America unites on one issue: Reining in datacenters

Data Center Watch found 43% of 74 selected gubernatorial candidates across 36 states supported a ban or moratorium on new data centers, while 51% favored further regulation or conditional development; about 5% had no stated position. In Texas, Democrat Gina Hinojosa, who supports a temporary moratorium, defeated incumbent Greg Abbott and takes office in January. Separately, 43 of 46 municipalities with data-center restriction measures had approved them at publication, indicating material permitting and policy risk for developers, though the article notes that targeted restrictions do not necessarily oppose all development.

Analysis

The investable signal is not a near-term collapse in compute demand; it is a higher and less predictable cost of converting demand into permitted, powered capacity. Political pressure can lengthen approval timelines, raise community-benefit and mitigation costs, and strand land or grid deposits before it materially constrains installed capacity. That makes pipeline-heavy projects more vulnerable than operating sites with secured power and permits. Existing operators such as Equinix and Digital Realty could gain scarcity value if customers shift toward leasing, but that benefit is conditional: delays that persist long enough can also cap their expansion and weaken customers’ ability to deploy AI. Microsoft, Amazon, Alphabet, and Meta face a potential capex-timing and location constraint, not evidence of weaker end demand.

The contrarian point: election rhetoric and municipal votes are leading indicators, not proof of binding statewide rules or a uniform national moratorium. Governors’ authority varies, and local permitting, utility interconnection, and water approvals may matter more than campaign positions. In the next few weeks, sentiment may pressure exposed infrastructure names; over 1–3 months, watch Texas transition decisions and actual permitting/audit outcomes. Over 6–18 months, the key question is whether projects are cancelled or merely relocated, delayed, or redesigned. Falsifiers include continued permits and energized capacity growth in contested markets, or policy language that exempts already-permitted projects.

Key Decisions for Investors

• Avoid broad, outright shorts on data-center operators: sentiment is not yet evidence of demand destruction, and supply friction may support rents on existing permitted capacity. Prefer a relative-value watchlist: Equinix and Digital Realty versus AI infrastructure exposures whose returns depend more heavily on new-site delivery; validate regional pipeline and valuation exposure before sizing.
• Use Texas as the near-term catalyst, not the election result alone. Do not add to a restriction thesis until the incoming administration issues specific rules or changes permitting, water, or power policy. Track permit approvals, interconnection queues, project cancellations, and audit findings; a pause limited to review or grandfathered projects would weaken the bearish case.
• For hyperscaler exposure, distinguish capex timing from AI monetization. A credible delay to powered capacity could shift spend between owned builds and leased capacity rather than erase it. Reassess Microsoft, Amazon, Alphabet, and Meta only if management guidance or disclosed capacity plans show sustained deployment slippage—not on campaign rhetoric alone.
• Risk/reward is asymmetric by asset: speculative projects lacking permits or secured power face downside from delay and sunk costs, while operating facilities may benefit from scarcity. The thesis is falsified if contested jurisdictions continue bringing substantial new capacity online without material delays, or if developers successfully relocate projects with limited cost or schedule impact.