The Invisible Trillion · How a Neighborhood Dies

What the Industry Says, and Why It Doesn’t Fully Survive

The industry has answers to all of this, and a fair-minded reader should hear them, because some of them have real force.

Data centers are necessary infrastructure. True. Modern life, AI included, depends on them, and no serious person wants to abolish them. The argument is about how and where they’re built, not whether — and treating it as a fight over their existence misreads it.

They bring jobs and tax revenue. Sometimes. The jobs are real but few; a hyperscale building might permanently employ a hundred or two, fewer per dollar than almost any other use of the land. The tax revenue is real but frequently offset, sometimes more than offset, by the subsidies handed over to land the project. Whether a given deal nets out positive depends entirely on its terms, and the blanket version doesn’t survive a look at them.

The industry invests in mitigation. Sometimes. Some facilities are well designed; some aren’t. And the investment tends to show up where there’s organized opposition — reactively, after the lawsuits and the coverage and the pressure — rather than as standard practice everywhere.

They comply with the regulatory limits. Largely true, and largely beside the point, for the reason the energy chapter gave: the limits were written before this industrial form existed, calibrated against other problems, and mostly never updated. Complying with an outdated limit is not the same as doing no harm.

The people complaining are NIMBYs. This is the move that does the most work, so it’s worth pausing on. NIMBY — “not in my back yard” — is a way to dismiss a specific objection without engaging it. The useful test, when you hear it, is whether the person using the word would take the project in their own neighborhood. The residents’ concerns — the noise, the air, the water, the bills, the property values, the children — aren’t opposition to all development. They’re opposition to this development, in this place, with these costs, which they didn’t agree to and have few good ways to refuse.

Remote siting isn’t economically feasible. This is the most interesting one, and it deserves a straight answer. At the margin it’s true: building out in the country costs more — fiber, workforce, logistics. What it leaves out is who pays the cost of building close in. The cost of the remote site is paid by the developer. The cost of the residential site is paid by the residents, in everything this chapter has described.

Those are externalities — costs pushed onto other people — and when they’re large, the efficient arrangement is for the party creating them to carry them. Require developers to do that — stricter siting, real mitigation, full payment for the grid, honest regulation — and some close-in sites stop penciling out while others survive with protections attached. The buildout costs a bit more and lands more fairly. The “infeasibility” is about margins, not physics.

All 19 chapters of The Invisible Trillion · Library · All courses