 The project became known as the Pebble Bed Modular Reactor (PBMR). In 1993, the South African national power utility Eskom obtained the license to the technology from the German HTR GmbH, who were no longer actively pursuing the technology.482 A company was launched in 2000 and included amongst its international investors some big international players of the time like British Nuclear Fuels, and PECO, Exelon, and later Westinghouse.483

While claiming to have completed a successful viability study for the commercialization of the PBMR, the program encountered a range of obstacles that slowed the projected progress considerably. In 2005, the environmental advocacy group Earthlife Africa successfully overturned an environmental impact assessment that had endorsed a new PBMR testing facility.484

More telling though, despite the involvement of many of the country’s nuclear scientists and extensive government financial support, progress in operationalizing the concept was too slow. Investors withdrew from the program due to uncertain progress and ballooning costs, and the funding ultimately had to be carried by the South African Government. In the words of one analyst (in 2010):

In 1998, they were saying that they would have the demo plant online in 2003 [at a cost of ZAR2 billion (US$1998330 million)]. The final estimate was that the demo plant would be online in 2018 and it would cost 30 billion rands (US$[2010]4 billion).485

Government funding to the program was eventually terminated in 2010. It is estimated that up to that point the PBMR had used up ZAR8.67 billion (US$20101.18 billion) of taxpayers’ money.486

Considering the failure history of the PBMR development, it is surprising to see a PBMR-400 termed “preliminary design” referenced in various recent publications.487

Revival of Controversial Nuclear Newbuild Plans

During the nine-year term of office of former President Jacob Zuma (2009–2018), South Africa embraced a grand strategy to build new nuclear plants with a combined capacity of 9.6 GW.488 Had this materialized, this would have been the largest public works exercise in the country’s history, and it was at the time estimated to cost ZAR1 trillion (US$2011140 billion).489 This mega-project became very controversial in South Africa, not only because of the prohibitive costs (ZAR1 trillion then even exceeded the country’s total tax revenue collected in a year), but also the construction and operating contract was in practice (though not formally) awarded to Russia’s Rosatom490. The initiative was eventually stopped by public pressure against then President Zuma (who drove this project relentlessly) and legal intervention that pinpointed major irregularities491 in that constitutionally required public consultation and parliamentary debate had not been carried out before the far-reaching decisions had been taken.

The build was initially conceptualized as part of the national Integrated Resource Plan for Electricity published in 2011.492 It was proposed as a suitable intervention to mitigate a large increase in electricity demand expected due to projected economic and population growth and massive electrification of previously neglected black rural areas and urban neighborhoods, as well as the planned closure of some of the oldest coal plants. It followed an unsuccessful attempt at initiating nuclear newbuild in 2008 that attracted bids from Areva and Westinghouse, which was however cancelled at the end of that year due to financial shortfalls.493 It was also a period when renewables were still far more expensive than they are now, and before the Fukushima disaster placed major brakes on international nuclear rollouts. This ambitious nuclear newbuild program proposed in 2011 initially attracted the interest of potential bidders from France, South Korea, China, Japan, the U.S., and Russia.494

By 2014 it was becoming clear that Russia’s Rosatom was the front runner. For example, unlike the designs then available from other bidders, the requested 9.6 GW exactly matched the capacity of eight of Rosatom’s VVER 1200 MW reactors. Rosatom also established an office in Johannesburg in 2012.495 On the political front, the Russian government was increasingly expressing their expectation to be awarded this megaproject, in the opinion of some observers as an acknowledgement by the South African government of closer ties recently established through South Africa’s admission into the international BRICS (Brazil, Russia, India, China, South Africa) grouping.496

Russia’s strategy in trying to secure this project followed the same pattern that led to Rosatom nuclear newbuilds at Rooppur in Bangladesh, El Dabaa in Egypt, and Akkuyu in Turkey (see sections on Bangladesh, Egypt and Turkey in Potential Newcomer Countries, and past WNISR editions). While details were never finalized, there were reports that Rosatom was conceptualizing a financing model for South Africa similar to the financing schemes of the abovementioned projects (variants of the Build-Own-Operate approach497).

South Africa’s president, Jacob Zuma, was an ardent supporter of the nuclear plan and in 2014 saw it as defining his legacy. He also appeared to have struck a deal with the Russian president, Vladimir Putin, that the bid would go to Russia.498 At the 2018–2022 Zondo Commission into State Capture, the former Minister of Finance, Nhlanhla Nene declared that pressure was exerted on him by President Zuma during a state visit to Russia in 2015 to sign a declaration that would have bound the South African government to a financial commitment and thus practically award the project to Russia. The Zondo Commission explicitly concluded that the refusal by the Finance Minister to do so was the leading reason why he was dismissed as Minister in December 2015.499

While the President’s enthusiasm for the nuclear plan and his favoring of the Russian bidders may just have been influenced by political considerations, there were also numerous reports of benefactors and businesspeople close to the President who were in positions that would benefit from nuclear newbuild.500 In particular, the Gupta family acquired the Shiva uranium mine in anticipation of greater demand for nuclear fuel with the additional future plants.501

The matter was thereafter taken up in court by two non-governmental organizations, Earthlife Africa and the South African Faith Communities Environment Institute [SAFCEI] who argued that the agreement with Russia was illegal. Presiding judge Lee Bozalek went further:

Bozalek’s judgment effectively declared all government’s efforts to procure nuclear energy null and void. In addition to declaring South Africa’s agreements with Russia, the US and South Korea unlawful and unconstitutional, he also ruled that government’s 2013 and 2016 determinations to procure nuclear energy will be set aside. 

The judgment also determines that the request for proposals and information to start procuring nuclear energy were unlawful, unconstitutional and therefore set aside.502

That 2017-court ruling led to the termination of the entire initiative (see also WNISR2017 and subsequent editions). President Zuma did try to revive the Russian deal, notably by appointing one of his most trusted lieutenants, former State Security Minister David Mahlobo, to the Energy portfolio in 2017.503 By then however considerable opposition and protests had been building up against the President with the Russian deal seen as one of the key pillars of what has been termed as “State Capture” in South Africa. At the end of 2017, at the ruling ANC’s elective conference, Zuma’s preferred successor suffered a narrow defeat to the current South African president, Cyril Ramaphosa.

Ramaphosa met Vladimir Putin in July 2018, and communicated to the Russian leader that South Africa no longer intended pursuing the new nuclear build.504 This ended the saga, temporarily.

South Africa’s Enduring Electricity Crisis

South Africa’s power utility Eskom has in recent years been forced to institute rolling power blackouts due to its inability to meet national electricity demand at all times. These outages have been progressively deteriorating as breakdowns worsen at its fleet of coal power stations, which in 2022 still accounted for 85 percent of electricity production with wind and solar providing just 4.5 percent and 2.9 percent respectively.505 At times, the outages amounted to as much as 10 hours per day,506 and since the start of 2023, some level of power cuts have occurred almost every day507.

The power supply crisis has turned into the country’s single most important and emotional discussion point, and the cost to the economy has been estimated by the South African Reserve Bank as between ZAR204 million (US$11.7 million) a day when the power shortfall is 3 GW, and ZAR899 million (US$51.7 million) a day when the power shortfall is 6 GW.508 The power crisis is recognized as one of the issues that is going to dominate the national discourse for some time still and is expected to shape the outcome of the 2024 national elections.509 Recognizing the severity of the electricity crisis, the government went as far as declaring it a national State of Disaster on 9 February 2023 (terminated two months later).510

The Dwindling Performance of South Africa’s Power Plants

The electricity crisis in South Africa is largely caused by increasingly frequent technical problems at the country’s large fleet of coal power stations. The electricity availability factor (the percentage of the power that can be delivered at any particular time relative to the total capacity) has been steadily decreasing over the years. In the first half of 2023, it fell below 50 percent on some days.511

The Koeberg nuclear power plant generated 10.12 TWh in 2022, a drop of 17 percent over the previous year, bringing its share in the national electricity mix to 4.9 percent, down 1.1 percentage points. The last time both units were operating together was just before Koeberg-1’s shutdown for lifetime extension upgrades on 10 December 2022. Since then, Koeberg-2 experienced short interruptions twice. This means that Koeberg-2 has unexpectedly lost generating capacity three times since it was reconnected to the grid in August 2022 following its lengthy shutdown and mid-2023.512 (See Figure 45). Meanwhile, Koeberg-1, which was initially expected to undergo a six-month refurbishment outage, is now set to resume production after eleven months, in November 2023. 513

Koeberg’s Troubled Operational History

Ever since their grid connection in the middle of the 1980s, the operational history of the two Koeberg reactors supplied by France has been roller-coaster (see Figure 44). It took Unit 1 until 1997 to achieve a full year with a load factor exceeding 80 percent. Since then, in nine of the 25 years, its load factor remained below 70 percent.

In the first 15 full years of operation from 1986–2000, Unit 2 has seen only three times a load factor beyond 80 percent. Since then, in nine of the 22 years, its load factor remained below 70 percent.

Not surprisingly, the lifetime load factors of both units remain at just over 70 percent very modest by international comparison. The average annual load factor of the world nuclear fleet has been around 80 percent over the past five years.514 Small programs with less than five units frequently experience lifetime load factors in the high eighties.

The 18-month period between January 2022 and July 2023 has seen a range of events that have impacted the operation of the Koeberg reactors that are now respectively in their 40th and 39th year of operation since startup.

Chronology of Events January 2022–July 2023

18 January 2022 – Koeberg-2 is shut down for major works to meet requirements for lifetime extension, including the replacement of steam generators and routine refueling.

24 January 2022 – Koeberg-1 unplanned shutdown; reportedly, a technician cut a valve on the active Koeberg-1 rather than the shutdown Koeberg-2. The plant is down for almost a day, and power is slowly ramped up again over the following two days.

7 August 2022 – Koeberg-2: after nearly seven months offline, much longer than the projected outage period of five months and without having achieved the main purpose of the outage (steam generator replacement), the reactor starts being powered up again; normal power output is again reached on 14 August.

19 August 2022 – Koeberg-2 suddenly powers down because the control rods developed a “mechanical problem” (not further specified).

25 August 2022 – Koeberg-2 is restarted, and full power is achieved on the same day.

3 September 2022 – Koeberg-2 trips during a control-rod test. The outage lasts three weeks.

25 September 2022 – Koeberg-2 is restarted achieving full power the following day.

10 December 2022 – Koeberg-1 is shut down for lifetime-extension refurbishing (including refueling). The outage is then scheduled to last until June 2023. As of early October 2023, return to service is scheduled for November 2023.

17 February 2023 – Koeberg-2 trips “while replacing a failed electronic turbine protection module”. The system returns to full operation 24 hours later.

15 April 2023 – Koeberg-2 shuts down due to “problems with its feedwater pumps”. 70 percent of its capacity is restored on 17 April, and full power on 19 April.

14 June 2023 – Koeberg-2 experiences a 30-percent power loss (reasons unclear); full power is restored four days later.

Sources: Various, compiled by WNISR, 2023

1. Figure 44 | Historical South African Nuclear Reactor Performance, 1984–2022

Source: IAEA-PRIS, 2023

Note: Grid connection dates: Koeberg-1 on 4 April 1984, Koeberg-2 on 25 July 1985.

1. Figure 45 | Recent South African Nuclear Power Plant Performance

Source: Eskom, 2023515

The Current South African Electricity Plan

South Africa’s most recent official electricity management and development roadmap, the Integrated Resource Plan for Electricity 2019 (IRP 2019),516 was published in the Government Gazette in October 2019. The plan estimates future electricity demand and projects an outline of new generating capacity (only specifying technologies) and plant closures for each year until 2030. It marked a significant move away from nuclear energy. In particular, the 9.6 GW build listed in the previous IRP from 2010 had been removed. Instead, the IRP2019 explicitly advocates a 20-year life extension of the two units at Koeberg to 2044.

The IRP2019 also includes an ambiguous reference to potential future 2.5 GW of nuclear:

Decision 8: Commence preparations for a nuclear build programme to the extent of 2 500 MW at a pace and scale that the country can afford because it is a no-regret option in the long term.517

These 2,500 MW appear to have been a late addition to the document.518

The IRP2019 timeline lists no addition of nuclear generating capacity until 2030. Despite this, the now combined Ministry of Mineral Resources and Energy has been quick to get the ball rolling, seeking to lay the ground for a round of expressions of interest. In 2021, the Ministry had foretold it would issue a request for proposal in late FY2021 in order to finalize the procurement in 2024.519 This has however not materialized to date, yet in May 2023, Minister Mantashe maintained that the procurement would be completed by the 2024 deadline, while announcing that a request for proposals would be launched in Q4 of FY2023.520 At the same time, the initiation of new bidding rounds for renewable energy projects is now two years behind schedule.

Lacking Government Action on Renewables and Divisions in the Ruling Party

Other than the disturbingly frequent breakdowns at the coal power stations, the electricity crisis has also been exacerbated by a reluctance from critical sectors in government (notably the Department of Mineral Resources & Energy—DMRE) to vigorously drive the expansion of solar and wind projects—which are exceptionally well suited to the South African weather.521

An episode that well illustrates the partisan position of influential politicians is the dismissal in February 2022 from the board of the National Nuclear Regulator (NNR) of Peter Becker, a leading member of the South African anti-nuclear activist group Koeberg Alert. The reason given by the responsible DMRE Minister Gwede Mantashe is that

If you resist nuclear and you [are] a board member, I fire you, simple. You can’t be [on] a board of something you’re not advocating for.522

Peter Becker, who had been nominated by several community groups and appointed in June 2021 by Minister Mantashe himself to serve as a civil society representative on the NNR board, subsequently successfully challenged his dismissal in a court of law that rendered its judgment in early 2023.523 In May 2023, the same judge rejected an application filed by the Minister and NNR to bring the case before the Supreme Court of Appeal.524

In contrast to this, President Cyril Ramaphosa has at numerous points expressed strong support for a vastly expanded renewable energy rollout, as well as promoting the Just Energy Transition initiative, which seeks to develop renewable energy plants near the sites of coal plants earmarked for closure and the retraining of coal sector workers. He is strongly endorsing the major investment in this program.525 The Just Energy Transition Implementation Plan envisages the investment of US$8.5 billion over the five-year period 2023–2027 pledged by several developed nations for initiatives that will promote the decarbonization of the electricity sector, as well as the introduction of green hydrogen and electric vehicle developments.526

South African Nuclear Sector Developments

After an aborted start in 2022, the past year has seen work begin at Koeberg to implement the plant maintenance and upgrades required to secure the plant’s life extension to 2044. There have been controversies surrounding the termination of a fuel supply agreement with the United States. And, in view of South Africa’s worsening electricity supply shortages and increasing periods of rolling power cuts, lobbying for the construction of new nuclear plants has intensified (even if this obviously does not represent a short-term option to address the power crisis).

Koeberg’s Lifetime Extension

The Koeberg refurbishment project is now effectively over a year behind schedule and looks increasingly set to drag on well past the 31 July 2024 deadline when its operating license expires. This has various regulatory implications, as well as aggravating the electricity crisis.

The lifetime extension of Koeberg beyond its initially projected 40-year operating span has been treated as a given for some time. Major operations to replace a variety of components, in particular the plant’s six steam generators, have been planned for over a decade (see previous WNISR editions). As the South African electricity crisis has grown more acute, the need to keep large electricity producing facilities such as Koeberg going for as long as possible has been considered increasingly crucial. A 20-year extension of the ageing nuclear plant has accordingly been recommended in the most recent IRP. In order to approve the extension, the South African National Nuclear Regulator requires a series of maintenance operations and instrumental replacements to be carried out. The most significant of these is the replacement of the six steam generators.

The projected costs of these upgrades was quoted in 2010 to be ZAR20 billion (US$2010 2.7 billion).527 There are now signs that the final costs are going to be considerably higher, though no revised figure is being put forward by Eskom.528 It has been claimed that some of these additional costs are to be covered by the Koeberg maintenance budget, the justification then being that these are routine replacements part of normal plant operations.529

The steam generator replacements of Koeberg-2 were scheduled to coincide with the unit’s refueling outage between January and June 2022.530 While never fully explained by the utility, reports indicate that work could not proceed as planned because the utility failed to construct the storage facilities for the contaminated old steam generators in time. Indicative of the disorganized manner in which the project management has proceeded is an incident—the second time in as many months—during the Koeberg-2 shutdown, a technician mistakenly cut a valve of the running Koeberg-1 while intending to execute this same action on the equivalent valve of the inactive Koeberg-2.531 The steam generator replacement operation was then postponed, with Koeberg-2 eventually being returned to service in mid-August 2022 with the original steam generators (i.e. 7 months after the start of an outage originally scheduled for 5 months), though two more outages occurred in the following weeks.532

On 10 December 2022, Koeberg-1 was taken offline to start refurbishment work, including the replacement of its steam generators.533 The six months projected for this operation proved insufficient, and Eskom’s completion time estimate then became October 2023. In July 2023, South Africa’s Electricity Minister already indicated that he was “worried and extremely upset” about the delayed completion of this project, highlighting that the Koeberg-1 shutdown was now likely to extend into the period in which Koeberg-2 was to finally be fitted with its new steam generators.534 In August 2023, Koeberg’s current acting Chief Nuclear Officer indicated that the new steam generators had now been moved into position.535 Regarding the lengthy time overruns, he stated that

In a nutshell we were overly optimistic in terms of what we thought we could achieve and, in hindsight, if we could have done it differently we would have scheduled a [much] longer time for this intervention.536

As of August 2023, the projected completion day for the operation has now been moved further back to November 2023.537

A further sign that the work appears not to be going as planned is that one of Eskom’s most senior officials, the Chief Operating Officer, who had been retained to manage the project, has suddenly left after less than a month in this role without reasons given for his departure.538 Koeberg is also only under temporary general nuclear management since the departure of the previous Chief Nuclear Officer in July 2022.539

Fuel Supply Uncertainties

The U.S.-South Africa so-called 123 agreement that enabled Westinghouse to supply nuclear fuel to Koeberg lapsed on 4 December 2022. The reasons for why this agreement was not renewed prior to lapsing has not been fully explained. One given reason is that South Africa has hinted on its intention to fabricate its own fuel again, triggering proliferation concerns in the U.S. Others have speculated the move would be an indicator of U.S. displeasure and growing mistrust with the South African state’s comparatively close relationship with Russia at a time of global polarization due to the war in Ukraine.540

The nuclear fuel supply arrangement is currently split between Westinghouse, which provides the fuel assemblies for Koeberg-1, and Framatome, which supplies Koeberg-2. Koeberg-1 is being refueled during the current outage, but there is now no certainty regarding future fuel supplies.541 In March 2023, the South African Minister of Mineral Resources and Energy stated that the U.S. and South Africa were actively engaging with the aim of reinstating the fuel-supply agreement, and that permission had now been granted to Westinghouse to produce and deliver the next round of nuclear fuel.542

At the end of July 2023, South Africa’s NECSA signed a Memorandum of Understanding to build stronger bilateral collaboration with Russia’s TVEL.543 Nuclear fuel purchase and production is highlighted in particular, and thus establishes Russia as a likely preferred replacement to the current French and U.S. suppliers. This is set to enhance perceptions that South Africa is gravitating away from its traditional French and U.S. trade partners and towards closer links with the Russian state and is now exhibiting a pro-Russian bias at times of heightened geopolitical instabilities induced by the war in Ukraine.544

Renewed Talk of Nuclear Newbuild in South Africa

South Africa’s grave power crisis has spawned numerous propositions of how to effectively expand the missing electricity capacity. This has included numerous suggestions that this can be best achieved through new nuclear plants. The lobbying for new nuclear has been driven by a variety of figures across the political spectrum. The most influential nuclear promoter has been the Minister of Mineral Resources and Energy. His position is shared by many other senior figures in the ruling ANC, although there appear to be widely diverging viewpoints on the energy policy direction within the party.

The third-largest party, the Economic Freedom Fighters, has advocated for the building of new nuclear plants, expressly stating that these should be built by Russia.545 Other figures have warmed to suggestions to import small modular reactors, vocal proponents of which include the former leader of the largest parliamentary opposition party (Democratic Alliance)546 and the Afrikaner lobby movement Afriforum547.

Rosatom has been quick to come forward as a potential international partner by actively marketing its nuclear power ships as a relatively fast solution to the country’s electricity crisis.548 Russia’s only operating small reactors on a barge have been disappointing however with close to 13 years construction time and very low load factors since commissioning (see section on Russia in chapter on SMRs).

Navigating South Africa Through Its Worsening Power Crisis in 2023

Despite experiencing far more severe electricity shortages in 2023 than in recent years, there have been early signs that South Africa is starting to adapt and implementing policies that could extract the country out of the crisis. The long neglected domestic solar industry has been booming as large organizations, businesses, and households have scrambled to escape the deepening power cuts by installing rooftop solar systems.549 The building of medium-scale solar and wind farms has also been boosted by the removal of restrictive regulations that greatly complicated the development of private facilities in the 1–100 MW capacity range.550

On the political front, recognizing that the dire state of electricity supply presented a threat to the ruling party re-election bid at the 2024 national elections, the president created a new ministerial post, the Minister of Electricity, appointed in March 2023 to specifically intervene and implement concrete measures to mitigate the electricity crisis.551 While this appointment has been met with some cynicism given that few believe that there are short-term solutions to the energy crisis, it has allowed the President to assert the government is trying to do what it can to alleviate the difficult situation.

To the surprise of most analysts, South Africa has been able to get through its coldest months in 2023, when electricity demand is at its annual peak, with smaller power shortfalls than anticipated.552 This is to some extent due to lower electricity demand on the grid than projected, and partly due to the much greater penetration of solar devices, with recent reports claiming a 349 percent increase in solar rooftop installations between March and June 2023.553

South Korea Focus

South Korea (the Republic of Korea) is the fifth largest nuclear power producer in the world. It operates 24 reactors and has one reactor in LTO, including 22 Pressurized Water Reactors (PWRs) and three Pressurized Heavy Water Reactors (PHWRs). As of September 2023, there are three reactors—all Korean made APR-1400 type—under construction. So far, two reactors, Kori-1 and Wolsong-1, have been closed, in 2017 and 2019 respectively. In April 2023, then the oldest reactor, Kori-2, was shut down after 40 years of operation; however, as it is expected to be restarted, it is considered in LTO.

The Yoon government is upping support for the nuclear industry whose state-controlled flagship enterprise Korean Electric Power Company (KEPCO) has cumulated an unprecedented net debt of US$149 billion.

Nuclear Power Plant Name Changes

There used to be four nuclear power plant sites in South Korea, namely Kori, Wolsong, Hanbit, and Hanul. However, on 3 January 2017, the Kori plant was divided into Kori in Busan and Saeul in Ulsan.554 And in November 2022, almost six years later, Korea Hydro & Nuclear Power (KHNP) officially changed the names of the reactors at the Saeul site. The operating Shin-Kori-3 & -4 became Saeul-1 & -2, and Shin-Kori-5 & -6 under construction became Saeul-3 & -4.

Until the restructuring of January 2017, Kori was destined to be the largest nuclear power plant complex in the world upon completion of Shin-Kori-4 (later Saeul-2). Yet, after the separation of Kori and Saeul, the title of the world’s largest nuclear power plant has been kept by the Hanul site. At Hanul, in addition to the Hanul-1 to -6 and Shin-Hanul-1, Shin-Hanul-2 is waiting for an operating license approval from the Nuclear Safety and Security Commission (NSSC), South Korea’s nuclear regulator.

Hanul, Largest Nuclear Power Plant in the World

Once Shin-Hanul-2 is connected to the grid, Hanul will become one of only two sites in the world hosting eight reactors (the other one being Bruce in Ontario, Canada). The global average number of reactors per site is 2.5 while the Korean average is double with five reactors per site—seven for Hanul, six for Hanbit, five each for Kori and Wolsong, and two for Saeul.

The total installed capacity of the eight Hanul units will be 8.65 GW net, larger than that of Bruce with 6.36 GW and 1.5 times larger than Europe’s largest nuclear site, Zaporizhzhya in Ukraine with six units totaling 5.7 GW (currently occupied by the Russian army). As Ukrainian nuclear facilities have been militarily targeted, there is a growing concern in the South Korean society about the risk of densely located nuclear reactors being attacked since the Korean Peninsula still technically remains in a state of war between North and South Korea with an armistice agreement since 1953.555

Increased Nuclear Power Generation

According to IAEA-PRIS, in 2022, nuclear power generation increased by 11.3 percent to 167.5 TWh net and provided 30.4 percent of the electricity in the country, up 2.4 percentage points from 2021.556

The increase of the nuclear power generation in 2022 compared to the previous year was mainly due to increased performance of some reactors as well as the start-up of Shin-Hanul-1.The load factor of Hanbit-5 increased from 18.7 percent in 2021 to 99.8 percent in 2022, generating an additional 7 TWh.557 Hanbit-5 was back online after experiencing a prolonged outage from 26 October 2020 to 22 October 2021 due to safety issues related to an automatic trip of the reactor and “faulty welding” of the reactor vessel head penetrations.558

The new APR-1400 reactor Shin-Hanul-1 contributed 3 TWh to the generation increase in 2022. The seventh Hanul unit was first connected to the grid on 9 June 2022 after almost 10 years of construction starting on 10 July 2012 with years of delay.

Hanbit-4, which was in long-term outage (LTO) since May 2017 finally restarted on 11 December 2022. Local residents, NGOs, and nearby city councils were opposed to the restart because they claimed that the safety reviews and repair work on the 140 voids identified in the concrete containment walls and corrosion on containment liner plates were not thorough enough. However, the national nuclear regulator, the Nuclear Safety and Security Commission (NSSC), decided to grant permission to restart without consultation of the local people which is not legally required in South Korea.559

KEPCO’s Financial Crisis

In 2022, the state-owned company KEPCO, builder, owner, and operator of South Korea’s nuclear power plants, filed a record operating loss of KRW32.6 trillion (US$202225.2 billion) and its net debt jumped by 32 percent to reach unprecedented KRW192.8 trillion (US$149 billion). KEPCO’s CEO resigned over the results in May 2023. Under the principle of “selling all available properties” KEPCO announced the sale of its Seoul headquarters building in the heart of Seoul, along with 44 buildings owned by the company.560 Investor trust has been eroding for a while. KEPCO stocks lost 70 percent of their value over the past seven years (see Figure 46). The downward trend did not change following a new, ultra pro-nuclear administration taking office in mid-2022.

1. Figure 46 | Korea Electric Power Corporation Stock Value

Source: Yahoo Finances, 2023

Nuclear Policy Under the Moon and Yoon Administrations

In June 2022, President Yoon and his administration pledged KRW1,000 billion (US$2022774 million) in investments by 2025 “to rebuild” the industry, a sum that corresponds to 0.5 percent of KEPCO’s net debt.561 The current administration also means to allocate KRW400 billion (~US$2022310 million) for the development of SMRs.562

On 25 July 2023, the Ministry of Environment (ME) announced first estimates of the country’s 2022-greenhouse gas emissions (GHG). In the press release, ME attributed the year-on-year decrease in GHG emissions to Yoon’s energy policy changes, quoting nuclear power among the main drivers563.

However, the nuclear production increase was not related to policy changes. The differences between the Moon and Yoon administrations’ nuclear policy consist in the implementation of lifetime extensions of existing reactors (Yoon) or not (Moon) and the possible launch of new reactor constructions beyond the ones already underway (Yoon) or not (Moon). Therefore, the differences did not influence nuclear power generation between 2021 and 2022.

Moon’s nuclear phaseout policy established in 2017 is often misrepresented. For example, South Korea’s country profile-page on the World Nuclear Association (WNA) website states that “the previous government’s policy was to phase out nuclear power over a period of 40 years”.564 In reality, Moon’s nuclear policy guaranteed the design lifetime—without extension—of the existing reactors and those under construction, and it planned no additional reactors after Saeul-3 and -4. Because the design lifetime of Saeul-3 and -4 is 60 years, the planned complete nuclear “phaseout” in South Korea was scheduled around 2085 considering that Saeul-4 was planned to be completed by 2025. Compared to the nuclear phaseout policy in Germany in 2023 and in Taiwan by 2025, the South Korean policy was in fact rather a nuclear program limitation than a phaseout strategy.

It will not be before June 2025 that Yoon’s pro-nuclear policy approach could substantially influence actual nuclear power generation, as that is when the oldest operational reactor Kori-2 is scheduled to be restarted. Kori-2 was shut down on 8 April 2023 upon expiration of its 40-year license and is to undergo inspection and refurbishment work over several years to allow restart and lifetime extension.565

Proactive Lifetime Extension

In order to prevent a reactor from being out of operation for a long time due to safety reviews and refurbishment for lifetime extension, the NSSC amended the Enforcement Decree of the Nuclear Safety Act to the effect that the operator can submit a safety assessment report for lifetime extension five to ten years—rather than two to five years—prior to the operating license’s expiration date.566 With the amendment, the number of nuclear reactors whose lifetime extension are likely to be applied for during the Yoon administration’s 5-year term (10 February 2022–10 February 2027) increased from 10 to 18 reactors.

More Newbuild Planned

The incumbent Yoon administration announced its intent to revive the previously abandoned construction of the 9th and 10th reactors at the Hanul site, Shin-Hanul-3 and -4. The two new units are scheduled to be completed by October 2032 and 2033.567 The Tenth Basic Plan for Long-term Electricity Supply and Demand (BPE, 2022~2036) was issued in January 2023; it included the construction of the two Shin-Hanul units.

1. Table 12 | 2022, 2030 and 2036 Electricity Mix in South Korea

Sources: WNISR, based on data from KOSIS and MOTIE, 2023568

(a) NRE: New and Renewable Energy. New energy in South Korea includes Integrated Gasification Combined Cycle (IGCC) and fuel cells.

As shown in Table 12, the nuclear share in the electricity mix is planned to continue to increase up to 34.6 percent in 2036. The share of fossil fuels in the 2030 electricity mix of the 10th BPE by the Yoon administration is not very different from the share of fossil fuels in the previous governmental plan (the 2021 Nationally Determined Contribution or NDC) of Moon’s administration. However, Yoon’s administration decreased the 2030-renewable-share target from 30.2 percent to 21.6 percent as the result of increasing the nuclear share from 23.9 percent to 32.4 percent.

South Korea was the 8th largest electricity producer in the world in 2022, surpassing Germany since 2020.569 It is remarkable that the country generated more electricity than Germany, considering that Germany has a 1.6 times larger population. Especially the industrial and service sectors have kept increasing their consumption and represent disproportionate shares of overall consumption. While in most industrialized countries, electricity consumption has stagnated or declined over the past decade or so, South Korea has seen a steady increase over most of this period. In April 2023, Germany completed its nuclear phaseout. Under the current policy, South Korea would rely on nuclear power generation at least until close to the end of the 21st Century.

Yoon Administration in Search of New Sites

On 18 July 2023, the Yoon administration announced that it would develop the 11th BPE for 2024–2038 early. Also, MOTIE said that the key direction of the next BPE would be the increase of nuclear power capacity in the framework of its climate change policy, in particular to cover new electricity demand from industry, such as the expansion of a semiconductor cluster in Yongin.570 If new nuclear capacity was included in the 11th BPE, it would be the first time since the 7th BPE in 2015, which featured the plan for Shin-Hanul-3 & -4.

Two further sites were envisaged for nuclear newbuild before. Samcheok in Gangwon Province and Yeongdeok in North Gyeongsang Province were officially designated as greenfield sites for nuclear reactor construction in 2012.571 However, following the Fukushima catastrophe, local opposition to the project grew. Consequently, Samcheok residents organized a local referendum in October 2014, resulting in 84.9 percent voting against the project.572 The people in Yeongdeok also organized a local referendum in November 2015, resulting in 91.7 percentage of votes against the new nuclear reactors, but it was invalidated as voter turnout at 32.5 percent remained just below the legally required one third of eligible voters.573 While the consultations had no legal weight, during the Moon administration, due to the long civil movement against the plan, the site designations were officially cancelled in 2019 for Samcheok and 2021 for Yeongdeok.

Therefore, the Yoon administration likely needs to find one or several new sites. Local people in Ullu-gun in Ulsan where the Saeul nuclear plant is located are already divided over new nuclear construction. Social conflicts among the local people over nuclear power plant projects are likely to erupt again in South Korea.574

Efforts to Boost Nuclear Exports

In August 2023, the South Korean Financial Services Commission boosted financing support for exporting companies—including the nuclear power industry with an unknown share—by around 50 percent to a total of KRW23 trillion (~US$18 billion).575

This is only the latest in a number of actions translating the government’s efforts to help the ailing export and nuclear sectors. KEPCO is still in the course of finalizing its only foreign construction project so far, i.e., the delivery of four reactors to the United Arab Emirates (UAE). The Barakah project was supposed to demonstrate the feasibility of the implementation on-time on-budget of a nuclear power plant construction in a newcomer country. The project is three years behind schedule and the extent of cost overrun is unknown.

Further reactor export projects by the Korean nuclear industry are fragilized by an ongoing litigation in a case brought against KHNP by Westinghouse in October 2022. Westinghouse claims KHNP requires U.S. approval to export its APR-1400 technology and infringed its intellectual property rights by failing to do so. The case is also being reviewed by the Korean Commercial Arbitration Board since August 2023, as previous negotiation efforts failed to settle the dispute. Damages claimed by either side amount to several hundred million US dollars.

On 25 August 2022, six months into the Ukraine invasion, KEPCO with its subsidiary KHNP and Rosatom with its subsidiary Atomstroyexport JSC signed a US$2.25-billion contract to provide around 80 buildings and structures at four units of El Dabaa nuclear power plant and supply related equipment and materials. Rosatom is the contractor for the El Dabaa plant. No information on financial aspects is publicly available concerning Barakah and El Daaba. KEPCO’s 2022-Annual Report indicates:

The contracts with purchasers state that the disclosure of information related to UAE and Egypt Eldaba nuclear power plant construction projects such as contract date, contractual completion date, rate of progress, unbilled construction, impairment losses, etc. is not allowed without consent from the purchasers. The purchasers did not agree to disclose such information. Accordingly, the [KEPCO] Group did not disclose such information…576

While the current administration has clearly set a different political agenda for nuclear power than its predecessor, the outlook remains highly uncertain—as well in the country as concerning the country’s overseas ambitions—as the dire financial state of the flagship company KEPCO leaves little or no room for major investment expansions.

United Kingdom Focus

As of mid-2023, the United Kingdom (U.K.) operated nine reactors, two less than the previous edition of the WNISR, with the closure of the two units at Hinkley Point B on 6 July 2022 (B-2) and 1 August 2022 (B-1) respectively. This follows the closure of the two reactors at Hunterston in November 2021 and January 2022, and two units at Dungeness officially closed in June 2021 (last power generation in 2018, see Figure 47). In total, 36 nuclear reactors have been closed in the U.K., the second largest number of any country behind the United States (see Decommissioning Status Report). This includes all 26 Magnox reactors, two fast breeders, one small Steam-Generating Heavy Water Reactor (SGHWR) and seven Advanced Gas Reactors (AGRs). There are now 5.8 GW of nuclear capacity in operation, with 7.8 GW awaiting decommissioning.

In 2022, nuclear plants generated 47.7 TWh, an increase for the first time in six years, producing 14.7 percent of electricity, down from a maximum share of 28 percent in 1997.577

The electricity mix in the U.K. has changed rapidly over the past decades, as seen in Figure 48. The most significant trend was the rapid increase in the use of renewable energy—from 2.8 percent at the turn of the century to 41.5 percent in 2022.578 The total contribution of renewable energy to the power mix—including biomass and hydro—saw a significant increase of 1.9 percentage points over the previous year due to 3.8 GW of new wind and solar capacity.

1. Figure 47 | U.K. Reactor Startups and Closures

Source: WNISR with IAEA-PRIS and EDF Energy, 2022-2023

Type of Reactors:

AGR: Advanced Gas Reactors; FBR: Fast Breeder Reactor; PWR: Pressurized Water Reactor; SGHWR: Steam Generating Heavy Water Reactor

While Great Britain—including England, Scotland, and Wales, but not Northern Ireland—has left the E.U. Internal Energy Market, electricity trade continues with E.U. Member States. Despite Brexit, electricity trade is increasing as new interconnectors become operational. Most recently, in 2021, a new connection was made with Norway, the North Sea Link, a 1.4 GW ~720 km cable, which follows on the back of new interconnectors to France in 2020 and Belgium in 2019.579 As of 2022, there were seven cables with a total capacity of 7.4 GW, and while these allow power to flow both ways, the British market has historically been a net importer.580 However, due to the ongoing generic problems in the French nuclear fleet, the falling production of electricity in France led to the U.K. becoming a net exporter of electricity for the first time in forty years in 2022 with a positive trade balance of 5.3 TWh.581

1. Figure 48 | Electricity Generation by Source in the U.K., 2000–2022

Source: U.K. Government, DUKES 2023582

Closure of the Advanced Gas-cooled Reactors (AGRs)

Managing reactors as they age is a constant problem for any technology design, and the AGRs are no exception. In recent years, issues with the core’s graphite moderator bricks have raised concerns. Keyway Root Cracks (KWRC) were unexpectedly found at the (now closed) Hunterston B reactors in 2016. This can lead to the degradation of the keying system, a vital component that houses the fuel, the control rods, and the coolant (CO2). Their cracking or distortion could impact the control rods’ insertion or the coolant’s flow. There are also issues of graphite erosion, and several of the AGRs are close to the erosion limits that the Office for Nuclear Regulation (ONR) has set. ONR has said, “most of the AGRs will have their life limited by the progression of cracking”, as replacing the graphite bricks is impossible.583

Besides the small unit at Windscale/Sellafield, 14 AGRs were built (see Figure 47), operating at seven stations. Until mid-2021, Hinkley Point B and Hunterston B were due to operate until 2023, while Dungeness B was due to operate until 2028. However, by early 2022, the situation had dramatically changed, with EDF officially closing Dungeness B-1 and -2 in June 2021, Hunterston B in January 2022, and then Hinkley Point B in July/August 2022.584

Hartlepool and Heysham A were due to close in 2024; Electricité de France (EDF) delayed closure by two years in March 2023,585 but the Office for Nuclear Regulation (ONR) has said that while a plant life extension did not require formal approval, EDF would need to produce updated safety cases for the plants, which will be assessed by the regulator.586 In late 2021, the closure of the last two units (Torness and Heysham B), previously due in 2030, was brought forward to 2028.587 (See Table 13)

1. Table 13 | Status of U.K. EDF AGR Nuclear Reactor Fleet (as of 1 July 2023)