Flip a light switch in France and there is a 70% chance the electricity flowing through the wire was produced by splitting atoms. Not burning anything. Not spinning a windmill. Splitting uranium atoms so precisely that the entire output of a billion-watt power plant hangs on the position of a few bundles of metal rods, adjustable by inches.
A nuclear reactor cannot explode like a bomb. The fuel is enriched to only 3-5% uranium-235. A weapon requires 90% or higher. The physics of low enrichment make a nuclear detonation physically impossible, regardless of operator error.
Most people picture a nuclear reactor as a contained catastrophe: a glowing, unstable device always on the verge of a mushroom cloud. Hollywood reinforced this with decades of meltdown thrillers. In reality, a nuclear reactor is a very precise heat source. It does exactly what a coal furnace does (boil water to spin a turbine) but with a fuel so energy-dense that a single ceramic pellet the size of your fingertip replaces an entire ton of coal. The challenge was never making atoms split. Uranium does that on its own. The challenge was making them split at exactly the right rate, no faster, no slower, for decades at a time.
The process starts with uranium-235, a naturally occurring isotope that makes up only 0.7% of mined uranium. Reactor fuel is enriched to 3-5% U-235, then formed into small ceramic pellets of uranium dioxide (UO2) and stacked inside tubes of zirconium alloy called fuel rods. A typical pressurized water reactor (PWR) core contains about 157 fuel assemblies, each holding 264 fuel rods, for a total of over 40,000 rods standing about 4 meters tall.
When a slow-moving neutron strikes a U-235 atom, the atom absorbs it, becomes momentarily unstable as U-236, and then fissions: it splits into two lighter atoms (fission fragments like barium-141 and krypton-92), releases about 200 million electron-volts of energy as heat, and ejects 2 to 3 fast neutrons. Those new neutrons are moving too fast to efficiently split more U-235. They must first be slowed down, or moderated, by colliding with the hydrogen atoms in the surrounding water. Once slowed to thermal speed (about 2 km/s instead of 20,000 km/s), each neutron can strike another U-235 atom and repeat the process. This is the chain reaction.
The key number is k, the neutron multiplication factor. When exactly one neutron from each fission event causes one more fission (k = 1.0), the reactor is critical: the chain reaction sustains itself at constant power. When k drops below 1.0, the reaction fades. When k rises above 1.0, power increases. Operators control k by inserting or withdrawing control rods made of neutron-absorbing materials like boron carbide or a silver-indium-cadmium alloy. Push the rods in and they absorb neutrons, dropping k below 1.0 and reducing power. Pull them out and more neutrons reach fuel, raising k and increasing power. The entire output of a 1,000-megawatt reactor is governed by how many inches of control rod sit inside the core.
How operators actually control a nuclear reactor
In practice, operators rarely move control rods by large amounts during normal operation. A reactor running at full power has its rods withdrawn to a carefully calculated position where k equals exactly 1.0. To increase power, they withdraw rods by a few centimeters, letting k rise to perhaps 1.0005. The neutron population grows, fission increases, water temperature rises, and steam output climbs. Once the desired power level is reached, operators reinsert the rods just enough to bring k back to exactly 1.0 at the new, higher power level.
There is a built-in safety feature that makes this easier than it sounds. PWRs have a negative temperature coefficient: as the water heats up, it expands and becomes a less effective neutron moderator. Fewer neutrons get slowed down, fewer fissions occur, and the reaction naturally dampens itself. If something goes wrong and the core overheats, physics itself applies the brakes before any human needs to act. This is not an engineered safety system; it is a consequence of how water and neutrons interact. For the ultimate emergency, a SCRAM drops all control rods into the core by gravity in about 2 seconds, shutting the chain reaction down entirely.
But heat does not stop immediately. Even after shutdown, radioactive fission products in the fuel continue to decay, generating about 7% of full power as decay heat. For a 1,000 MW reactor, that is 70 MW of heat with no chain reaction running. This is why reactors need cooling even after they are shut down, and why the loss of cooling water (not a runaway chain reaction) is the actual danger scenario in reactor accidents.
The cost of splitting atoms
Nuclear power produces zero carbon during operation, but it creates radioactive waste that remains dangerous for thousands of years. Every kilowatt of nuclear electricity comes with a waste management obligation that outlasts every human institution that has ever existed.
The energy density that makes nuclear remarkable is also what makes its waste problem unique. A 1,000 MW reactor produces about 20 tonnes of spent fuel per year. That is remarkably little compared to the 3 million tonnes of coal a similarly sized fossil plant burns. But spent nuclear fuel contains fission products that emit intense radiation and heat for years, plus transuranic elements (plutonium, americium) that remain radioactive for tens of thousands of years. Spent fuel first sits in cooling pools at the reactor site for 5 to 10 years, then gets sealed in dry cask storage: steel cylinders inside concrete shells. All the spent fuel from 60+ years of American commercial nuclear power would fit on a single football field, stacked less than 10 meters high. The volume is small. The timescale is the problem.
There is also the economics. Nuclear plants cost $6 to 12 billion and take 7 to 12 years to build in Western countries. Once running, they produce electricity for 2 to 3 cents per kilowatt-hour (fuel is cheap when a pellet replaces a ton of coal), and they run 93% of the time for 40 to 80 years. The upfront cost is staggering; the lifetime cost per kilowatt-hour is competitive with anything. Whether that tradeoff makes sense depends on how a society values carbon-free reliability versus construction risk and waste duration.
Every nuclear reactor on Earth is doing the same thing: splitting uranium atoms one neutron at a time, using water to slow those neutrons down and carry the heat away, and relying on a few hundred pounds of neutron-absorbing metal to keep a billion-watt chain reaction from growing by even a fraction of a percent. The engineering is complex. The principle is not. A neutron hits an atom. The atom splits. Heat comes out. Control how many neutrons reach how many atoms, and you control enough energy to light a city from a building the size of a warehouse, using fuel you could carry in the back of a pickup truck.