A cheap Russian drone struck Chornobyl's confinement shelter in February 2025, raising fresh safety concerns about the site of the world's worst nuclear accident.
The Subcommittee for the Long-Term Management of Historical Mistakes had, by all accounts, been doing a reasonably good job of its primary function, which was to ensure that the mistakes of the past remained sufficiently contained so as not to inconvenience the present. The confinement shelter at Chornobyl was the physical manifestation of this function - a massive, expensive, and incredibly complicated architectural shrug, designed to say to the laws of physics, “We know what happened in 1986, and we have decided to put a very large lid on it.”
The problem with any system designed for containment is that it assumes the primary threat will be something predictable, like a structural failure or a particularly ambitious leak. It does not, however, account for the Committee Problem applied to modern warfare. In a modern conflict, the primary threat is often a piece of technology so inexpensive and so fundamentally unspecialised that it bypasses the entire logic of the containment system.
The recent strike on the shelter by a cheap Russian drone is a perfect example of a process producing an outcome that no individual participant would have endorsed. On one side, you have the Russian military apparatus, which is optimising for the cost-effective delivery of kinetic energy to a specific set of coordinates. On the other, you have the international community of nuclear safety experts, who are optimising for the preservation of a delicate, multi-billion-dollar equilibrium. Neither side is specifically trying to trigger a radiological event that would render much of Europe a very unpleasant place to visit; that would be an inefficient use of resources. However, when you take a system designed for high-level stability and subject it to a process designed for low-level, high-frequency disruption, the result is a catastrophic loss of stability that neither side actually requested.
The drone itself represents the ultimate failure of the planning permission principle. The shelter was built to withstand the weight of history and the pressures of radioactive decay, but it was not, in its technical specifications, designed to withstand the arrival of a piece of plastic and circuitry that costs less than a mid-range smartphone. It is the equivalent of building a vault to protect the Crown Jewels from a siege, only to have the vault compromised by a particularly aggressive pigeon.
The tragedy of the situation is that the system is working exactly as it was evolved to work. The containment structure is doing its job of containing the 1986 disaster, and the drone is doing its job of being a cheap, effective tool of disruption. The gap between these two functions is where the danger lives. We have created a world where the most sophisticated safety protocols in human history can be rendered moot by a piece of hardware that was likely assembled in a factory that also produces much of the world’s disposable cutlery.
The uncertainty regarding the extent of the damage is, in itself, a classic bureaucratic byproduct. When a system is too complex to be easily repaired and too important to be ignored, the standard procedure is to enter a state of perpetual, highly-documented observation. We will monitor the radiation levels, we will issue reports on the integrity of the steel, and we will hold meetings to discuss the possibility of further meetings, all while the actual physical reality of the situation remains stubbornly indifferent to our paperwork. The safety of the site is currently being contested not because we lack the data, but because the data is being produced by a process that is increasingly unable to distinguish between a controlled environment and a combat zone.