The Atom Returns: Why the World's Most Feared Energy Source Is Its Best Hope

In September 2024, Microsoft signed a twenty-year power purchase agreement to restart Unit 1 of the Three Mile Island nuclear plant in Pennsylvania. The restart will cost Constellation an estimated $1.6 billion. The facility whose partial meltdown in 1979 had become the defining image of nuclear danger in America was being brought back to life — not by an energy company, but by a software company that needed electricity for artificial intelligence. Microsoft’s data centres consume power at a scale that would have seemed fantastical a decade ago: a single hyperscale facility draws a hundred megawatts or more, around the clock, every day, regardless of wind speed or cloud cover. Neither solar panels nor wind turbines could guarantee that. Google followed weeks later with a deal for Kairos Power’s small modular reactor. Amazon signed for nuclear-adjacent capacity in the same state. In the space of a few months, three of the most powerful technology companies on earth had placed their bets on the energy source that polite opinion had spent forty years declaring dead.

The irony is rather thick. The same Silicon Valley that lectures the world about sustainability discovered that when you need power at two in the morning on a windless January night, you need something rather more reliable than good intentions and a solar farm. The Germans, meanwhile, had spent a staggering fortune on their Energiewende — the most expensive energy transition in human history — shut their last nuclear reactors, and found themselves importing Russian gas to keep the lights on. The French, who generate roughly seventy per cent of their electricity from nuclear power at a carbon intensity less than half of Germany’s, were too busy drinking affordable wine under affordable electric lighting to comment.

The atom never really left. It simply waited for the world to exhaust every fashionable alternative before knocking on the door again.

The Promise and the Fear

The civilian nuclear age began with ambition bordering on euphoria. Dwight Eisenhower’s “Atoms for Peace” address to the United Nations in December 1953 promised a future in which nuclear energy would power cities, desalinate oceans, and lift the developing world out of poverty. The International Atomic Energy Agency was established in 1957. Britain opened Calder Hall in 1956; the United States launched Shippingport in 1958. By the early 1970s, nuclear power was growing faster than any energy source in history.

The nation that embraced the atom most decisively was France. After the 1973 oil crisis — OPEC quadrupled the price of crude overnight, and the French economy, almost entirely dependent on imported oil, shuddered — Prime Minister Pierre Messmer announced what became known as the Plan Messmer. France would build enough nuclear reactors to achieve near-total energy independence. Over the next fifteen years, France constructed fifty-six pressurised water reactors — a rate of roughly one every three months. The programme succeeded because it made one decision that Britain and the United States could not: it standardised on a single reactor design, built in series, with a single national utility managing the fleet. By the mid-1980s, nuclear provided over seventy per cent of French electricity. Britain and America, by contrast, built expensive one-off designs through adversarial regulatory and legal systems. France built an energy system. Britain and America built arguments.

By 1990, nuclear supplied approximately seventeen per cent of global electricity from 413 reactors in thirty-one countries. The technology was proven, the economics were competitive, and the growth trajectory suggested nuclear would dominate electricity generation within a generation. Then something changed — not in the physics, but in the politics.

Chernobyl, on 26 April 1986, killed thirty-one people directly and caused an estimated four thousand to sixteen thousand excess cancer deaths over subsequent decades. But the true casualty — the one that would cost the world four decades of clean energy — was public trust. Spencer Weart, in The Rise of Nuclear Fear, traces how the atom was never evaluated on its engineering merits alone. From the beginning it was entangled with apocalyptic imagery — the mushroom cloud, the invisible poison, the mutant future. The RBMK reactor at Chernobyl had no containment building and a positive void coefficient that made it inherently unstable at low power; the operating culture treated safety protocols as suggestions. These were specific Soviet failures, not inherent features of nuclear technology. What mattered, though, was the image: a burning reactor, an invisible cloud drifting across Europe, and a word — Chernobyl — that became shorthand for technological hubris. Germany’s Green Party gained decisive momentum. Italy voted to phase out nuclear. Sweden announced a phase-out, later reversed.

Fukushima Daiichi, on 11 March 2011, killed zero people from radiation. The United Nations Scientific Committee on the Effects of Atomic Radiation has confirmed no attributable radiation deaths. The earthquake and tsunami that caused the meltdowns killed approximately twenty thousand people — but it was the reactors, not the wave, that dominated global headlines. Angela Merkel, a physicist by training, ordered a full nuclear phase-out for Germany by 2022. Japan shut all fifty-four of its reactors. The perverse arithmetic went unexamined. Nuclear power, according to Our World in Data, causes 0.03 deaths per terawatt-hour of electricity produced. Coal causes 24.6 — eight hundred times as many. Oil causes 18.4. Natural gas causes 2.8. Even solar and wind, when you include manufacturing and installation accidents, kill more people per unit of energy than nuclear does. It is, by every statistical measure, the safest form of electricity generation ever devised. This is not a close call. It is not a matter of interpretation. It is a numerical fact as solid as the half-life of uranium-235.

Germany’s nuclear exit is estimated by researchers at the National Bureau of Economic Research to have caused approximately 1,100 additional deaths per year from increased air pollution, as coal and gas plants filled the gap left by the shuttered reactors. These deaths — real people, from real particulate matter, in real German cities — attracted no headlines and no protests. They were statistical deaths, invisible in a way that a reactor breach is not. The anti-nuclear movement correctly identified that nuclear power carries risks. It incorrectly concluded that those risks were worse than the alternative — which was never a pastoral utopia of wind turbines and wildflowers, but the continued burning of fossil fuels at industrial scale.

The Renaissance

Something shifted around 2022, and it shifted on five fronts simultaneously.

The first was climate arithmetic — and it had finally caught up with what France had known for fifty years. Modelling exercises, including those summarised in the IPCC’s Sixth Assessment Report and the IEA’s Net Zero roadmap, now state plainly that most credible pathways to limiting warming require significant nuclear expansion. France and China had reached that conclusion through national calculation long before any supranational body endorsed it. France’s grid emits roughly 56 grams of CO₂ per kilowatt-hour; Germany’s emits approximately 385 — nearly seven times as much, for the same unit of electricity.

The second was energy security. Russia’s invasion of Ukraine in February 2022 exposed, with brutal clarity, how dependent Europe had become on imported natural gas. Nuclear fuel, by contrast, is energy-dense and easy to stockpile; a single fuel load lasts eighteen to twenty-four months, and the fuel itself is sourced from geopolitically stable nations — Australia, Canada, Kazakhstan. A country with a nuclear fleet does not check the weather forecast or the gas price every morning.

The third was the insatiable appetite of artificial intelligence. Training a single large language model can consume tens of megawatts for weeks. Hyperscale data centres draw five hundred megawatts to a gigawatt — equivalent to a large city — and they need that power continuously. These facilities cannot tolerate intermittency. When a data centre loses power, it loses revenue measured in millions per hour and customers measured in the hundreds of millions. Microsoft, Google, and Amazon did not turn to nuclear out of ideology. They turned to it because they needed kilowatt-hours, twenty-four hours a day, regardless of the weather.

The fourth was new technology. Small modular reactors — compact designs of fifty to three hundred megawatts — promise factory manufacturing, standardised components, and dramatically shorter construction times. Rolls-Royce SMR, GE Hitachi’s BWRX-300, and TerraPower’s Natrium are all advancing through regulatory approval. China’s HTR-PM, a high-temperature gas-cooled reactor at Shidao Bay, became the world’s first operational Generation IV reactor in 2023.

The fifth was fusion. In December 2022, the National Ignition Facility in California achieved fusion ignition for the first time — producing more energy from a fusion reaction than the lasers put in. Commercial fusion remains distant, but the milestone shifted the conversation from “always thirty years away” to “which decade.”

At COP28 in Dubai, twenty-two nations announced individual pledges to triple nuclear capacity by 2050 — a diplomatic echo of choices that serious countries had already made on their own. The summit did not lead; it followed. The atom had returned to the mainstream because nations acting in their own interest had proved the case.

Winners, Losers, and One Spectacular Own Goal

France simply never stopped. The fleet built under the Plan Messmer gives it the lowest carbon intensity of any major European economy and among the lowest industrial electricity prices on the continent. President Macron announced six new EPR2 reactors, with an option for eight more. France has been quietly right for fifty years while louder nations were expensively wrong.

China is building nuclear capacity at a pace that dwarfs anything in the West. It has fifty-seven operational reactors, twenty-eight under construction, and plans for roughly 150 by 2035. It builds a reactor in five to six years — half the Western timeline. China’s nuclear programme is, in scale and speed, the closest modern equivalent to France’s Plan Messmer.

The United Kingdom has approved Sizewell C and established Great British Nuclear as a dedicated delivery body. Hinkley Point C, the first new reactor in a generation, has been plagued by budget overruns — but it will eventually produce 3.2 gigawatts of reliable, zero-carbon baseload.

The United States has extended operating licences for its existing fleet, included nuclear production tax credits in the Inflation Reduction Act, and brought the two new AP1000 reactors at Vogtle online — the first new American reactors in three decades.

Japan is cautiously restarting its fleet after the post-Fukushima shutdown. Of its fifty-four reactors, twelve have returned to service and more are in the approval pipeline. Prime Minister Kishida declared that Japan would maximise nuclear energy use, including next-generation designs — a remarkable reversal for the only country to have experienced nuclear weapons in war.

South Korea, which derives roughly thirty per cent of its electricity from nuclear, reversed a short-lived phase-out policy and is now planning new reactors. Its APR-1400 design has been exported to the UAE — four reactors at Barakah, operational and generating power in a country that previously depended entirely on natural gas.

And then there is Germany, which future energy historians will study with the same bewildered fascination that military historians reserve for the Maginot Line. On 15 April 2023, Germany shut its last three nuclear reactors — in the middle of an energy crisis triggered by dependence on Russian gas. The reactors were functioning perfectly. They were producing roughly six per cent of German electricity at near-zero marginal cost and zero carbon emissions. They were shut because a political commitment made in 2011, in the emotional aftermath of Fukushima, had acquired the force of religious doctrine. The Green Party refused to countenance even a temporary extension. Coal plants were switched on. Electricity sector emissions rose. BASF, the world’s largest chemical company, announced it would shift investment to China, where energy was cheaper and more reliable. The Energiewende spent approximately five hundred billion euros and delivered a grid that is more expensive, more carbon-intensive, and more dependent on the weather than France’s nuclear fleet. Future generations will note, with some astonishment, that the country which shut its zero-carbon reactors during a climate emergency did so in the name of environmentalism.

The Atom Does Not Care

The nuclear renaissance is not an ideology. It is physics, economics, and arithmetic. Decarbonising the global energy supply without nuclear is, in the assessment of virtually every credible energy modelling exercise, either physically impossible or so ruinously expensive that no electorate would tolerate the cost. The nations that understood this early — France in the 1970s, China in the 2010s — have the cheapest, cleanest grids. The nations that delayed — Germany above all — are paying in higher emissions, higher electricity prices, and the slow migration of energy-intensive industry to countries that made better choices.

The challenges are real and should not be minimised. NuScale’s first small modular reactor project was cancelled due to cost overruns — a reminder that the technology is promising but not yet proven at commercial scale. Hinkley Point C’s budget has roughly doubled. The Olkiluoto 3 reactor in Finland was delivered fourteen years late. Western democracies have largely lost the institutional muscle memory for large nuclear construction: the supply chains have atrophied, the skilled workforce has aged out, and the regulatory systems — designed in an era of abundant caution — add years to timelines that China manages in months. These are serious obstacles. But the direction of travel is unmistakable. The atom is returning because the alternatives — not the alternatives in a policy paper, but the alternatives in the real world of intermittent renewables and geopolitical gas dependency — have been tried and found wanting.

Physics does not negotiate with ideology. The atom does not care whether you approve of it. It simply offers the densest, most reliable, lowest-carbon energy source available to human civilisation — and waits, with the patience of a half-life measured in millennia, for the world to stop arguing and start building. Send that to someone who still thinks nuclear is a dead end.