The infrastructure powering the AI boom is generating a resistance movement that Big Tech did not see coming.
A March 2026 Gallup poll, the first time the organization has measured public opinion on AI data center construction, found that seven in ten Americans oppose building these facilities in their local communities. Nearly half say they are strongly opposed. Just 7% express strong support.
For context: 53% of Americans oppose nuclear power plants being built in their communities. Opposition to nuclear power has not exceeded 63% since Gallup began tracking the question in 2001. AI data centers are polling worse than nuclear facilities on the same measure.
Energy expert Robert Bryce has cataloged nearly 1,200 examples of rejections or restrictions of energy projects over his career. He told Financial Sense in May that he has never seen anything like the current backlash against AI data centers.
The numbers back that up. In all of 2025, there were 49 rejections or restrictions of data center projects across the United States. In the first four months and one week of 2026, there have already been 79. Two came in a single week, one in Ferguson, Missouri, and one in Wichita Falls, Texas. In Maine, lawmakers passed the first statewide moratorium on new AI data centers, though the governor vetoed it. In Utah, residents are actively fighting plans for what would be the largest data center in the world, proposed for Box Elder County.
The opposition is bipartisan. Democrats focused on environmental impact and Republicans skeptical of Big Tech are arriving at the same conclusion from entirely different starting points.
What communities are objecting to is concrete and local. Energy costs going up. Water consumption by facilities that require significant cooling. Noise. Land use. The broader distrust of the companies building these structures. A Gallup follow-up survey asked Americans to explain their opposition. Environmental impact was the dominant concern, with 46% saying they worry a great deal about data center effects on the environment.
There is another factor that sharpens the local calculus considerably. Data centers tend to create relatively few permanent jobs for the communities hosting them. The infrastructure serves a global AI ecosystem. The disruption is local.
Bryce predicts the backlash will not stop the buildout. More than 700 data centers are currently under construction across the United States. The financial momentum is too large. But the regulatory and political environment is shifting in ways that will slow project timelines, increase siting costs, and complicate expansion plans for hyperscalers and their enterprise customers.
For executives making AI infrastructure decisions, this introduces a real planning variable. Siting new capacity now involves community relations, regulatory risk, and local political dynamics that hyperscalers have historically been able to avoid. That is changing quickly. Organizations treating AI infrastructure like any other major capital project, with stakeholder engagement and environmental planning built in from the start, will have fewer surprises than those still operating as if public opposition does not exist.
