Americans Prefer Nuclear Plants Over AI Data Centers in Their Communities, Gallup Poll Finds

A March Gallup survey found that 70% of Americans oppose building AI data centers in their local area, with 48% strongly opposed. Only 27% support such projects, and just 7% feel strongly in favor. This marks the first time Gallup has polled on data center construction, capturing a sentiment that has intensified as AI infrastructure expansion accelerates across the country.

The opposition spans political affiliations, though Democrats express somewhat stronger resistance than Republicans. The survey of 1,000 adults reveals a striking public relations challenge for an industry pouring billions into compute capacity. According to Business Insider, environmental concerns drive much of this anti-AI sentiment, with residents worried about noise, water consumption, and the transformation of rural landscapes. The Register notes that AI and its supporting infrastructure have developed a serious PR problem that could complicate future development plans.

Americans now view nuclear power plants more favorably than data centers as local neighbors, a remarkable shift from Cold War-era anxieties. The Gallup data shows data centers lag behind nuclear facilities in public acceptance, reversing decades of nuclear fear rooted in disasters like Three Mile Island and Chernobyl. Fortune observed that the old Duck and Cover PSAs have given way to a new era where atom-splitting feels less threatening than server racks.

This preference reflects both evolving attitudes toward nuclear energy and growing awareness of data center impacts. Gizmodo reported that nuclear's improved safety record and carbon-free baseload power have rehabilitated its image, while data centers' constant noise, visual blight, and resource demands have soured communities on hosting them. The comparison suggests Americans have weighed relative risks and found the invisible threat of radiation less immediate than the very visible disruption of industrial-scale computing facilities.

Specific concerns center on environmental and quality-of-life impacts. CNET highlighted that data centers consume extensive electricity and water while negatively affecting local communities. In Saline, Michigan, residents opposed a $16 billion Oracle and OpenAI facility developed by Related Digital, worried about losing farmland and stressing local resources. These conflicts illustrate how abstract national AI ambitions collide with concrete local interests.

The opposition pattern follows classic NIMBY dynamics but with unusual intensity. Tom's Hardware pointed to previous surveys showing 47% opposition to data centers, indicating sentiment has hardened considerably. Engadget noted that AI companies are spending astronomical sums to build infrastructure as quickly as possible, yet this urgency runs directly into community resistance. The mismatch between corporate timelines and democratic consent processes threatens to slow or derail projects worth billions.

The polling data signals tough times ahead for site developers and their corporate customers. Startup Fortune argued the finding turns AI infrastructure into a local trust and permitting problem, not merely a technology spending race. Companies can no longer assume rural communities will welcome massive facilities based on promised tax revenue and jobs. The Register warned that developers face a serious challenge in securing locations for the server capacity needed to power generative AI expansion.

This resistance arrives at a critical moment. Tech companies are racing to acquire compute capacity to train larger models and serve growing user bases. Morningstar's coverage emphasized that nearly half of Americans oppose neighborhood data centers, while 38% support them, a gap that gives political cover to local officials who block projects. The industry may need to fundamentally reconsider siting strategies, community engagement, and perhaps even facility design to address concerns about water use, noise, and visual impact.

The renewed interest in nuclear power partly stems from the same AI-driven demand surge that sparks data center opposition. Commonfund's analysis noted that U.S. electricity demand is rising after long stagnation, putting pressure on power grids already strained by data center expansion. This has fueled interest in small modular reactors as a potential solution for powering AI infrastructure without carbon emissions.

The irony is apparent: the same computing demands that make data centers unwelcome could make nuclear more acceptable if framed as powering them. Yet the Gallup data suggests Americans distinguish between the power source and the facility itself. They may accept nuclear generation in principle while still rejecting the data centers it would support. This complicates arguments that nuclear-powered AI facilities will overcome local opposition. NewsNation's coverage of the poll underscored the stark choice Americans face between competing infrastructure priorities.

The industry faces a choice between overcoming community resistance or finding ways to reduce local impact. Forbes AI's coverage highlighted the nuclear comparison as a warning sign that traditional siting approaches may fail. Some companies may pursue offshore or otherwise isolated locations, though transmission costs and latency constraints limit this option. Others might invest heavily in community benefit agreements, though 48% strong opposition suggests money alone may not suffice.

Longer term, the opposition could accelerate interest in distributed computing, smaller edge facilities, or more efficient hardware that reduces facility footprints. The political dimension matters too, as Business Insider noted the cross-party nature of opposition means neither party reliably supports data center development. State and local officials in both parties may find blocking projects popular. For an industry accustomed to moving fast and breaking things, democratic consent may prove the most immovable constraint on AI's physical expansion.