Advantages and disadvantages of nuclear energy

This entry gathers verifiable facts about nuclear energy, for and against, each with its figure and its primary source, without taking a side or closing with a balance. Cost and build time, capacity factor, radioactive waste, and accident risk each have their own entry in this wiki and are summarized here without repeating them; what follows in detail is what has no entry of its own: land use, the gap between a reactor's license and its real lifetime, the decommissioning fund, dependence on imported fuel, load following, and employment.

Upfront capital is high and build time is long: in the West, the first units of each design ran well over their original budget and schedule. Once running, though, the U.S. fleet had a capacity factor above 92% in 2023, according to the U.S. Energy Information Administration (EIA): the investment is already made, so each additional kilowatt-hour is cheap. That asymmetry between spending heavily upfront and almost nothing afterward is what shapes the levelized cost of nuclear electricity. The detail behind those figures is in Capacity Factor and the Economics of a Reactor.

Radioactive waste is the environmental cost most often raised against nuclear energy: how much volume it generates, how dangerous it stays over time, and what happens to it while most countries still have no operating deep geological repository. This wiki covers it in Radioactive Waste.

The risk of a severe accident is the other heaviest potential cost, and the industry's and regulators' answer is to stack several independent physical barriers on the principle that none of them needs to be perfect on its own. This wiki covers it in Safety Barriers and Defense in Depth.

The power it delivers takes up relatively little land: between 6.7 and 13.8 km² (2.6 to 5.3 sq mi) per 1,000 MW of capacity, according to the U.S. Department of Energy (DOE), a range that runs from the plant site alone up to adding transmission lines, water supply, and rail access; neither end includes the footprint of uranium mining or waste disposal, which happen elsewhere and on a different timescale.

A commercial reactor's license in the United States runs for up to 40 years and is renewed in 20-year increments; since 2019 the Nuclear Regulatory Commission (NRC) also approves subsequent license renewal, allowing operation up to 80 years, and by September 2026, 25 units at 13 plants had already received it, most recently Hatch 1 and 2 in June 2026. At Barakah, each of the four units already holds a 60-year operating license from startup, according to its operator, ENEC. Real-world lifetimes are more uneven: per Reactor Atlas's own data as of September 2026, of the 416 reactors connected and running in the world today (not counting the 26 suspended ones, 18 of them in Japan), 189 (45%) are past 40 years old, and the oldest has been running since 1969; but of the 235 reactors that have permanently closed with a recorded closure year, the median lifespan was 32 years, well short of those 60-to-80-year licenses.

Permanent shutdown also carries a specific cost, set aside in advance. The NRC requires every U.S. operator to hold a decommissioning fund of at least $290 million for a pressurized-water reactor and $370 million for a boiling-water reactor, in 1999 dollars adjusted annually; those funds build up over the plant's entire operating life, not as a bill that arrives only at the end.

Most countries that operate reactors do not produce their own fuel. The United States, with the world's largest fleet, imported 99% of the uranium concentrate (U₃O₈) its reactors used in 2023 — from Canada, Australia, Russia, Kazakhstan and Uzbekistan — and bought 27% of its enrichment services abroad from Russia that same year, according to the EIA; the U.S. Congress banned Russian uranium imports starting in August 2024 (Public Law 118-62), with waivers the DOE can grant if no other source is available. That physical dependence coexists with low price exposure: fuel is only 15 to 20% of the levelized cost of nuclear electricity, against capital that runs 60% or more, per the OECD comparison cited in Capacity Factor and the Economics of a Reactor.

A reactor's classic design runs best at constant power, but that is not an absolute physical limit. French reactors can make up to two daily reductions of up to 80% of rated power, according to a 2024 French Senate report using EDF and RTE (the grid operator) data; in practice, the fleet's modulation averages around 7% of rated power, with a load-following volume of about 30 TWh a year, ranging between 20 and 55 depending on the year. That capability depends on each reactor's design and on whether the operator and the regulator enable it: not every fleet in the world uses it the same way.

Building, operating and decommissioning a 1,000 MWe reactor together generate about 50,000 person-years of direct employment over its full life, according to a 2018 study by the OECD Nuclear Energy Agency: about 12,000 person-years over the 10 years of construction (roughly 1,200 people), 30,000 over 50 years of operation (roughly 600 people a year), and 8,000 in decommissioning and waste management. Adding indirect employment, generated by purchases from suppliers, and induced employment, from the spending of those wages in the local economy, the total runs to about 200,000 person-years per reactor; that near-fourfold multiplier over direct employment is an artifact of the input-output model the study uses, not something specific to nuclear energy — any sector of the economy can report a similar one with the same method.