Infrastructure · By Amaya Kavya · 2026-07-14 · 11 MIN

Why the World Is Choosing Nuclear Again, Despite Chernobyl and Fukushima

Two of the worst industrial accidents of the last century were nuclear, and yet support for nuclear power is rising, more than thirty countries have pledged to triple it, and tech giants are switching old reactors back on. This is what the numbers actually say, which fears are real, and how the mood turned around.

In 2019 the last working reactor at Three Mile Island was switched off. It was not the reactor that melted down in 1979. That one, Unit 2, has been dead for over forty years. Unit 1 next door had run quietly and safely the whole time, and it was closed only because it was losing money. Five years later, in September 2024, its owner announced it would be switched back on. The buyer of the electricity is Microsoft, which needs round the clock power for the data centres behind its artificial intelligence. The plant that stands on the site of America's most famous nuclear scare is being brought back to life to run the future.

That tells you something has shifted. Nuclear power handed the world three accidents everyone remembers: Three Mile Island in 1979, Chernobyl in 1986, and Fukushima in 2011. For decades those names were enough to end an argument. And yet support for nuclear power is now higher than it has been in years. In 2023 more than twenty countries signed a pledge to triple the world's nuclear capacity by 2050. In 2024 heads of government met in Brussels for the first summit ever held on nuclear energy alone. In the United States, a clear majority now wants more of it.

So why, given the disasters, do people keep choosing nuclear? The reason is not that anyone forgot. It is that the fear and the facts slowly drifted apart. This is what the numbers say, which of the fears are real, and how the mood turned around.

What the three accidents actually did

Start with the accidents, because they are the whole reason for the fear.

Three Mile Island, in March 1979, was a partial meltdown of one reactor. Very little radiation escaped. Study after study since has found no measurable rise in cancer around the plant. Nobody died. It frightened people because it was new, because the operators themselves did not understand what was happening for hours, and because a film about a nuclear meltdown had come out that same month. The fear was real. The harm was not.

Chernobyl, in April 1986, was the real disaster. During a botched safety test the reactor exploded. It had no proper containment building, the kind every Western plant is wrapped in, so when the core caught fire it threw radioactive material straight into the air and across Europe. Two people died in the explosion. Twenty-eight firefighters and plant workers died of acute radiation sickness in the weeks that followed. The clearest longer harm was thyroid cancer in children, much of it avoidable if the Soviet state had not hidden the accident and had simply stopped people drinking contaminated milk. In 2005 a review led by the World Health Organisation estimated that radiation could eventually cause up to about 4,000 deaths among the most exposed groups, the emergency workers and evacuees and people living in the worst hit areas. It was an awful event. It was also the product of a reactor design and a culture of secrecy that no other country used.

Fukushima, in March 2011, is the one most people get wrong. A magnitude nine earthquake and a wave up to fifteen metres high hit the north east coast of Japan. The wave killed around 18,000 people. It also knocked out the power that cooled the Fukushima Daiichi plant, and three reactors melted down. The radiation, though, killed almost no one. Years later the Japanese government recognised a single death, a plant worker who developed lung cancer, as caused by his exposure. The United Nations scientific committee found no rise in cancer in the public that could be tied to the accident. What did kill people was the panic and the evacuation. More than 2,000 deaths in Fukushima prefecture were later classed as disaster related, mostly frail and elderly people whose health gave way after they were moved out of homes and hospitals. At a climate summit in 2021 the head of the IAEA, Rafael Grossi, said plainly that no one died from radiation at Fukushima. Some people in the room laughed, because it clashes so hard with what everyone assumes. It is broadly what the evidence shows.

Put the accidents in a wider frame and the picture flips. Researchers work out deaths per unit of electricity produced, counting both accidents and the slow harm of pollution. On that measure coal kills about 25 people for every terawatt hour of electricity. Oil and gas are lower but still bad. Nuclear, even with Chernobyl and Fukushima folded in, kills fewer than 0.1. That puts it roughly level with wind and solar, and hundreds of times safer than the coal it usually replaces. The reason is not complicated. Coal and gas kill quietly, through the soot and fumes people breathe every day, and those deaths never make the news. Nuclear kills rarely and all at once, so it feels far more dangerous than it is.

The fears that do not hold up

The dread outlived the accidents, and it settled into a set of beliefs that sound obvious and mostly are not.

The first is waste. High level nuclear waste is real, and it stays dangerous for a very long time. But the amount is tiny. All the used fuel that American reactors have produced in more than sixty years would cover a single football field to a depth of a few metres. It is solid, it is counted, and it sits in sealed steel and concrete casks, not sloshing around in the open. Set that against the ash and the carbon dioxide a coal plant pours out every hour. And there is now somewhere to put it. In Finland a repository called Onkalo has been dug 430 metres down into rock nearly two billion years old, built to hold the fuel safely for a hundred thousand years. It ran its first full trial in 2024 and is due to start taking waste around 2026. Sweden and France are building their own.

The second is radiation in daily life. People picture a plant leaking poison into the neighbourhood. In normal running a nuclear plant gives off almost nothing. You take in more radiation from a single chest X ray, a long flight, or the natural rock under your house. And oddly, living next to a coal plant exposes you to more radiation than living next to a nuclear one, because coal carries traces of uranium and thorium that go up the chimney in the ash.

The third is the bomb. A power reactor cannot explode like a nuclear weapon. The fuel is nowhere near concentrated enough, and the physics does not allow it. The link between civilian power and weapons is a genuine worry that inspectors watch closely, which is part of why the IAEA exists, but the reactor keeping a city's lights on is not a bomb waiting to go off.

Why people actually want it now

Fear fading is not the same as wanting more. Several solid reasons pushed people towards nuclear rather than just away from being scared of it.

The first is reliability. A nuclear plant runs almost all the time. In the United States reactors run at about 92 percent of their maximum output across a year, the highest figure of any source and nearly three times that of wind or solar. It does not care whether the wind drops or the sun sets. As grids lean harder on weather dependent renewables, that steady around the clock output looks more valuable, not less.

The second is carbon. Nuclear makes electricity with almost no carbon dioxide. For anyone serious about cutting emissions while keeping the lights on, it is one of very few sources that is both low carbon and always available. That is why the climate talks, which used to leave it out entirely, now name it.

The third is energy security. When Russia invaded Ukraine in 2022 and gas prices across Europe went haywire, the worth of power you make at home, from fuel that lasts years rather than arriving by pipeline, stopped being an abstract idea. Countries that had planned to close reactors paused. A few reversed.

The fourth is new demand, and it is the one changing the mood fastest. Data centres for artificial intelligence swallow enormous, constant amounts of power. The big technology firms have gone straight to nuclear to get it. Microsoft's deal to restart Three Mile Island Unit 1 is the loud example. Google and Amazon have both signed agreements to back a new generation of smaller reactors. When the richest companies on earth start buying nuclear to power what they think comes next, public opinion tends to follow the money.

And it has. A 2024 survey by the Pew Research Center found 56 percent of Americans in favour of expanding nuclear power, up 13 points in four years, and rising among both Republicans and Democrats.

The meetings that moved it

The change did not only happen in power markets. It happened in rooms full of governments, and the run of those meetings is a story in itself.

It began with a speech. In December 1953 President Dwight Eisenhower stood at the United Nations and gave the "Atoms for Peace" address, arguing that the same science behind the bomb could power and heal the world instead. Two years later, in 1955, the first big international conference on the peaceful use of atomic energy met in Geneva, where scientists from rival blocs actually compared notes. In 1957 that momentum created the International Atomic Energy Agency, the United Nations body that to this day both promotes nuclear power and polices it for weapons.

For a long stretch the meetings that followed were about fear rather than growth. After Chernobyl, countries wrote the Convention on Nuclear Safety in 1994, a binding promise to hold each other to shared standards. After Fukushima, ministers gathered at the IAEA in Vienna in 2011 to tighten those rules again.

Then the mood turned. For years the big United Nations climate conferences barely said the word nuclear. At COP28 in Dubai in December 2023 that broke. More than twenty countries, among them the United States, France, Britain, Japan and the United Arab Emirates, signed a declaration to triple nuclear capacity by 2050, and for the first time nuclear was named in the summit's main stocktake as part of the answer to climate change. The group has since grown past thirty. A few months later, in March 2024, Belgium and the IAEA hosted the first Nuclear Energy Summit in Brussels, the first gathering at head of government level held on nuclear power alone. Around thirty leaders and ministers turned up to say the same thing: build more.

What the argument is really about now

None of this makes nuclear perfect, and pretending otherwise would be its own kind of dishonesty. The strongest case against building it today has almost nothing to do with safety. It is money and time. New plants are hugely expensive, and they have a habit of running years late and far over budget. Flagship projects in France, in Finland and in the American state of Georgia all came in painfully behind schedule and billions above their first price. That, not radiation, is the real reason some countries still hesitate. The great hope now is the small modular reactor, factory built and in theory cheaper, but that is mostly still a hope. Very few are actually running. Anyone selling nuclear as quick and easy is not being straight with you.

But the thing people feared most about it, that it would quietly kill them in large numbers, turned out to be the thing it does least.

In Finland, deep under the forest, workers are lowering the first canisters of spent fuel into rock that has sat undisturbed since before there was complex life on Earth, then planning to seal them in and let the ground hold them for longer than our species has existed. It is a strange, patient answer to a problem people once said had no answer at all. Back in Pennsylvania, on the very site of the accident that scared a generation, engineers are checking the pipes on a reactor that is about to run again. Both of those things are true at the same time, and both are why the fear finally began to lose.

Sources

Delvewire