Author: Brian Merchant
Compiled and organized by BitpushNews
Data centers protests have swept across the United States. From Vermont to Oklahoma, from Indiana to California, communities are organizing to halt the tech industry’s frantic expansion of data centers in their backyards. This week, New York’s legislature passed a one-year moratorium on data center construction, which has now been sent to the governor’s desk for signature. The governor of Illinois has also suspended tax incentives for data centers. Few issues have stirred such political passion or unity; the rare bipartisan consensus emerging in the United States for 2026 is a shared rejection of data centers and hostility toward artificial intelligence (AI).
This article presents an exclusive report from a data scientist who, through deep data analysis, precisely identified who is obstructing data center construction and how successful these protests have been.
If you think I'm exaggerating, take a look at a recent survey released by Heatmap, which polled over 4,000 Americans about their attitudes toward data centers and whether they support building such facilities near their homes.
The results show that public sentiment toward data centers has become irreversibly negative. Polls reveal that 55% of Americans “strongly oppose” the construction of data centers in their regions. This represents a “record low, revealing a dramatic shift in public opinion toward the facilities underpinning the AI boom.”
Profile of the group expressing opposition
Opposition is particularly strong among Democrats, rural residents, and young people: among respondents aged 18 to 35, as many as 80% oppose data centers. (This aligns perfectly with the current overall public sentiment; other polls and countless anecdotes have long confirmed that Generation Z holds deep-seated hostility toward AI—just look at the torrent of boos that greeted commencement speeches this summer promoting AI.)
However, as readers are well aware, there has been ongoing debate and skepticism regarding the driving forces and nature of this growing boycott movement.
Some have firmly argued that opposition to data centers is merely conservative NIMBYism (“Not In My Backyard”) and is driven by affluent environmentalists wearing Patagonia. While the absolute number of Americans opposing data centers in the Heatmap survey suggests otherwise, the poll did not specifically examine these class-based factors.
If you want to challenge this view—as I, Astra Taylor, and Saul Levin have, arguing that the data center opposition movement is fundamentally rooted in working-class politics—then having solid data becomes crucial, and that’s where data scientists come in. After I published my report on “data center rebellion” (which relied on my original interviews and a review of national news coverage), researcher Geoff Holtzman reached out to share his quantitative analysis of the movement, centered on identifying who was truly participating in the protests.
Holzmann describes himself as a philosopher and data scientist who writes about quantitative propaganda and scientistic rhetoric, and he frequently publishes articles in his Science & Power newsletter. His peer-reviewed work has appeared in prestigious journals such as the Proceedings of the National Academy of Sciences (PNAS) and the American Journal of Bioethics. He also heard the widely circulated claim that the data center protest movement is dominated by affluent NIMBY groups, so he decided to investigate. He conducted a comparative analysis of current and proposed data center projects against U.S. Census data (Footnote 1), and he has agreed to exclusively share his findings here. He reached at least three striking conclusions.
1. The resistance rate to data centers in the poorest communities is nearly five times higher than in the wealthiest communities (19.0% vs. 3.8%).
(Note: These quartiles are calculated only for census tracts within the dataset, not for the entire United States.)
“The most frequent resistance comes from communities with a median household income between $8,000 and $72,000,” Holzman noted. “The communities with the lowest resistance rates have average annual household incomes between $133,000 and $250,000.”
This directly dispels the political myth that opposition to data centers is led by affluent, Patagonia-wearing upper-middle-class individuals; protests in low-income or working-class communities occur far more frequently than in affluent areas.
As Holzman stated: "Setting up computing centers in higher-income regions is significantly easier for tech companies, purely from a pragmatic standpoint, regardless of any moral or justice concerns."
He added, "In communities with low income and low education levels facing project proposals, the communities with the lowest income and lowest education levels resist the most fiercely." Meanwhile:
The level of protest from communities with high education and income levels is unusually low. Regarding the potential role of homeownership, we are not talking about old money resisting affordable housing—we are talking about people who may be living in affordable housing themselves.
In addition, Holzman’s data confirms that the data center resistance movement is working. We have seen numerous headlines about development projects being canceled or scaled back—just this week, under intense public pressure, Governor Ken O'Leary’s massive project in Utah was halved in size. Other projects have been canceled entirely.
According to Holzmann's analysis:
2. Proposed data centers that have recently faced opposition are more than five times as likely to be canceled or paused compared to those without protests (28.2% vs. 5.2%).
This is an astonishing figure. When new data center proposals face community opposition, nearly one-third of projects are ultimately canceled, paused, or shut down. This is an exceptionally impressive success rate and should further inspire organizers of anti-data center movements considering whether to launch a new round of resistance.
Finally, combining the insights from the previous two points, Holzman found:
3. The fact that projects in low-income regions have the highest cancellation rate can be fully explained by their higher resistance rates.
“In communities that resist, projects are six times more likely to be canceled than in those that comply,” Holzman noted. He added: “The rise in cancellation rates in low-income areas is entirely due to the high levels of protest in these communities. Therefore, continuing to propose projects in these areas may fuel greater public anger, provoke stronger resistance, and further increase the likelihood of project cancellations.”
I hope this data helps dismantle the arrogant misconception that data center resistance movements are led by affluent NIMBY elites. In fact, the vast majority of those standing up in resistance are working-class residents and communities. I also hope these findings provide a powerful tool for cities, residents, and organizers facing the challenges of data center development.
I again extend my sincere thanks to Holzmann for allowing me to publish these research findings on my blog. For those interested in studying or further validating his data, he has hosted the entire code repository on GitHub.
The United States has overall become the most resistant country to new data centers.
Data comes from the research firm Public First (thanks to Molly Taft, reporter at WIRED): How did the United States, as the epicenter of AI prosperity, become its own stumbling block? Our investigation offers several explanations.
– Knowledgeable opposition
The public is now more aware of what AI is, what it does, and what data centers are and what they do. When we conducted our AI survey five years ago, it was at best a niche interest. Today, we see clear growth in public awareness and understanding, along with more mature usage of tools, particularly among those aged 25–44. Our analysis of who understands AI needs to shift from “who has opened a large language model” to “who uses large language models in complex, integrated ways.”
Our survey shows that the U.S. ranks at a moderate level in self-reported awareness of data centers, higher than other “developed” markets. Given the widespread presence of data center construction in the U.S., this is not surprising.
This "informed opposition" makes it more averse to data centers than any other country surveyed, interestingly!
Note 1:
According to Holzmann: I used the 5-year American Community Survey data from 2020–2024, so the income figures are typically somewhat lower than you might expect. I needed to do this to obtain data at the census tract level; therefore, for the national median, I stuck with the same dataset.
Twitter: https://twitter.com/BitpushNewsCN
BitPush Telegram community: https://t.me/BitPushCommunity
BitPush TG subscription: https://t.me/bitpush