Is Public Support For Sale? What Americans Would Charge to Accept a Data Center Next Door
When people think about what it takes to build a data center, they usually think of four material resources: energy, land, water, and network infrastructure. Our research introduces a fifth constraint, one that is often overlooked but increasingly decisive: public sentiment.
Tech companies will have to reckon with public opinion to keep expanding - and public support is shaping up to be costly and scarce. To what extent can public support be secured in the same way that a tech company would secure a water rights agreement? In other words, is public support for sale? And if so, how much does it cost? To find out, we conducted a nationwide survey of more than 3,200 American adults First, we mapped the size and scope of public opinion toward data centers by asking participants to rate how they would feel about a data center being built near them, then describe their thinking in an open-response item. Next, we asked participants to name the price on their support.
Opposition is fierce. And it doesn’t map onto political party.
We asked a nationally representative sample of 3,234 Americans how they would feel about a data center being built in their local community. Two in three Americans (66%) oppose a data center being built in their local area, and less than 1 in 5 support one. This is a dramatic change from just as recently as 6 months prior to this survey, when a Pew survey reported 39% of the public in opposition to a local data center. A March 2025 survey documented only 29% opposition to a local data center, meaning that opposition has risen by more than 120% in just over a year.
66% of Americans oppose data centers in their local area
Surprisingly, we find that opposition doesn’t follow the typical party lines. In today’s political climate, Republicans and Democrats agree on very little, but distaste for local data centers appears to be one of those rare points of consensus. Democrats lead the opposition and Republicans are somewhat less opposed, but across party lines, a majority of respondents oppose a data center coming to their community.
Across party lines, most people oppose local data center builds.
The same holds across age groups. Opposition runs high among young adults, but it isn’t limited to them: across generations, most people oppose a local data center.
Opposition unites across generations.
We also asked participants to rate how much each of five factors influenced their view: quality of life for local residents, local economic effects, AI development, environmental effects, and effects on local resources. The vast majority weighed all five, and that was true for supporters and opponents alike. In other words, people weren’t driven by a single concern; they were weighing many at once. Whether they support or oppose, data centers are a hot-button issue that brings a wide range of costs and benefits to mind.
Data centers aren’t a one-button issue.
Next, we asked an open-ended question about why people feel the way they do about data centers. We found that supporters tended to talk about the potential economic benefits and other civic opportunities. For opposers, concerns about the environment were most prominently on their mind. Among supporters, the three most common words were: jobs, community, future. Among opposers: water, resources, and environment.
Republican and Democratic opponents also used largely the same words to describe their thinking. That is striking: the two parties have disagreed on environmental policy for decades, yet opposition to data centers has them speaking the same language.
Opposition yields uniquely bipartisan agreement.
Is local support for sale?
Next, we quantified the cost of local opposition. We asked participants who said they did not support a local data center to register their required price - in the form of a one-time payment - to gain their support. The item was formatted as an open-response item, meaning that participants could write in any number that they wanted, as long as it was a real number. This is where the depth of opposition really came through: 22% of participants reported a price-tag above $1 million, 174 people registered a price above $1 billion, and 7% respondents reported that nothing could buy their support. Some answers were so high that they rendered the average meaningless, so we report medians instead.
Support might be buyable - but some won’t budge at any price point.
Scatterplot graph of responses to the following survey question: “Imagine a tech company is interested in building a data center in your community. They are willing to negotiate terms with the local residents. They are in talks with your local government to agree on a deal that would give all local residents a one-time tax rebate. Under what conditions - in terms of a tax rebate - would you be willing to say that you support the construction of this data center in your community?” The x axis measures dollar amounts, with markers for “$10k,” “$50k,” “$100k,” “$250k,” “500k,” and “$1M.” A key to right of the graph is labeled “Number of respondents” and show four circles of increasing size that represent “50,” “100,” “150,” and “200” respectively. The graph shows that most responses fell between “$10k” and “$100k,” with the notable exception of about 200 respondents at “$1M.”
Public support is costly.
The median price required to buy opposers’ support was $10,000 in individual, one-time payouts. Scaled up to a median-sized county, buying support would cost a total of $240 million. And because the cost grows with every additional resident, the bill climbs steeply for data centers sited in larger counties.
While the median acceptable payment for a non-supporter to change their mind was $10,000, that increases as income increases: the median among those that make between $75K and $150K is $25,000; the median for the upper-class is $50,000. Which means that a data center seeking support in a more populated and wealthier county - such as in Southern California or Northern Virginia - would have to pay a much heftier payout.
The wealthier the opposition, the more their support costs.
Public support for the data center buildout can’t be taken for granted. It is scarce, and it is expensive. Two in three Americans oppose a data center in their community, and that opposition cuts across party lines and generations. It is also multidimensional: people weigh many concerns at once, from water and power to quality of life, so no single concession is likely to win them over. In some places, support may be for sale, but at a steep price that rises with population and income, and a meaningful share of opponents say no amount would be enough. For tech companies and policymakers, public sentiment belongs on the list alongside energy, land, water, and network infrastructure: it is a resource that has to be secured before ground is broken.
Source: NSI survey fielded June 7–8, 2026.
Americans want investments in their local community when a data center comes to town.
And they really like when it goes directly to their pockets.
Public opposition to data centers has already stopped billions of dollars worth of projects, and stalled many more. Data center developers are starting to realize that they can move faster with communities on their side, and they are increasingly making bids for public support. Meta launched an ad campaign in a bid to paint data centers in a community-oriented, locally-grown light. Amazon wants the public to know that they will build data centers “responsibly” and in a way that benefits the locals. And a developer in Pennsylvania went so far as to directly offer locals $10,000 one time payouts if they allowed a data center to be built in their area. Tech companies and developers are interested in investing money in a bid to earn the public’s support, but where do the people want that money to go?
To find out, we ran a 4-condition randomized survey experiment with a pre- post- design to test what type of investments the American public would prefer. Launched in early July 2026, we collected a sample of 1,991 American participants.
We based the size of the investment on our first study: a median price of $10,000 per household, or about $240 million for a median-sized county.
First, at the beginning of an unrelated survey study, we asked: “Overall, would you favor or oppose the construction of more data centers in your local community to support artificial intelligence, or AI, technology in the U.S.?” and allowed participants to respond on a 5-point Likert scale from strongly oppose to strongly support. Next, after completing their participation in the other unrelated survey, we brought them back and introduced a hypothetical data center’s proposition: “A technology company has proposed building a data center for AI applications in your local community.” Then, each participant saw one of these four scenarios:
Control condition: The technology company would operate under a sales tax exemption, as is the standard for most previous data center projects.
Taxes condition: Under a negotiated agreement, the company would provide a $240 million investment in your local community, distributed as a $10,000 payment to each household.
Pro-environmental efforts condition: Under a negotiated agreement, the company would provide a $240 million investment in your local community, distributed to environmental initiatives.
Social capital condition: Under a negotiated agreement, the company would provide a $240 million investment in your local community, distributed as a schools, libraries, and community centers.
Then, we asked: “Considering the negotiated terms, would you favor or oppose the construction of this data center in your local community?” Because we measured each person’s view before and after, we could see how much each type of pledge changed minds.
Investment pledges always boost support. The strongest boost comes when money goes straight to local residents’ pockets.
We tested three ways companies might invest $240 million in communities to win support for local data center development. While every type of investment nudged support upward, direct cash payments were the only type of investment that yielded more supporters than opposers.
Only Cash Moves the Town
There is an important caveat though. No investment was enough to yield average support, because at least 1 in 10 people still sit on the fence. So direct cash payments can yield more supporters than opposers, but doesn’t completely flip the town.
After learning that a prospective data center pledges to invest $240 million in the local community, the percentage of people in opposition fell by 10 to almost 20 points, and support nearly doubled.
Investment pledges lower opposition and nearly double support.
Our findings revealed some interesting nuances about how people think about investment pledges. For example, proximity to a current data center generally increases tolerance for a new data center build. But, that same proximity blunts the positive effect of the $240M pledge in local public goods. That is: an investment pledge in local public goods brings support up to the same level as those that already have a data center near them. But that same investment moves the needs half as much amongst those that already have a data center near them. This was not, however, the case for the other investment pledges.
Proximity blunts the $240M pledge in social capital.
Money can buy goodwill, but how it’s spent matters. Every investment pledge we tested softened opposition, but none matched direct payments to residents. Even so, not even direct payments won over a majority, and communities that already host data centers responded less to promises of new schools and libraries. For developers, the lesson is that data center siting should start with investments in the community at a baseline, and continue building goodwill from there.
Source: NSI survey fielded July 2 - 3, 2026.
Who Wants Data Centers?
We hear a lot about the people fighting data centers. Dozens of surveys, polls, and headlines announce that people really don’t like data centers. Two in three Americans do not want a data center built near them (that figure is so well established that dozens of surveys could be cited for it, including our own [LINK HERE]). Opposition is so fierce and ubiquitous that it is reshaping American politics and tearing American towns apart. Localities across the US are voting on data center moratoriums. The landscape of data center opposition is exceedingly well-mapped. So we flipped the question and asked the increasingly interesting question: who wants data centers?
In early July, we fielded a survey study with a nationally representative sample of 1,991 American adults. We asked all participants how they would feel about a data center being built near them, whether they have a data center in their local area, how often they use AI, and a number of demographic questions. We then randomly assigned each participant to evaluate one of four scenarios for a proposed local data center: three in which the company pledged a $240 million investment in the community, and one with no investment. We analyzed all of the demographic variables and condition assignments, and identified 4 key patterns that sketch a portrait of the data center supporter.
Those that already have a data center nearby are more tolerant of them
People who already have a data center within 5 miles of their home are 19 percentage points more likely to say they would support another one being built in their community. This might sound counterintuitive: a reader might expect that someone who already has a data center in their town would be less willing to have another one taking up space in the same area, compared to someone whose town currently has no data centers at all. However, this pattern is not new.
When Americans began opposing nuclear power, an overwhelming majority of Americans did not want nuclear power near their homes. But the communities that already lived near a reactor were the most willing to accept another one. Familiarity made the risks concrete and manageable rather than abstract and catastrophic, and host communities had seen the jobs and tax revenue firsthand. This pattern shaped the industry’s roll-out for the next few decades: nearly every new reactor built in the U.S. in recent decades has gone up at an existing plant site. Data centers may be following the same path. For people who already live near one, a data center is no longer a looming unknown. It’s a familiar building down the road whose costs and benefits they’ve already experienced.
There are important nuances that are worth mentioning. For example, we asked people to self-report whether they currently have a data center in their town. Many people may not know about data centers near them, and the ones that do may be more likely to have benefit from them, for example by being employed during their construction. Second, only 458 out of 1,991 participants (23%) reported living within near a data center, so our estimates for this group are less precise than for the full sample. Still, the pattern closely mirrors what happened with nuclear power, which gives us some confidence that the effect is real.
Daily AI users are more tolerant of data centers
People that use AI daily are 22 percentage points more likely to support a data center being built in their community. 34% of daily AI users support a local data center, compared with about 12% of everyone else. The gap is just as striking on the opposition side: about half of non-daily users strongly oppose a local data center, while only about a quarter of daily users do.
4 key features significantly predict local support for data centers
Two other features had predictive value: leaning Republican and being promised a cash payout from the data center (in our assigned investment condition). We combined these predictive variables to fit a linear model that predicts likelihood to be a supporter or non-supporter. Each of the four independently and significantly predicted support, and each added a similar boost.
A participant with none of the four had a predicted 15% chance of supporting a local data center. Being Republican raised that by 16 percentage points, living near a data center by another 16, using AI daily by 18, and being offered a personal payout by 13. No single factor gets a community past 50% on its own, but any three of the four do. And because the model is additive, meaning each factor’s effect is the same regardless of the others, their order does not matter. In other words, if you took a sample of participants who live near a data center and are daily AI users - but don’t lean Republican - and added a pledge of a personal payout, you could expect to see 62% support. A sample that has all four of these features would yield a predicted 78% chance of support. This is a massive change from the average levels of support, which hover around 20%.
However, there’s a catch: this group is exceedingly rare or nonexistent. Out of 1,991 people we surveyed, only 84 of them were Republican daily AI users that lived near a data center. That’s about 4% of our nationally representative sample.
A data center can’t change how a community votes or how often its residents use AI. It can choose to build where data centers already exist, and it can pay residents directly. Those two levers alone raise predicted support from 15% to 44%. That’s nearly triple the baseline, but still short of a majority. This means that choices about where data centers go, and policies about how they engage the communities around them, are among the few levers that can move public support.
The Data Center Backlash is Not About Data Centers
Public opposition to infrastructure projects is not new. Residents often resist facilities that bring local costs with no benefits to their communities, even when they acknowledge that there are national-level benefits. This pattern is referred to as “NIMBYism” (“not in my backyard”), and for decades it has shaped debates over projects ranging from affordable housing to nuclear power plants. Data centers are the latest facilities to trigger this kind of resistance. But is opposition to data centers just NIMBYism, or is something else driving it?
We ran a randomized survey experiment comparing data centers to three other facilities to identify the why behind the opposition: distribution warehouses, nuclear power plants, and landfills. Each of these facilities have been known to be the subject of NIMBYism, and they share some features with data centers but differ on others. This allows us to isolate the features that are driving the opposition to data centers. For example, distribution warehouses and data centers share an association with Big Tech (Amazon for warehouses, AI for data centers), but differ on the amount of resources they consume and the amount of permanent jobs they bring to the area. Accordingly, this allows us to isolate the effect of a Big Tech association from the effect of local economic benefits. Landfills take up space but are an obvious necessity. Nuclear power plants bring along environmental concerns and obvious hazards, but aren’t associated with big business.
show code
library(rempsyc); library(tibble)on <-"✓"; off <-""#### Table 1: Developer & technology ####tbl_developer <-tibble(Facility =c("Nuclear power plant", "Warehouse", "Landfill", "Data center"),`Big tech`=c(off, on, off, on),`Tied to AI`=c(off, off, off, on),`Newer / unfamiliar technology`=c(off, off, off, on),`Privately owned`=c(off, on, off, on),`Often run by a public utility or city`=c(on, off, on, off))nice_table(tbl_developer,title =c("Table 1", "Facility Features: Developer and Technology"))
Table 1
Facility Features: Developer and Technology
Facility
Big tech
Tied to AI
Newer / unfamiliar technology
Privately owned
Often run by a public utility or city
Nuclear power plant
✓
Warehouse
✓
✓
Landfill
✓
Data center
✓
✓
✓
✓
show code
#### Table 2: Costs & concerns ####tbl_costs <-tibble(Facility =c("Nuclear power plant", "Warehouse", "Landfill", "Data center"),`Requires a lot of land`=c(on, on, on, on),`Environmental concerns`=c(on, off, off, on),`Safety concerns`=c(on, off, off, off),`Property value concerns`=c(off, off, on, on),`High water use`=c(on, off, off, on),`Strains grid / raises utility bills`=c(off, off, off, on),`Health concerns`=c(on, off, on, off),`Heavy truck traffic`=c(off, on, on, off),`Noise`=c(off, on, off, on),`Odor`=c(off, off, on, off),`Eyesore`=c(on, on, on, on))nice_table(tbl_costs,title =c("Table 2", "Facility Features: Costs and Concerns"))
Table 2
Facility Features: Costs and Concerns
Facility
Requires a lot of land
Environmental concerns
Safety concerns
Property value concerns
High water use
Strains grid / raises utility bills
Health concerns
Heavy truck traffic
Noise
Odor
Eyesore
Nuclear power plant
✓
✓
✓
✓
✓
✓
Warehouse
✓
✓
✓
✓
Landfill
✓
✓
✓
✓
✓
✓
Data center
✓
✓
✓
✓
✓
✓
✓
show code
#### Table 3: Benefits ####tbl_benefits <-tibble(Facility =c("Nuclear power plant", "Warehouse", "Landfill", "Data center"),`Many permanent local jobs`=c(on, on, off, off),`Directly serves local residents`=c(on, on, on, off),`Main benefits go to people outside the community`=c(off, off, off, on),`Essential public service`=c(on, off, on, off),`Large local tax revenue`=c(on, off, off, on),`Produces something tangible`=c(on, on, on, off))nice_table(tbl_benefits,title =c("Table 3", "Facility Features: Benefits"),note ="Tangible outputs include power, goods, and waste removal.")
Table 3
Facility Features: Benefits
Facility
Many permanent local jobs
Directly serves local residents
Main benefits go to people outside the community
Essential public service
Large local tax revenue
Produces something tangible
Nuclear power plant
✓
✓
✓
✓
✓
Warehouse
✓
✓
✓
Landfill
✓
✓
✓
Data center
✓
✓
Note. Tangible outputs include power, goods, and waste removal.
We recruited a politically representative sample of 809 American adults and randomly assigned them to evaluate one of four facilities. We introduced the hypothetical facility using the prompts listed in the table below.
show code
tbl_conditions <-tibble(Condition =c("Nuclear power plant", "Warehouse", "Landfill", "Data center"),`Prompt`=c('Imagine a company is interested in building a nuclear power plant within 10 miles of your home to generate electricity for the regional power grid', 'Imagine a company is interested in building a distribution warehouse within 10 miles of your home to store and ship products to consumers and businesses', 'Imagine a company is interested in building a landfill within 10 miles of your home to manage household and commercial waste from the surrounding region', 'Imagine a company is interested in building a data center within 10 miles of your home to support AI technology'))nice_table(tbl_conditions,title =c("Table 4", "Four conditions and prompts"))
Table 4
Four conditions and prompts
Condition
Prompt
Nuclear power plant
Imagine a company is interested in building a nuclear power plant within 10 miles of your home to generate electricity for the regional power grid
Warehouse
Imagine a company is interested in building a distribution warehouse within 10 miles of your home to store and ship products to consumers and businesses
Landfill
Imagine a company is interested in building a landfill within 10 miles of your home to manage household and commercial waste from the surrounding region
Data center
Imagine a company is interested in building a data center within 10 miles of your home to support AI technology
Americans would rather live by a nuclear power plant than a data center
We asked all respondents how they would feel about their respective condition’s facility being built within 10 miles of their home. Each person was randomly assigned to evaluate just one of four facilities: a data center, a landfill, a nuclear power plant, or a distribution warehouse. Because no one saw more than one, respondents judged each facility on its own terms rather than ranking them against each other. This allows us to rigorously isolate the psychological drivers behind NIMBYism for each facility and the associations with the particular parameters of each type of facility.
Of the four facility types we tested, people liked distribution warehouses the most and landfills the least. Nuclear power plants have had a decades-long reputation for being notoriously difficult to site because resistance to them is so widespread and intense. But opposition to data centers now outpaces opposition to nuclear power plants. Americans would rather live by a nuclear power plant than a data center.
Data centers are the only facility where benefits are lowest at home
We also asked participants to report how much they expected each facility to benefit five groups: themselves, their neighborhood, their town or city, their state, and the country. In general, NIMBYism is a story about local costs outweighing local benefits, and that’s what we found. People’s willingness to accept a local facility was shaped in large part by whether they expected the facility to bring benefits to their town or city. Warehouses bring the most local benefits and are the most welcome. Data centers ranked the lowest on their potential to bring local benefits. At the national level, people acknowledge that data centers render benefits to almost the same extent as nuclear power plants. But data centers were the only facility whose expected benefits grew the farther they reached from home: low at the local level, higher at the state level, and highest at the national level. This is a key pattern that drives NIMBYism: people do not want to suffer local costs for a project that is going to render benefits away from home.
People are tolerant of having a warehouse nearby because they expect them to bring benefits locally. In particular, they expect jobs. We asked respondents to write down the thoughts they had as they evaluated their new facility, which allows us to capture the concepts that come to mind naturally. These free responses revealed that people were evaluating data centers and warehouses on a dimension of economic benefits to the town, and were evaluating nuclear power plants on a safety and risk dimension. 1 in 3 people spontaneously brought up “jobs” when evaluating a warehouse, and all of those responses were positive. By contrast, 15% of respondents mentioned “jobs” when evaluating data centers, and 1 in 5 of those responses expressed skepticism.
show code
tbl_jobsQutoes <-tibble(Warehouse =c("“[A warehouse] could create jobs and help the local economy”", "“Jobs!! Economic boost to local restaurants and businesses.”"),`Data center`=c('“[Data centers] provide no financial benefit to residents and provide no jobs”', '“AI will ... take away jobs”'))nice_table(tbl_jobsQutoes,title =c("Table 5", "Quotes from respondents free responses that mentioned 'jobs'"))
Table 5
Quotes from respondents free responses that mentioned 'jobs'
Warehouse
Data center
“[A warehouse] could create jobs and help the local economy”
“[Data centers] provide no financial benefit to residents and provide no jobs”
“Jobs!! Economic boost to local restaurants and businesses.”
“AI will ... take away jobs”
Data center opposition is driven by builder opportunism and a distrust in the federal government
We asked people to consider the type of company that was behind the facility and to rate how opportunistic they are: Would they would break promises and alter facts to get what they want, and are they willing to hurt the community to get ahead? Of the four builders behind the facility types we tested, data center developers seemed the most opportunistic and warehouse builders the least. This goes hand-in-hand with people’s expectations for locally rendered benefits: people view data center builders as opportunistic because they aren’t bringing local benefits; people view warehouse builders as more invested in the community because they expect jobs.
This is surprising because distribution warehouses bring Amazon to mind, and data centers bring AI to mind, both of which fall under Big Tech. It wasn’t long ago that Amazon was the go-to corporate villain. But now, it seems that there is a bigger, badder villain on the scene: AI infrastructure that is taking land and resources without offering jobs, and, in some people’s minds, will eliminate jobs. The perception that data center developers are opportunistic builders is a big driver behind opposition. Among people who saw data center developers as opportunistic, 90% did not support a local build. Among those who didn’t, 41% did not support a local build. Distrust of the developer helps explain why data centers face so much more opposition than warehouses.
Distaste for local data centers is also driven by Americans’ distrust in the federal government, Big Tech, and Big Business. We asked respondents how much trust they have in four American institutions: Big Tech, Big Business, the federal government, and their local government. Trust for the first three is very low across the board. People reported that they “somewhat distrust” those three institutions, an average of -1 on a scale from -2 to +2. That distrust doesn’t factor into their evaluations of landfills, nuclear power plants, or warehouses. But it matters big when they evaluate data centers. People who distrust the federal government are 32% more likely to oppose a local data center.
The data center backlash is not actually about the data centers
This is all pointing to the broader theme: the data center backlash is not actually about the data centers. When people evaluate landfills or warehouses or nuclear power plants, they are evaluating the pros and cons of that facility. When people evaluate data center propositions, they are passing judgment on the AI industry at large. And their distrust for the federal government translates to a sense of economic instability brought on by AI and failed to be regulated by the government. Which means that those that trust the federal government feel less threatened by the AI industry, and by-proxy, data centers.
Indeed, people only mentioned distance (“miles”) 1.5% of the time when they were thinking about a local data center. And the three most common words were water, AI, and environment - notably abstract concepts. In the three other conditions, respondents were talking about the actual facility in concrete terms. The most common words across the three other conditions were: “smell,” “miles,” close,” “safe,” and “traffic.” When people evaluated warehouses or landfills or power plants, they asked themselves: am I close enough to smell it, hear it, or sit in its traffic? Almost nobody did for data centers.
Underneath it all is a crisis of trust. People are uneasy about economic instability and the pace at which things are changing under their feet - and they don’t trust the federal government to steady them. They also don’t trust data center developers to act in their interest. Perceived opportunism and expected benefits go hand in hand: the people don’t expect data centers to render local benefits because they think data center developers are out for themselves, and vice versa. The two beliefs reinforce each other, which means either end has a lever to pull. Concrete, visible local benefits can build trust, and more trust makes promised benefits more believable. Offering residents some form of local benefit that promises to support local economic growth and offer some sense of economic stability will go a long way toward restoring trust and cooling opposition. But developers will need to convince people that they are serious when they offer to make an investment.
