France built 40 nuclear reactors in a decade. Here’s how they did it, and how the world can follow their lead today.
In the 1970s, the golden age of nuclear energy appeared to be coming to an end. Sharply escalating costs, regulatory upheaval, and public opposition had made industry boasts of electricity ‘too cheap to meter’ seem increasingly hubristic. After peaking in 1973, new US nuclear reactor orders would drop to zero in 1979, despite the oil shocks leading to a doubling of electricity prices in the intervening years. Meanwhile, poor design and planning resulted in the UK’s advanced gas reactor program running 50 percent over budget and years behind schedule.
France stands out as an exception. As much of the world was retrenching, France’s nuclear ambitions were accelerating. The construction of 37 of the 57 reactors currently operating in France began during the 1970s. Today, France has largely decarbonized its grid: 94 percent of French electricity is generated from low-carbon sources, with 70 percentage points of that coming from nuclear power.
France’s success in civilian nuclear power was by no means preordained. Throughout the 1960s, France was a nuclear laggard. Its industry, hamstrung by poor technology choices and institutional infighting, was the subject of domestic and international derision. The program’s eventual success was contingent on a politically courageous about-turn, regulatory focus, and a tax system that made local communities enthusiastic about hosting nuclear power stations.
Inauspicious beginnings
As France embarked on the daunting task of reconstruction after the Second World War, interim leader and Free French hero Charles de Gaulle was set on restoring his country to its preeminent position on the world stage. Before the newly restored French Republic had settled on a constitution, de Gaulle decided that it should have nuclear weapons.
France’s nuclear ambitions were split across two rival entities. The Commissariat à l’Énergie Atomique (CEA), established in 1945, focused on producing plutonium to support the country’s nascent nuclear weapons program. Meanwhile Électricité de France (EDF), created the next year, was in charge of operating France’s electricity grid.
In its early days, EDF was not focused on nuclear energy. It possessed little atomic expertise and had to focus its limited resources on rebuilding and rationalizing the devastated postwar electric grid at the same time as integrating 1,400 newly nationalized electricity companies. This left the CEA to pursue its own choice of reactor design. After some early experimentation, it settled on the gas-graphite design.
This design had two advantages from the CEA’s standpoint. First, it employed natural uranium (uranium-238) at a time when France didn’t have any domestic enrichment facilities. Second, the gas-graphite design didn’t rely on high-pressure systems or sealed fuel assemblies. This meant that individual fuel rods could be easily extracted while the reactor was still running, allowing the operator to harvest the weapons-grade plutonium that the fission reaction produced before it was contaminated with other isotopes.
The CEA finished Marcoule G1, its first full-scale reactor, in 1956. By this stage it had started to work with EDF. The director of EDF’s research division sat on the CEA’s steering committee and had persuaded them to add small power generating units to the Marcoule reactors to capture the heat produced by the reactor and use it to generate electricity. At the same time, this EDF director began to build support within his own organisation to pursue nuclear power.
The resulting reactor, finally completed in 1962, was an unhappy compromise. Deliberately throttled to optimize plutonium enrichment, it generated only about 40 percent of the electricity it was capable of producing (versus 60–70 percent for reactors in other countries at the time and 80–90 percent for reactors today). It provided 11 years of underwhelming performance before being decommissioned in 1973.
EDF1 did not mark the end of the early French nuclear program’s troubles. The unhappy marriage produced two more gas-graphite reactors on the same Chinon site: EDF2 (completed in 1965) and EDF3 (completed in 1966).
In 1965, American company Westinghouse, creator of the pressurized water reactor (PWR), embarked on a global public relations offensive. The Westinghouse sales pitch, which heavily emphasized lower capital costs, found favour with EDF. The company became convinced that the ‘sovereignty’ offered by the use of natural uranium was not worth the inefficiency of gas-graphite plants. A summer of negative press commentary, after the fueling of EDF3 was blighted by technical failures, only reinforced this.
For the next two years, a joint CEA-EDF study committee toured America, poring over designs and spreadsheets. EDF concluded that the pressurized water design would be 20–30 percent cheaper, even after paying for enrichment. The CEA contested these numbers and dug in their heels. EDF workers also went out on strike in 1969, in defence of the gas-graphite reactor, fearing that the switch to an American design would damage their employment prospects.
It was not until De Gaulle, who was set on an autonomous French design, left the presidency and was replaced by Georges Pompidou that change was possible. In 1970, after persuading the French cabinet and planning commission, EDF began building two pilot pressurized water reactors at Fessenheim, near the German and Swiss borders. EDF had successfully argued that supporting cheap, reliable electricity was more patriotic than building an inefficient all-French reactor. In 1972, after yet another round of technical failures, the government formally ended the gas-graphite program.
In 1973, global oil prices rocketed after a coalition of Arab countries implemented a total embargo against countries that had supported Israel during that year’s Yom Kippur War. Following Algerian independence, France had lost its major source of domestic natural gas production, so it was hit particularly hard. At this time, France met 70 percent of its energy needs through imported oil. By contrast, the UK and West Germany had access to abundant coal deposits, and the British knew that North Sea oil would come onstream from 1975. The US was the world’s biggest oil producer.
In response, France unveiled the Messmer Plan, named after then prime minister Pierre Messmer who was running the country while Pompidou battled with terminal illness. This committed France to a mass buildout of pressurized water reactors. EDF ordered 16 new reactors in 1974, doubling the country’s total reactor order book. Illustrating the improved efficiency, these new reactors had a total output of 14,400 megawatts versus 9,000 megawatts from the previous 16.
The Messmer Plan amounted to the fastest nuclear buildout ever, in both number of reactors built and capacity added. During the 1980s, France increased the number of reactors in commercial operation from 15 to 55. Even China, with the world’s most streamlined regulatory process and developed industrial base, has failed to match this record.
The most striking element of the buildout was speed, with new reactors consistently being built in approximately six years. EDF had a pre-screened list of suitable sites, and reactors were ordered in bulk with a standardized design. This removed the need for a convoluted site licensing process. Meanwhile, safety standards were set in near-total secrecy through a ‘technical dialog’ process between EDF, the CEA, and other government experts, dubbed ‘French cooking’ by Anglo-American observers. Before 1979, EDF was not even required to conduct an environmental impact assessment on individual construction sites.
Long-term planning gave suppliers the certainty they needed to build out teams and production lines. If engineers came up with potential improvements, they were noted and saved for a potential future reactor series, but EDF only made the smallest of changes to designs once they had been approved.
Framatome, a Franco-American consortium responsible for licensing Westinghouse’s technology, was responsible for much of the plant’s manufacture, but EDF wasn’t a passive customer. EDF maintained control of the engineering, much to Framatome’s irritation, and ran a rigorous cost control regime to compensate for the lack of traditional competition.
This rigorous standardization and long-term order book helped create a specialized supply chain, which introduced significant efficiencies into the manufacturing process. Framatome, for example, could invest in a specialized heavy component facility that could produce vessels, steam generators, pressurizers, and piping. The Saint-Marcel facility, opened in 1975, could use spare capacity for exports and has provided components for 106 reactors. They also built a 9,000-ton hydraulic forging press that produced four to five identical pressure-vessel shells per year.
In the 1970s, France was building nuclear reactors at one third to a half of the pre-interest costs that new reactors in the US had risen to amid toughening environmental and safety regulations.
The backlash that wasn’t
EDF was greatly aided by the weakness of its opposition. While the 1970s were a time of nuclear backlash in most major economies, French technocrats faced one of the world’s least potent anti-nuclear movements.
A big factor was structural. The constitution of the Fifth Republic, adopted in 1958, created a strong presidency and a weak legislature. France’s parliament had next to no legal oversight of nuclear policy and the key decisions sat in executive agencies that were indifferent to lobbying or public opinion. The early anti-nuclear movement in the US and UK successfully weaponized the legal and planning process, forcing lengthy public inquiries and court battles. This not only delayed projects, but forced an interminable public debate about nuclear energy that chipped away at popular support. The French system offered no such legal avenue.
There were a handful of prominent nuclear skeptics, like future president François Mitterand, but France’s lack of domestic fossil fuel production ensured there was a broad pro-nuclear consensus across the major political parties.
Communist parties and their affiliated unions were often a hot-bed of anti-nuclear sentiment. In many European countries with smaller nuclear workforces, communist movements aligned with environmentalists in the hope of winning over younger voters and opposed nuclear as part of a wider anti-NATO program. By contrast, France’s communist movement derived significant support from workers either directly employed by the nuclear industry or involved in its support infrastructure. This meant that the country’s powerful trade unions were either indifferent or actively hostile towards the nascent environmentalist movement. The French Communist Party’s daily newspaper derided anti-nuclear activists as opponents of progress who wanted to return to using candles.
While anti-nuclear sentiment was shut out of the French political system, the success of the buildout ensured that there less of it in France versus other western countries. In 1987, the year after the Chernobyl accident, France was the only major western country where more of the population supported nuclear energy than opposed it. In fact, support was greater than it had been a decade earlier.
Buying support
One of the biggest drivers of opposition to new infrastructure is the sense that communities bear the disruption of development while receiving little short-term benefit. For example, when a new power station is built in the UK, the local authority collects business rates, but most of the revenue is then sent back to the central government.
France’s business tax system has been effective at correcting this. Avoine, the commune that hosted the first EDF reactors, for example, saw annual revenue jump from tens of thousands of francs in the mid 1950s up to nine million a year by the time EDF2 and EDF3 were being brought online in 1964–1966. Locals dubbed their town ‘the Kuwait of Indre-et-Loire’.
The residents of Avoine were the beneficiaries of La Patente, a business tax based on location, activity type, and sometimes estimated turnover. La Patente was subsequently replaced in 1975 by the Taxe Professionnelle, which similarly directed industrial revenues towards the areas that hosted them. It taxed fixed industrial assets (i.e. reactors, cooling towers, machinery) along with wages (although this component was later dropped). This gave communities an economic stake in projects and, because it focussed on physical assets, provided predictable future revenue.
As a further sweetener, the French government began paying out a special class of subsidy from 1975 to municipalities that accepted nuclear power to cover the costs of any new infrastructure.
Avoine wasn’t the only mini-emirate. The northeastern town of Chooz has offered its residents free high-speed internet and a 120-channel TV subscription since 1999. In 2012, local councils in France paid an average of €35 in subsidies or benefits per local resident. In the 19 areas that hosted nuclear power, the average was €450.
Mayors who hosted reactors were able to pave roads and build schools, sports halls, ice rinks, and cultural centers without raising local rates. Until 2023, French communes charged a local residential tax on property. Nuclear municipalities were able to charge significantly less than their regional neighbors. Residents of Fessenheim paid nine percent a year, versus 13 percent for comparable neighbors. Residents of Avoine, near the Chinon nuclear power station, paid 0.1 percent versus a 12 percent regional average.
This transformed mayors into active lobbyists for new plants – rezoning land for industry, publicly praising EDF’s tax payments, and pushing the government to ensure that expansion schedules were kept on track. Meanwhile, EDF would also build new housing for workers and local infrastructure like schools, as well as sponsoring local cultural events.
Unlike Austria or Sweden, France never held a national referendum on nuclear power, but the town of Flamanville in Normandy held a local one in April 1975. Pro-nuclear councilors won the community over by campaigning on a platform of cutting local taxes and providing free electricity.
Two months later, Golfech in southwestern France held the country’s only other consultative referendum on nuclear power. This time, residents overwhelmingly voted against. EDF didn’t give up and signed an agreement with the regional authorities in February 1982. This committed EDF to paying 10 million francs per year during the construction phase and six million francs annually throughout the operational period of the power plant. EDF also promised to give priority to regional construction companies and local workers as well as commit to a number of environmental protection schemes. This significantly calmed local opposition, and construction began the same year.
To build popular support for the first nuclear reactors, EDF had also branded them as modern ‘chateaux’, suggesting that they were part of natural cultural revival. One historian who interviewed locals around the plant found that the reactors were a source of regional pride. According to one Chinon resident, the success of the area’s nuclear industry was explained by their superior temperament: ‘Brittany has not accepted [nuclear power]. But the Bretons are more chauvinistic, while we are more welcoming.’
This lobbying would continue even after plants were established. François Hollande, elected president in 2012, had pledged to shut down the two reactors in Fessenheim. While these reactors were reaching the end of their 40-year natural life, regulators had concluded that it was safe to extend their use for at least another decade after the completion of upgrades. The intended shutdown was fiercely, albeit unsuccessfully, opposed by a range of interests. Five trade unions joined forces, EDF engineers threatened to disobey instructions and not disconnect the reactor from the grid, and local officials warned of the damaging economic impact.
The end of French exceptionalism?
France’s nuclear program was not perfect. Construction costs doubled per kilowatt hour between the 1970s and late 1990s. But it still looked good by international standards, with costs rising by a factor of four in the US over the same period. And as time goes on, it looks better and better. New builds like Hinkley Point C in the UK, Vogtle 3 & 4 in the US, or Flamanville in France itself are many times more expensive per kilowatt hour of energy delivered than Civaux, the last reactor France built before freezing construction for a decade.
Rising costs in France were the result of a combination of bad luck and suboptimal planning. The original Messmer Plan envisaged as many as 170 reactors by 2000, which would have been able to meet electricity demand growing by seven percent a year. To take advantage of economies of scale, EDF embraced larger designs. New reactors moved from being 900 megawatts initially to 1,300 megawatts in 1975, and then eventually to 1,450 in 1984.
On the surface, it was reasonable to expect that bigger reactors would be cheaper per unit of electricity generated. For example, the control room and the security fencing should cost approximately the same for a 900 and a 1,300 megawatt reactor. EDF’s own estimates found that, for the same total output, a smaller series of 1,450 megawatt units would be 15 percent cheaper than a larger series of 600 megawatt reactors.
But these larger designs brought greater complexity. For example, the 1,300 and 1,450 megawatt reactors needed an extra coolant pump, steam generator, and piping circuit to remove heat safely from the larger reactor core. The engineering work required to integrate these systems did not scale linearly with increased output.
French nuclear power was then buffeted by two other forces.
First, the country had overbuilt compared to its electricity demand. By the mid-1980s, a sluggish economy meant that the steady growth in electricity demand that Messmer had forecast had evaporated. Many of France’s nuclear power stations were operating significantly below capacity. From 1984, the government limited new reactor buildout to one a year.
Second, France was not immune to the international context around nuclear regulation. The 1979 accident at the Three Mile Island nuclear power station in Pennsylvania accelerated a preexisting global move towards tighter safety regulations. This would intensify in 1986 after Chernobyl. New regulation forced EDF to retrofit reactors with new monitoring, ventilation, and containment systems. These changes built in more costs and disrupted standardization.
In 1991, the government froze new reactor orders until Flamanville-3 in 2004. Flamanville-3 was to be the flagship European Pressurized Reactor, a Franco-German collaboration between Framatome and Siemens. The two companies hoped that their modified version of the Westinghouse pressurized water reactor design could be an effective vehicle for driving up exports of nuclear technology and services.
Far from being a demonstration of French engineering prestige, Flamanville overran by €11 billion and 12 years. During the 13 year hiatus in France’s nuclear buildout, EDF and Framatome had licensed reactor designs to China and exported individual services like fuel fabrication, but along with their subcontractors, they essentially ‘forgot’ how to build new reactors. Supply chains were dismantled, industry didn’t keep up with new regulatory standards, and engineers were forced into early retirement or sought lucrative consulting contracts overseas.
As the Flamanville design was a collaboration, the reactor had been designed in accordance with stricter German nuclear standards, These, for example, required a fourth emergency safety and cooling system. In a world of tougher compliance and facing an emboldened post-Fukushima regulator, EDF struggled with a raft of compliance failures. The company lost years to paperwork, recertification, and work being redone at the regulator’s insistence. In the end, the European Pressurized Reactor wasn’t even an export success, with a slim order book of eight compared to 18 for the Westinghouse AP1000, 12 for KEPCO’s APR-1400, and 24 for Rosatom’s VVER-1200.
Where France led, the rest of the world can follow
Thanks to its rapid buildout in the 1970s and 1980s, nuclear power still provides about 70 percent of France’s electricity. France is comfortably Europe’s largest electricity exporter, supporting other European countries that rely on intermittent renewables, and has some of the lowest carbon power in the world. French homes and businesses enjoy significantly cheaper electricity bills than their neighbors without the addition of hundreds of Euros in green levies.
France provides a template for others to follow: a single buyer, streamlined regulation, a localized supply chain, building in fleets, and providing visible benefits to local communities. Parts of the French model have been successfully copied, most notably by China. In technology transfer agreements signed in 1992 and 1995, Framatome provided China with the documentation, plans, and computer code for the original 900-megawatt reactor. Chinese state-owned enterprises and research institutes reduced the number of impurities in the steel, replaced the analogue control systems with digital ones, and then set about localizing the supply chain and building a fleet. China now operates 22 of these reactors, some licensed and built in as little as five years.
France’s nuclear buildout was one of the country’s most remarkable achievements in the past fifty years. Yet France’s political elite flirted with abandoning it. Under Hollande, France swung in an anti-nuclear direction, with plans to reduce nuclear power down to 50 percent of the energy mix, fuelled by the post-Fukushima panic. When Emmanuel Macron took power in 2017, he was still committed to this target and closed down Fessenheim in 2020, despite calls from pro-nuclear groups to extend its life.
Following the surge in energy prices triggered by the Ukraine War, France reversed course yet again, and the country is planning six new reactors and exploring a potential small modular reactor program. As France attempts to position itself as a sovereign power in AI, its grid positions it well to welcome data centers and foreign direct investment. UK-based AI cloud provider Fluidstack cited France’s ‘abundant, carbon-free, and predominantly nuclear energy’ when it announced its plans for a French supercomputer. With broad-based public support ahead of a major sovereignty-focused infrastructure buildout, Messmer and Pompidou would be smiling.