Americans hate data centers. Tech needs them

More than 70% of Americans oppose data center construction, according to a Gallup and UMass Amherst poll, and Democrats ran anti-tech or anti-data-center ads in over 130 federal and gubernatorial races. The article says voter opposition may collide with Anthropic’s compute needs ahead of its anticipated IPO, while California’s proposed billionaire tax appears likely dead. Some data center stocks, Tesla and SpaceX fell on Wednesday, even as the Dow rose 1,500 points; the author frames the election as evidence of public anxiety about the AI boom.

Analysis

The investable signal is not a blanket halt to AI capex; it is a higher political and permitting cost for adding compute capacity. Local opposition can slow approvals, extend power-connection timelines, and raise the cost of delivered compute even if hyperscalers maintain spending plans. That favors operators with secured power, permitted sites, and scale, while disadvantaging developers whose returns depend on rapid greenfield buildout. For Anthropic (ANT), constrained or more expensive capacity could limit growth or pressure economics; the offset is that compute scarcity increases the value of model efficiency and access to long-term infrastructure partners. The article provides no company-level capacity, contract, or cost data to determine which effect dominates.

Near term, treat the election as sentiment and policy-risk repricing, not evidence of cancelled projects. Over 1–3 months, watch for concrete state or local permitting restrictions, utility interconnection decisions, and revised data-center project schedules. Over 6–18 months, sustained constraints could entrench incumbents with power and site access and shift value toward grid equipment, generation, and efficiency. Contrarian view: voter hostility may not translate into broad construction bans; opposition can instead concentrate projects among better-capitalized incumbents, potentially strengthening their position. The thesis weakens if permitting remains stable and ANT demonstrates secured compute capacity and improving unit economics; it strengthens with project delays, higher capacity costs, or adverse guidance.

Key Decisions for Investors

• ANT: Avoid treating political backlash alone as a short thesis or paying a scarcity premium at IPO. Reassess against disclosed compute commitments, capacity availability, cost trajectory, and cash needs; those data are missing here.
• Watch, rather than initiate a broad data-center short: track local approval delays, utility interconnection queues, and project cancellations over the next 1–3 months. A headline opposition signal without actual permitting friction is not enough.
• If project-level evidence confirms a widening gap between permitted/powered capacity and speculative development, consider a relative-value position favoring grid-equipment and power suppliers over development-heavy data-center exposure; verify issuer revenue exposure before selecting names.
• Falsifiers: stable permitting and project schedules, no deterioration in compute availability or cost disclosures, or evidence that efficiency gains offset capacity constraints. Revisit the thesis after ANT's first public financial and operating disclosures.