Sigma

US voters backed restrictions on AI data centers

2 min read

AI-generated image: US voters backed restrictions on AI data centers AI

Local referendums and gubernatorial elections showed that opposition to data centers has become a political force in the US. OpenAI is already changing its construction rules, and the industry is bracing for slower expansion of AI infrastructure.

In the midterm elections on November 3, Americans voted in more than 40 local referendums concerning data centers in around 10 states. In nearly all of them, they supported restrictions, moratoriums, or bans. According to a Newsweek tally, this was 40 out of 41 votes. Some media outlets report slightly different figures, due to different counting methods.

The most questions, around 21, were asked in Ohio. In Granville, the ban was supported by 82.45 percent of voters, in Alhambra, California, by 82 percent, and in Edgewater, Florida, by 77 percent. Opponents of the investments, as reported by PAP from Kentucky, point to higher energy and water bills, environmental strain, and few new jobs. According to a Gallup poll from April, 70 percent of Americans do not want a data center in their area.

Data centers were also a campaign issue for candidates. In Texas, Democrat Gina Hinojosa became governor, defeating Greg Abbott by 53 to 46 percent. She announced that she would halt approvals for new data centers until legislation protecting residents is passed. In Michigan's 7th district, Will Lawrence won a seat and wants a federal moratorium. Elon Musk's America PAC spent $340,000 against him, as reported by Wired. Musk, whose companies are building data centers including in the Memphis metropolitan area, wrote on X after the election that data centers in space now look appealing.

On November 5, OpenAI published the document “Compute Campus Policy Blueprint.” In it, the company calls data centers “compute campuses” and commits not to require officials to sign non-disclosure agreements (NDAs). It also says it will cover full energy costs and disclose water and electricity consumption. OpenAI plans to spend $600 billion on computing power by 2030.

According to Goldman Sachs, electricity prices in the US rose by 6.9 percent in 2025, and data centers account for 40 percent of the increase in power demand. Slower infrastructure construction and more expensive energy may affect the prices of cloud services and AI models used by companies, including those outside the US, if they rely on services from अमेरिकी providers.

Key takeaways

The summary could not be loaded. Refresh the page and try again.