France Built Reactors, Germany Shut Them, and America Argued in a Side-by-Side

Same exam, same question, three students. In the fall of 1973 the Arab oil embargo handed every wealthy democracy a pop quiz: how do you keep the lights on when the oil can be turned off by somebody else? One student went home and built. One student went home and, over several decades, un-built. And one student stood in the hallway arguing about the answer key until the bell rang.

I think that makes for the nearest thing energy policy has to a controlled experiment. Not a perfect one. Countries differ in too many ways to run a clean test. But it is a natural experiment, and when three comparable nations make three different choices from the same starting point, the results are worth reading carefully.

My claim is that the results favor the country that standardized and repeated, hurt the country that quit working plants for reasons other than engineering, and embarrass the country that could not make up its mind. Let me show you the evidence and then give the other side its due.

The shock they all faced

In October 1973, Arab oil producers imposed an embargo after the Yom Kippur War, and oil prices roughly quadrupled over the following months. The details of what that did to American drivers are in Eight Days Into His Term, Reagan Scrapped Oil Price Controls for Good. For our purposes here, the key point is that France, Germany and the United States all concluded that depending on imported fuel to make electricity was a vulnerability. All three turned, at least at first, toward nuclear power.

What they did next diverged.

France: pick a design and build it again and again

France in 1973 imported nearly all of its oil and had little coal of its own. In March 1974, Prime Minister Pierre Messmer announced a program, known ever since as the Messmer Plan, to build a large fleet of nuclear reactors in a hurry.

Here is how France did it, in five features.

1. One utility. The state-owned Électricité de France, EDF, was the buyer and operator. There was one customer, one owner and one set of requirements.
2. A small number of designs. The French reactors were built to a few closely related designs, based on a pressurized water reactor design licensed from the American company Westinghouse. Instead of custom-building every plant, France built siblings.
3. Repetition. Crews built the same thing over and over, so they learned. Twins and triplets at the same site shared parts, drawings and workers.
4. Speed. Much of the fleet came online in roughly two decades starting in the late 1970s.
5. Public buy-in. The program ran as a national project with consistent political support across governments of different parties.

The result was that nuclear power came to supply well over half of French electricity for decades, at times near three-quarters. France became a net exporter of electricity to its neighbors.

I owe you the other side of that ledger. The fleet is old and has had trouble. In 2022, corrosion problems found in some plants' piping led to the shutdown of a large part of the fleet at once, and France had to import power during an energy crunch. And France's newest reactor, the Flamanville 3 EPR, took construction from 2007 to well over a decade late. Even the champion cannot build a one-off project on time. That is, in a way, my point: the strength of the French model was repetition, and when France left the pattern, it ran into the same troubles as everyone else.

Germany: the Energiewende

Germany's story begins with the same worries but a different politics. Anti-nuclear sentiment was strong from the 1970s, and the Chernobyl disaster of 1986 deepened it. In 2000, a coalition government under Chancellor Gerhard Schröder reached an agreement with the utilities to phase out nuclear power over time. That same year Germany passed the Renewable Energy Sources Act, which guaranteed payments for wind and solar power and kicked off a build-out of those sources. The broad program of shifting the country's energy system is called the Energiewende, "energy transition."

In 2010, Chancellor Angela Merkel's government extended the lifetimes of the remaining reactors. Then came Fukushima. After the March 2011 earthquake and tsunami in Japan, her government reversed course, shut the eight oldest reactors, and set a schedule to close the rest. The last three, Isar 2, Emsland and Neckarwestheim 2, were scheduled to shut at the end of 2022. After Russia invaded Ukraine on February 24, 2022, and gas supplies became uncertain, the government let the three run a few months longer. They closed on April 15, 2023.

What did Germany use to fill the gap? Wind and solar grew. But a country cannot run on those alone when the wind drops and the sun sets, so Germany leaned on coal, including lignite, and on natural gas. A good part of that gas came from Russia, through the Nord Stream 1 pipeline that opened in 2011. When the war in Ukraine disrupted that supply, the country found itself scrambling. German households have also paid some of the highest electricity prices in Europe for years. I will not quote a figure that changes with every season, but the pattern is clear.

Germans made their choices democratically, and I do not look down on that. Many were sincere in their fear after Fukushima. My argument is not about their motives. It is about the outcome: they gave up a large source of steady, low-carbon power and replaced part of it with fossil fuels while claiming a climate benefit. That is a trade a conservative may fairly criticize.

America: build a few, cancel many, argue about the rest

The American story is told in more detail in Why America Stopped Building Reactors and What the NRC Actually Does, so here is the short version. The United States built its fleet mostly in the 1970s and 1980s, with several vendors and many customized designs. It canceled a great many planned reactors as costs, interest rates and licensing delays rose and demand growth slowed. It then went decades without finishing a new one.

The 2000s brought talk of a "nuclear renaissance." The payoff was meager. The V.C. Summer project in South Carolina was abandoned in 2017 after Westinghouse went bankrupt. Two new units at Plant Vogtle in Georgia were finished years late and billions over budget, the first in 2023 and the second in 2024. What the country did do well was keep existing plants running: the NRC allows reactors to renew their licenses, taking them from an initial 40 years to 60, and in December 2019 it approved the first renewals to 80 years, for the Turkey Point units in Florida.

Not all closures were for engineering reasons, either. Some plants shut down for market or political reasons. The Indian Point plant in New York closed its two operating units in 2020 and 2021 after a long political fight with the state.

On fuel, America has long imported most of the uranium its reactors burn, and relied on foreign enrichment services, including Russian. In 2024 Congress passed a law banning imports of Russian uranium, with waivers through a transition period. The point for our comparison is that a country can have a nuclear fleet and still be exposed on fuel if it neglects the domestic supply chain.

The table

Here is the side-by-side, in words rather than current figures, since numbers in this field change and I'd rather give you the direction than an out-of-date digit.

The cost row deserves a caveat. Most analyses of the two countries' histories find that costs stayed comparatively stable in France during the large build-out while they rose in the United States, but cost accounting across countries is notoriously tricky, and I would encourage you to read the studies before you quote a number.

The arithmetic of replacing a plant

When Germany shut reactors, it did not simply swap one machine for another. Here is a hypothetical to show why. Take a 1,000-megawatt reactor running at 90 percent of its full output over a year. That makes 1,000,000 kilowatts × 8,760 hours × 0.90 = 7,884,000,000 kilowatt-hours.

Now suppose you want to replace that energy with wind turbines that, in the location you choose, produce at 35 percent of their nameplate rating over the year. You need nameplate capacity of 7,884,000,000 ÷ (8,760 × 0.35) = 7,884,000,000 ÷ 3,066, which is about 2,571,000 kilowatts, or roughly 2,571 megawatts. For solar panels producing at 25 percent, the figure is 7,884,000,000 ÷ (8,760 × 0.25) = 7,884,000,000 ÷ 2,190 = 3,600,000 kilowatts, or 3,600 megawatts.

Those capacity percentages are assumptions I chose to illustrate the method, and real sites vary. The point is not that wind and solar are bad. It is that replacing a steady plant takes a much larger pile of equipment, plus transmission, plus a plan for the hours when the wind is calm and the sky is dark. In Germany that plan involved coal, gas and imports. Any honest comparison has to price the whole system, not only the panel.

The strongest objection

Let me give the other side its best argument, and it is a good one. France is a centralized, unitary state with a state-owned utility, a technocratic elite and no interest in being sued by every county. The American system is federal, private, litigious and proud of it. You cannot transplant the French model to a country where utilities answer to dozens of state regulators and shareholders. Meanwhile, Germany's choice reflected voters' preferences after real accidents elsewhere, and a democracy gets to choose what risks it accepts. And France's own record shows trouble in old plants and delay in new ones.

I accept most of that. I do not propose that America copy France's state ownership. A conservative should not want a government monopoly on electricity. What I take from France is the method, not the ownership. Standardize the design, certify it once, build it repeatedly, and hold the rules steady while you do. A private American consortium can do that. The ingredient that matters is repetition, and repetition does not require a ministry. It requires a regulatory system that does not change the rules in the middle of the job.

What this means for a family

I am not here to tell you wind and solar are worthless. They are real, they are growing, and in many places they are cheap to build. My point is about what stands behind them on a windless night in January. The grid has to balance supply and demand every second, as How the Power Grid Really Works and Why Supply Must Match Demand Every Second explains, and firm power that you can count on in any weather is not a luxury. It is the thing that keeps the heat on.

That is why the choice Germany made in closing working plants matters beyond Germany. It is also why America's habit of closing plants early, and its failure to finish the waste promise described in Yucca Mountain and the Nuclear Waste Promise Washington Broke, deserve attention. And for a broader grading of the options, see All of the Above, Graded on What a Family Actually Needs From Power. For the money side, Trace Your Family's Energy Dollar From the Meter and Pump to the Grocery Bill follows your own dollar.

Who each model suits

If you value speed and a single decision-maker, the French model has the track record, with the caveat that it is a state enterprise. If you value democratic choice above all and are willing to accept the costs, Germany's path is a coherent one, though the costs are real. If you value private enterprise and competition, the American system has the most potential, but only if Washington stops moving the goalposts. The American model is the one with the most to gain, because it has the most room to improve.

The bottom line

• France, Germany and the United States faced the same oil shock in the 1970s and chose different nuclear paths: a standardized national build-out, a long phase-out ending in April 2023, and a mix of costly one-off projects, cancellations and license extensions.
• The lesson of France is repetition. The lesson of Germany is that closing working low-carbon plants has costs. The lesson of America is that changing rules in the middle of construction is expensive.
• You do not need a state utility to copy the French method. You need standard designs, a license that is certified once, and rules that stay put.
• Ask your representatives whether they will protect the plants still running, push for standard designs, and keep firm power in the mix when they talk about a reliable grid.