Voters across the country delivered a sweeping rebuke to data center development in Tuesday’s midterm elections, approving restrictions on the massive facilities that have become central to the AI boom and electing officials who have promised to rein in their construction.
The results pose a new challenge to the building boom that has helped power U.S. economic growth and stock market gains, bringing in trillions of dollars in investments, in addition to a large amount of backlash.
Some 41 measures related to restricting or permitting data centers across ten states were up for consideration in the midterms, with Ohio alone accounting for half, per Ballotpedia, an election-focused nonprofit. In all but one of those measures, voters opted to restrict (or to not approve) data center construction, though four measures remained officially undecided as of Wednesday afternoon. Data-center skepticism crossed party lines, with restrictions passing in traditionally red states like Kansas and Florida, as well as liberal strongholds like California and New Jersey.
Many of the AI-focused measures would temporarily restrict the construction of new, large data centers, while others would restrict their location to industrial areas and place limitations on their environmental impacts. Five of the petitions or referendums approved Tuesday would allow data center construction as long as the facilities used fewer than 25 megawatts. Data centers built for AI purposes often require more than 100 megawatts.
Across the country, voters also elected candidates — such as incoming Texas Governor Gina Hinojosa — who promised to rein in data centers’ construction or put strict guardrails around their use. NBC’s exit poll found more than half of voters said AI and data centers were “a major factor” in their vote for Congress. Approximately 76% of respondents said that they would oppose the construction of a data center in their community.
In Maine, for example, voters in the southern town of Eliot agreed to ban any movement on data centers for 180 days, while Perry, Ohio residents passed an ordinance that will cap the amount of noise data centers can emit.
Brendan Steinhauser, CEO of The Alliance for Secure AI and a former Republican strategist, said that the election results delivered a resounding message.
“Data centers were on the ballot and they got trounced,” he said. “The moratoria were very popular, data centers were very unpopular, and it was clear across different states and different communities.”
Experts and pollsters attribute the backlash to a variety of root causes: frustration with growing income inequality and the billions of dollars tech tycoons stand to reap from AI; worries about catastrophic risks from AI; and more prosaic concerns about noise and health impacts.
Justin Smith, the owner of a residential construction company in Andover, Kansas, who has advocated for stronger guardrails on data center construction in county deliberations, said Tuesday’s results showed a clear desire for responsible AI development.
“People just don’t want these things directly next to their homes,” he told NBC News, noting that he viewed AI as a national security issue and that data centers should be built — just not right next to people’s homes. “The whole crux of the issue is people just want a proper amount of buffer.”
His county’s board of commissioners recently moved forward with regulations that Smith said did not adequately protect residents. In nearby El Dorado, Kansas, voters overwhelmingly passed a moratorium on new high-impact data centers and industrial-scale batteries on Tuesday.
But a trillion-dollar question now looms: If the U.S. builds fewer data centers, will the economy suffer?
For much of President Donald Trump’s second term, data center investments have been at the heart of the U.S. economic growth story — and have driven large stock gains. By some measures, approximately half of all economic growth has come from AI-related investments into data centers, while as much as 75% of the run-up in stocks in the first half of 2026 was attributable to tech-stock gains.