Washington, D.C.’s metropolitan area has long hummed with data centers. The region, which encompasses much of Northern Virginia, has become known as Data Center Alley, home to more data centers than anywhere else in the world. But the data center boom, driven by the rise of AI and the race to build the infrastructure powering it, is changing the geography of these energy-intensive, warehouse-like facilities.
Data centers have arrived in rural America.
The Daily Yonder analyzed crowd-sourced data on the locations of data centers, how they connect to existing grid infrastructure, and where communities are fighting back.
Though the majority of in-progress data centers are in metropolitan areas, our reporting and data analysis show that large data centers, characteristic of the current boom, represent outsized investments in rural areas. Smaller populations often mean smaller tax bases and fewer government officials, leaving rural communities with fewer resources to negotiate deals with developers or weather the tax revenue fallout after facility closures.
The Geography of the Data Center Boom
The following map shows the number of data centers in both metropolitan and non-metropolitan, or rural, counties. As of June 2026, approximately 30 of the 529 data centers currently in operation are located in rural counties, representing about 6% of operating data centers. At the time of this analysis, 16 rural data centers were either recently approved or currently under construction, representing about 12% of all in-progress projects.
The Daily Yonder uses a county-level definition of rural derived from the Office of Management and Budget (OMB). All counties outside of a Metropolitan Statistical Area (MSA) are classified as rural for the purposes of this analysis.
Information on the locations of data centers comes from FracTracker Alliance, an organization providing the public with interactive tools to understand oil, gas, petrochemical, and data center development across the country.
While there are many other datasets out there tracking data center development, few of them are publicly accessible, making a complete analysis difficult. We chose FracTracker because of the organization’s methods for obtaining data and the monthly frequency with which it updates the database. We saw that frequency as a key metric for tracking this fast-evolving issue. FracTracker obtains information from crowd sourcing, media announcements, partner datasets, and public records. While the organization is primarily focused on proposed data centers, it is slowly adding existing data centers to the database.
That means that the data used in this analysis is preliminary. Because it is a work in progress, FracTracker encourages readers to share new and updated information about data centers in their own communities via a submission form.
How Do the Daily Yonder’s Findings Compare to Others?
Analyses that rely on other rural definitions and datasets yield different results. For instance, a recent analysis by the Pew Research Center relied on data from Data Center Map, an industry database. Pew Research found that more than 1,500 new data centers are planned nationwide, with 67% slated for rural areas.
Yet Pew Research uses a slightly broader interpretation of rural than the Daily Yonder. Pew Research interprets rurality as it is defined by the U.S. Census Bureau’s urban-rural classification, which includes areas on the outskirts of metropolitan counties that have low population density. These communities are not captured in the OMB definition that the Daily Yonder uses.
Still, as data centers expand beyond metropolitan areas, their development and reception in Census-defined rural places often resemble patterns found in non-metropolitan counties defined by the OMB.
So, although figures may vary by rural classification method, the general trend is that the presence of data centers is a growing concern in rural communities, particularly because the complex dynamics of data center development show up differently in smaller municipalities.
“It is fundamentally a power imbalance for a city attorney in a medium or small-sized city to be able to go up against Amazon’s lawyers, right? That’s not a fair negotiation,” said Asad Ramzanali, former deputy director for strategy at the White House Office of Science and Technology Policy under the Biden Administration.
An example of that imbalance can be found in Port Washington, Wisconsin, where Vantage Data Centers is building a $15 billion hyperscale data center campus to serve OpenAI and Oracle as part of OpenAI’s $500 billion Stargate campaign to develop AI infrastructure across the U.S.
The investment in Port Washington, a town of around 13,000 on the shores of Lake Michigan, represents 50% of the community’s tax base, according to Prescott Balch, a local advocate who advises Wisconsin communities on data center negotiations. Balch is a retired software developer and former technology executive at U.S. Bank. He lives in Caledonia, Wisconsin, a rural village south of Milwaukee, and helped organize his neighbors to get a Microsoft data center proposal cancelled in October 2025, which he opposed in part because of the project’s financial impact.
“If fifty percent of your tax revenue dries up, then everybody’s tax bills double overnight,” Balch said. “That’s a disaster in no uncertain terms. Nobody would run an investment portfolio or a retirement portfolio that way.”
Data Center Development Maps Onto Existing Infrastructure
The following map displays non-metropolitan data centers on top of transmission lines with voltages at or above 345 kilovolts (kV). Data centers’ electricity needs often drive their construction in relative proximity to power grid infrastructure that can support their operations.
But not all transmission lines are created equal. The transmission lines relevant to the data center conversation are those with higher voltage classes, typically 345kV and up, since they deliver the most substantial power, according to Anna Haensch, an associate research professor at the University of Wisconsin-Madison’s Data Science Institute.
Zoom in on the map above to see how rural data centers track onto major transmission line networks. The majority of the facilities are five miles or closer to high-voltage transmission lines. According to our analysis, the median distance from a rural data center to one of these major transmission lines was about four miles.
In Wisconsin, for example, all of the data centers in the FracTracker dataset are in close proximity to 345kV transmission lines. Observe the same trend in southern Minnesota, where a Goodhue County District Court Judge recently ordered developers to halt a hyperscale project near Rochester until an environmental lawsuit is resolved.
According to the FracTracker data, there are more than 113 data centers proposed in rural counties. In regions like the mid-Atlantic and Appalachia, many of these proposed data centers track alongside existing transmission line networks that stretch across Pennsylvania and Virginia into West Virginia and Ohio. In Texas, many facilities are proposed or have already been built in close proximity to the state’s 345kV network. There, locals from across the political spectrum have been fighting to stop the development of a 765kV transmission line in rural Central Texas that will power oil, gas, and data centers in West Texas.
But in some regions without the right infrastructure, data centers aren’t being constructed, even if the local or state political environment is favorable to development. Take Northern Wisconsin, for example. There, Haensch said, fewer data centers have been proposed because the rural region lacks high-voltage infrastructure.
“If a data center wanted to build there, then they would have to be paying for laying down transmission lines, which is already really expensive, but the social cost of it is so tremendously high, because then you’re talking about blasting through people’s farms and claiming right of way.”
The Growing Resistance to Data Center Development
According to our analysis of data from the Data Center Opposition Report, a coalition of organizations and academics tracking resistance to data centers, there are approximately 129 groups in rural counties that are fighting the rise of data centers.
Opposition movements are largely concentrated in regions where there is more data center development, like the Washington, D.C., metro area and adjacent rural communities.
But across the country, opposition is growing. Recent polling shows that a majority of Americans are concerned about data center development. In May, a Gallup poll found that seven in 10 Americans oppose constructing AI data centers in their area, with nearly half being strongly opposed. In June, a Reuters/Ipsos poll found that 77% of Americans are concerned that AI data centers will make electricity more expensive. That same poll found that 57% of Americans, including two-thirds who identify as Democrats and half who identify as Republicans, said they’d oppose building a data center in their community.
“These data centers[…], there’s really a rush to build them as quickly as possible, and communities often feel like they’re being forced upon them without their consent and bypassing standard processes for approving projects,” said David Krueger, one of the authors of the Data Center Opposition Report and an AI professor at the University of Montreal who has been studying AI safety for more than a decade.
Some groups encompass large geographic areas, like a whole state or region within a state. A group might therefore be based in an urban area, but include organizers in surrounding rural areas. These groups are not without success. In rural communities, some resistance groups have brought together broad coalitions to cancel or pause proposed data centers.
“Generally, it’s quite non-partisan or bipartisan,” Krueger said. “We found a pretty even distribution of these sorts of groups in more or less conservative or liberal areas. I think that’s probably the main thing that stands out.”
Sentinel — Human
This analysis demonstrates complex synthesis of varied data and quoted sources, reflecting a high degree of human journalistic structuring rather than purely algorithmic generation.
