The 2026 US midterm elections transformed construction of data centers into a significant local political issue, potentially making the rollout of AI infrastructure less predictable, slower and more expensive. An NBC exit poll cited by multiple outlets found that nearly 80 percent of respondents opposed having a data center built in their neighborhood, and almost 60 percent said the issue was an important or decisive factor in their vote. Those figures suggest these projects are no longer only technological or investment decisions but also matters of energy, environment and community consent.
Where did resistance succeed?
Reporting by TechCrunch and post‑election analyses show candidates critical of rapid data center expansion performed well in several states, including Ohio, Texas, Georgia, Michigan, Maine, Nevada and South Carolina. In Ohio, multiple local communities voted for restrictions even though the state had previously been a target for large tech investments.
What regulatory steps might follow?
A Brookings Institution analysis warns that the post‑election political landscape could produce stricter rules at the federal, state and local levels. Possible measures discussed include:
- imposing construction moratoria;
- limiting automatic, preapproved development rights;
- rolling back tax incentives;
- increasing pressure around nondisclosure agreements;
- expanding community consultation requirements.
For example, in Prince William County, Virginia, officials plan to end the automatic approval that previously took effect after a 90‑day waiting period, giving local decision‑makers greater influence over what projects may proceed.
Who pays for grid upgrades?
At the center of the controversy is electricity. AI data centers place large and often rapidly growing demands on the grid, requiring new network investments, generation capacity and backup solutions. Brookings highlights that a likely regulatory question is whether technology companies should be required to pay for the additional energy demand and network upgrades they cause, rather than having those costs partially appear on local consumers’ bills. Today those rules are not uniformly mandatory and in many places remain voluntary or apply mainly to future projects.
Taxes, permits and legal challenges
Regulatory risk extends beyond power costs. Brookings notes Senator Ron Wyden (D) has been working on an excise tax proposal that would target gross revenues of large data centers of at least 25 megawatts, with proceeds intended to support workers displaced by digitalization. Some states may also reconsider sales‑tax exemptions for data center equipment, sharpening debates between economic development incentives and tax fairness.
Legal challenges are already appearing. Both TechCrunch and Brookings highlight the case of xAI’s Memphis data center, which uses gas turbines for power; the NAACP has filed a lawsuit alleging violations of the Clean Air Act and health risks to nearby communities. Such actions illustrate how data centers are increasingly framed as industrial facilities with visible environmental and public‑health impacts, not merely invisible digital infrastructure.
Impact on investment timelines and costs
Analysts say the most immediate consequence for tech companies will be reduced predictability. If states or municipalities impose moratoria, tighten zoning, end automatic permitting, levy new taxes, or require developers to bear network upgrade costs, building large AI data centers could become slower and costlier. The Verge’s recent roundup notes that major tech firms continue to spend heavily on AI infrastructure, while investors increasingly question when the record‑sized data center investments will pay off.
This does not necessarily halt the US AI investment wave, but it does reshuffle the risks: alongside chips, power and financing, a fourth factor now matters — securing the consent and approval of local communities and regulators.
Lessons for observers abroad
From the perspective of Europe and Hungary, the elections underline a clear lesson: AI infrastructure cannot be separated from energy policy, water management, grid investment and local social acceptance. Building a data center is simultaneously a technological project and a long‑term industrial infrastructure undertaking, and developers must win community buy‑in in addition to securing chips, power and financing.
(Reporting synthesized from TechCrunch, Brookings and The Verge; election polling cited: NBC exit poll.)
