Americans overwhelmingly voted against data centers in the midterms

Of 45 local ballot measures on data centers across 10 states, 33 had passed by early Wednesday and six more were projected to pass, with measures ranging from bans and moratoriums to energy-use limits and approval requirements. In Ohio, where 257 data centers operate, restrictions had won or were projected to win on all 20 ballots; Urbana and Hubbard voters backed outright bans by 76% and 77%, respectively. The Associated Press declared Democrat Sherrod Brown the winner of Ohio’s Senate race after he campaigned against data centers, and an NBC exit poll found 56% of voters said AI and data centers were a crucial or major factor in their congressional vote.

Analysis

Local permitting risk is becoming a financing and valuation variable for greenfield data-center capacity, not just a community-relations issue. The second-order pressure is on developers and power providers whose investment cases rely on forecast load arriving on schedule: delayed approvals can defer utility upgrades, weaken regional load-growth assumptions, and leave site or interconnection spending idle. Existing facilities with secured power and permits may gain scarcity value if new capacity becomes harder to build, while AI infrastructure demand itself is more likely to relocate than disappear.

The vote is a meaningful signal, but not proof of a nationwide construction stop. Local measures differ in scope, and state or federal preemption, negotiated community benefits, or siting shifts could blunt the impact. Near term, watch project-level cancellations and permitting timelines rather than extrapolating from election results to AI spending. Over 1–3 months, utility load forecasts, interconnection queues, and developer guidance should reveal whether this is changing capex decisions; over 6–18 months, persistent local restrictions could concentrate growth in jurisdictions with available power and clearer approval paths.

Contrarian angle: a broad short of AI infrastructure may overprice the risk. The likely transmission is higher cost and slower deployment at the margin, with incumbents and power-secured sites relatively advantaged—not necessarily lower aggregate demand. The thesis weakens if approvals proceed under revised terms or load forecasts remain intact; it strengthens with project deferrals, revised utility forecasts, or state-level restrictions.

Key Decisions for Investors

• Favor a relative-value stance: prefer established data-center operators with permitted, powered capacity over developers more dependent on unpermitted greenfield projects. Verify each operator’s site pipeline, power commitments, and jurisdiction mix before sizing.
• Avoid treating utilities as an automatic short. Monitor regional load forecasts and planned transmission or generation capex; reduce exposure where expected growth depends heavily on projects facing local opposition, and reconsider if utility guidance or project schedules are cut.
• Over the next 1–3 months, track local implementation rules, state preemption proposals, interconnection milestones, and company commentary on permitting. A formal moratorium, material project deferral, or downward load-forecast revision would be a stronger bearish catalyst than the vote signal alone.
• Do not chase a broad AI-infrastructure selloff absent evidence of cancellations or slower customer commitments. The main falsifier for the scarcity-beneficiary thesis is a sustained return of approvals and on-time capacity additions across contested markets.