Communities Resist Forced Datacentres Over Energy Strain

This summer’s disputes in Scotland and India benefit, by the industry’s own reckoning, an enormous population: every person who will eventually use an AI system, plus the shareholders of the big tech companies building the infrastructure, plus governments hoping to attach their names to growth figures. Set against this vast diffuse constituency stands a much smaller, sharply defined one: the residents of the villages and towns where the machines actually sit, who lose water, grid capacity, quiet, and in some cases the land itself. The arithmetic looks lopsided in favour of the many until you ask the four questions the calculus insists on asking of any pleasure or pain: how intense, how certain, how near, and how long.

Begin with the harms, because they are the easiest to measure and the ones everyone else prefers to leave vague. A community near a proposed site in Scotland facing higher electricity costs and strained supply suffers a pain that is immediate, certain, and continuous - it does not depend on any future event going well. A village in India watching its water table drop because a cooling system has first claim on it suffers something with the same properties. These are not probabilistic pains awaiting confirmation. They start the day the machines switch on.

Now the benefits. The claim that datacentres are essential for cashing in on AI is, notice, a claim about the future - about growth that has not yet arrived, tax revenue not yet collected, jobs not yet created in the numbers promised. Even granting every optimistic projection, this pleasure is diffuse across millions of eventual users and uncertain in its arrival, whereas the pain is concentrated in hundreds or thousands of households and certain in its arrival. A pleasure spread thin over a distant population, discounted by uncertainty, does not outweigh a pain delivered whole and immediate to your actual neighbour. This is not sentiment. It is simply what happens when you multiply magnitude by probability by proximity rather than simply counting heads.

The honest advocate for datacentres will object that I am rigging the comparison by comparing present certainty to future promise, and that every piece of transformative infrastructure looked like a bad bargain to the people living next to it while it was being built - the railway, the power station, the reservoir. This is a fair objection and deserves a fair answer. The answer is that a railway, once built, delivered its benefit to the very people who bore its noise and its smoke; the local farmer got a station. The datacentre delivers its cooling costs to the village and its output to a data centre customer three thousand miles away who will never know the village’s name. The asymmetry is not between present cost and future benefit - it is between local cost and non-local benefit, which is a different and much less forgivable arrangement.

Consider a specific image: a family outside Mumbai checking a borewell that used to reach water at thirty feet and now does not, while forty miles away a server farm processes requests for a chatbot summarising emails for someone in Ohio. Nobody designed this to be cruel. It is simply what happens when the party that captures the pleasure and the party that absorbs the pain are never made to negotiate directly, because the government stands between them promising jobs that arrive, if at all, on a different clock.

The reform this suggests is not prohibition - a legislator who forbade datacentres outright would sacrifice a real, if uncertain, future good to avoid a real, certain present one, which is its own kind of miscalculation. The reform is to make the local population a genuine party to the transaction rather than a cost absorbed on its behalf. Energy and water drawn from a community should be metered, priced, and returned to that community in tariffs and infrastructure before a single server rack is switched on - not promised as a future dividend of growth, but delivered as a present offset for present pain. Scotland and India are, in this sense, running the same experiment with different weather: testing whether a government can extract consent from a village by describing a happiness that will belong, in the end, almost entirely to somebody else.