Voters Restrict Data Centers in 40 of 41 Ballot Measures as AI Stocks Slide

Voters approved restrictions on data centers in 40 of the 41 local measures decided on Nov. 3, according to Ballotpedia, turning AI infrastructure into a ballot issue across the country. Two days later, the worst stock-market session since April 2025 erased about $3.7 trillion of U.S. market value by Bloomberg's estimate, with chipmakers hit hardest. The week taught the AI industry how much of its buildout depends on permission, and how much permission now costs.

What did the votes decide?

The Nov. 3 measures ranged from outright bans and moratoriums to 25 MW capacity caps, and Ohio accounted for 21 of them. NBC News' exit poll found 76% of voters opposed building a data center in their community. In Texas, governor-elect Gina Hinojosa campaigned on pausing new data centers until lawmakers pass protections for electricity customers, water and local communities — a state that MLQ's tracker credits with 23.2% of planned AI campus capacity. CoreWeave's CEO wrote to Hinojosa's transition team asking to finish projects that already hold permits.

Company reactions split in tone. Meta president Dina Powell McCormick said building data centers "is going to be harder." OpenAI CFO Sarah Friar told CNBC the company would keep investing "to meet that customer demand" and is still weighing a 2027 IPO; OpenAI also promised to pay full power costs at campuses it owns. The backlash spread abroad: protesters broke into Microsoft's planned Amsterdam site, and a Sydney developer dropped a data center from its plans. Local consent, once a formality, is now a line item.

How bad was the selloff?

Markets first cheered the election result — the Dow jumped 1,569 points, or 3%, on Wednesday. Thursday reversed everything: the S&P 500 fell 4.4%, the Nasdaq 5.5%, and the Dow 1,833 points, the worst day for all three since April 2025. The S&P technology sector fell 7.3%, and the Philadelphia Semiconductor Index dropped about 12%, its second-largest one-day fall on record. Sandisk, Micron, AMD and Marvell each lost about 17%. Nvidia alone shed more than $440 billion of the Magnificent Seven's $1 trillion-plus loss. Bloomberg described the mechanics: retail investors sold first, then hedge funds, volatility-control funds and leveraged ETFs.

The Nasdaq closed Nov. 6 at 24,579.42, down 3.4% for the week and its fourth straight weekly decline, leaving chip valuations at their lowest since 2019. One loss was company-specific: Marvell fell 12.7% on Nov. 2 after agreeing to buy Lumentum in an all-stock deal valued at $115 billion including debt, and ended the week down 28.1% — its worst week since 2001 — cutting the fixed-ratio offer to about $635 a Lumentum share.

Why is money getting more expensive?

The October jobs report made capital cost the week's second theme. Employers added 165,000 jobs against the 59,000 economists expected, and unemployment fell to 4.1%. CME FedWatch odds of a quarter-point hike at the Dec. 8-9 meeting rose to 76.8% from 66.9% a week earlier. San Francisco Fed President Mary Daly said "we're going to have to tighten policy further" and called the tech repricing "painful" but "orderly."

The tightening reached AI through credit. CoreWeave's 9.625% bond due 2032 now yields more than 16%, and its five-year credit default swaps topped 1,289 basis points on Thursday, according to Deutsche Bank data cited by CNBC. Goldman Sachs counted $296 billion of hyperscaler debt issued this year, and default-insurance costs rose for Nvidia, Broadcom and Amazon as well.

What price will buyers pay?

Anthropic became the first big test of public appetite for frontier AI. Its amended S-1, filed Nov. 2, showed third-quarter revenue of $16.96 billion, up from $11.56 billion in the second quarter, a $23.29 billion net loss driven largely by fair-value charges, and about $895.4 billion of purchase and lease commitments — substantially all for compute and infrastructure. Bloomberg reported that prospective buyers now discuss $1.2 trillion to $1.3 trillion, down from the $1.4 trillion to $1.5 trillion communicated to bankers, with the roadshow expected next week.

OpenAI got a similar price check. It raised $11 billion at a $1.25 trillion valuation in a first close that could reach about $20 billion — having sought $30 billion at about $1.4 trillion in September. Microsoft sent the most direct signal from a buyer: it declined an option to more than double Nscale's Texas contract, is winding down Oracle AI servers rented since 2023, and, per Bloomberg, tightened oversight of Nebius, IREN and Nscale while debating whether to sign new neocloud contracts at all. That leaves exposed the neoclouds that borrowed against Microsoft contracts — up to about $43.8 billion of Nscale's statements of work, and about two-thirds of CoreWeave's 2025 revenue. Microsoft's own demand is not falling: Azure revenue rose 48% last quarter on total revenue of $92.3 billion.

Did demand actually crack?

The week's results said no. AMD's third-quarter revenue rose 45% to a record $13.4 billion, data center revenue nearly doubled to $8.5 billion, and the company guided to about $17 billion for the fourth quarter. Micron earned $37.70 billion on $54.23 billion of revenue. TrendForce expects conventional DRAM contract prices to rise another 10%-15% this quarter and NAND prices 15%-20%. Goldman Sachs raised its 2030 global data center capacity forecast to 273 GW from 217 GW, and Google committed $17 billion to its first Wyoming data center near Cheyenne, powered by a 702 MW AEP fuel cell plant, with air cooling to spare local water. Demand held; only the price of everything around the chips moved.

The week of Nov. 9 brings the answers: CoreWeave reports Tuesday, Nebius, Cisco, Applied Materials and Cerebras on Thursday, October CPI lands Tuesday as the last inflation print before the Fed's December meeting, and the Senate holds a Monday cloture vote on a bill that would make data centers pay for grid upgrades.