What Actually Happens Inside a Nuclear Reactor? A 5-Minute Explanation
The physics is stranger and more elegant than you probably think.
Most people have a rough mental image of a nuclear reactor: something enormous, vaguely ominous, surrounded by cooling towers breathing white steam into the sky. But what’s actually happening inside? The honest answer is: a chain of events so small it’s invisible, producing heat so intense it could power a city, controlled by a system so precise it has to work perfectly every second of every day. That’s the deal. And once you understand it, you can’t look at a uranium pellet the same way.
The core idea is not complicated. Split an atom, capture the heat, boil water, spin a turbine, make electricity. That’s it. A nuclear power plant is, at its simplest, a very exotic way to boil water. The complexity is in the “split an atom” part, and that’s where things get interesting.
The atom that started it all
The star of the show is uranium-235, a particular variety of uranium with 92 protons and 143 neutrons packed into its nucleus ☢️. It’s not very common. As the U.S. Nuclear Regulatory Commission explains, only about 0.7% of naturally occurring uranium is U-235 — the rest is mostly the heavier and far more stable uranium-238, which simply refuses to cooperate with the process we need. So before uranium can fuel a reactor, it gets enriched — that is, the proportion of U-235 is artificially boosted to around 5%. Not weapons-grade by a long shot, but enough to sustain a reaction.
The enriched uranium is then pressed into small ceramic pellets, roughly the size of a sugar cube 🍬. These pellets get stacked inside sealed metal tubes called fuel rods, and more than 200 of those rods get bundled together into a fuel assembly. A full reactor core holds hundreds of these assemblies. What you’re looking at is thousands upon thousands of tiny ceramic cylinders, each one carrying an almost absurd amount of potential energy:
A single uranium pellet contains roughly the same energy as one tonne of coal
A typical reactor needs about 27 tonnes of fresh uranium fuel per year
A coal plant of comparable size would burn through more than 2.5 million tonnes of coal for the same output
The fuel pellets themselves are not radioactive in the dangerous, touch-it-and-die sense — they become hazardous only once the fission process begins
Think about that energy density for a second. It’s almost offensive how much power is crammed into something you could hold in your palm. That’s the fundamental promise of nuclear energy.
The chain reaction: a controlled domino effect
Here’s where the physics gets genuinely beautiful ⚡. When a slow-moving neutron strikes a U-235 nucleus, the nucleus absorbs it and becomes briefly, violently unstable. It splits — typically into two smaller atoms like barium and krypton — and in doing so, it releases a burst of heat and two or three additional free neutrons. Those neutrons go on to hit other U-235 atoms, which also split and release more neutrons, which hit more atoms, and so on. This is the nuclear chain reaction.
The U.S. Energy Information Administration puts it clearly: the chain reaction is self-sustaining as long as enough fissile material is present. Left completely uncontrolled, it would be catastrophic. Inside a reactor, it’s kept in a careful, deliberate balance where exactly one neutron from each fission event triggers one more fission event. Engineers call this the critical state, and keeping a reactor in it is the entire point.
The math of criticality:
If fewer than one neutron per fission causes another fission, the reaction fizzles out (subcritical)
If exactly one neutron sustains the reaction, you have stable, controlled power generation (critical)
If more than one neutron triggers further fissions, the reaction accelerates (supercritical, which is what you absolutely don’t want)
This is worth pausing on if you’ve ever worried about a reactor “going off like a bomb.” It can’t. The geometry of the fuel, the concentration of U-235, and the moderator all conspire to make an uncontrolled weapons-grade explosion physically impossible in a commercial reactor. The scenarios that have caused real harm — Three Mile Island, Chernobyl, Fukushima — were runaway heat events, not nuclear explosions. Different problem. Still serious, but categorically different.
The unsung hero: the moderator
Here’s a detail most people never hear about 🔬. The neutrons that fly out of a fission event are fast. Too fast, actually. U-235 atoms don’t readily absorb high-speed neutrons — they prefer the slow, lazy kind. So the reactor includes a moderator, a material whose job is to slow the neutrons down without absorbing them.
The most common moderator is plain water. As the MIT Nuclear Reactor Laboratory explains, when a fast neutron bounces around inside water, it gradually loses energy in collisions with hydrogen atoms until it slows to what engineers call a thermal neutron speed. At that speed, it becomes far more likely to cause fission when it encounters a U-235 nucleus. This is why the reactor core is submerged in water — it’s not just for cooling. The water is doing two separate jobs at once.
The main types of moderators used in reactors worldwide:
Light water (ordinary H₂O): the most common, used in about 90% of reactors globally
Heavy water (deuterium oxide, or D₂O): absorbs fewer neutrons, allows natural uranium to be used as fuel without enrichment
Graphite: used historically in some reactor designs, including the now-infamous RBMK at Chernobyl
The choice of moderator shapes almost everything else about a reactor’s design: fuel type, coolant, operating pressure, and safety characteristics. It’s one of the first design decisions a reactor engineer makes.
Control rods: the brakes on the whole system
So you’ve got neutrons flying around, splitting atoms, releasing heat 🌡️. How do you turn it down? Or off? The answer is control rods, usually made from materials like boron or silver, which are excellent at absorbing neutrons. Insert the control rods deeper into the core, and you’re soaking up neutrons before they can cause further fissions — the reaction slows. Pull them out, and the reaction speeds up. All the way in and the chain reaction stops entirely.
This is, in essence, how an operator controls the output of a nuclear reactor. It sounds almost too simple, and in a way, it is:
Control rods inserted fully → reactor shuts down (called SCRAM, the term for an emergency shutdown)
Control rods partially inserted → steady-state power output
Control rods partially withdrawn → power level increases
Control rods combined with moderator adjustments → fine-grained power control
What you might find interesting is that modern reactor designs increasingly lean on passive safety systems — mechanisms that work automatically using gravity and natural convection, no operator action required. Small modular reactors, the technology this publication covers closely, are built around this principle. The physics itself provides the safety net, not just the engineering. If you’re curious about how SMR designers are baking passive safety into their reactor concepts, that’s a topic worth exploring in depth.
From heat to electricity: the boring part that matters enormously
Once the reactor core generates heat, the rest of the process is basically a very sophisticated steam engine ♻️. The details differ by reactor type, but the principle is identical across the board.
The IAEA breaks this down cleanly: the heat warms a coolant, typically water, which either directly produces steam or transfers its heat through a steam generator to a separate, non-radioactive water loop that produces steam. That steam drives a turbine, which spins a generator, which produces electricity. The steam then cools back into water and the cycle repeats.
The two dominant reactor types handle this slightly differently:
Pressurized Water Reactors (PWRs): the water in the reactor stays liquid under high pressure (it never boils), then heats a separate water loop that does boil into steam. As of 2024, PWRs account for 74% of all operating reactors worldwide — around 308 units globally
Boiling Water Reactors (BWRs): the water boils directly inside the reactor vessel and feeds steam straight to the turbine, cutting out the middleman
Both types sit within a pressure vessel — a thick steel container that houses the fuel assemblies, moderator, and coolant. Surrounding that is the containment structure, typically a reinforced concrete shell designed to contain radioactive material if something goes wrong. And surrounding that is the reactor building. The engineering is layered, deliberately and thoughtfully. Multiple barriers, not just one.
Have you ever wondered why nuclear plants produce so little waste relative to fossil fuels? A reactor generates no carbon dioxide during operation. And while spent nuclear fuel is genuinely challenging to manage, according to the World Nuclear Association the entire global nuclear industry produces a comparatively tiny volume of high-level waste relative to coal, oil, and gas, which dump their waste directly into the atmosphere.
Why any of this matters for SMRs
As of 2025, there are 417 commercial nuclear reactors operating globally, and the physics inside every single one of them is identical to what’s described above 🌍. Uranium-235 splits. Neutrons slow down. Heat transfers. Steam spins turbines. Electricity flows.
What small modular reactors are trying to do is take that same physics and apply it in a more compact, factory-built package — one that can be deployed faster, scaled incrementally, and placed in locations where a conventional gigawatt-scale plant would never fit. The fission process doesn’t care about the size of the container. U-235 splits whether the reactor weighs 10,000 tonnes or 500. That’s the elegance of the approach.
Here’s what distinguishes SMR physics from conventional reactor physics:
Smaller core volume means inherently lower power output but also simpler thermal management
Passive cooling is more practical at small scale, since natural convection can do what pumps do in large reactors
Higher power density in some advanced designs, using fuels like HALEU (high-assay low-enriched uranium) at enrichment levels above the standard 5%
Alternative coolants — liquid sodium, molten salt, helium gas — are more viable in SMR designs because the engineering challenge scales with reactor size
The next time you read about a new SMR startup promising a reactor that “runs for years without refueling” or “can’t melt down,” the physics above is the underlying story. Sometimes those claims hold up. Sometimes they need scrutiny. Knowing how a reactor actually works is the best tool you have for telling the difference.
So here’s the question worth sitting with: given that the core physics of nuclear power has been understood since the 1940s, why does it still feel so mysterious to most people — and what would it take for public understanding to finally catch up with the technology?