How the US midterm elections reflected resistance to data centers | News 2026 US midterm elections

In September, a poll found that 70% of Democrats and 57% of Republicans supported limiting the construction of new data centers in the United States.

But that opposition was put to the test in Tuesday’s midterm elections, with analysts looking at key elections to see if public reaction would translate into votes.

This is indeed the case. On Tuesday, at midterms, 41 measures related to data centers were voted on, spread across the country.

Voters are on track to reject data center development in all but one contest, according to Data Centers on the Ballot, an election monitoring project run by Redbud Consulting.

Two of the races are still counted but are expected to result in majorities to reject or limit data centers.

Only Harwood, North Dakota, a small town of fewer than 800 people, voted against such limits. The rejected measure would have required independent reviews of large industrial projects such as data centers before they could be built in the region.

Experts say Tuesday’s midterm elections indicate that the debate over data centers has become a bipartisan issue, closely tied to concerns about the cost of living and fears about job stability.

A construction boom

In April, the Pew Research Center estimated that there were already 3,000 data centers in the country, with more than 1,500 more under construction.

The construction boom is largely due to the growth of the artificial intelligence (AI) sector. This industry has fueled concerns about a changing economic landscape, experts say.

“The uncontrolled acceleration of AI development raises concerns at all levels,” Bhaskar Chakravorti, dean of global business at Tufts University, told Al Jazeera.

“People are worried about their jobs and what that might mean from a broader economic perspective. But closer to home, people could see the physical manifestations of AI in data centers being built in their backyards, or they were hearing about it from others.”

Not only are voters concerned that AI could replace human labor, but they also fear the downstream costs of data centers, according to Chakravorti.

Data centers consume large amounts of water and energy, which critics say can drive up electricity costs for nearby communities.

“Whether rising energy costs are due to broader factors, such as the war in Iran or tariffs, AI’s competition for space on the power grid is certainly part of its unpopularity,” Chakravorti said.

Voters, he added, were seeing the results “in their energy bills.”

A 2026 working paper from the Federal Reserve Bank of Dallas estimates that since last year, wholesale electricity prices have increased 6.9% thanks to data centers.

James Mercer, founder of Metro Colo Advisory, an independent data center consulting firm, told Al Jazeera that these anticipated costs had mobilized voters in many areas.

“What they’re pushing away is a closer building that needs as much power as a small town, when they suspect their own electric bill will be enough to cover it,” Mercer said.

“This is why local resistance has grown so quickly: counties putting new projects on hold, zoning fights, and data centers becoming a ballot issue.”

A topic of discussion during election period

This popular resistance has led several candidates to embark on high-profile races to place AI and data center development at the heart of their campaigns.

Among them is Democrat Sherrod Brown, who flipped a key Senate seat in Ohio, the state that accounted for the largest share of data center ballot initiatives this year.

Some 21 votes were taken regarding the state’s data centers, out of a total of 47 referendums and measures that took place across the country in 2026.

Brown, a former senator, has made incumbent Republican Jon Husted’s support for data center development a key point of attack.

“One guy led the charge on this. That guy is responsible, and that’s Jon Husted,” Brown told NBC News earlier this year.

In Texas, Democrat Gina Hinojosa called for a moratorium on approving new data centers until stronger protections are put in place for local communities.

Although his stance worried tech investors eager to expand in the Lone Star State, Hinojosa ultimately defeated incumbent Greg Abbott in the gubernatorial race.

“While Abbott currently has a moratorium on data centers in Texas, the prevailing view is that this will end once he is re-elected,” Dan Ives, managing director of Yorkville Ives Bank, said in an analyst note.

“Hinojosa is a known anti-data center politician and negative toward AI development, making concerns about the future of data center development in Texas a real issue for tech investors.”

Another Democratic candidate-elect, Sen.-elect James Talarico, also flipped a Republican-held seat by pledging to strengthen data center regulations.

Texas is at the center of growing concerns about water demand from AI infrastructure.

The city of Corpus Christi, for example, faced severe drought conditions, with its reservoir falling to just 8% of its full capacity earlier this year.

The city has been described as being at risk of running out of drinking water. Talarico proposed adopting closed-loop water systems to reduce data center consumption. He also advocated for data center operators to fund local infrastructure improvements.

He won Corpus Christi by a margin of 12 percentage points.

In another key Senate race, Democrat Abdul El-Sayed called for the creation of a federal agency to oversee the AI ​​industry. He defeated his Republican opponent with more than 52 percent of the vote.

Impending pressures

With Democrats poised to control both houses of Congress, the AI ​​industry is likely to face increased scrutiny from lawmakers.

“(We are) more likely to get real regulation and real oversight,” Alex Turner, a former researcher who resigned from Google DeepMind, told Al Jazeera.

“I look forward to advising any member of Congress who wants to take the risks of AI seriously, regardless of party affiliation.”

Congressional leaders increasingly view AI as a threat to national security and public safety, after a series of alarming security failures over the past six months.

In one incident, OpenAI autonomous agents escaped from a controlled testing environment and infiltrated tech startup Hugging Face. In another case, AI agents accessed bank accounts without authorization, according to a report from the publication Axios.

In late October, Anthropic’s Claude AI model also accessed voter data in Georgia.

Once the 120th Congress is sworn in on January 3, Democrats are expected to use the legislature’s oversight committees to conduct investigations into AI-related malfeasance.

On Tuesday evening, as it became clear that Democrats would gain a majority in Congress, Sen. Ed Markey of Massachusetts promised that new investigations were underway.

“They are going to have to plant new forests for all the subpoenas we issue,” the Democrat wrote on the social media platform X.

Another congresswoman, Yassamin Ansari of Arizona, posted on social media her support for more restrictions on AI growth.

“From day one, a priority on the Democratic agenda must be a pause on the advanced development of AI, the regulation of this out-of-control industry, and the banning of artificial superintelligence,” Ansari said.

Tech companies themselves responded to the midterm results with proposals aimed at allaying voters’ concerns.

After the election, OpenAI released a framework titled Setting a New Standard for Responsible AI Infrastructure. In it, the company commits to “prioritizing closed-loop or low-water cooling” and “paying our own energy and infrastructure costs.”

Days before Tuesday’s competition, OpenAI was also looking to rebrand its data centers as “computing campuses.”

Political change also affects Wall Street. Investors increasingly see tighter AI oversight on the horizon, and some of the sector’s biggest growth stocks have stumbled.

On Thursday, for example, chipmaker Nvidia lost $450 billion in market value in a single trading day.

Despite Thursday’s losses, Nvidia rebounded on Friday, gaining 4% in early trading, although it remained down 6.8% over the previous five days.

Data center operators such as CoreWeave fell on Thursday before climbing 6.6% on Friday. The company remained down 16 percent over the previous five days.