Nuclear — much energy from little fuel

Cost, construction, uranium, water, safety and continuous electricity without slogans.

A plant is not just its reactor

Why a heavy nucleus releases so much energy when it splits

Fission harnesses a fact of nuclear physics with no chemical equivalent: when a uranium-235 nucleus absorbs a neutron, it becomes unstable and splits into two lighter nuclei, releasing several new neutrons and an amount of energy enormous compared with any chemical reaction. That energy comes out mostly as kinetic energy of the fragments — that is, heat — exactly the same kind of energy produced by burning coal, except the amount released by each split nucleus is millions of times greater than that released by breaking a single chemical bond of carbon. The neutrons released in each fission can, in turn, trigger new fissions in nearby nuclei: this is the chain reaction that a reactor keeps under controlled, stable conditions, neither growing unchecked nor shutting down, using control rods that absorb the excess neutrons.

From the mine to the fuel, before there is any electricity

Before the reactor produces anything, uranium has to be mined, concentrated, converted and enriched — that whole process is developed in this section's "From uranium to fuel" leaf. Only once the fuel has been manufactured into small ceramic pellets, assembled into rods inside a steel vessel, does the phase most people associate with "nuclear energy" begin: the heat of fission boils water, the steam turns a turbine, and the turbine spins a generator — the same final mechanical principle as a coal or gas plant, only the heat source is completely different.

Why the building weighs as much as the reactor itself

A nuclear plant is not just the reactor core: around it sit large amounts of steel and reinforced concrete forming several successive containment barriers, redundant and diverse cooling systems (so that a single failure cannot disable all of them at once), extremely redundant control instrumentation, and an operation subject to almost permanent independent inspection. That structural robustness is not bureaucratic excess: it is the direct counterpart of handling a material that remains dangerous long after it has generated its last unit of electricity, and it is also the main reason a nuclear plant is expensive and slow to build, even though it is cheap to operate for decades afterward.

Energy density: the trait that explains almost everything else

A single gram of fully fissioned uranium-235 releases roughly as much energy as burning several tonnes of coal. That difference in scale — orders of magnitude, not a simple percentage improvement — is the physical explanation for why a nuclear plant can generate electricity continuously, with a compact land footprint, without needing a constant stream of fuel trucks like a thermal plant. It is also why nuclear plants have, in practice, the highest capacity factors of any generation technology — running close to their maximum almost all year — because they depend neither on weather nor on a continuous flow of fresh fuel: refuelling happens only every one or two years.

Why "how much does nuclear cost?" has no single answer

The cost of a nuclear plant depends brutally on whether it is a design already built several times in the same country (predictable cost) or the first example of a new technology (considerable risk of cost overrun and delay, as shown by the Vogtle case developed in this leaf's data tab). Also weighing in are the cost of financial capital — the interest rates at which a decade-long project is financed —, a country's regulatory stability, and whether a local industrial supply chain already exists or has to be built from scratch. None of those factors is a physical property of the reactor: they are economic and institutional decisions, which is why countries with nearly identical reactors end up with very different final costs.

Inseparable advantages and obligations

In favor: dispatchable production, high density, low life-cycle emissions and very little fuel.

To watch: a low-probability, high-consequence accident, upfront investment, construction time, cooling, proliferation, decommissioning and waste custody.

A fair comparison includes the system it replaces, the country's institutional capacity, and the real financial cost. Defending nuclear without discussing waste is incomplete; rejecting it by comparing only its worst-case scenario with another source's ideal operation is incomplete too.

The numbers, with their source

Comparative table (nine primary sources)

Life-cycle emissions, mortality and land footprint, with the source for each range. These are medians from peer-reviewed studies, not fixed values: the specific technology, country and manufacturing electricity mix all shift each figure.

Sources for this table: IPCC AR6, WG3 and its earlier harmonization by NREL (NREL Life Cycle Assessment Harmonization); Our World in Data — Safest and cleanest sources of energy; Our World in Data — Land use per energy source.

What is specific to fission

Fuel density. One kilogram of enriched reactor uranium releases, over its full cycle, on the order of a million times more energy than one kilogram of burned coal. That density is the reason for the minimal land footprint and the very low volume of fuel that has to be moved, transported and stored compared with any thermal plant of equal power.

Capacity factor. Nuclear plants operate with the highest capacity factors of any technology: sustained above 90% in the United States, compared with typical figures of 40-55% for combustion thermal plants, 30-40% for onshore wind and 20-25% for solar PV. This means an installed nuclear gigawatt delivers, on average, far more annual energy than an installed gigawatt of almost any other source — a key fact when comparing nameplate power with actual generation.

The cost that triggers overruns. The Vogtle case (Georgia, USA) illustrates the financial risk of new nuclear construction: budgeted at 14 billion dollars, it ended up costing more than 30 billion, with 8 years of delay against the original schedule, partly due to using a reactor design (AP1000) with no prior construction experience in the country and the bankruptcy of the main contractor (Westinghouse, 2017). It is not representative of the whole industry — South Korea and other countries build reactors on time and on budget using already-proven designs — but it shows why "the cost of nuclear" depends brutally on whether it is a repeat design or a first-of-a-kind.

Cooling water. A typical nuclear plant withdraws from the environment on the order of 160 liters of water per kWh generated (similar to coal, more than combined-cycle gas), although most of it returns to its source after cooling; net consumption — water that does not return, evaporated in cooling towers — is around 2 liters per kWh. Solar PV and wind need no cooling water because they boil nothing.

What this table does not say

None of the figures on this page include the cost of managing used fuel over the millennia it remains radioactive — that is covered in the waste leaf — nor the risk of a low-probability, high-consequence severe accident, which the "deaths per TWh" statistic dilutes across all the energy ever produced without capturing the social fear or the cost of a permanent territorial evacuation.

Sources

• The IAEA explains the nuclear fuel cycle.
• The OECD Nuclear Energy Agency analyzes costs, financing and system value.
• The IPCC assesses fission alongside other low-carbon electricity options.
• NREL harmonizes life-cycle emissions across technologies: Life Cycle Assessment Harmonization.
• Our World in Data compares mortality and land use by source: Safest and cleanest sources of energy.
• Coverage of the Vogtle cost overrun: E&E News and the follow-up assessment by The Current.