67% of US AI Data Centers Are Heading Rural — and Towns Are Resisting

Sixty-seven percent of planned US data centers are targeting rural land — a sharp reversal from the current geographic footprint, where 87% of existing centers sit in urban areas. Pew Research data, reported by the Financial Times, captures the scale of the pivot AI infrastructure buildout is now driving. The consequences are already visible at the town level: Archbald, Pennsylvania, a community of roughly 7,000, is facing six proposed data center campuses that collectively cover approximately 14% of the borough's total land area.

What the Source Actually Says

The Washington Post's April 26 investigation into Archbald provides the human frame for the Pew statistic. The borough — less than four square miles — has become an unlikely front in the tension between tech industry land demand and small-town planning authority. Six separate proposed campuses, if built, would collectively consume a land share that is visible at a glance on any map of the town. Residents have organized in opposition, citing concerns about noise, water consumption, power grid load, and the permanent conversion of open and productive land.

The FT's coverage of the Pew data adds the systemic dimension. The rural targeting trend is not accidental: land in rural areas costs less, power rates are lower, and large contiguous tracts are available where urban density makes them impossible. The gap between the existing infrastructure distribution (87% urban) and the planned pipeline (67% rural) signals an industry-wide calculation that rural sites represent scalable capacity in a way that urban land no longer can.

The resistance forming in places like Archbald represents a test of that calculation. Community opposition does not automatically stop projects, but it introduces permitting delays, political risk, and, in some jurisdictions, local referenda. Multiple rural counties have already enacted or proposed moratoriums on large-scale data center development.

Strategic Take

AI compute scale now depends as much on land-use politics as on capital and hardware. Teams planning infrastructure at scale should treat zoning boards, community approval timelines, and local elected officials as project risks on par with power procurement. Rural sites with existing community support, water rights, and favorable zoning are undervalued options in a market that is only beginning to discover how contested rural land can become.

#archbald-pennsylvania

67% of US AI Data Centers Are Heading Rural — and Towns Are Resisting

AI Intelligence Newsletter