Datacenter Resistance

I read another article today about the general resistance to datacenters. Build Where You're Wanted The data-center backlash is being treated as a political problem. It's a siting problem — and siting problems have answers nobody in the fight is offering. The opposition is real now, and it isn't partisan. Roughly seven in ten Americans say they don't want a data center near them. Organizers ran 142 demonstrations across 42 states in a single weekend. Counties are passing moratoriums; states are auditing their own connection queues; the federal government and the states are fighting over who gets to say no. Whatever else you think about it, the resistance has arrived, and it is not going back in the box. The discourse around it has hardened into two positions. One says build faster, preempt the states, one rulebook, the race is everything. The other says stop, ban it, our water and our grid are not yours to spend. Both camps treat community resistance as a political variable — something to be overpowered or something to be mobilized. There's a third position, and it's nearly empty. It belongs to the operator who looks at every objection a community raises — the water, the power, the loss of local control — and sees not a protest to manage but a design input that arrived too late. That's the position worth occupying, because unlike the other two, it has solutions in it. The objections are the specification Here is how a data center gets sited today. Power availability, decided first. Land, fiber, tax abatement — the first-order inputs, locked up front. And then, somewhere well down the list, added in pen after the plan is drawn, water. Community impact later still, if at all. That ordering is the whole problem. Every grievance driving the backlash is a first-order input that the industry has been treating as an afterthought. The basin was never in the room when the site was chosen. Neither was the ratepayer. So the project arrives fully formed, and the community discovers it's been handed the costs of a decision it was never part of. Move those inputs to the front of the list and most of the fight never starts. That isn't public relations. It's engineering sequencing. What follows is what it looks like on each of the three axes people are actually angry about. Geography: choose the ground The single most underrated variable in whether a project gets built is where you decide to try. Approval timelines swing from same-day to more than a hundred days, and the biggest driver isn't the developer's persuasiveness in the hearing — it's the jurisdiction. Same building, same company, different county, opposite outcome. The best operators have already internalized this: they win by not fighting. They land-bank several sites, keep the optionality, deploy into the ground that welcomes the load, and walk away from the parcels that would cost them two years and a recall election. The clearest version of this is the campus rising on decommissioned federal land — a former enrichment site with power already on it and no residential grid to strain, no town hall with standing to lose. That project has no consent fight because it was sited where no consent fight exists. That's not luck. That's the strategy layer sitting above execution: the discipline of choosing a battle you don't have to have. The honest caveat, and I'll hold myself to it: frictionless ground is finite. Pre-zoned industrial corridors, friendly grids, federal parcels — they saturate. When Texas, the friendliest ground in the country, turns around and freezes its own connection queue, the escape route is narrowing faster than the models assume. Choosing the ground buys time. It does not buy immunity. But time, in a buildout racing its own demand curve, is worth a great deal. Water: stop consuming it, start producing it The industry's water playbook is defensive. Efficiency metrics trending down. Net-accounting frameworks that tally water saved elsewhere against water spent on site. Replenishment pledges and stewardship reports. Nearly all of it organized around one idea: consume less, or offset what you consume. That is the timber industry arguing it cuts fewer trees than you think. It isn't wrong. It just doesn't change anything. The reframe is this. Producing fresh water at scale — desalination, atmospheric capture, advanced purification, wastewater reclamation — is bottlenecked almost everywhere by two inputs: energy and heat. Pulling salt from seawater or moisture from air takes enormous power, and the most efficient purification methods are thermally driven, hungry for a steady source of waste heat. Now look at what a data center physically is. A machine that consumes power at gigawatt scale and rejects heat as its primary byproduct. It is a standing surplus of the exact two things water production is starved of. An industry that co-located water production with its buildout wouldn't be defending its consumption. It would be a net producer of fresh water — designed so that a facility gives back more than it takes, measured not in liters on a compliance sheet but in people served. Consume X, add 2X. I'll be clear that the multiplier is a provocation, not a policy target, and that the hard parts are real: brine disposal, the energy cost of capture, the economics of purification. The point is the direction, not the number. But the direction reframes the entire fight. The community's strongest water objection assumes the data center is a drain. What if the buildout became the largest fresh-water innovation program in history? You don't argue your way out of a drought. You engineer your way into being the reason the basin has more water than it did before you came. Power: quit competing with the town for its own grid The sharpest objection communities raise is the one that shows up on their bill. Independent analysis has residents in the most data-center-dense region paying tens of dollars more a month by 2040. That is the real engine of the backlash — not noise, not aesthetics, but a distant company profiting while the neighbor's electricity gets more expensive. Local cost, distant benefit. Self-generation removes that objection at the root. A facility that brings its own power — behind-the-meter generation, stranded gas that was flaring anyway, existing capacity on federal or industrial land, and on the horizon, behind-the-meter nuclear — never competes with the town for the town's grid. The "you raised my rates" fight has no fuel if the load never touches the residential system. It's telling that when Texas froze everything else, the projects it carved out were the ones bringing their own power and the ones sitting outside the shared grid entirely. The regulator drew the map. The escape is self-generation and grid-jurisdiction arbitrage. Same honest caveat applies. Self-generation is also a relocation, not a cure — the off-grid and stranded-power routes will crowd and the easy ones will go first. But a project that carries its own power has subtracted the community's most concrete complaint before the first hearing is ever scheduled. What this actually solves, and what it doesn't None of this defeats consent. That's the part I won't oversell, because it's the top rung of a ladder I've spent a year describing — the one constraint that answers to neither money nor engineering. You cannot buy a town. You cannot prepay a moratorium. When permission becomes the binding constraint, capital and cleverness both stop working. What siting, water, and power do is let you avoid needing the fight. Build where you're wanted. Give the basin back more than you take. Carry your own power so no one's bill moves. Do those three things and most projects never reach the hearing that kills the ones that skipped them. And even that is relocation, not removal — the friendly ground depletes, the escape routes saturate, and the advantage erodes for everyone as the easy paths run out. That's the nature of a bottleneck. You don't remove it. You move it somewhere the answer is still yes. Which is the whole point, and the question the industry keeps declining to ask. Everyone is working out how to win the consent fight. Almost no one is working out how to build something a community would fight to keep. The second problem is harder. It's also the only one with a durable answer.