Just 1 in 9 Americans Support AI Data Center Construction in their Local Communities

The latest national University of Massachusetts Amherst Poll has found that just 11% of Americans say they would support the construction of an artificial intelligence data center in their local community, while 65% oppose their construction.

Nearly half of the 1,000 respondents surveyed said that they strongly oppose the local construction of data centers. Just 5% strongly support their local construction, while about 1 in 4 respondents said that they neither support nor oppose their construction.

Republicans are more receptive than Democrats, but even Republicans oppose a data center in their community by nearly 3-to-1, 52% to 18%. Among the strongest MAGA supporters—the group most receptive to data centers—just 25% support them being built nearby, compared with 42% who oppose.

Among the reasons given by respondents for their opposition to local data center construction were environmental or natural resource use, land use and local disruptions, general distrust of AI and the industry and the impact on utility bills and costs.

Additionally, the poll found that 57% say that President Trump has handled the issue of AI poorly—a 7-percentage point increase since March of this year—while just 23% say he has handled the issue well.

“In an increasingly polarized America, where generational, class, racial, and partisan groups often find themselves on opposite sides of our most pressing political, economic, and cultural debates, it is striking to find an issue on which Americans are largely in agreement. Opposition to AI data centers has emerged as one of these rare areas of consensus,” says Tatishe Nteta, provost professor of political science at UMass Amherst and director of the poll. “Nearly two-thirds of Americans oppose the construction of AI data centers in their local communities, and this opposition cuts across the political and demographic spectrum. Majorities of Harris voters (81%) and Trump voters (52%), Democrats (76%) and Republicans (52%), and liberals (84%) and conservatives (51%) all oppose the construction of AI data centers in their communities.”

“The remarkable thing is that we can’t find a constituency that actually wants these data centers,” Raymond La Raja, professor of political science at UMass Amherst and co-director of the poll. “Republicans are more receptive than Democrats, but even Republicans oppose a data center in their community by nearly 3-to-1, 52% to 18%. Among the strongest MAGA supporters—the group most receptive to data centers—just 25% support them being built nearby, compared with 42% who oppose. The politics of AI may be somewhat polarized, but resistance to putting its infrastructure next door crosses the partisan divide.”

“In order to better understand this emerging movement and the public opinion behind it, we asked respondents to tell us in their own words why they feel the way they do about AI data centers,” says Alexander Theodoridis, associate professor of political science and co-director of the UMass Poll. “This gives us an opportunity to see what Americans are thinking, uninfluenced by choices we put in front of them. We then used the latest methods in AI text-as-data analysis to classify these responses. The responses*
tended to cite more local or community considerations like environmental or natural resource use (28%), the impact on utility bills and costs (10%), and land use and local disruptions (15%). A smaller proportion of respondents (13%) talked about their general
distrust of AI and the industry. So, a big chunk of Americans may not oppose data centers per se, but they do not want them in their backyard.”

“The conventional wisdom is that strong public opposition to data centers reflects a more general negative reaction to generative AI, but our poll results suggest this isn’t exactly right,” notes Jesse Rhodes, professor of political science at UMass Amherst and co-
director of the poll. “When we asked people to explain why they opposed data centers, relatively few expressed more general opposition to generative AI. Instead, most people expressed more specific local concerns—that data centers will wreak environmental
havoc by polluting fresh water, increasing fossil fuel use, and degrading farmland and forests. They are concerned that during a period of dramatic climate change data centers make a bad situation worse. So, although Americans may have serious misgivings about
generative AI, their strong opposition to the siting of data centers shouldn’t be interpreted as a stand-in for these more general fears.”

Conversely, Rhodes notes that the few supporters of data centers have distinctive views. “While opponents cite negative environmental and aesthetic impacts as well as broader distrust of the generative AI industry, supporters argue that data centers are essential for economic development,” he says. “One argument is that data centers will spur local job creation and boost local economies, particularly in depressed areas. But data center boosters also argue that more data centers are needed to ensure American economic dominance, especially in competition with China. For these individuals, data centers are a key to the United States’ long-term economic competitiveness.”
 
“Taken together, these findings suggest that the debate over AI data centers is, at its core, a debate over costs and benefits,” Nteta says. “Opponents are overwhelmingly focused on the tangible costs that these facilities impose on their communities, while supporters point to the broader economic and strategic benefits of AI development. Americans may agree that AI represents the future, but they appear far less certain about who should pay the price for building it.”

*See toplines and crosstabs below for response results.

“Americans aren’t giving President Trump particularly high marks for his handling of artificial intelligence,” La Raja says. “Just 23% say he has handled AI well, compared with 57% who say he has handled it poorly, and the public has grown more negative since March. The share saying Trump has handled AI ‘not well at all’ has risen from 36% to 43%. AI may be a new issue, but it is fast becoming one in which politicians are judged negatively on it.”

“In recent weeks, Americans’ fears about the rapid development of artificial intelligence have been underscored by a series of troubling developments,” Nteta explains. “Hundreds of OpenAI agents broke containment and launched a coordinated cyberattack against the open-source platform Hugging Face, while Jacob Coxon, a former OpenAI and Anthropic researcher, resigned from Anthropic, declaring that the company was not ‘acting responsibly’ and warning that many in the industry believe AI ‘could kill us all by the end of the decade.’ These developments prompted Anthropic CEO Dario Amodei to call for the industry to slow the pace of AI development and put greater emphasis on safety."

• Field Dates:                 August 21 –26, 2026
• Sample:                        1,000 Respondents
• Margin of Error:         3.5%

The respondents were matched to a sampling frame on gender, age, race, and education. The sampling frame was constructed by stratified sampling from the full 2023 American Community Survey (ACS) 1-year sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file).

The matched cases were weighted to the sampling frame using propensity scores. The matched cases and the frame were combined and a logistic regression was estimated for inclusion in the frame. The propensity score function included age, gender, race/ethnicity, years of education, region, and home ownership. The propensity scores were grouped into deciles of the estimated propensity score in the frame and post-stratified according to these deciles.

The weights were then raked (iterative proportional fitting) on 2024 presidential vote choice and post-stratified using a four-way stratification of gender, age (4-categories), race (4-categories), and education (4-categories), to produce the final weight.

Read more about the UMass Poll's methodology

Read more about the contributing members of the UMass Poll team

View the toplines and crosstabs from the September 14, 2026 National Public Opinion Poll release regarding AI.

AI Data Center

Support for Construction of an AI Data Center

Do you support or oppose the construction of an artificial intelligence data center in your local community?

AI Data Center Support by Demographic Group

Open-Ended Responses on AI Data Centers

In a sentence or two, please explain why you (support/oppose/neither support nor oppose) the construction of an artificial intelligence data center in your local community?

Thematic Analysis of Open-Ended Responses

Data. 1000 respondents who expressed responses to the open ended question, “In a sentence or two, please explain why you (support/oppose/neither support nor oppose) the construction of an artificial intelligence data center in your local community?”

Method. Inductive thematic coding was conducted across all respondents. Five substantive themes were derived, plus a sixth residual category for responses that state no interpretable reason (Don’t know/Missing). Each response is assigned to exactly one theme. All quotations are verbatim, including original spelling, capitalization, and profanity.

Theme 1. Environmental Harm and Depletion of Natural Resources (N=281)

Respondents in this theme object to data centers because of their perceived negative environmental impacts on fresh water, electricity and the capacity of the power grid, air quality, farmland, forest, wildlife habitat, and climate.

“Data centers are a huge drain on resources such as water, which is already a major issue in our area. Draining our water reserves significantly with a data center would be catastrophic to our region.”

Theme 2. Higher Utility Bills and Economic Burden on Residents (N=101)

Respondents in this theme object because, in their view, data centers raise the cost of living for people who receive none of the benefits. The most common claim is that electricity and water rates rise after a data center comes online, sometimes reported as first-hand experience with an existing facility rather than as anticipation. Related claims are that local property values fall, taxes rise, tax abatements and utility-rate support subsidize a profitable industry that does not need them, and promised jobs fail to materialize in numbers that would offset negative impacts.

“Because ever since a data center has been built in my local area the prices of water and electricity have more than doubled and is draining my funds to support my family.”

Theme 3. Local Disruption, Land Use, and Poor Community Fit (N=145)

This theme highlights objections based on perceived negative effects of data centers on local quality of life. The most frequent complaint is noise — the continuous mechanical drone reported by people living near existing centers, described repeatedly as constant, aggravating, and inescapable. Alongside noise are visual blight, light pollution, traffic and congestion, dust, and health worries attributed specifically to proximity.

“Noise issues, energy issues and visual site are three issues that are major concerns for a AI data center to be in the area I live in.”

Theme 4. Rejection of Artificial Intelligence and Distrust of the Industry (N=134)

Respondents in this theme oppose the data center as a proxy for the technology and the companies behind it. Some responses are flatly affective — “I hate AI,” “I detest AI” — while others make more specific claims that AI will displace workers, that it degrades human thinking and creativity, that it is unreliable and error-prone, that it enables scams and deepfakes, or that it is being deployed far faster than it is being tested. A second strand targets the industry and the state rather than the technology, casting data centers as surveillance infrastructure, as threats to privacy and the Fourth Amendment, as an oligarchic grift enriching a handful of firms, or as an enterprise operating in a near-total regulatory vacuum.

“I think that AI needs to be better regulated and a lot more developed and proven safe before any massive full scale deployment”

Theme 5. Economic Benefits, Technological Necessity, and Competitiveness (N=115)

Respondents in this theme support generative AI for a variety of reasons. These respondents believe data centers bring jobs, tax revenue, business investment, and prosperity. One sub-theme is that data centers are necessary infrastructure for important technology and that, therefore, local communities should capture economic benefits. A second distinct sub-theme is national competitiveness, framed in terms of US economic competition with China.

“Benefits to numerous industries. Financial benefits to communities. Important to keep the US competitive.”

Theme 6. Unsure, Indifferent, Ambivalent, or Uncodeable (N=224)

This residual category contains every response that does not contain a substantive theme. Five different types of statements are collected: self-professed ignorance; claims of indifference; fatalistic statements to the effect that data centers will be built regardless of the individual’s preferences; ambivalent statements acknowledging both pros and cons; and uninterpretable and off-topic responses.

“I don't really know what the pros and cons would be.”

Trump AI Issue Performance

How well would you say that President Trump has handled artificial intelligence?

Trump AI Issue Performance by Demographic Group

Demographic Breakdown of Survey Respondents