Data Centers Are Moving Into Rural America. The Counties Aren't Ready.
A new mapping finds rural counties host roughly 6% of operating data centers but about 12% of in progress ones, and they negotiate the deals from the weakest seat at the table.
A new mapping finds rural counties host roughly 6% of operating data centers but about 12% of in progress ones, and they negotiate the deals from the weakest seat at the table.
The geographic center of America's data-center build-out is no longer just the Northern Virginia cluster long known as Data Center Alley. A crowd-sourced mapping project from the Daily Yonder and the FracTracker Alliance finds that rural counties now host roughly 6% of operating data centers but about 12% of in-progress projects, a near-doubling of rural share at exactly the moment when the facilities themselves are getting bigger, hungrier for power, and harder to walk away from.
That asymmetry is the real story. Warehouse-scale server facilities, the physical buildings that house AI training runs, cloud capacity, and hyperscale compute, need land, water, and grid headroom that metropolitan counties rarely have to spare. Developers are increasingly knocking on rural doors, often in counties with thin tax bases and part-time economic-development offices.
The catch is who shows up to that meeting on the county side. Rural counties have smaller tax bases and fewer government staff, which weakens their leverage to negotiate host agreements, power-purchase terms, or decommissioning bonds. The people sitting across from developers are usually stretched across multiple responsibilities. The Daily Yonder analysis treats this gap as the structural variable. A metro county can absorb a bad host agreement or a closure shock. A rural county, where one facility can represent a meaningful share of the local tax base, cannot.
How the map was built matters here. FracTracker Alliance, an organization that tracks oil, gas, and industrial-development siting, compiled the underlying map from crowd-sourced submissions and active reporting on opposition campaigns. The dataset is not a complete census; coverage depends on which communities know to submit their own sites, and the Daily Yonder piece flags that self-selection caveat. Even so, the rural-versus-metro split is consistent enough across operating and in-progress categories to be treated as structural rather than artifactual.
The opposition piece is real and not reducible to NIMBY shorthand. The same mapping that tracks facility locations also tracks organized community pushback, and the rural-versus-metro split is sharper there: opposition campaigns cluster in non-metro counties that have the most to lose and the least institutional capacity to negotiate. That is a civic response to a fiscal exposure, not a preference for the status quo.
The federal government's Metropolitan Statistical Area classification, the Office of Management and Budget's framework for sorting counties by population and commuting ties to urban cores (everything outside an MSA falls into the non-metro bucket), is doing quiet work in this story. The rural share of operating facilities is roughly half the rural share of in-progress ones, a 6%-to-12% swing that will reshape which counties carry the operational, fiscal, and grid load of the AI build-out if it holds.
What to watch: whether in-progress projects reach operation without renegotiation of host-agreement terms, and whether the eventual tax-abatement cliffs trigger the first wave of rural-county revenue shocks. Daily Yonder's mapping is the spine to keep coming back to as those answers arrive.