The United States is in the middle of a major data-center build-out, as technology companies seek more capacity for artificial intelligence. Corporate groups and tech billionaires backed Republican candidates friendly to those projects. In the 2026 midterms, Democrats won control of the House and Senate, and voters rejected data centers in 40 of 41 local ballot measures across 10 states. The lone exception was Harwood, North Dakota.
That combination puts political risk closer to the center of AI infrastructure planning. It does not establish a nationwide halt. It does show that projects face a less predictable path through local approvals and federal oversight.
Voters rejected 40 of 41 data center ballot measures
According to CNBC’s analysis, voters sided against data centers in every measure except one in Harwood. The tally describes local ballot outcomes, not a referendum on every proposed facility. Still, the consistency gives project opponents a clear electoral result to cite when seeking restrictions or additional review.
Permitting already depends on decisions by state and local governments. The R Street Institute expects data-center moratoriums to become more onerous as governments add requirements before issuing permits. That forecast is not proof that every proposal will be delayed or rejected. It points to a concrete risk for developers: more conditions can make the timing and cost of a project harder to predict.
Opposition also surfaced in statewide races. Anti-data-center messaging featured in Democratic wins in Texas and Ohio, according to CNBC. Texas is a major center of planned construction, and Gina Hinojosa campaigned for a pause on new projects until tougher rules are adopted. In Ohio, plans for a campus involving OpenAI, SoftBank and SB Energy could face resistance from the governor’s office.
The ballot measures do not prove that all planned construction is in jeopardy. They do give elected officials evidence that opposition can win votes, making local approval a political question as well as a land-use process.
A Democratic Congress can intensify scrutiny
Democrats’ control of the House and Senate gives them power to challenge President Donald Trump’s administration through investigations, spending negotiations and legislation, and to block presidential nominees, according to VentureBeat’s election analysis.
A newly Democratic Congress is expected to examine technology, cyber and surveillance policy, including data centers, Nextgov/FCW reported. Hearings can put company decisions under public scrutiny. Spending negotiations and confirmation fights can also shape federal priorities and the officials responsible for implementing them.
The more likely near-term effect is oversight than a comprehensive federal AI statute. The R Street Institute expects investigations and subpoenas to be more likely than broad AI legislation. That leaves infrastructure developers facing two distinct pressures: greater attention in Washington and variable permitting requirements across states and localities.
The election also dealt a blow to tech billionaires and corporate groups that supported Republicans friendly to data-center projects. CNBC reported that voters voiced frustration with the AI data-center build-out. Political spending did not overcome that concern. It may make the relationship between project sponsors and elected officials more visible to voters and regulators.
Local resistance changes the risk calculation for AI infrastructure
The election’s effect on edge-tech industries is indirect but material. Data centers supply the computing capacity behind AI services, and their construction depends on site approvals, permits and infrastructure. More uncertainty at those stages can affect developers, operators and the companies relying on new capacity.
Investors reacted to the results, although market moves alone do not establish why a stock rose or fell. Business Insider reported that CoreWeave shares fell 9% and Nebius Group fell more than 6% in Wednesday trading. That is evidence of market concern, not proof that the election changed either company’s project economics.
In Texas, Hinojosa’s proposed pause offers a route from campaign opposition to possible policy. In Ohio, the proposed OpenAI, SoftBank and SB Energy campus could encounter resistance from the governor’s office. Neither example establishes that a project will be stopped. Both show how a change in political leadership can put a project’s schedule and approval process under new scrutiny.
For developers, the practical response is to treat political and permitting risk as part of site selection, rather than as a communications problem after a site is chosen. That is analysis, not a reported industry-wide spending shift. The election results give companies reason to reassess where approvals are most uncertain and what conditions local governments may demand.
Political insulation may favor developers with deeper pockets
Over the next 24 months, I expect the largest infrastructure developers to place more weight on predictable permitting and local support when choosing sites. They are unlikely to abandon expansion solely because of the ballot results: demand for compute remains a central driver of investment. The likelier adjustment is to pursue projects where the route to approval looks clearer, while spending more time and resources on permitting and local engagement in contested places.
The mechanism is straightforward. A local rejection can prompt officials elsewhere to demand more review, stronger conditions or a pause before approving new construction. That uncertainty makes early site assessment and sustained engagement more valuable. It can also favor firms that can afford to evaluate multiple locations and absorb delays while a project moves through approvals.
That advantage could reinforce the position of established companies. A smaller developer with fewer sites and less capital may have less room to wait out a prolonged process or redirect a project. Larger firms can spread that risk across more potential locations. This is a plausible consequence of uneven permitting costs, not a demonstrated outcome of the 2026 votes.
The same pressure may change where projects are built. If contested areas impose longer reviews or additional conditions, companies can compare those costs with locations offering a more predictable approval process. That could redistribute investment among regions without reducing total demand for data-center capacity. The ballot results themselves do not show that a relocation has happened.
The forecast is testable. By the end of the next 24 months, look for evidence that major developers are changing site choices, timelines or public permitting plans in response to local opposition. If expansion continues at the same pace in contested areas without more attention to approval risk, the political-insulation thesis will have little support. If projects shift toward places with clearer permitting paths, local resistance will have changed the map of the build-out even without a national law.
Permitting becomes part of the capital allocation decision
A data center’s investment case depends on more than the cost of equipment and the expected demand for computing. The time and conditions required to secure approvals also matter. Added reviews or mandates can delay a project, while uncertainty can make competing locations more attractive.
That pressure may raise the relative advantage of companies able to fund legal, permitting and government-relations work. It could also make new projects harder for smaller developers to finance. The result would be a more concentrated field, though the election results alone cannot establish that this is already happening.
The larger shift is in how companies weigh project risk. The industry still has strong reasons to build compute capacity, but a low-cost site with uncertain political support may be less attractive than a more expensive site with a clearer approval process.
Data center ballot measures point to a more contested build-out
The 2026 midterms did not end the U.S. AI infrastructure boom. They made the political conditions around that boom harder to ignore. A Democratic Congress can increase federal scrutiny, while local voters and state officials can add friction to individual projects.
The next phase of the build-out will depend on more than access to capital, power and equipment. The 40-to-1 ballot pattern makes local approval a strategic constraint for data-center developers, and a potential advantage for companies able to plan around it.
