The Doomsday Myth: Nuclear Weapons Will Not End the World — and That Should Worry You More

Estimated reading time: 16 minutes

On 1 March 1954, twenty-three Japanese fishermen watched the horizon catch fire. Their tuna boat, the Daigo Fukuryū Maru, was working waters roughly 130 kilometers east of Bikini Atoll. Crewman Ōishi Matashichi later described the moment in one image: a yellow flash poured through the porthole.

Nobody had warned them. They finished hauling their gear. Then a fine white ash began to fall across the deck.

It fell for hours. The men brushed it from their hair and kept working. The ash was pulverised coral, lifted by the fireball and returned to earth. Within days, every man aboard showed signs of acute radiation sickness. The boat’s radio operator, Aikichi Kuboyama, died on 23 September 1954.

The test that outgrew its own forecast

The ash came from Castle Bravo, a United States thermonuclear test at Bikini Atoll. It released roughly 15 megatons. Its designers had expected a small fraction of that. The forecast failed, and the fallout plume ran hundreds of kilometres downwind.

The Marshallese paid the heaviest price. Fallout reached Rongelap, home to 64 people, and Utrik, home to 157. American authorities evacuated more than 230 residents from Rongelap, Rongerik and Utirik after the exposure had already begun.

Children on Rongelap played in flakes that looked like snow. Nobody had told them what the snow was.

How a bad test became a global story

On 31 March 1954, Atomic Energy Commission chairman Lewis Strauss met reporters in Washington. Pressed on how large a hydrogen bomb could grow, he confirmed that one could take out a city.

That sentence travelled further than any technical report. Within months, Japan had a mass anti-nuclear movement and a monster film about a creature woken by an H-bomb test. Within a year, Bertrand Russell and Albert Einstein had signed a manifesto warning of the end of the human race.

The fear was earned. Castle Bravo poisoned real people, and Washington spent years minimising it. But something else grew alongside the fear: a story about what nuclear weapons do that has drifted steadily away from the evidence.

Why we are publishing this now

That story matters again, because Europe now has to make decisions about nuclear weapons that it avoided for thirty years.

New START, the last treaty limiting American and Russian strategic forces, expired on 5 February 2026. For the first time since the early 1970s, no binding cap exists on either arsenal. In May, the eleventh Nuclear Non-Proliferation Treaty Review Conference closed in New York without a consensus document — the third consecutive failure. The next review is scheduled for 2031.

Meanwhile, France has changed its doctrine, Germany has opened talks, and nine European states are now in structured nuclear dialogue with Paris.

Those decisions depend on what citizens believe nuclear weapons actually do.

What we checked, and how

Myth one: a single bomb erases a country

Blast damage grows with the cube root of yield. That single fact governs almost everything about nuclear planning.

Increase a weapon’s explosive power eightfold, and you roughly double the radius of severe blast damage. You do not multiply the devastation by eight. Physics imposes diminishing returns on size.

This is why arsenals stopped chasing enormous yields decades ago.

What Hiroshima actually tells us

The Hiroshima bomb released roughly 15 kilotons. Local authorities recorded around 140,000 deaths in Hiroshima and 74,000 in Nagasaki before the end of 1945. Those numbers are catastrophic by any human measure.

They are also city-scale numbers, not national ones. Japan did not become uninhabitable. Both cities function today as ordinary Japanese cities, with trams, universities and commuter traffic.

Myth two: today’s warheads are bigger than ever

The direction of travel here surprises most readers. The bombs got smaller.

A 1997 US National Academy of Sciences review recorded the shift plainly. The average yield of an American nuclear weapon fell from above one megaton at its peak to a little over 200 kilotons.

The number nobody expects

Declassified US Department of Energy data puts the total explosive yield of the American stockpile at its 1960 maximum: 20,491 megatons. Historian Alex Wellerstein, who built the widely used NUKEMAP effects tool, notes that this equals roughly 1.4 million Hiroshima bombs.

Global first-strike destructive potential peaked above 15,000 megatons in the early 1980s, according to Our World in Data’s compilation. At that peak, first-strike weapons could have directly destroyed more than 40 per cent of the world’s urban land. Today the same measure sits above 2,500 megatons, covering roughly 7 per cent.

Why accuracy shrank the bomb

Enormous yields existed to compensate for enormous inaccuracy. Early missiles missed by kilometres, so planners bought certainty with megatons.

Guidance improved. Once a warhead could land close to its aim point, the extra yield bought nothing. The Soviet 50-megaton device of 1961 was a political demonstration, not a weapon anyone intended to field.

Myth three: the land stays dead for ten thousand years

Standard weapons-effects references describe a rule of thumb for fission fallout decay. For every sevenfold increase in time, the radiation dose rate falls roughly tenfold.

Seven hours after a detonation, the dose rate has dropped to about a tenth of its one-hour value. Two days later, it sits near one hundredth. Two weeks later, near one thousandth.

Airbursts, ground bursts and the Chernobyl comparison

Burst height changes everything. An airburst fireball never touches the ground, so it lifts little soil and produces comparatively little local fallout. Hiroshima and Nagasaki were airbursts. A surface burst pulls debris into the fireball and creates heavy, persistent local contamination.

Chernobyl is the more instructive comparison for long-term contamination. A reactor core holds tones of fuel and years of accumulated long-lived fission products. A fission weapon consumes on the order of a kilogram. The inventories are not comparable.

What the survivors’ own data records

The Radiation Effects Research Foundation has followed Hiroshima and Nagasaki survivors since 1950. It is the largest long-term radiation study ever conducted.

Its thirteenth mortality report tracked 86,572 people with individual dose estimates over 47 years. Researchers recorded 9,335 solid cancer deaths. They attributed roughly 440 of those — about 5 per cent — to bomb radiation. Among 31,881 non-cancer deaths, they attributed around 250.

Every one of those deaths is a person. The numbers still sit far below what most people assume.

The finding almost nobody knows

Researchers have also followed roughly 77,000 children of survivors since 1945. They have observed no increased risk of malignancies or other diseases from parental exposure. The foundation continues to investigate, noting the cohort remains relatively young.

The popular image of generations of mutated descendants has no support in the data.

Where myth-busting must stop

Correcting exaggerations is not the same as minimising harm, and we want to be exact about that line.

Rongelap was evacuated, resettled, and abandoned again. Marshallese families still live with displacement, thyroid disease and compensation disputes seventy years after Castle Bravo. The atolls carry measurable contamination today.

Hiroshima recovered. Bikini did not. Both facts are true, and the difference between them is mostly burst height, soil chemistry and political neglect.

Myth four: nuclear war means human extinction

This is the hardest section to write honestly, because scientists have argued about it for forty years and still do.

The nuclear winter hypothesis emerged in 1982 and 1983. Paul Crutzen and John Birks published first; Richard Turco, Owen Toon, Carl Sagan and colleagues followed in Science. Their argument is straightforward. Cities burn, firestorms loft soot into the stratosphere, and sunlight fails for years.

The step everyone fights about

The dispute is not really about climate physics. Both camps agree that large quantities of stratospheric soot would cool the planet severely.

The fight concerns whether that soot gets up there.

In 2018, a Los Alamos National Laboratory team led by Jon Reisner modelled an urban firestorm in detail. Their simulation showed soot largely failing to reach the stratosphere. They predicted far smaller climatic effects than earlier work.

Alan Robock, Toon and Charles Bardeen replied in 2019. They argued the modelled fire was not typical of a nuclear attack on a dense city. In 2021, a Berkeley team led by David Romps identified a further problem: the Los Alamos simulation had omitted water vapour, which helps drive plumes upward.

Who is arguing, and from where

Institutional affiliation tracks the disagreement closely, and that deserves stating plainly.

University and civilian-agency scientists generally find severe climatic effects. Scientists at the US National Nuclear Security Administration laboratories — Los Alamos and Lawrence Livermore, which design American weapons — generally find milder ones.

Los Alamos climate scientist Manvendra Dubey told a National Academies committee in 2023 that researchers need to be very even handed when we talk about it. Critics counter that preferring moderate estimates by default is itself a position, not neutrality.

What the National Academies concluded

Congress commissioned an independent review, and it landed on 25 June 2025.

The National Academies of Sciences, Engineering, and Medicine published Potential Environmental Effects of Nuclear War. Its verdict was deliberately unsatisfying. The committee found too many uncertainties to say how any given scenario would reshape climate, ecosystems or food supply.

It agreed on the conditional: heavy stratospheric soot means severe global cooling. It declined to adjudicate whether the soot gets there.

Five billion is not eight billion

The most widely cited catastrophic estimate comes from Lili Xia, Robock and colleagues, published in Nature Food in 2022.

Their models estimate that more than two billion people could die of famine after an India–Pakistan war. A full US–Russia exchange could kill more than five billion. Even a small regional war cut modelled global calorie production by 7 per cent within five years.

Note what that scenario is, and what it is not. It describes the collapse of industrial civilization and the worst famine in human history. It does not describe extinction.

Why the difference is not academic

We think this distinction carries real political weight, and we want to argue for it directly.

If nuclear war ends everything, then nothing after the first detonation matters. Civil defence becomes theatre. Damage limitation becomes fantasy. Arms control becomes symbolic.

If nuclear war instead kills a large fraction of humanity, then every marginal decision matters enormously. Warhead counts matter. Alert levels matter. Sheltering guidance matters. The gap between two billion dead and five billion dead is three billion people.

Myth five: nothing an ordinary person does would matter

Public-health agencies across NATO states give the same core instruction. Go inside, stay inside, stay tuned.

The reasoning follows directly from the decay curve. The most dangerous fallout is also the shortest-lived. A person who shelters in a basement or a building’s interior core for the first day avoids the worst of it.

The instinct to flee immediately is the dangerous one. Roads offer no shielding at all.

Fatalism as a political convenience

Here we should be blunt about who benefits from apocalyptic framing.

A public that believes nuclear war ends the world does not scrutinise procurement budgets. It does not read posture reviews. It does not ask why 2,100 to 2,200 warheads sit on high alert, launchable within minutes.

Despair is not the same as opposition. It is quieter, and it asks nothing of anyone.

The numbers that should actually worry you

Now consider what the current data shows, stripped of apocalyptic framing.

SIPRI counted approximately 12,187 warheads worldwide in January 2026. Roughly 9,745 sat in military stockpiles for potential use. About 4,012 were deployed with missiles and aircraft.

Compare those figures with January 2025: 12,241 total, 9,614 in stockpiles, 3,912 deployed. The headline total dropped. The usable stockpile and the deployed force both grew.

Why the total keeps falling anyway

The declining headline number reflects dismantlement of retired American and Russian warheads, nothing more.

SIPRI notes that annual dismantlement is slowing. Its analysts expect new warheads entering stockpiles to outpace retired warheads leaving them, probably soon. At that point the reassuring downward line reverses.

China now holds around 620 warheads and expands faster than anyone else. Even at 1,000 by 2030, it would hold roughly a quarter of the American or Russian stockpile.

Europe’s decision, taken in public

Europe is now doing something it has not done since the 1980s. It is debating nuclear deterrence openly, and the debate has moved fast.

On 2 March 2026, Emmanuel Macron spoke at the Île Longue submarine base. He announced the first increase in France’s nuclear stockpile since 1992. He also unveiled dissuasion avancée, or forward deterrence.

The framework offers European partners consultation, participation in nuclear exercises, visits to strategic sites and the possible temporary deployment of French nuclear-capable aircraft on allied soil.

Who has joined, and who has not

By June 2026, nine European states had entered structured bilateral discussions with Paris: the United Kingdom, Germany, Belgium, the Netherlands, Denmark, Poland, Sweden, Greece and Norway.

Poland’s prime minister, Donald Tusk, framed the shift in blunt terms, saying it would ensure that our enemies will never dare to attack us.

The Baltic states and Finland are absent from the original framework. Those are the countries sharing a land border with Russia. Analysts at the Atlantic Council read that absence as a real limit on French extended deterrence.

The proliferation risk hiding inside the offer

Chatham House reads Macron’s initiative partly as a preventive move, and we find that reading persuasive.

If European states doubt collective deterrence, some will consider building their own weapons. Tusk has publicly mused about acquiring them. A nuclear-armed Poland would strain the EU’s non-proliferation position badly.

Offering partners a role inside the French force is, among other things, a way of discouraging new national programmes.

Testing, and the taboo nobody has broken yet

One more thread runs through all of this, and it is unresolved.

On 30 October 2025, Donald Trump announced on social media that he had directed the Pentagon to begin nuclear weapons testing on an equal basis with other powers. No state except North Korea has conducted an explosive nuclear test since the 1990s.

What that instruction means in practice remains contested inside the US government. Environmental groups have raised concerns about renewed plutonium pit production and about contamination at test sites.

What we are watching next

Three dates deserve space in any European reader’s calendar.

First, the United Nations Scientific Panel on the Effects of Nuclear War is due to report in 2027. It will be the first UN study of its kind since 1988. China voted for it; the United States abstained.

Second, the next NPT Review Conference falls in 2031. Conference president Do Hung Viet summed up May’s failure with dark economy: My goal was to make everyone equally unhappy.

Third, watch whether Washington and Moscow agree any successor arrangement to New START.

The conclusion we reached

Nuclear weapons are not a doomsday machine. They are something more difficult to think about.

They are instruments capable of killing hundreds of thousands of people in minutes and billions more through famine, while leaving the planet habitable and the survivors to sort out what remains. That scenario has no comforting shape. It cannot be dismissed as fantasy, and it cannot be filed away as the end of everything.

Castle Bravo taught two lessons in 1954. The first was that these weapons harm people far beyond any target. The second took longer to learn: that the forecasts were wrong, in both directions, and that getting them right is the whole job.

Europe now has to make choices about deterrence with New START gone and the NPT process stalled. Those choices will be worse if the public debate runs on myth.