Data Centres, Tax Breaks, and the Familiar Smell of a Bad Deal

There’s a story doing the rounds this week about Illinois Governor JB Pritzker moving to suspend tax incentives for data centres. He’s pausing new applications through the state’s commerce department after the legislature sat on its hands instead of putting guardrails around AI-specific facilities. It’s not a full revocation, just a pause on processing. Incremental. Cautious. Very much the kind of move that gets called “a start” by people who are trying to be generous.

The reaction online was pretty predictable. A lot of people asking, in various registers of exasperation, why these tax breaks existed in the first place.

It’s a fair question, and the honest answer is: they made sense once, in a specific context, and then the world changed faster than the legislation did.

The original logic was straightforward enough. You have underserved regional communities. You want to attract investment and jobs. A data centre brings construction activity, a handful of decent-paying permanent roles, and a large chunk of property added to the local tax base. The Illinois incentive was apparently tied to state sales tax on hardware purchases, not local property taxes, and required at least twenty full-time jobs paying above the county’s median wage. For a struggling rural town, twenty well-paying jobs is not nothing.

What nobody fully accounted for was the AI boom. Traditional compute and storage facilities were relatively benign neighbours. They didn’t consume electricity at a scale that could reshape a region’s grid. They weren’t particularly loud outside the building. They were large, yes, but manageable. Then came the hyperscale AI training facilities, which are a completely different animal. We’re talking about infrastructure that can consume the output of a power station. The tax incentive that made reasonable sense for a conventional data centre becomes a hundred-million-dollar giveaway when applied to one of these things, and the community gets almost nothing in return that it didn’t already have.

I’ve seen this pattern before. Not with data centres specifically, but with the general shape of it. A policy gets written for one set of conditions. The conditions change. The policy doesn’t. And by the time anyone in authority notices the gap, there are contracts signed and facilities half-built and lawyers involved.

What gets under my skin is the jobs argument. It kept coming up in the discussion around this story, and it doesn’t hold up to much scrutiny. Once a large data centre is operational, the on-site headcount is genuinely small. Security, a maintenance crew, a few engineers. The parking lot is often nearly empty. Initial construction does employ people, sometimes a lot of them, but those workers frequently come from specialist contractors based elsewhere, not from the local community. And when the ribbon gets cut, most of them leave.

Compare that to what a factory or distribution centre of equivalent footprint would bring. More jobs, more consistently, spread across a wider skill range. Not glamorous, but real.

To be clear, I’m not against data centres. The internet runs on them. Some of the research being done with large-scale compute is genuinely interesting, even if I’m sceptical of the more breathless claims. The infrastructure has to go somewhere. The question is who pays for it, and who absorbs the cost when the maths doesn’t work out the way the pitch deck promised.

The deal that should be on the table is simple in principle, even if it’s complicated in practice: you want to build here, you pay your share, you fund the grid upgrades your facility requires, and you don’t push those costs onto residential customers. If a data centre operator can’t make that work financially, then the business model depends on public subsidy, and that subsidy should be a deliberate democratic choice, not a legacy incentive that nobody updated.

What Pritzker is doing is modest. A pause, not a reversal. He can’t unilaterally kill the program, so he’s slowing it down through the one lever his office controls. It’s the kind of move that looks small but might matter, depending on whether the legislature actually does anything with the time it’s been given.

I genuinely don’t know if they will. Legislatures tend to move at their own pace, and the lobbying capacity of large tech companies is not trivial. But the political temperature on this has clearly shifted. When the default response from ordinary people is “why were we giving them anything in the first place,” that’s a different environment than the one these incentives were written in.

That’s something, at least. Not much. But something.