Growing local opposition to AI data center development is transforming previously popular projects into political liabilities for elected officials.

Data center development is becoming a campaign liability for politicians who once championed it as an economic development win.

At least 18 incumbent officeholders lost elections between August 2025 and August 2026 after taking documented action on a data center project or becoming entangled in a political fight over its development, according to a Data Center Knowledge review of election results, government records, campaign materials, and local reporting.

The cases vary in conclusiveness. In some instances, defeated officials directly attributed their losses to votes on data centers. In others, challengers elevated data centers to a prominent campaign issue alongside electricity costs, taxes, land use, and local control.

Data Center Knowledge counted an incumbent only when the official lost a completed election following a documented vote, approval, incentive, zoning action, or public advocacy tied to a data center project or policy controversy. The review does not treat chronological proximity alone as proof that the issue determined the outcome.

Related: Land Banking Explained: A Novel Strategy for Data Center Expansion

The tally includes one officeholder, Bridget McCandless of Missouri, who ran for mayor while serving as an incumbent council member. This is a count of documented losses, not a statistical measure of how frequently data center involvement causes an incumbent to lose.

Incumbents have generally been reluctant to publicly blame data centers for their defeats. Their opponents have been considerably less hesitant to draw the connection.

Utah produced the clearest examples. State Senate President J. Stuart Adams lost his Republican primary in June to Stephanie Hollist after becoming closely associated with the proposed Stratos data center project in Box Elder County.

Two Box Elder County commissioners, Lee Perry and Boyd Bingham, also lost their Republican primaries after voting to advance the project.

The proposed Stratos project has faced opposition over potential water and environmental impacts. The commissioners voted to authorize the state’s Military Installation Development Authority to establish the Stratos project area, granting the state power to override county zoning and land-use regulations.

Hollist made opposition to Stratos a central pillar of her campaign.

Perry was unusually direct about the political fallout following his loss.

“Do I think that the data center vote cost me the election? Yes, I do,” Perry told The Salt Lake Tribune after conceding his primary race.

Adams, who chairs the MIDA board, also faced mounting pressure over the project and moved late in the campaign to soften his stance on its proposed footprint.

Together, the Utah cases provide the review’s clearest evidence that data center politics can extend far beyond local zoning disputes and threaten senior state officials.

Related: Praetorian’s Virginia Megasite Could Accommodate Data Center Development

Missouri produced another cluster—and one of the most revealing cases. In Festus, four incumbent city council members—Jimmy Collier, Brian Wehner, Bobby Venz, and Jim Tinnin—lost their elections amid a fierce battle over a proposed $6 billion data center project.

Collier, Venz, and Tinnin voted in favor of the development agreement, while Wehner voted against it. All four lost.

This makes it difficult to frame the backlash as simple punishment for a “yes” vote. The controversy also centered on the city’s development process, transparency, and the manner in which the agreement was approved.

In Independence, council members Jared Fears and Bridget McCandless also lost their elections amid a dispute over a proposed data center and associated tax incentives. Fears later acknowledged the project factored into his defeat.

The Missouri cases demonstrate that political exposure can stem from the project itself, the approval process, or simply an official’s association with a broader development fight.

The pattern extends well beyond Missouri and Utah. In North Carolina, David Batts defeated four-term Edgecombe County Commissioner Donald Boswell after campaigning explicitly against the proposed Kingsboro data center. Batts had previously warned commissioners that residents would “primary” them over the project. His campaign signs read: “Vote no for data centers.”

In Virginia, Democrat John McAuliff branded Republican Delegate Geary Higgins “Data Center Geary” and leveraged data centers as a major attack line in his successful challenge.

Maryland presented a different variation of the backlash: officials who opposed a pause on new data center development still became targets for candidates arguing the county had mishandled growth. Calvert County commissioners Todd Ireland, Mark Cox Sr., and Earl “Buddy” Hance lost their Republican primaries after challengers criticized their opposition to a proposed pause on new data center development and the county’s broader management of growth. Frederick County Council President Brad Young also lost his Democratic primary after the council approved a zoning expansion that opened more than 2,600 acres to data center development.

Oregon state Senator Janeen Sollman lost her Democratic primary in a race that featured opposition to data center development in the Hillsboro area. Labor and education issues also played significant roles, making the data center connection less direct. Georgia Public Service Commissioner Tim Echols lost his reelection bid in a contest dominated by electricity rates and Georgia Power. While data centers factored into the broader debate over who should fund new generation and grid infrastructure, the evidence linking the issue directly to his defeat remains weaker.

These cases differ in evidentiary strength, but collectively they reveal a shift in political incentives: candidates now treat data centers as shorthand for wider grievances regarding power costs, tax policy, land use, and local control.

Mike Casey, president of Tigercomm, an energy technology public relations firm, said the industry has consistently underestimated the necessity of campaigning for projects.

“Most tech companies weren’t built to campaign for what they need to build,” Casey told Data Center Knowledge. “But winning a community’s permission to build in its midst is the opposite. It’s one county at a time, it’s heavily retail, and it’s person-driven.”

Casey noted that developers accustomed to rapid permitting are now navigating a “high-friction permitting environment” and scrambling to build community-engagement operations after encountering widespread resistance.

“Too many developers are trying to bulldoze projects through instead of skillfully campaigning for them,” he said.

The result, he added, is “a community relations problem metastasizing into a widespread political crisis.”

The Data Center Coalition’s response has largely focused on clarifying the industry’s economic contributions. Dan Diorio, the coalition’s executive vice president of state policy and government affairs, stated in an email to Data Center Knowledge that the industry recognizes the need for “continued efforts to better explain the role of data centers to the public through community engagement and stakeholder education.”

Diorio pointed to a PwC report estimating that the industry supported 5.5 million jobs, contributed $927 billion to U.S. GDP, and generated $204 billion in federal, state, and local taxes in 2024. He also emphasized that data centers are committed to covering the energy they consume and acting as responsible partners in the communities where they operate.

Jeff Thramann, founder and CEO of LT350, a company developing distributed data center infrastructure, argues that communication alone cannot resolve the backlash.

“I think fundamental change is required because the impact of huge data centers on a single community is real,” Thramann said.

His solution is to distribute computing capacity across smaller facilities rather than concentrating hundreds of megawatts at individual sites.

“If 300 MWs of capacity is needed in one location, the impact on that geography compared to spreading 1 MW across 300 grid circuits is incredibly different,” he said.

Thramann observed that the public draws a sharp distinction between massive centralized campuses and smaller computing facilities that exert little visible pressure on a community.

“A data center can range from an enormous, centralized facility to a closet in an office,” he said. “The public does not care about the closet because it changes nothing in a community, but they do care about the giant facility that changes an entire community.”

Data centers are also appearing in paid campaign advertising well beyond the local jurisdictions that approve them. Campaigns in Ohio, Pennsylvania, Wisconsin, Texas, and Michigan have attacked candidates or elected officials over data center development, electricity costs, tax incentives, or ties to developers.

Those races fall outside the 18 completed losses documented in this review. They demonstrate that campaigns increasingly view data centers as a decisive political wedge issue.