How Nuclear Weapons Work — and Why They Haven't Been Used Since 1945

At 8:15 a.m. on August 6, 1945, a uranium bomb exploded roughly 600 meters above Hiroshima. Less than one kilogram of its uranium fissioned, yet the blast matched about 15,000 tons of TNT. Three days later a plutonium bomb struck Nagasaki, and no nuclear weapon has been used in war since. Here is how nuclear weapons work, from chain reactions to hydrogen-bomb staging, followed by the evidence for each explanation of the 81-year gap: deterrence, a norm against use, and a fair amount of luck.

Key takeaways

• A fission weapon works by assembling more fissile material (uranium-235 or plutonium-239) than can sustain a chain reaction, then holding it together for a few millionths of a second. The Hiroshima bomb fissioned less than a kilogram of uranium.
• A hydrogen bomb adds a second stage. X-rays from a fission "primary" compress a fusion "secondary," and the compression, not a spark, ignites the fuel.
• Modern strategic warheads are not a thousand times more powerful than the Hiroshima bomb. The common US submarine warheads are roughly 100 and 475 kilotons, about 7 and 32 times the Hiroshima yield.
• Three explanations compete for why nuclear weapons have not been used since 1945: mutual deterrence, a norm against use, and luck. The evidence supports all three in part, and none can be proven, because the counterfactual cannot be observed.
• As of early 2026 the Federation of American Scientists estimates about 12,187 warheads worldwide, down from a peak of about 70,300 in 1986. New START, the last treaty limiting the US and Russian strategic arsenals, expired on February 5, 2026.

How a chain reaction turns a lump of metal into an explosion

Start with the scale of the thing. A ton of TNT is defined as 4.184 billion joules, which means one kilogram of TNT releases about 4.2 million joules. Fission of one kilogram of plutonium-239 releases roughly 17 kilotons of TNT equivalent, according to the physics tables kept by the Nuclear Weapon Archive. That is about 17 million times more energy per kilogram than the chemical explosive. The reason is that chemical reactions rearrange electrons, which are held by forces worth a few electron volts per bond, while fission rearranges the nucleus. A single fission releases about 180 million electron volts.

Only a few kinds of nuclei fission readily enough to build a bomb: uranium-235 and plutonium-239. When one of these nuclei absorbs a neutron, it splits into two lighter fragments and releases two or three more neutrons. If those neutrons strike other nuclei, the process repeats. The physicists who worked this out in the late 1930s, among them Lise Meitner, had identified the first necessary ingredient of a bomb, and the Chicago Pile-1 experiment in 1942 showed that a sustained, controlled chain reaction was possible.

Picture a room whose floor is covered with thousands of mousetraps, each set with two ping-pong balls resting on the trigger. Toss in one ball and it springs a trap, which launches two balls, which spring two more. Within a few generations the whole room is a blur of airborne balls. A chain reaction is the same arithmetic. If each fission triggers, on average, more than one further fission, the number of fissions per generation grows exponentially, and a generation in a bomb lasts on the order of billionths of a second.

Why size, shape, and purity matter

Whether the arithmetic works depends on how many neutrons find another nucleus before they escape through the surface of the lump. Think of a campfire on a windy beach. A small pile loses so much heat from its surface that it never catches. A big pile keeps most of its heat inside. For the same reason a bigger sphere of fissile metal is more likely to catch its own neutrons, because its volume (where neutrons are made) grows faster than its surface (where they leak out).

The smallest amount of material that sustains a chain reaction is the critical mass, and it is not a single number. The Nuclear Weapon Archive lists about 10.2 kilograms for a bare sphere of pure plutonium-239, and 52.25 kilograms for Godiva, a bare sphere of uranium enriched to 93.71 percent uranium-235. A reflector that bounces escaping neutrons back in lowers those figures substantially. In one cited example a tungsten-carbide reflector cut the critical mass of a uranium-235 assembly to 16.5 kilograms. Density matters as well, which is the whole idea behind the second bomb design described below.

The engineering problem is therefore not just getting enough material. A bomb must bring the material together from a safe, subcritical state into a supercritical one faster than the first fissions can blow it apart. Once the chain reaction starts, the heat is enormous, and the metal begins to expand within microseconds. If the assembly is too slow, the weapon "fizzles": it releases a small fraction of its potential yield and scatters the rest.

Gun-type versus implosion: two answers to the assembly problem

The first bombs solved this problem in two different ways, and the difference explains a great deal about nuclear proliferation.

Feature	Little Boy (Hiroshima)	Fat Man (Nagasaki)
Date	August 6, 1945	August 9, 1945
Fuel	About 64 kg of enriched uranium (average about 80% uranium-235)	About 6.2 kg of plutonium-239
Assembly	Gun-type: one uranium piece fired into another with cordite	Implosion: 32 shaped explosive charges compress the core
Material that fissioned	Less than 1 kg	About 1 kg (roughly 16% of the plutonium)
Yield	About 15 kilotons (13 to 16 in published estimates)	About 21 kilotons
Tested beforehand	No	Yes, in the Trinity test on July 16, 1945

Sources: Atomic Heritage Foundation for dates and Nagasaki yield; Los Alamos estimates as summarized on the Wikipedia pages for Little Boy and Fat Man.

The gun design is crude but reliable. Because uranium-235 is scarce (the K-25 enrichment plant produced about 30 kilograms of enriched uranium a month, and one bomb needed more than 60), the Manhattan Project did not spend any of it on a test. Designers trusted the mechanism enough to use the weapon untested. The price of that simplicity was efficiency: less than a kilogram of the uranium fissioned. A quick check shows how little. Seven-tenths of a gram of mass converted to energy, multiplied by the speed of light squared, gives about 63 terajoules, which is 15 kilotons.

Plutonium made in a reactor is a different problem. It contains some plutonium-240, which fissions spontaneously often enough that a slow gun assembly would be likely to start the chain reaction too early. Implosion solves this by squeezing a subcritical sphere with a symmetrical shock wave from explosive lenses, compressing it to a much higher density within a tiny fraction of a second. The compressed core becomes supercritical and fires much more efficiently, about 16 percent against under two percent for the uranium gun. The cost was complexity. Perfect symmetry is hard, which is why the Trinity test preceded Nagasaki, and why a country's first plutonium weapon is usually a copy of a proven design. The Soviet Union's first test, on August 29, 1949, produced about 22 kilotons, and the RDS-1 device followed the Fat Man approach after Soviet intelligence had gathered extensive information about it.

How a hydrogen bomb gets its extra power

Fission has a ceiling. The more material you pile into a fission weapon, the harder it is to keep it from blowing itself apart before it burns, and the more dangerous the stockpiled core becomes. Fusion has no such ceiling. In fusion, light nuclei such as deuterium and tritium (hydrogen isotopes with one and two extra neutrons) combine into helium and release energy, the same reaction that powers the Sun. The catch is temperature. Positively charged nuclei repel each other, and overcoming that repulsion takes tens of millions of degrees. A fission explosion is the only practical match for that.

The design that does the job is called the Teller-Ulam configuration. A thermonuclear weapon has two stages inside a shared casing. The first stage, the primary, is a boosted fission bomb. Its output is mostly X-rays, and X-rays travel at the speed of light, faster than the debris and shock wave of the blast. The casing channels that radiation onto the second stage, the secondary, a column of fusion fuel (usually lithium-6 deuteride) wrapped in a heavy tamper. The radiation heats the outer layer of the secondary so violently that it blows off, and the recoil crushes the fuel inward. Neutrons turn lithium-6 into tritium, which fuses with deuterium. A hollow plutonium "spark plug" at the center, compressed until it goes critical, adds heat. A uranium-238 tamper often fissions when struck by fusion neutrons, and in typical designs fission still supplies about half of the total energy, which is why big hydrogen bombs make so much fallout.

A thermonuclear weapon works less like a bigger bomb and more like a diesel engine. A diesel has no spark plug; it squeezes the fuel until compression alone makes it burn. The primary is the piston, delivering a very short, very hard squeeze, and the secondary is the fuel that ignites under it.

The yield ladder

The first full-scale test of the idea, Ivy Mike on November 1, 1952, produced 10.4 megatons at Enewetak Atoll and erased the island of Elugelab. It used cryogenic liquid deuterium and filled a building, so it was a proof of principle, not a weapon. Later designs used solid lithium deuteride and fit on missiles. The record explosion came on October 30, 1961, when the Soviet Union detonated Tsar Bomba, about 50 megatons, over Novaya Zemlya. The device was designed for roughly 100 megatons, and the yield was halved by replacing a uranium-238 tamper with lead to limit fallout, a change credited to Andrei Sakharov. For more records of this kind, see the biggest explosions in history.

Explosion or warhead	Yield	Multiple of Hiroshima (15 kt)
Hiroshima, Little Boy (1945)	about 15 kilotons	1
Nagasaki, Fat Man (1945)	about 21 kilotons	about 1.4
W76 warhead (Trident II)	100 kilotons	about 7
W88 warhead (Trident II)	475 kilotons	about 32
Ivy Mike (1952)	10.4 megatons	about 690
Tsar Bomba (1961)	50 megatons	about 3,300

Sources: yields from the pages cited above, plus Wikipedia's W88 entry, the Ohio-class submarine entry, Ivy Mike, and Tsar Bomba. Multiples are rounded.

The table corrects a common claim. Today's deployed strategic warheads are not hundreds or thousands of times more powerful than the Hiroshima bomb. A Trident II D5 missile can carry up to eight W88 warheads (475 kilotons each) or up to 12 W76 warheads (100 kilotons each). The Navy's 14 Ohio-class ballistic missile submarines were built with 24 missile tubes, and four tubes on each boat were deactivated in 2017 under New START, leaving 20. What changed since the 1950s is not the largest yield but the delivery: many warheads per missile, each aimed at a different target, with accuracy good enough that a smaller warhead can destroy a hardened target.

There is a physical reason designers stopped pursuing giant yields. The radius of blast damage grows roughly with the cube root of yield. A bomb 3,333 times larger than the Hiroshima weapon, which is the Tsar Bomba ratio, damages a radius only about 15 times larger. Beyond a certain size, a bigger bomb is a poor use of material, and several smaller bombs cover more ground.

What a detonation does, layer by layer

A nuclear explosion is not one effect but several, arriving in sequence. For an air burst near sea level, roughly half the energy goes into the blast and shock wave, about 35 percent into thermal radiation (light and heat), about 5 percent into prompt radiation (gamma rays and neutrons), and 5 to 10 percent into residual radiation from fallout, according to the summary on Wikipedia's page on the effects of nuclear explosions. The split varies with design and burst height, and a "neutron bomb" shifts it toward prompt radiation.

Effect	What it does	When it arrives
Thermal pulse	Burns exposed skin, ignites flammable materials, and can start mass fires	Within seconds, before the shock wave
Blast wave	Crushes structures, throws debris and people	Seconds after, depending on distance
Prompt radiation	Gamma rays and neutrons damage cells; shielded by distance and mass	Within the first minute
Electromagnetic pulse	A high-altitude burst produces a pulse lasting about a millisecond that can induce damaging voltages in long conductors and unshielded electronics	Instant
Fallout	Radioactive dust, mostly from soil or debris drawn into the fireball, falls downwind	Hours to weeks

The human toll in 1945 shows the layers at work. Estimates of Hiroshima's death toll run from 90,000 to 166,000 over the four months after the bombing, according to the Atomic Heritage Foundation, and roughly half the deaths came on the first day. Nagasaki's toll was between 40,000 and 75,000 immediately, and may have reached 80,000 by the end of 1945. Ranges this wide reflect real uncertainty about records, not carelessness, and they are one reason serious sources give ranges instead of single numbers.

Fallout depends on how the weapon is used. An air burst high enough that the fireball does not touch the ground makes little local fallout, because the radioactive material stays aloft and disperses. A surface burst draws up soil that carries radioactive material back down, which is why Wikipedia's summary describes fallout as the primary source of ionizing radiation exposure from a large weapon. The Hiroshima and Nagasaki bombs were air bursts. For a scenario-level view of a large exchange, see our explainer on what would actually happen to Earth in a nuclear war. For how fission is used peacefully, and how enrichment overlaps with weapons work, see Is nuclear energy safe?

From monopoly to standoff: how the arms race unfolded

The history matters because it explains why the world's two largest arsenals ended up shaped the way they did.

Date	Event	Why it matters
July 16, 1945	Trinity test, New Mexico	First detonation of an implosion device
August 6 and 9, 1945	Hiroshima and Nagasaki	The only combat use so far
August 29, 1949	First Soviet test (RDS-1, about 22 kt)	End of the US monopoly after four years
November 1, 1952	Ivy Mike	First multi-stage thermonuclear test
October 30, 1961	Tsar Bomba	Largest explosion ever detonated
October 14 to 28, 1962	Cuban Missile Crisis	Closest direct superpower confrontation
1968 to 1970	Non-Proliferation Treaty opened and entered into force	Basis of the nonproliferation regime
1986	Global stockpile peaks at about 70,300 warheads	Estimate from the Federation of American Scientists
December 8, 1987	INF Treaty signed	First treaty to eliminate a class of nuclear missiles

The Cuban Missile Crisis is the best-documented example of how close things came. According to the State Department's Office of the Historian, a U-2 flight photographed missile sites under construction in Cuba on October 14, 1962, and President Kennedy announced a naval quarantine on October 22. On October 27, a U-2 was shot down over Cuba and Khrushchev's second message tied any deal to the removal of US Jupiter missiles from Turkey. Kennedy answered the first, more conciliatory message instead, and Attorney General Robert Kennedy told the Soviet ambassador that the Jupiter missiles would be removed, though not as part of a public deal. Khrushchev announced the dismantling on October 28. The Jupiters left Turkey in April 1963.

The long arc after Cuba is a story of managed competition. The Non-Proliferation Treaty, which opened for signature in 1968 and entered into force in 1970, now has 191 states parties, according to the UN Office for Disarmament Affairs. Arms control then pared back the numbers. The Intermediate-Range Nuclear Forces (INF) Treaty, signed December 8, 1987, banned ground-launched ballistic and cruise missiles with ranges of 500 to 5,500 kilometers, and the two sides destroyed 2,692 missiles by June 1, 1991. The estimated global stockpile fell from about 70,300 in 1986 to about 12,187 in early 2026, a reduction of roughly 83 percent, even though the remaining weapons are more accurate and more reliable than those they replaced.

The collapse of the Soviet Union created a different problem: nuclear weapons stationed in states that were suddenly independent. In the Budapest Memorandum, signed on December 5, 1994, Belarus, Kazakhstan, and Ukraine agreed to join the NPT as non-nuclear states, and Russia, the United States, and the United Kingdom pledged to respect their borders and refrain from threatening or using force against them. Whether those pledges amounted to "guarantees" is contested. US officials describe the memorandum as a political commitment that is not legally binding, while the Ukrainian and Russian texts used the word for guarantees. Ukraine and many other governments argue that Russia breached the memorandum in 2014 and again in 2022.

Why deterrence works, and where it strains

Deterrence is the idea that you can prevent an attack by promising a retaliation so costly that no gain could justify it. Nuclear deterrence adds a stark feature. Because a single nuclear weapon can destroy a city, the threat does not depend on winning a war. It depends on being able to do unacceptable damage afterward.

That is why second-strike capability, the ability to absorb a nuclear attack and still retaliate, carries so much of the weight. If one side could destroy the other's weapons in a surprise first strike, it would have an incentive to strike first, and the other side, knowing this, would have an incentive to launch first in a crisis. Both sides therefore spread their forces across land-based missiles in hardened silos, bombers that can be launched on warning, and submarines that hide in the ocean. A submarine like those of the Ohio class is a deterrent because nobody knows where it is.

Imagine two people standing in a room with gasoline up to their ankles, each holding a lit match. Each knows that dropping the match would burn both of them. Neither is safe unless the other believes the threat, and neither can be sure the other will not slip. The standoff is stable only as long as both sides fear the fire more than they want the room, and it demands constant attention. The logic is closely related to the formal models in game theory, where the outcome depends on what each player expects the other to do.

Deterrence has well-known weak points, and each has a real-world analogue.

• Asymmetric opponents. A nuclear state facing a non-nuclear state cannot be deterred by a retaliation the other side cannot carry out, and non-state groups have no cities to threaten.
• Tactical weapons. Lower-yield weapons designed for battlefield use are seen by some analysts as more usable, and therefore as lowering the threshold for first use.
• Conventional-nuclear entanglement. Early-warning radars and command systems serve both conventional and nuclear forces, so an attack on one could be read as an attack on the other.
• Accident and misperception. A system built on instant response leaves little time to check a warning.

Does the nuclear taboo explain the gap?

A second explanation is cultural. Political scientist Nina Tannenwald popularized the idea of a "nuclear taboo" in a 1999 article and developed it in her 2007 book, The Nuclear Taboo: The United States and the Non-Use of Nuclear Weapons Since 1945. She defined it as a de facto prohibition against the first use of nuclear weapons, a shared understanding that using them is illegitimate and immoral.

The argument has critics. The authors of one strategy textbook point out that there are no truly hard cases in which a nuclear state chose not to use its arsenal when it had a lot to gain. A 2024 paper by Joshua Schwartz argues that the taboo does not exist, pointing to cases where policymakers seriously considered nuclear use and to polls showing majority or near-majority support for nuclear use in some circumstances. Work by Dill, Sagan, and Valentino found that public support for nuclear use in the United States, the United Kingdom, France, and Israel rises when the weapon is most effective and falls when civilian casualties are large, which is not the pattern a pure taboo would predict. The fair reading is that the norm is real as a political constraint on leaders, but it may be weaker and more conditional than its name suggests.

How much was luck?

The third explanation is the least comfortable. A record of near-misses shows that the system sometimes worked because of an individual's judgment, or because a crude safeguard held.

Incident	What happened	What it shows
Goldsboro, North Carolina (January 24, 1961)	A B-52 broke up in flight and dropped two hydrogen bombs, each with a predicted maximum yield of 3.8 megatons. One bomb's parachute opened and it landed nearly intact. The other broke apart in mud and its secondary was never recovered	In one bomb, a single low-voltage arm/safe switch is described by a Sandia analysis and an engineer as the main barrier against detonation
Soviet submarine B-59 (October 27, 1962)	During the Cuban crisis, the submarine's officers debated firing a nuclear-armed torpedo. Accounts differ on the details, but Vasili Arkhipov is credited with arguing against it	Command-and-control rules for nuclear weapons at sea were looser than many assumed. Historians dispute the exact vote
Soviet early-warning false alarm (September 26, 1983)	The Oko satellite system reported one US missile launch, then four more. Lieutenant Colonel Stanislav Petrov judged it a computer error and chose to wait for corroboration. Investigators later traced the alarms to a rare alignment of sunlight on high-altitude clouds and the satellites' orbits	A single officer's skepticism stood between a sensor fault and an escalating decision
Able Archer 83 (November 1983)	A NATO command exercise coincided with intense Soviet suspicion that the West might be preparing a first strike	Exercises and real alerts can be hard to tell apart under stress

Sources: Goldsboro crash and 1983 false alarm incident summaries; B-59 accounts vary by source.

No single event on this list proves that war was near. Several involved safeguards that worked as designed. But the pattern matters: when a sensor failed, a person made the right call; when a bomb fell, a switch held. A stable system does not usually depend on so many individual decisions going the right way. That is why careful analysts say the absence of nuclear use since 1945 is consistent with deterrence, with norms, and with luck, and that no available evidence cleanly separates the three. We cannot rerun history to see what would have happened in a world with weaker deterrence or fewer taboos, which is why the claim "deterrence works" is best read as a well-supported hypothesis and not a proven fact.

The landscape in 2026

As of the start of 2026, the Federation of American Scientists estimates about 12,187 nuclear warheads worldwide, of which about 9,745 are in military stockpiles and the rest are retired weapons awaiting dismantlement. The figures below are estimates, and the FAS page itself notes that several countries' numbers are highly uncertain.

Country	Military stockpile	Total inventory
Russia	4,400	5,420
United States	3,700	5,042
China	620	620
France	290	370
United Kingdom	225	225
India	190	190
Pakistan	170	170
Israel (undeclared)	90	90
North Korea	60	60

Source: Federation of American Scientists, Status of World Nuclear Forces, 2026 estimates. For the ranking context, see countries with the most nuclear weapons.

Three features stand out. First, the United States and Russia together hold about 86 percent of all warheads in the total inventory, about 83 percent of the military stockpiles. Second, China's arsenal is widely described as the fastest-growing: the FAS table above puts it at about 620 warheads in early 2026, and new missile silo fields are under construction. Third, the numbers for North Korea and several others are low-confidence estimates, so the order of countries in the middle of the table is less certain than the order at the top.

Arms control is thinner than it was

The treaty structure that restrained the largest arsenals has been unraveling piece by piece.

Agreement	What it did	Status in 2026
New START (2010)	Limited the US and Russia to 700 deployed launchers, 1,550 deployed warheads on them, and 800 total launchers	Expired February 5, 2026
INF Treaty (1987)	Banned ground-launched missiles of 500 to 5,500 km range	The United States withdrew on August 2, 2019, citing Russian deployment of the 9M729 (SSC-8) cruise missile; Russia denied violating it
Non-Proliferation Treaty (1970)	Bargain: nuclear-weapon states disarm in principle, others forgo weapons	191 states parties
Comprehensive Test-Ban Treaty (1996)	Bans all nuclear explosive tests	Not in force; 188 signatories and 179 ratifications; Russia revoked its ratification in November 2023; the United States and China signed but have not ratified
Treaty on the Prohibition of Nuclear Weapons (2017)	Bans possession and use for its parties	Adopted July 7, 2017 by a vote of 122 to 1 with one abstention; in force since January 22, 2021

Sources: Arms Control Association for New START limits and expiry; UN Office for Disarmament Affairs for the 2017 treaty; the Wikipedia pages on the INF Treaty and the CTBT.

The expiry of New START matters because it ended the last legally binding limits on the two largest arsenals and also the data exchanges and inspections that came with them. Reporting around the expiry described Washington and Moscow as free to choose their next steps, and, according to the Arms Control Association, officials in both capitals had floated staying below the old limits for a time as an unstable and informal arrangement. US officials have said that, absent a change in adversary arsenals, the United States may need to consider increasing deployed numbers. The Arms Control Association's analysis noted that no successor agreement was in place. Modernization programs continue in several countries, including new delivery systems such as hypersonic weapons, and new missile silos in China.

What a large nuclear war would do to food

The 2022 study by Lili Xia, Alan Robock, and colleagues, published in Nature Food, is the most-cited recent attempt to model the global consequences. The researchers used climate, crop, and fishery models to simulate six scenarios of soot injected into the stratosphere by firestorms: five smaller India-Pakistan wars and one large war among the United States, NATO, and Russia. Soot above 5 teragrams (5 million metric tons) led to mass food shortages in almost all countries. Global average caloric production fell about 7 percent within five years in the smallest scenario, and about 90 percent three to four years after the largest. The authors estimated more than 2 billion deaths from an India-Pakistan war, and more than 5 billion from a US-Russia war, mostly from famine, not from the blasts themselves. These are model results with large uncertainties, and they assume a specific number of urban targets burning, but they changed the policy debate by showing that even a regional war could have global effects.

Three claims that sound right and are not

"Modern hydrogen bombs are a thousand times Hiroshima." The largest weapons ever tested were. The weapons on missiles and submarines today are mostly in the range of 100 to 475 kilotons, which is roughly 7 to 32 times the Hiroshima yield. They are more dangerous in aggregate and in delivery, not in single-warhead size.

"The Cold War arsenals are still the baseline." Global inventories are down about 83 percent from the 1986 peak. The danger today comes less from the raw count than from the loss of verification, the arrival of new delivery systems, and the larger number of nuclear-armed states.

"Deterrence has been proven." What has been observed is the absence of use. Different theories predict that outcome, and the near-miss record means that at least some of that absence depended on individual judgment and a few simple safeguards.

Frequently asked questions

Q: How much material does it take to make a nuclear weapon? A: The critical mass of a bare sphere is about 10.2 kilograms for plutonium-239 and about 52 kilograms for uranium enriched to roughly 94 percent uranium-235, and reflectors and compression reduce the amount needed. The real barrier is not the quantity but producing the material, which requires enrichment plants or reactors and reprocessing facilities, and that infrastructure is detectable.

Q: What is the difference between an atomic bomb and a hydrogen bomb? A: An atomic bomb, such as those used in 1945, gets all its energy from fission. A hydrogen bomb uses a fission primary to compress and ignite a fusion secondary, and fission still supplies about half the energy in typical designs. The two-stage design has no practical upper limit on yield, which is why the Soviet Union could test a 50-megaton device in 1961.

Q: Could terrorists get a nuclear weapon? A: The main barrier is fissile material, which is tightly guarded and hard to produce. A stolen state weapon would be protected by physical security and, in many designs, by coded locks. A radiological "dirty bomb," which spreads radioactive material with conventional explosive, is a much lower technical hurdle but does not produce a nuclear explosion; it would cause contamination and panic, not mass blast casualties.

Q: What is the difference between tactical and strategic nuclear weapons? A: Strategic weapons are built to hit an adversary's homeland (cities, industry, long-range forces) and are usually delivered by intercontinental missiles, submarines, or heavy bombers. Tactical or non-strategic weapons are lower-yield weapons meant for use against military forces in a theater. The line is blurry, and some critics argue that tactical weapons make nuclear use seem more thinkable.

Q: How many countries have nuclear weapons? A: Nine: the United States, Russia, China, France, the United Kingdom, India, Pakistan, North Korea, and Israel, which does not publicly acknowledge its arsenal. The NPT recognizes five as nuclear-weapon states. India, Pakistan, and Israel never joined the treaty, and North Korea has withdrawn.

Q: What happened to Soviet nuclear weapons after 1991? A: Nuclear weapons were stationed in Russia, Ukraine, Belarus, and Kazakhstan. Between 1993 and 1996, the three non-Russian states gave up the weapons on their soil, and the 1994 Budapest Memorandum offered them security assurances in return. The United States helped fund dismantlement and security through the Nunn-Lugar Cooperative Threat Reduction program.

Sources

• Federation of American Scientists, Status of World Nuclear Forces (2026 estimates and the 1986 peak)
• Nuclear Weapon Archive, critical masses and fission energy
• Atomic Heritage Foundation, Hiroshima and Nagasaki
• Office of the Historian, US Department of State, the Cuban Missile Crisis
• Arms Control Association, Life After New START
• UN Office for Disarmament Affairs, Non-Proliferation Treaty and Treaty on the Prohibition of Nuclear Weapons
• Xia, L., Robock, A., et al. (2022). Global food insecurity and famine from reduced crop, marine fishery and livestock production due to climate disruption from nuclear war soot injection. Nature Food, 3(8), 586-596. doi:10.1038/s43016-022-00573-0

What this means

Nuclear weapons are physically simple to describe and extremely hard to build well: a lump of metal that sustains a chain reaction for a few millionths of a second, then a second stage that borrows the first stage's X-rays to ignite fusion. The 81 years without use are real, but the evidence does not tell us which of deterrence, taboo, or luck deserves the credit, and that uncertainty is itself the strongest argument for treating the surrounding machinery of treaties, verification, and crisis communication as load-bearing. When that machinery weakens, as it did when New START lapsed, the explanations that depend on human judgment under stress matter more, not less.

Related Reading

• How democracy works and why it is under threat
• How vaccines work and why some people refuse them
• How pandemics spread and what stops them
• How space exploration works and what is next
• What is the stock market and how does it actually work?