AivexaNewsSearch
AI news for builders and product teamsChecked every hour

Meta warns that election backlash against data centers could make AI expansion harder

Collected Nov 5, 2026

Meta President Dina Powell McCormick said at the WSJ Tech conference on Wednesday that Democratic gains in the midterm elections will make it harder for companies to build data centers, and that the industry has fallen short in explaining what those investments bring to host communities. Election-season opposition to data centers, she said, was bound up with how people feel about AI more broadly.

Powell McCormick framed the physical facility as the most tangible face of the technology. A data center is something residents can point to and reject, she said, while the deeper driver is sentiment toward AI itself. Meta's stated remedy is to show value twice over: the local benefits of a project and the usefulness of the technology.

She tied that case to a broader pitch, arguing the stakes lie in who builds systems that could cure cancer or widen access to education. She described CEO Mark Zuckerberg as betting on people over machines and cast democratized, fair and transparent access as the way to keep anyone from being left behind.

On scale, Powell McCormick said Meta has built data centers for roughly 20 years to support the compute needs of its social products. The company now operates facilities in 23 states, and five are planned as projects of one gigawatt or more, including one in Texas. For context, a gigawatt is a large concentration of power demand; campuses of that size typically require dedicated transmission and generation arrangements, which is often where local opposition concentrates.

The executive was direct about the outlook. Building has already been difficult, she said, and she expects it to get harder, though she added that workable models are becoming visible and that states want them. Her expectation is more scrutiny of projects and more weighing of whether residents actually support them.

She laid out the playbook Meta uses to close local deals: paying for its own electricity, directing tax surpluses generated by a facility toward teachers and first responders, keeping the workforce local, and securing bipartisan backing. In practice these are the levers available to a large buyer negotiating with a county or state, since a hyperscaler cannot offer much beyond tax revenue, construction jobs and grid upgrades. The mention of self-funded power is notable because electricity supply and rate impacts are frequently the sharpest local complaint; whether such terms satisfy critics is a separate question the remarks do not settle.

Meta's AI ambitions run through its consumer agent, Muse, which she said has 10 million users today and is expected to scale toward billions. Even if that audience does not materialize as hoped, she argued the compute will not be wasted: the company receives daily inquiries about leasing capacity, giving it a route to monetization either way. That is a hedge worth noting. Renting out GPU capacity turns a stalled product bet into an infrastructure business, though it puts Meta in competition with cloud providers and depends on the same power and permitting constraints she just described.

The political dimension is the new element here. When candidates campaign against data centers, approvals that once moved through planning commissions become electoral questions, and timelines stretch. For teams planning training runs or inference capacity, siting risk becomes a schedule risk that is harder to model than chip supply. This suggests the constraint on AI expansion is shifting from silicon toward permits, power contracts and local consent, and companies that treat community relations as a cost center may find themselves behind those that negotiate revenue-sharing terms up front.

Why it matters: Developers and product teams relying on Meta-hosted or leased compute, or on AI capacity generally, face longer and less predictable infrastructure timelines as local politics enters siting decisions. Companies that can demonstrate tangible community benefits, and that hold fallback options such as leasing out capacity, are likely to be more resilient, though that is inference rather than anything Meta confirmed.

Read at TechCrunch · AI

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

Meta warns that growing political opposition to data centers could threaten future AI investments, as communities push back against massive infrastructure projects.