AI Data Centers Lost 40 of 41 Ballot Votes: The Backlash Is Heading for State Capitols
On Tuesday, American voters did something the tech industry did not see coming: they voted against the buildings that make artificial intelligence possible, and they did it almost everywhere at once.
Forty-one local ballot measures on data centers were on the ballot across ten states on 3 November. When the counting finished, 40 of them had gone against the industry, according to Ballotpedia's tracker. Bans, moratoriums, zoning blocks, new public-vote requirements — voters approved nearly all of them. The single exception was Harwood, North Dakota, where voters rejected extra review requirements for large industrial projects by 325 votes to 105. Counting six votes decided earlier in the year, 46 of the 47 data-center measures decided in 2026 have favoured limits.
Ohio was the centre of the revolt. The state had 21 measures on the ballot, more than any other, and every one of them went against data-center development. In Wilmington, four rezoning proposals tied to data-center projects were rejected by roughly 80% to 20%. In Ashville, more than 83% of voters backed a challenge to the village's negotiated terms with an EdgeConneX facility. The pattern repeated around Columbus — Pataskala, Grove City, Sunbury, Granville — and across the country: in Alhambra, California, 82% voted to ban data centers from the city outright; in El Dorado, Kansas, voters approved a ten-year ban by 71% to 29%; in Joliet, Illinois, 79% said no in an advisory vote; in Morgan County, 88% backed restrictions.
The striking part is that this was not a partisan spasm. In a CNBC voter poll, 56% of voters said AI and data centers were a crucial or major factor in their vote. An August poll by Embold Research found 75% of registered voters opposed a new data center near their home. The objections voters actually cite are concrete, not ideological: electricity prices, water use, and tax breaks for facilities that employ few people once built. Kansas defined a "high-impact" facility as one needing more than 10 megawatts of peak electricity or more than 100,000 gallons of water a day — thresholds a single large AI training site can cross many times over.
The revolt also won races, not just referendums. In Texas, the state with the most planned AI data-center capacity, Democrat Gina Hinojosa beat Governor Greg Abbott after promising a moratorium on new data-center construction. In Georgia, Keisha Lance Bottoms won the governor's race after calling for a pause. In Ohio, Amy Acton defeated Vivek Ramaswamy. The irony is sharpest in Ohio: OpenAI announced in August that it had secured 8 gigawatts of computing capacity at the PORTS-Pike Technology Campus, described as the largest planned AI data-center campus in the world — in the same state where voters most decisively rejected the industry this week.
What happens next is the interesting question. More votes are coming: on 1 December, Clarksville, Arkansas votes on bonds of up to $50 billion and $5 billion for data-center projects, and a ban vote follows in North Dakota on 15 December. In Ohio, campaigners are working on a state constitutional amendment that would ban new data-center construction, which could reach the ballot next year. The fight is moving from town halls to state capitols, and the industry's problem is that its usual arguments — jobs, investment, tax revenue — are precisely the ones voters have just spent a year hearing and rejecting.
I have spent this year writing about what AI can do. This week voters answered a different question: what AI costs the people who live next to it. The industry spent billions building the infrastructure for the next decade of computing. It forgot to ask the neighbours. That oversight now has a number attached to it: 40 to 1.