Voters Reject Data Centers While Musk's Half-Billion Bet Fizzles

Voters approved nearly every local ballot measure restricting or freezing data center construction in this week's midterm elections, handing hyperscalers a political problem that money did not solve. Elon Musk, one of the largest data center builders in the country, donated roughly half a billion dollars to Republican candidates and appears to have bought very little with it. His companies now sit headquartered in a state that, as TechCrunch's Sean O'Kane put it, has turned "about as blue as it could possibly turn."

The scope of the rebuff was broad rather than universal. Ballot measures on data centers appeared in a number of localities, not everywhere, but the pattern was consistent. Where restrictions or moratoriums were on the ballot, almost all of them passed. In Texas, governor-elect Gina Hinojosa campaigned explicitly on reining in data centers and defeated incumbent Greg Abbott; she has said the issue is a priority once she takes office.

What changed for the builders?

The consensus view in the tech world, O'Kane noted, is that Democratic power means "a lot of roadblocks ahead for data centers," especially in states like Texas. Anthony Ha, speaking on TechCrunch's Equity podcast, argued the sentiment runs deeper than one party's fortunes. Anti-AI and anti-data center feeling is "bipartisan in a lot of ways," he said, and Democrats proved more effective at harnessing it. He expects the next two years, and potentially beyond, to be more difficult for hyperscalers and data center companies, with the issue likely to persist into 2028.

Kirsten Korosec was careful about causation. Data centers did not produce the "blue wave" on their own. They did, however, become "an increasingly grassroots issue, in part because of how hyperscalers and companies have gone about" getting projects approved, she said. The industry's habits supplied the fuel:

• Secretive negotiations, often requiring NDAs from elected and appointed officials
• Withheld information about project scale in environmentally sensitive areas
• Buildout in the Southwest, where water scarcity sharpens every objection

Republicans, independents and Democrats alike read those practices as wrong, in Korosec's account. It was a motivating issue, not a singular one.

OpenAI discovers policy flexibility

The most concrete corporate response came from OpenAI, which released what it calls a compute campus policy blueprint. The company says it will stop requiring NDAs and will share its plans, operating with less secrecy. Korosec drew the lesson plainly: when regulators provide certainty, or even signal that direction, corporations respond. Absent that pressure, they do not volunteer.

Ha flagged the naming as strategy. Data centers have become "such a toxic phrase" that the industry is attempting a rebrand: not data centers anymore, but compute campuses. Whether the new vocabulary takes is an open question.

O'Kane's verdict was blunt. "This is ridiculous. Come on. It's not going to change anything," he said of OpenAI's effort. For a company that claims to have abandoned side quests, he argued, it is an odd use of time that will not move the needle. Korosec offered a dry suggestion in the same spirit: if the goal is making threatening things seem cute, OpenAI should create a friendly compute campus avatar and present that instead of a power- and water-consuming facility. Ha proposed a simpler test — if these are such lovely campuses, the companies should build their own offices next door.

A second punch from Wall Street

The election may be only half of the problem. Ha pointed to a shift in how Wall Street treats hyperscalers and AI companies, with investors no longer eager to fund enormous spending on new data centers and compute. The financial incentive to build is weakening at the same moment the regulatory path is narrowing.

The bind is structural. Companies that want to reach the scale they envision will still need to build data centers somewhere. But the likely landscape ahead offers harder permits on one side and unrewarding capital markets on the other, and OpenAI's sudden policy conversion suggests the industry itself has started pricing that in.