Spent Fuel and Waste: The Honest Picture

When a fuel assembly leaves a reactor after its three cycles, about 95% of it is still uranium. The rest is the problem and the resource: roughly 3–4% fission products (the split fragments — intensely radioactive, mostly short-lived) and about 1% plutonium and other transuranics (mildly radioactive but very long-lived). This inversion — the hot stuff fades fast, the long-lived stuff isn't very hot — is the key to thinking clearly about waste. Spent fuel's radioactivity falls by orders of magnitude in its first decades; what remains after a few centuries is dominated by isotopes whose hazard is real but vastly lower, on timescales (tens of thousands of years) that are more a question of geology and institutions than of engineering.

The quantities surprise everyone. The entire US commercial program — six decades, the world's largest fleet — has produced on the order of 90,000 metric tons of spent fuel. All of it, stacked as assemblies, would cover a single football field to a depth of about ten meters. A reactor supplying a city of a million people produces roughly three cubic meters of spent fuel a year; the coal plant it displaces would produce hundreds of thousands of tons of ash (itself mildly radioactive) plus millions of tons of CO₂ dispersed into the sky. Nuclear is the only energy source that fully contains its waste — which is both its burden and, arguably, its virtue.

Where is it? At the plants. Fresh spent fuel spends its first ~5 years in the spent-fuel pool, deep racks under 12 meters of water that handle both cooling and shielding (the water glows faint blue — Cherenkov radiation — one of the industry's genuinely beautiful sights). Once decay heat falls enough, assemblies move into dry casks: welded steel canisters inside concrete overpacks, cooled passively by air, sitting on guarded pads. Casks are licensed for decades, ride out floods and projectiles in testing, and have accumulated a clean safety record across 30+ years. Every operating plant on this map, and several sites where the reactor itself is long demolished, hosts a cask pad — de facto interim storage, nationwide.

What the US hasn't done is the last step. The scientific consensus solution is deep geologic disposal — engineered canisters in a stable formation hundreds of meters down — and the law (the Nuclear Waste Policy Act of 1982) made it federal responsibility, funded by a fee ratepayers already paid for decades. The designated site, Yucca Mountain in Nevada, was studied exhaustively and then halted by politics in 2010; the licensing is suspended, the money sits collected, and the federal government pays damages to utilities for fuel it was contractually obliged to take beginning in 1998. Meanwhile Finland has simply done it: its Onkalo repository, 400+ meters down in granite, is entering operation as the world's first — proof the problem is solvable where consent and persistence align. The live US motion is consolidated interim storage (licensed private sites in Texas and New Mexico, litigated) and a federal restart of consent-based siting.

Two closing nuances. Spent fuel can be reprocessed — France does, recycling plutonium into fresh fuel — but the US chose not to in the 1970s on proliferation grounds, and at today's uranium prices reprocessing is uneconomic; several advanced-reactor developers (Oklo most vocally) nonetheless pitch their designs as eventual spent-fuel burners. And the category matters: this page is about high-level waste; the much larger volumes of low-level waste (gloves, filters, components) are routinely disposed of in licensed near-surface facilities and are not the hard problem.