Any aggregation of radioactive elements is inherently dangerous, always, no matter what safeguards, simply as a material that is hazardous to health. Most radioactive hazards associated with fissile materials represent health risks with atypically, even for poisonous elements, pernicious qualities. Not many people would be sanguine with health-impacting quantities of fissile materials in their environment, unlike say, lead or arsenic, both quite awful but a whole different scale of awful.
Any intentional aggregation of fissile material is intrinsically a risk to human health and safety. Nuclear advocates (many of whom would qualify as frothing fanbois) dismiss such concerns claiming they know of a "safe" reactor design, some of which are quite clever and are, indeed, intrinsically quite safe when operated correctly. Such fanbois will sometimes make the entirely valid point that the rate of death from even older reactor designs is way below the mortality rate induced by burning coal (true) and that nuclear power is much less carbon intensive than fossil fuels (also true).
There are compelling arguments about a diversity of sources, above the reliability of fissile power, about the longevity of investments, and about immunity to a range of possible weather (storms, extended rain) or geopolitical constraints (international fuel shipments), all of which (and more) have merit and should be considered carefully.
I consider four counter arguments that should be considered when contemplating the relative merit of fissile power generation over other methods:
1) Current advanced nuclear power plants have a projected LCOE of $0.11/kWh (DOE) standard new at $0.118-$0.192/kWh (lazard) while solar currently runs $0.05-$0.06/kWh (lazard) and $0.086 (NREL) for battery-backed 24/7 solar, declining to $0.047/kWh by 2050 (NREL). The economic justification for nuclear power is weak in all but a few geographic locations occupied by humans.
2) Uranium is a finite resource, the cost of which is rising and currently $86.8/kg (tradingeconomics). The total proven proven reserves are sufficient for about 2 years of total human consumption in standard light water fission reactors. Breeders and advanced systems have significantly different total reserve values, but such systems come with significant risks, either in proliferation (breeders) or technological (advanced reactor designs). While it is an entertaining speculative argument to discuss technologies that might extend fuel supplies, solar panels and storage (PSH or grid-scale battery) are well proven and widely available today.
3) National security policy: defense against external antagonist: any centralized civilian facility is a target and nuclear reactors, even intrinsically safe ones, are a particularly tempting target. While containment structures are designed to be robust against, for example, the impact of a large passenger aircraft, they are not robust against "bunker buster" style munitions. Consider the consequences of a relatively trivial drone impact with the Chernobyl sarcophagus. As these sorts of attacks are democratized by the proliferation of low cost, long range, difficult to attribute munitions they will inevitably become more common and the only practical mitigation is decentralization. While nanoreactors such as the Zeus (1-20 MW) might make targeting more difficult, they're still vastly more expensive than a few more drones making them juicy asymmetric targets for any NSAG and proliferate the next concern:
4) National security police: defense against extremist state takeover or occupation: "safe" nuclear reactors assume operation remains in the hands of moderately responsible, reasonable people, an assumption not supported by recent history. There is no fission design that does not rely on a concentration of fissile material, intrinsically radioactive. There is no such concentration that cannot be widely dispersed with a properly engineered application of conventional high explosives. Every nuclear reactor, anywhere in the world, is a pre-emplaced nuclear munition. To quote the (poor) AI transcription of Serhii Plokhy from the Economist's Intelligence podcast https://www.economist.com/podcasts/2026/04/24/an-explosion-still-echoing-chernobyl-at-40 "The taboo on the occupation and military attacks on nuclear sites that existed before 2022 had been broken. It is gone. Now, the war in Ukraine is the war of drones more and more. And the non-nuclear country. Can very easily go nuclear by attacking somebody's nuclear facilities like nuclear power plants." Chernobyl is occupied, Zaporizhzhia is occupied, the Ukrainians got close to taking Kursk; if an occupying force is forced out, might they be tempted to salt the earth they're forced to leave with fissile debris? Might one of the many political parties animated by an Armageddonist eschatology consider a nuclear reactor an inexpensive and readily available doomsday munition, a sort of clarion call to whatever deity they are waiting for? Even if they're wrong about the level of doom a conventional detonation of a fissile core might achieve, the consequences would be far more lasting and terrifying than a fuel depot or even grid-scale lithium battery fire.
