Making electricity: thermal, nuclear, hydro, wind and solar: Nuclear electricity

3.1 Fission heat raises steam, and a little fuel goes a very long way

When the nucleus of a uranium-235 atom absorbs a slow neutron, it splits into two lighter nuclei and releases two or three more neutrons plus a great deal of energy, mostly as the motion of the fragments. Those fragments slam into the surrounding fuel and turn their motion into heat. If, on average, one of the released neutrons goes on to split another nucleus, the chain reaction sustains itself at a steady rate.

Natural uranium is only 0.7% uranium-235; the rest is uranium-238, which does not split readily. Most reactors therefore use fuel enriched to 3 to 5% uranium-235, pressed into ceramic pellets and stacked in metal tubes. Water flowing through the core does two jobs. It carries the heat away, and it slows the neutrons down, which makes them far more likely to cause fission. Most of the world's reactors are light-water reactors, a design inherited from the US naval programme of the 1950s. In a pressurised water reactor, the most common type, the core water is kept under such high pressure that it does not boil; it passes its heat to a separate loop that makes steam for the turbine.

The energy density is extraordinary. Complete fission of one kilogram of uranium-235 releases about 80 TJ, the heat of some 3,000 tonnes of coal. Real fuel is mostly uranium-238 and is removed long before all its fissile atoms are used, but even so a kilogram of enriched fuel typically delivers around 3,500 to 4,000 GJ of heat over its time in the reactor, about 150,000 times as much as a kilogram of coal. A 1,000 MW reactor needs roughly 20 to 30 tonnes of fresh fuel a year; a coal plant of the same size needs around 3 million tonnes of coal.

The catch is temperature. Water-cooled reactors produce steam at only about 280 to 320 C, much cooler than in a modern coal plant, so their efficiency is only about 33%. Two thirds of the reactor's heat goes to the condenser, which is why nuclear plants have the largest cooling needs of all.

• Jaffe and Taylor, The Physics of Energy (2018) · via Smil Mastery Course, M7 'Nuclear electricity: an unfulfilled promise'
• Smil, Energy at the Crossroads: Global Perspectives and Uncertainties (2003) · via Smil Mastery Course, M7 'Nuclear electricity: an unfulfilled promise'

3.2 Smil calls nuclear power a successful failure

In 1954 Lewis Strauss, head of the US Atomic Energy Commission, predicted that the children of his day would enjoy electricity too cheap to meter. Smil treats that line as one of the great energy myths of the post-war era. Nuclear electricity has been technically impressive, but it never became cheap, and it never became the dominant source of power.

The achievements are real. A well-run reactor operates at a capacity factor above 90%, meaning it produces more than nine tenths of what it would if it ran flat out all year. It emits almost no CO₂ in operation, needs very little land and fuel, and its fuel is cheap and easy to stockpile for years. France built most of its fleet within about 15 years after the 1973 oil crisis and has since drawn around two thirds or more of its electricity from fission.

The problems were just as real. Reactors are enormous, complex projects whose capital cost dominates the price of their electricity, and many Western projects have run years late and billions over budget. Three accidents, Three Mile Island in 1979, Chernobyl in 1986 and Fukushima in 2011, hardened public opposition and led to new safety rules that raised costs further. Most countries still have no permanent repository for high-level waste. Finland is the exception: in October 2026 it granted the world's first operating licence for a deep geological store, Onkalo.

Smil also tells the story of the fast breeder reactor, which was supposed to replace light-water reactors by making more fissile fuel than it consumed, multiplying the usable uranium resource by tens of times. Despite large national programmes, breeders never became commercial. In his book on hype and failure in invention he uses them as a leading example of a technology confidently expected to dominate that did not.

His overall stance is balanced. Nuclear power is valuable for a low-carbon grid, and smaller, factory-built reactors might change its economics, but he warns against expecting a rapid renaissance on the strength of designs that have not yet been built at scale.

• Smil, Energy Myths and Realities: Bringing Science to the Energy Policy Debate (2010) · via Smil Mastery Course, M7 'Nuclear electricity: an unfulfilled promise'
• Smil, Invention and Innovation: A Brief History of Hype and Failure (2023) · via Smil Mastery Course, M7 'Nuclear electricity: an unfulfilled promise'
• Finland grants licence for world's first nuclear waste repository (CP24/AP, 22 October 2026) · web

3.3 Nuclear output hit a record in 2025 while its share kept shrinking

Nuclear's share of world electricity peaked in the mid-1990s at about 17%. Since then reactors have kept running and new ones have been built, but total generation has grown faster, so the share has fallen by half. According to Ember's Global Electricity Review 2026, nuclear supplied about 8.9% of world electricity in 2025. The IEA's Electricity 2026 report notes that nuclear generation nevertheless reached an all-time high in 2025, helped by reactor restarts in Japan, stronger output from France's fleet after its corrosion problems, and new units in China and India.

The centre of gravity has moved east. The IEA expects China alone to provide about 40% of the growth in nuclear output to 2030, building reactors on a steady programme at costs and schedules far below recent Western projects. Several countries have slowed or reversed planned phase-outs, and technology companies have signed contracts to buy nuclear electricity for data centres.

India has a small fleet of about 8.8 GW, which supplied 3.1% of its electricity in 2024-25, built largely on its own pressurised heavy-water reactor design, which runs on natural uranium and uses heavy water to slow neutrons. That choice was forced by decades of exclusion from international uranium trade. India also pursued the three-stage programme meant to lead, through fast breeders, to using its large thorium reserves; the 500 MW prototype fast breeder at Kalpakkam, begun in 2004, reached first criticality only in April 2026, an example of Smil's point about how slowly such technologies mature. The government's Nuclear Energy Mission targets 100 GW by 2047, and the SHANTI Act of December 2025 lets private companies build and operate plants, but that would require building at a pace India has never managed.

For an investor, nuclear output is reliable and nearly carbon-free, but returns depend on construction discipline, regulation and long-term contracts or state backing through a decade-long build.

• Ember, Global Electricity Review 2026 (April 2026) · world shares 2025, https://ember-energy.org/countries-and-regions/world/
• IEA, Electricity 2026 (February 2026) · web
• Smil, Numbers Don't Lie: 71 Stories to Help Us Understand the Modern World (2020) · via Smil Mastery Course, M7 'Nuclear electricity: an unfulfilled promise'
• PIB, SHANTI Bill backgrounder (December 2025) · 8.78 GW installed, 3.1% of generation in 2024-25, 100 GW by 2047, private participation: https://static.pib.gov.in/WriteReadData/specificdocs/documents/2025/dec/doc20251219739001.pdf
• Department of Atomic Energy, 'Prototype Fast Breeder Reactor at Kalpakkam attains first criticality' (April 2026) · first criticality 6 April 2026, https://dae.gov.in/prototype-fast-breeder-reactor-at-kalpakkam-tamil-nadu-attains-first-criticality

See it in the table

Check your understanding

Why is the efficiency of a light-water nuclear plant only about 33%?

Nuclear generation reached a record in 2025, yet its share of world electricity is about half its mid-1990s peak. How can both be true?

What is the main reason nuclear electricity has proved expensive in many Western countries?