Data centers had a bad election night. In dozens of local ballot measures this week, voters backed restrictions and moratoriums on new data center projects, and almost all of those measures passed. TechCrunch AI reports that the backlash went beyond the ballot itself. Voters also elected politicians who campaigned on reining in data centers, including Texas governor-elect Gina Hinojosa, who beat Greg Abbott.
For an industry that has spent two years assuming it could build compute wherever it found cheap power and land, that assumption no longer holds.
The Musk problem
On TechCrunch’s Equity podcast, Sean O’Kane pointed to the irony at the center of all this. Elon Musk, one of the biggest data center builders and the largest donor to the Republican Party, “seems to have gotten basically nothing out of donating half a billion dollars.” Worse for him, his companies are now headquartered “in a state that has, remarkably, turned about as blue as it could possibly turn.”
This matters because Texas was supposed to be the safe harbor. Light regulation, fast permits and a friendly governor made it the obvious place for big AI infrastructure bets. Hinojosa has openly called data center restrictions a campaign promise, so that advantage now looks shaky.
Why voters got angry
Kirsten Korosec doesn’t think data centers caused the “blue wave.” She does think the industry created its own opposition. According to TechCrunch AI, she pointed to how hyperscalers got projects approved:
- Secrecy: NDAs required of elected and appointed officials
- Missing information: the full scale of projects often wasn’t disclosed
- Environmental blind spots: big builds in sensitive areas, including the water-scarce Southwest
Korosec said Republicans, independents and Democrats “all saw this as wrong.” Anthony Ha made a similar point. Anti-data center sentiment crosses party lines, but Democrats did a better job of turning it into votes. That’s why this is more than a one-cycle story.
OpenAI blinks, and rebrands
The most telling reaction came quickly. OpenAI released what it calls a “compute campus policy blueprint,” which says it won’t require NDAs anymore and will share its plans more openly.
Korosec’s take is sharp. For years the industry insisted it couldn’t operate any other way. Now, “we’re already seeing companies like OpenAI all of a sudden go, ‘Actually, we can change our policies.'” Her point is that companies respond when regulation becomes a real possibility, and often not before.
The name change got less applause. Ha noted that “compute campus” is partly an attempt to rebrand a phrase that has become toxic. O’Kane didn’t hold back: “It’s not going to change anything.”
I’m with O’Kane on this one. Calling it a campus won’t change how a local town hall reacts to a facility that draws huge amounts of water and power. Dropping the NDAs might.
What the next 1-3 years look like
Here’s where I expect this to go.
- Local permitting becomes the new bottleneck. Chips and power have been the constraints people talked about. County boards and ballot initiatives will join that list. Expect more moratoriums, longer reviews and real conditions attached to approvals.
- Transparency becomes the price of entry. OpenAI moved first, and others will follow. Disclosing water use, power draw and project scale up front will likely become standard, whether companies want it or not.
- 2028 turns this into a national issue. Ha raised the question tech companies are probably worried about most: does this carry into 2028 and beyond? If it worked this year, candidates will run on it again.
- Geography shifts. Builders will look for places that actually want them and will pay for that welcome with community benefit deals, local hiring and water commitments.
What to do now
If you’re building on AI infrastructure or planning your own compute footprint, a few things are worth doing now:
- Build in delays. Capacity plans that assume fast permitting in Texas or the Southwest need a backup.
- Watch your cloud providers. Local fights can slow regional capacity and push up prices, so it’s worth spreading workloads across regions.
- Get ahead on disclosure. If you build or co-locate, publish your water and energy numbers before someone forces you to.
- Treat the community as a stakeholder. Projects that win local support will get built, and secret ones increasingly won’t.
The AI buildout isn’t stopping. It just has to win over the people who live next to it now. The full Equity conversation is available at TechCrunch AI.