Midterm backlash against AI data centers sets stage for new siting limits

Voters made data centers a ballot issue

Midterm elections on November 3 delivered a broad rebuke to the AI infrastructure buildout, according to TechCrunch. Local measures restricting data center construction passed in multiple states, and candidates who campaigned against rapid buildout won races from Texas to South Carolina. Nationally, Democrats took full control of Congress while Republicans kept the White House, a divided arrangement that makes sweeping federal action unlikely but hands new leverage to critics of the buildout.

An NBC exit poll cited by TechCrunch found that more than half of voters expect AI to do more harm than good, close to 80 percent oppose data centers in their own neighborhoods, and nearly 60 percent described data centers as a crucial or major factor in their vote. The argmin newsletter, in a post that reached the front page of Hacker News, counted close to fifty ballot measures dealing specifically with data center construction and reported that all appeared likely to pass in the restrictive direction. The same author pointed to earlier polling in which respondents said they would rather live next to a coal plant than a data center.

The local results

Ohio was the clearest flashpoint. TechCrunch reports that the county where Meta had planned a 700,000-square-foot facility approved a ban on data centers, and that communities across the state voted down buildout measures, even as OpenAI has explored a $500 billion data center in Ohio backed by more than $100 billion in financing from Nvidia. Morgan Stanley analysts expect the state's governor-elect, Amy Acton, to pose a major risk to those plans.

In Texas, governor-elect Gina Hinojosa made a statewide moratorium a centerpiece of her campaign; the argmin post noted she becomes the first Democrat to win the Texas governorship in over three decades. Governor Greg Abbott has also called for some limits. TechCrunch lists additional wins for data center opponents in Maine, Michigan, Nevada, Georgia and South Carolina. Aniketh Koneru, a tech and policy researcher, told TechCrunch that "money does not buy outcomes," referring to the millions spent by pro-AI super PACs this cycle.

What changes now

The near-term consequences analysts expect are mostly procedural: more construction moratoriums, and requirements that data centers pay for their own grid upgrades instead of passing those costs to nearby residents. Goldman Sachs predicts construction delays in Georgia and Texas as a result of Democratic gains.

In Congress, Democratic control improves the prospects of a bill from Sen. Bernie Sanders and Rep. Alexandria Ocasio-Cortez that would impose a national moratorium on new data centers until Congress adds guardrails on electricity costs, environmental harm and community approval, though such a bill could face a veto. A narrower proposal, the GRID Savings Act from New Mexico Sen. Martin Heinrich, would stop grid infrastructure costs from being passed on to communities during large AI buildouts.

Maia Woluchem of the research group Data & Society told TechCrunch she expects movement on energy regulation, land rights and environmental policy regardless of federal action, and noted that congressional Democrats may examine the administration's rollbacks of energy protections, as the New York Times reported they might. Nadine Smith of Color of Change said she anticipates executive orders aimed at bypassing Congress, along with heavy lobbying, and pointed out that even among Democrats a blanket freeze is contested over national security concerns.

Why resentment built

The backlash is driven by process as much as by the buildings themselves. TechCrunch points to outdated zoning rules, citing Amazon's data center in Gilroy, California, which was built largely without residents' knowledge under a 45-year-old ordinance, and to nondisclosure agreements that bar local officials from discussing proposed projects. Billions in tax breaks and promised jobs have not offset fears about pollution from gas turbines, like those powering xAI's Memphis facility, or about rising utility bills. The argmin author's argument goes further: data centers have become a legible stand-in for the AI industry itself, so whatever usage metrics suggest, the election measured something closer to dislike. As Smith put it, families should not be "subsidizing the richest corporations in history through their utility bills."

Why it matters

For cloud providers and AI labs, the midterms convert community opposition from a public relations problem into a siting constraint. County bans, state moratoriums and grid-cost rules shrink the map of viable locations and stretch timelines in precisely the markets, including Ohio, Georgia and Texas, where capacity was planned. If operators must fund their own substations, meet clean air standards and disclose their power use, the economics of hyperscale expansion change. With a Democratic Congress scrutinizing energy policy while state-level movements keep growing, the cheapest assumption behind any capacity roadmap, that power and permits will simply be there, no longer holds.