7 in 10 Americans oppose data centers being built in their communities

A Gallup survey found 7 in 10 Americans oppose data centers being built in their communities, with opposition especially intense among Democrats. The shift signals rising public resistance to AI-related infrastructure and could increase political and permitting friction for data center development. The article is sentiment-negative for the sector, but the immediate market impact is likely limited.

Analysis

The important read-through is not simply higher local resistance to data centers; it is a rising political cost of physical AI infrastructure. That shifts the bottleneck from compute demand to permitting, power interconnects, water usage, and community absorption, which should raise project timelines and capex for anyone trying to scale training or inference capacity inside the U.S. The first-order winners are not the hyperscalers themselves, but the adjacent beneficiaries of a slower build cycle: grid equipment, power generation, and industrial automation names that can monetize the same AI capex without facing the same land-use backlash.

Second-order, this is a sentiment headwind for “AI infrastructure” as a crowded trade because the market has been valuing a smooth conversion of model demand into domestic physical buildout. If opposition hardens into local zoning friction, the market may need to reprice the pace of revenue realization for power, cooling, switchgear, and construction contractors. That argues for dispersion rather than a broad short: the more exposed a company is to greenfield U.S. data-center delivery, the higher the execution risk over the next 6-18 months.

The contrarian view is that public opposition often delays rather than cancels projects, and that delay can actually be constructive for infrastructure vendors by stretching backlogs. In addition, some of the compute spend may shift to regions with friendlier permitting, existing industrial sites, or even overseas capacity, which would reduce the local political burden but not the aggregate AI capex cycle. The bigger risk to the bearish case is policy intervention at the state level to fast-track strategic infrastructure once grid reliability becomes a binding issue.

Near term, the catalyst set is local: zoning fights, permitting hearings, utility interconnect delays, and election-cycle rhetoric around water and power consumption. Over a multi-quarter horizon, the key variable is whether hyperscalers respond by pushing more spend toward modular, smaller-footprint deployments and higher utilization of existing campuses; that would blunt the opposition narrative while still supporting the AI capex super-cycle.

Key Decisions for Investors

• Buy a basket of grid and power-enablement names versus a short basket of U.S.-centric data-center developers on a 3-6 month horizon; the trade works if permitting friction widens buildout timelines by even 1-2 quarters.
• Prefer long positions in industrial electrical and cooling suppliers over broad AI-infrastructure beta; these names can keep backlog growth while avoiding the political overhang attached to new land development.
• If available, buy 6-12 month put spreads on the most U.S.-exposed data-center REIT / infrastructure names into election-season headlines; the downside is capped, while delayed project starts can pressure near-term multiple assumptions.
• Avoid adding to crowded long exposure in pure-play AI infrastructure until local permitting sentiment stabilizes; the risk/reward is poor if consensus is pricing uninterrupted domestic capacity expansion.
• Watch for a rotation long utilities with large transmission and substation exposure if data-center opposition pushes more projects toward existing grid-connected sites; that is a cleaner second-order beneficiary than the headline AI beneficiaries.