Amazon budgets about $200 million a year to win over its US data center towns

Allianz's $130 billion measures project value held up, with delays and outright blocks counted together ^[6]. Spread across the more than 75 projects in the count, it comes to at most about $1.7 billion each ^[16]. A delayed campus and a cancelled one are different losses, so what opposition actually costs the industry is some fraction of the headline figure. Delay hurts most where space is scarce. CBRE put Northern Virginia vacancy at 0.3% in the first quarter, with power shortages, zoning and local resistance holding back new supply ^[13].

Spread evenly, Amazon's pledge is about $200 million a year ^[14]. The full five-year sum equals 0.77% of what Allianz counted as stuck in a single quarter ^[17].

The spending lines up with the poll's complaints. Respondents to the UMass Amherst survey named environmental harm, resource use, land disruption, distrust of AI and higher utility costs ^[5]. Amazon's money goes to free community college, trades training and home energy-efficiency upgrades in data center counties ^[3]. The upgrades address the utility bill and the training addresses the jobs case. On water and pollution, Amazon has chosen to argue its case in public, through Garman's post ^[2].

Lancaster is the closest comparison. Amazon's national pledge is 50 times the $20 million that three developers put up there ^[18], but the Lancaster developers also accepted clean-power, water and noise conditions on a single site ^[10]. Neither account of Amazon's pledge says how many counties will share the money or describes operating limits of that kind ^[3].

If approvals in Amazon's counties speed up, $200 million a year is a cheap fix ^[14]. If the money is spread across too many counties to move a zoning vote, it changes little. The third outcome turns on the disclosure. WRI projects that data centers could draw as much as 20% of US electricity by 2035, up from 5.9% now ^[8], and an annual Amazon report ^[7] would put the company's part of that growth on paper every year.

I think the disclosure is the larger of the two commitments. Cash can be topped up quietly. A published water figure stays published, and next year's will be compared with it. The view is wrong if Amazon soon adds site-level cash like NorthPoint's proposed household payments ^[9], because that would show the national program had not bought enough consent on its own.