Here is a number that should land differently depending on which side of a zoning hearing you sit on: 71% of American adults oppose building AI data centers in their area. Nearly half (48%) are strongly opposed. Only 7% strongly favor them. That's from Gallup's first ever survey on the subject, conducted in March 2026. For an industry spending billions on expansion, this is not background noise.

Gallup's first poll on data center attitudes finds that Americans oppose local AI data center construction at rates exceeding their opposition to nuclear power plants (53%), with environmental and resource concerns driving half of the objections. The results are methodologically sound (n=1,000 telephone interviews, ±4 points) and the opposition spans every demographic group, though partisanship and gender widen the gaps. The caveat worth chewing on: Gallup asked about "AI data centers," not data centers generally, which means the framing conflated two separate anxieties into one number.

Two surveys, actually. The primary poll ran March 2 to 18, 2026: 1,000 adults aged 18 and older, reached by telephone (80% cell, 20% landline), with a margin of error of ±4 percentage points at the 95% confidence level. The question asked whether respondents favor or oppose "the construction of AI data centers" in their area.

A follow up ran April 1 to 15 via Gallup's online panel (2,054 members) and asked open ended questions about why people favor or oppose the facilities. That panel sample is self selected (people who signed up for Gallup's panel), which is a step down in rigor from the random digit dial primary survey, but perfectly reasonable for understanding motivations.

The headline is stark. Seven in ten oppose. But the texture matters more than the topline.

Why people oppose: 50% cited excessive resource consumption (water, energy, land). Another 22% raised quality of life concerns, and 20% worried about cost increases (primarily utility bills). Noise and pollution (16%) rounded it out. These are not abstract objections. They are tangible, local, and personal.

Why people support: Job creation. That's essentially it. 55% of the (small) pro data center group cited economic benefits. The value proposition is straightforward but thin.

The nuclear comparison: Opposition to data centers (71%) now exceeds opposition to nearby nuclear power plants (53%). Since Gallup started asking about nuclear in 2001, nuclear opposition peaked at 63%. Data centers have blown past that ceiling on their first at bat.

Demographics: Every major group opposes. But women register stronger opposition than men (55% vs. 43% strongly opposed). Democrats are more strongly opposed than Republicans (56% vs. 39% strongly), with independents at 48%. People who express concern about environmental quality are overwhelmingly opposed (78% vs. 52% among the unconcerned).

This is Gallup. The methodology is standard and transparent. Random digit dialing with an 80/20 cell to landline split reflects current phone usage. The sample size of 1,000 is adequate for national estimates. The margin of error is honestly reported. And importantly, this is Gallup's first ever measure of data center attitudes, which means there is no prior result to bias the framing. It is a clean baseline.

The open ended follow up survey (asking why people hold their views) is a good design choice. It prevents Gallup from putting words in respondents' mouths. When half of opponents independently name resource consumption, that's a signal, not an artifact of a leading question.

Three things to flag.

The "AI" in the question is doing work. Gallup asked about "AI data centers," not data centers generally. AI is politically charged in 2026 in ways that generic cloud computing is not. Would the numbers look different if the question said "cloud computing facilities" or just "data centers"? Almost certainly. How different? We don't know. But the framing choice likely inflated opposition by associating the facilities with a technology that carries its own anxieties.

NIMBY dynamics are universal. As Forbes noted in its analysis, polling on local infrastructure opposition tends to run hot. People oppose landfills, highways, wind farms, and affordable housing in their backyards at rates that don't always predict policy outcomes. The 71% number is real, but it measures sentiment, not political power. Data centers are getting built regardless, often in communities with less organized opposition.

The support case is underexplored. Only 27% of respondents favored data centers, but the survey doesn't tell us much about intensity of support among local officials, landowners, and the construction industry. These are the stakeholders who actually show up to zoning meetings. A poll of the general public may not capture the political dynamics that determine whether a facility gets approved.

The gender gap (12 points in strong opposition) deserves more attention than it's getting. Reporting from The Current notes that women are disproportionately leading grassroots resistance movements against data centers, from Memphis to Amarillo, often framing their opposition around health impacts and generational concerns. This mirrors patterns from other environmental justice movements. If data center opposition becomes a mobilized political force (not certain), it will likely be organized by women at the local level.

The partisan gap is real but narrower than you might expect. Even among Republicans, 39% are strongly opposed. Data center opposition is emerging as a genuinely bipartisan sentiment, which is unusual in 2026 American politics and makes it harder for the tech industry to navigate through conventional political channels.