How Nuclear Reactors Work

A single kilogram of uranium fuel contains roughly as much energy as 3,000 tonnes of coal. That’s not a typo. The physics that makes nuclear energy possible, the binding energy inside atomic nuclei, operates at a completely different scale than burning anything. The reason nuclear power remains controversial isn’t really about the energy density. It’s about what happens when something goes wrong, and a handful of very public disasters that have shaped perception ever since.

The short answer

A nuclear reactor is basically a very expensive kettle. It uses nuclear fission, not fusion, to generate heat. That heat turns water into steam, which spins a turbine, which generates electricity. The difference from a fossil fuel plant is the source of the heat: instead of burning coal or gas, you split uranium atoms apart.

The process is clean in terms of emissions, produces a tiny amount of waste by volume, and has prevented countless deaths from air pollution. But it carries a reputation shaped by two disasters that have almost nothing in common with modern reactor designs.

The full picture

Fission, not fusion

When people hear “nuclear,” they often think of hydrogen bombs. Those use fusion, where two light atoms slam together to form a heavier one, releasing enormous energy. That is what powers the sun.

Reactors do the opposite. They use fission, where a heavy atom like uranium-235 is hit by a neutron, splits apart, and releases more neutrons along with energy. Those neutrons then hit other uranium atoms, causing more splits. This is a chain reaction.

Here is the part that surprises most people: the fuel in a reactor is not concentrated enough to explode. You need over 90% uranium-235 to make a bomb. Reactor fuel is typically 3-5% enriched, per International Atomic Energy Agency (IAEA) standards. You could pile all the fuel from a reactor together and it would not detonate. More on this later.

The chain reaction under control

Inside a reactor core, uranium pellets are arranged in fuel rods. When a neutron hits a uranium atom and splits it, the reaction releases heat and more neutrons. If those neutrons keep hitting more uranium, you get an accelerating chain reaction. Uncontrolled, that is a bomb. Controlled, that is power.

Two things keep the reaction in check:

Control rods are made of materials that absorb neutrons without splitting, like boron or cadmium. They slide in and out of the reactor core. Push them in, they soak up neutrons, and the reaction slows down or stops. Pull them out, the reaction speeds up. Operators adjust them constantly to maintain a steady power level.

A moderator, usually water or graphite, slows down the neutrons. Fission releases fast neutrons, but uranium-235 splits most efficiently when neutrons are moving at moderate speeds. The moderator acts as a brake, slowing neutrons to the right speed to keep the chain reaction going.

Together, control rods and the moderator give operators precise control. You can throttle a nuclear plant up and down much like a gas plant, though most run at full capacity because they are most efficient there.

Turning heat into electricity

Once the reactor is producing heat, the rest works like any thermal power plant. Water circulates through the core and absorbs the heat. That hot water is kept under high pressure so it does not boil, and it transfers its heat to a separate water loop. That second loop turns to steam, which drives a turbine, which spins a generator, which produces electricity.

It really is a kettle, just enormously complex and expensive. A typical reactor produces about 1,000 to 1,600 megawatts, enough to power a large city. The newest designs, called small modular reactors or SMRs, produce around 50 to 300 megawatts and can be built in factories and shipped where needed.

Keeping it cool

The heat inside a reactor core is intense, and something always needs to carry it away, even when the reactor is shut down. This is where cooling systems come in.

Most reactors use one of two designs. Light water reactors, the most common type, use ordinary water as both the moderator and the coolant. Water flows through the core, absorbs heat, and transfers it to the steam cycle. Pressurized water reactors keep the primary water under such high pressure that it does not boil, even at over 300 degrees Celsius.

After the Fukushima disaster in 2011, the US Nuclear Regulatory Commission (NRC) and other international regulators required all reactors to have passive cooling systems that can operate without electricity. Some use gravity-fed water tanks that automatically flood the core if temperatures rise. Others rely on the natural circulation of air. The goal is simple: even if every power source and pump fails, the reactor must still be able to cool itself.

Why reactors cannot become bombs

This is worth repeating because it is one of the most persistent misconceptions. A nuclear reactor physically cannot detonate like a nuclear weapon.

The key concept is critical mass. You need a certain amount and concentration of fissile material to sustain a chain reaction powerful enough for a bomb. Weapons-grade uranium is over 90% enriched. Reactor fuel is 3-5%. There is simply not enough uranium-235 packed together tightly enough to cause the kind of runaway reaction that creates an explosion.

What can happen, and what did happen at Chernobyl, is a steam explosion or a meltdown where the reactor overheats and the fuel melts through its containment. That is catastrophic, but it is not a nuclear detonation. The physics simply do not allow it.

What went wrong: Chernobyl

In April 1986, Reactor Number Four at the Chernobyl nuclear plant in Ukraine experienced a catastrophic power surge during a safety test. The reactor design, known as an RBMK, had a dangerous flaw: it could become unstable at low power levels, and the control rods were not fully effective at shutting down the reaction.

When operators pulled out too many control rods to run a low-power test, the reactor went haywire. A combination of design flaws, operator error, and a rushed test led to a steam explosion that destroyed the reactor building and ejected radioactive material across Europe. The UN Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) has continued to study the long-term health consequences.

The RBMK design has been phased out everywhere. Modern reactors include negative void coefficient designs, meaning if the coolant boils, the reaction naturally slows down instead of speeding up. Chernobyl could not happen in a Western reactor built to modern standards.

What went wrong: Fukushima

In March 2011, a massive earthquake off the coast of Japan triggered a tsunami that overwhelmed the Fukushima Daiichi nuclear plant. The reactors shut down automatically when the earthquake hit, but the tsunami flooded the backup diesel generators that powered the cooling systems. Without cooling, the reactor cores overheated, the fuel melted, and hydrogen explosions occurred.

The death toll from the disaster was tragic but complex. Approximately 19,000 people died from the tsunami itself, according to Japan’s National Police Agency. According to the World Health Organization and UNSCEAR assessments, zero deaths are directly attributed to radiation exposure from the plant. The long-term health effects, while real, have been far smaller than initial fears predicted.

The lesson was clear: reactors needed to survive beyond-design events. New plants are built on higher ground, with waterproof barriers, and with cooling systems that do not require electricity. The Fukushima-style disaster is extremely unlikely to repeat.

The actual risk: numbers that surprise people

Here is the part that challenges most people’s intuition. Nuclear is statistically one of the safest energy sources on the planet.

According to analysis by Our World in Data (based on data from Oxford University and peer-reviewed mortality research), the death toll from nuclear power, including Chernobyl and Fukushima, is approximately 0.03 deaths per terawatt-hour of electricity generated. For comparison, coal kills roughly 24 deaths per terawatt-hour. Oil kills about 18. Natural gas kills about 3.

That means for every person who dies from nuclear power, roughly 800 people die from coal. Even solar and wind, which are much safer than fossil fuels, still have higher death tolls per unit of energy than nuclear, primarily from falls and accidents during installation and maintenance.

The reason is simple: nuclear accidents are spectacular and memorable. A mine collapse or a smog episode kills people quietly and constantly, so they do not make the same emotional impact. But the numbers do not lie.

What about the waste?

The waste question is legitimate, but it is also misunderstood. Nuclear waste is highly radioactive, and it needs to be stored safely. However, the volume is surprisingly small.

According to the Nuclear Energy Institute (NEI), all the spent fuel from decades of powering the entire United States would fill a standard football field stacked about 10 yards deep, roughly 83,000 metric tonnes total. A typical coal plant produces more radioactive waste in a single year, because coal contains trace amounts of uranium and thorium that get released into the atmosphere when burned.

Spent fuel is initially stored in pools of water, then moved to dry casks made of steel and concrete. Newer designs, like fast reactors and molten salt reactors, can actually recycle used fuel and reduce the waste by over 90%. France’s state-owned utility EDF already recycles its nuclear fuel commercially through the ORANO reprocessing facility at La Hague, extracting energy from material that other countries throw away.

The waste that concerns people most, the long-lived radioactivity, decays over hundreds of years to safe levels. The waste is a problem, but it is a manageable problem with a known solution, and the volume is tiny compared to the waste produced by any fossil fuel industry.

The next generation: why small modular reactors change the math

One of the biggest criticisms of nuclear power is construction cost and time. Large plants like the new Hinkley Point C in the UK or the Vogtle expansion in Georgia have run billions over budget and years behind schedule. This isn’t incompetence: it’s the nature of one-of-a-kind megaprojects built by teams that haven’t done it before, on new sites, subject to changing regulations.

Small modular reactors (SMRs) are an attempt to break this pattern. Instead of building massive bespoke plants, SMRs are factory-manufactured, standardized reactors with electrical output of 50-300 MW (versus 1,000+ MW for conventional plants). The idea is that you build the components in a factory, where quality control is consistent and workers learn by repetition, then ship them to the site for assembly.

Companies like NuScale (US), the only SMR developer to have received NRC design approval as of 2025, Rolls-Royce (UK), and Kairos Power are in various stages of development and regulatory approval. China has already deployed the world’s first commercial SMR at sea, powering an offshore oil platform. The economic theory is that mass production brings down costs the same way it did for aircraft and automobiles.

SMRs also have better passive safety characteristics. Many designs rely on natural convection and gravity rather than active pump systems to cool the reactor in an emergency. If power is lost (the failure mode that led to Fukushima), newer designs simply shut down and cool themselves without human intervention.

The first US SMRs are expected to come online in the early 2030s. If the cost projections prove correct, they could make nuclear economically competitive for smaller grids and industrial applications, not just large national power systems.

Common misconceptions

“A nuclear meltdown would destroy a huge area.” The worst-case exclusion zone from a nuclear accident extends maybe 30 kilometers. The Chernobyl exclusion zone is about 2,600 square kilometers. That sounds large until you compare it to the area degraded by coal mining, oil drilling, or the sheer scale of land lost to fossil fuel extraction and pollution.

“We should just use renewables instead.” Renewables are essential and growing fast, but they have a problem: the sun does not always shine and the wind does not always blow. Nuclear provides what is called baseload power, reliable round-the-clock generation that does not depend on weather. The countries with the lowest carbon emissions typically combine nuclear with renewables, not one or the other.

“Nuclear is too expensive.” The capital cost is high, but the fuel cost is near zero and the operating costs are low. Over the lifetime of a plant, nuclear is competitive with coal and increasingly competitive with natural gas, especially when you factor in the health and climate costs of fossil fuels. New small modular reactors are expected to cut construction costs dramatically.

Why it matters

Nuclear power is not perfect. It is expensive to build, takes years to construct, and carries genuine public fear that cannot be dismissed. But it is the only proven technology that can generate massive amounts of reliable, zero-carbon electricity at scale, right now, with a safety record that rivals or exceeds any other energy source.

The climate crisis is not a future problem. It is here, and it is being driven by fossil fuels that are killing millions of people per year through air pollution alone. Nuclear is not the only answer, but ignoring it because of fear, when the numbers tell a clear story, is a mistake that costs lives.

The next time someone tells you nuclear is dangerous, ask them to compare the death toll to coal. Then ask what they think is actually driving the fear.