Why Data Center Construction Is Facing a Growing Public Backlash

Data center construction is becoming a harder sell in the United States because AI companies are promising massive economic benefits while many communities see rising electricity bills, water use, secrecy and weak local accountability. A new Data & Society report says opposition is being driven less by ideology than by a widening trust gap between the public and the companies racing to build the infrastructure behind artificial intelligence.

The report, released Tuesday, comes as frontier AI labs and their backers are planning trillions of dollars in computing infrastructure, even as surveys show more than 60% of Americans want new data center building limited in their communities. The mismatch is now showing up in real-world project delays, with Data Center Watch estimating that $68 billion in projects were disrupted by local opposition in the second quarter of 2026 alone.

What Data & Society found in Pennsylvania helps explain why the backlash is spreading. Over 18 months, researchers interviewed 44 residents across the state and concluded that resistance is rooted in lived experience: industrial promises that have not always delivered, local officials who increasingly appear out of reach, and a sense that the AI boom is being imposed rather than negotiated.

Why are data center projects drawing more resistance?

Data center projects are facing resistance because many residents think the costs are landing locally while the gains are captured elsewhere. The issue is not just aesthetics or land use; it is about power prices, property values, water consumption, public health fears and mistrust of the companies pushing the projects.

The AI industry argues that more computing capacity is essential to improving models and keeping pace with demand. That case has persuaded investors and executives, but it has not translated into broad public enthusiasm for the buildings, substations, transmission lines and cooling systems required to support them.

In Pennsylvania, the report suggests that skepticism is intensified by the state’s history. Coal, oil, steel and fracking each arrived with claims of prosperity and modernization. For many residents, those eras also left behind pollution, injury, broken promises and communities that felt used up once the economic cycle moved on.

“People in Pennsylvania especially have incredible experience with industrial change,” said Maia Woluchem, a program director at Data & Society and a co-author of the report. “They have deep lineages of industrial trauma, enough to sniff when things don’t feel right to them.”

The report does not claim that every data center project is harmful in the same way or that every community will react identically. Instead, it shows that local objections are increasingly shaped by a broad suspicion of large industrial schemes, especially when the people promoting them sound certain that resistance is irrational or outdated.

What the new report found in Pennsylvania

The Data & Society study is based on ethnographic fieldwork conducted between 2024 and 2026. Rather than measuring emissions, jobs or tax revenue, the researchers wanted to understand how residents experience the debate over data center construction and how they decide whom to trust.

That distinction matters. The AI industry often frames the debate as a technical question about energy, land and investment. The report argues that, on the ground, the conflict is social and political: Who gets consulted, who benefits, who bears the risk, and whether communities believe the people making promises.

Researchers said the industry’s preferred message — that AI expansion is inevitable — often backfires. Residents may hear inevitability not as a neutral forecast but as an attempt to foreclose public discussion before it even begins.

“One of the strengths of ethnography and doing research on the ground for an extended period of time is piercing the bubble of AI hype,” said Livia Garofalo, a Data & Society researcher who worked on the report. “People are making sense of the dissonances between AI being the next inevitable industrial revolution, and fears about AI.”

That tension is central to the report. AI is presented in one context as transformative infrastructure comparable to past industrial revolutions, yet in another as a speculative technology whose social costs remain uncertain. Residents are being asked to accept both frames at once, and many are not persuaded.

How did Pennsylvania go from AI boosterism to new restrictions?

Pennsylvania shifted from enthusiastic promotion of AI infrastructure to tighter oversight in less than a year. In July 2025, Governor Josh Shapiro highlighted a summit celebrating $90 billion in investment tied to data centers and AI infrastructure. By the following year, the state had moved in the opposite direction.

Shapiro later signed an order adding new requirements for data center projects and removing them from a regulatory fast track. That reversal suggests state leaders saw enough local concern to conclude that accelerated approval could create political and practical problems.

The report uses that change as evidence that skepticism is not a fringe concern. Even officials who had once helped market the sector as a job-creating opportunity were forced to acknowledge that approval processes and community confidence were becoming harder to separate.

Who is backing the projects, and why?

Labor unions are among the strongest supporters of data center construction because they see the projects as a source of near-term work and a way to rebuild membership. Even when unions acknowledge that the facilities may not employ many people once operating, the construction phase and related development can still matter in regions that have lost industrial jobs.

That support is important because it gives developers a credible local ally. But the report suggests it does not erase broader concerns. In communities where people are worried about utility rates, property values or land use, job promises may not be enough to overcome suspicion.

The researchers also argue that the politics of the issue cannot be reduced to a simple pro-growth versus anti-growth divide. Instead, the coalitions forming around data center fights are fluid and often surprising.

Why the opposition is described as “post-partisan”

The report’s authors say resistance is best understood as post-partisan because it crosses standard political categories. Some opponents are motivated by traditional anti-development instincts. Others object to AI as a technology, to the corporations behind it, to environmental impacts or to public-health concerns that are still vague but emotionally powerful.

That mix makes the movement difficult for industry to discredit. If opponents were all driven by the same ideology, developers could tailor a single counterargument. Instead, they are facing a patchwork of objections that vary by neighborhood, county and local political culture.

One of the clearest unifying themes is distrust of process. Residents told researchers they were frustrated by nondisclosure agreements, limited public disclosure and the feeling that elected officials had stopped answering questions once a developer appeared.

“These fights are incredibly organic,” Woluchem said, adding that the debate is forcing people with different views to talk about what kind of future they want and what kind of present they are willing to accept.

How are data center developers shaping local debates?

Developers and their allies are trying to frame data centers as a public good, but the report says their advocacy often resembles a coordinated influence effort. The researchers singled out the Data Center Coalition, a trade group that has promoted projects through websites such as Pennsylvania Connects and spent at least $19,000 lobbying the state government last year.

The exact scale of that effort may be small relative to the industry’s overall investment plans, but the symbolism matters. To critics, it reinforces the idea that developers are trying to shape public opinion while keeping key details out of view.

That impression is particularly damaging in places where local residents already feel sidelined by technical language, complex utility planning and closed-door negotiations. If people think the entire process is designed to move ahead regardless of what they say, even a generous jobs proposal may not restore trust.

What local residents say concerns them most

The report identifies several recurring worries that are practical rather than abstract:

• Higher household electricity bills if data center demand strains the grid.
• Lower property values near large industrial sites.
• Water consumption and the strain on local resources.
• Potential energy-related emissions from supporting infrastructure.
• Public-health questions tied to large-scale industrial development.
• Closed-door tactics, including nondisclosure agreements and limited transparency.

These concerns are not always easy for companies to answer because they are not all solved by a single mitigation plan. A project can promise tax revenue while still worrying residents about long-term utility costs or the cumulative effect on local land use.

That is one reason the debate has become so combustible. The industry often talks about AI in terms of abstraction, scale and strategic necessity. Residents experience it as a zoning fight, a utility fight and a trust fight all at once.

Why trust has become the central issue

Trust is now the core challenge for the AI infrastructure buildout because the public is being asked to accept a new kind of relationship with powerful firms whose products already shape daily life. In the view of the Data & Society researchers, the problem is not simply that communities dislike the buildings themselves. It is that the whole system feels extractive.

Garofalo said the contradiction is especially visible in consumer AI use. People may interact with chatbots in moments of loneliness or need and experience them as helpful, reassuring or emotionally responsive. At the same time, they know the service is commercial, data-driven and tied to companies that profit from the interaction.

“It is an intimate and relational trust that is now being inserted in people’s lives by these industries,” Garofalo said. “I’m alone in my house, and I’m talking to a chatbot, and I’m feeling loved and cared for, and yet I know that this is a commodity, and I know I’m being extracted upon.”

That sense of being simultaneously comforted and exploited helps explain why infrastructure disputes are not separate from product debates. The same firms selling AI tools are also asking communities to accept enormous physical facilities needed to power them. For many residents, those two asks now feel connected.

How big is the backlash to new data centers?

The backlash is large enough to affect investment timelines, permitting strategies and state politics. Surveys cited in the report and broader industry tracking suggest opposition is no longer confined to environmental advocates or a few anxious suburbs. It has become an organizing issue in towns and counties across the country.

Public concern is also rising quickly. The report notes that long-running survey data show sentiment growing more negative, which means the industry is not just facing current resistance but a trend that may harden over time if trust does not improve.

That trajectory matters because data center developers are planning at a scale that requires many local approvals. Unlike a software product that can be rolled out remotely, a data center needs land, water, power, substations, transmission upgrades and a permitting path that often depends on elected officials and planning boards.

If even a modest share of communities conclude that these projects are not worth the trade-offs, the industry’s growth assumptions could face bottlenecks. The more local opposition spreads, the more project economics become uncertain.

What comes next for the AI infrastructure boom?

The immediate future is likely to feature more fights, not fewer. AI companies still need compute, investors still expect growth and developers still see data center construction as one of the clearest physical manifestations of the AI economy. But the political environment around those projects has changed.

The Data & Society report suggests that winning over communities will require more than generic claims about jobs and innovation. It will require transparency about utility impacts, more meaningful public participation and a willingness to treat residents as decision-makers rather than obstacles.

That is a difficult shift for an industry built on speed. It is also a necessary one if companies want to avoid the pattern now visible in Pennsylvania and elsewhere: hype at the top, suspicion at the local level, and a widening gap between corporate certainty and public consent.

For now, the central lesson is that data centers are no longer just an engineering issue. They are a social test of whether a technologically ambitious industry can gain permission from the people expected to live next to its footprint.

Timeline of the Pennsylvania data center debate

The state’s recent path shows how quickly attitudes can change once public scrutiny increases.

1. 2024: Data & Society researchers begin interviewing Pennsylvania residents as the data center debate develops.
2. July 2025: Governor Josh Shapiro headlines a summit celebrating $90 billion in AI and data center investment.
3. 2025-2026: Local skepticism grows as residents raise concerns about bills, water use, land use and secrecy.
4. 2026: The governor signs an order adding new requirements and removing data center projects from a fast-track process.
5. Q2 2026: Data Center Watch says $68 billion in projects are disrupted by local opposition.
6. Sept. 22, 2026: Data & Society publishes its report on the roots of community resistance.

That sequence suggests the argument over AI infrastructure is moving from theory to governance. What began as a story about future technology is now a story about who gets consulted, who bears risk and whether communities believe they can slow the pace of change.

If the AI boom continues, the question will not only be how much compute the industry can finance. It will also be how many communities it can persuade to host the concrete, steel and power systems that make that compute possible.