Analysis
As of November 4, 2026, Amazon, Microsoft, Alphabet and Meta are projected to spend $720 billion to $740 billion combined on AI data centers and compute this year, and the midterm elections just handed that buildout its sharpest political setback yet.
Where the backlash hit hardest
Data center proposals appeared on more than 40 ballot measures across 13 states this cycle, and voters rejected 38 of them, according to 404 Media, which cites the tracking site datacentersontheballot.com. Six cities voted to ban data centers outright and twelve capped the megawatts new facilities could draw, including a ban passed in Emporia, Kansas.
The biggest single result came in Texas, where Democrat Gina Hinojosa defeated Republican Gov. Greg Abbott, who was seeking a fourth term, by more than 700,000 votes after campaigning on a moratorium on new data center development until lawmakers add resident protections. Texas already hosts more than 300 working data centers with plans to double that figure. Abbott had accepted more than $9 million from data center-related donors and announced his own project "pause" before the vote, though regulators kept approving new facilities anyway, 404 Media reported. Smaller races told the same story: a Yukon, Oklahoma mayor lost a recall after backing a hyperscale project, and Asheville, North Carolina and San Marcos, Texas both elected mayors who ran explicitly anti-data-center campaigns.
“Texas already hosts more than 300 working data centers with plans to double that figure.”
Elon Musk spent roughly $500 million on the midterms, including about $224 million in Texas alone, according to 404 Media. Musk's reaction on X afterward, per the same report: "Data centers in space looking pretty appealing right now."
The capex math behind the backlash
The political heat is landing on an industry that has only accelerated its spending, per Fortune:
- Oracle -- carries more than $125 billion in debt and other notes payable and says it needs to raise another $45 billion to $50 billion in 2026; shares sank 8.76% when it priced a $20 billion stock-and-bond sale last week.
- SpaceX -- runs its own AI compute arm and expects $35 billion to $40 billion in capex in the quarter ending December alone.
- Meta -- plans a $1 billion, 1.4-gigawatt data center in Piqua, Ohio, the kind of project now triggering backlash directly: residents there voted to recall the mayor and two city commissioners who had backed it.
- The sector broadly -- Pulse has tracked roughly $138 billion in debt-backed AI financing deals already lined up as of late October, and Data Center Watch counted at least 45 projects worth about $68 billion disrupted by local opposition in the second quarter alone, before this month's elections piled on.
What Big Tech is already conceding
The industry saw this coming. Before the election, Microsoft pledged to cover power-infrastructure costs for host communities and stop seeking local property-tax breaks, while Amazon, Microsoft and Google all dropped requirements that local officials sign NDAs around new projects. Amazon's "Built Together" program now pledges more than $1 billion over five years directly to host communities. Macquarie's analysts told Fortune they expect the practical effect to be less about outright blocked projects and more about communities extracting more value -- tax revenue, infrastructure spend, local hiring -- from operators who still need the land and the power grid.
What the headline misses: no hyperscaler has actually cancelled a project because of this month's results. Hinojosa's moratorium is a campaign pledge, not an enacted law, and it carries no legal force until the Texas legislature convenes. Oracle, Amazon and Microsoft are all still signing multi-year compute contracts even as local fights multiply -- Microsoft alone added 24 new data centers last quarter, and Amazon is guiding to 40%-plus cloud revenue growth in 2027 and 2028. The near-term effect looks more like slower permitting and higher community-benefit costs than a stopped buildout, a distinction that matters for anyone underwriting AI infrastructure debt on the assumption capacity keeps coming online on schedule.
Texas's legislature next convenes in regular session in January 2027, leaving hyperscalers roughly two months to lock in local approvals under the current rules before Hinojosa's pledge has any chance of becoming law.