Key facts
- More than 60% of Americans favor limiting new data centers.
- Data center projects worth $68 billion were disrupted by local opponents in Q2 2026.
- A new report from Data & Society details opposition to data centers based on 18 months of ethnographic fieldwork.
- Researchers interviewed 44 Pennsylvanians between 2024 and 2026 for the report.
- Trade unions are a significant source of support for data center projects due to potential job creation.
- The Data Center Coalition, a trade group, has lobbied the state government and promoted projects.
Frontier AI labs are planning trillions of dollars in data center projects, believing that scaling computation is key to advancing artificial intelligence. However, public opinion surveys indicate that over 60% of Americans favor limiting new data centers, with this sentiment rapidly growing. In the second quarter of 2026 alone, data center projects valued at $68 billion faced disruptions from local opposition, according to Data Center Watch.
AI companies and investors are struggling to reconcile the increasing demand for their products with the resistance to the necessary infrastructure. The industry's attempts to attribute opposition to foreign interference are seen as unconvincing. A new report from the nonprofit Data & Society offers a detailed examination of the factors driving this opposition, based on 18 months of ethnographic research involving interviews with 44 Pennsylvanians between 2024 and 2026.
The report's authors, critical of how AI has been promoted, aim to understand residents' experiences with the debate surrounding data center construction. Livia Garofalo, a researcher at Data & Society, noted that ethnographic research helps to cut through the hype surrounding AI, as people grapple with the dissonance between AI being an inevitable industrial revolution and their fears about it.
Initially, data centers were promoted as drivers of economic development. In July 2025, Governor Josh Shapiro celebrated $90 billion in data center and AI infrastructure investments. However, a year later, he signed an order that introduced new requirements for data center projects and removed them from a fast-track regulatory initiative, signaling a shift in approach.
The report suggests that the industry's message of AI's inevitability has not resonated well with Pennsylvania residents. The state's history of industrialization, including coal mining, oil extraction, steel manufacturing, and fracking, has left residents wary of utopian industrial promises. Maia Woluchem, a program director at Data & Society and co-author of the report, stated that Pennsylvanians' experiences with industrial change and "industrial trauma" make them suspicious of such proposals.
Despite this opposition, trade unions remain a key source of support for data center projects due to the potential for job creation, viewing them as a means to rebuild their ranks, even if the long-term employment impact is limited. The researchers describe the efforts of AI proponents as a propaganda campaign, citing the Data Center Coalition's websites and lobbying activities.
Resistance to data centers has grown in diverse and unexpected ways. The opposition is characterized as "post-partisan," with a coalition forming from various groups. These include traditional NIMBYs, individuals concerned about AI technology and corporate interests, environmental groups worried about water usage and emissions, and those with general public health concerns.
Material interests are a significant driver of opposition. Residents are concerned about the impact of data centers' high electrical usage on their power bills, potential decreases in property values, and the tactics used by data center developers, such as requesting non-disclosure agreements from local officials. Woluchem highlighted instances where community members noticed increased developer presence at meetings and a subsequent lack of communication from their local council members on these issues.
This varied set of motivations for anti-data center activism poses a significant challenge for the AI industry, as there is no simple solution to address these widespread concerns. The core issue appears to be a lack of trust, exacerbated by the intimate and relational nature of interactions with AI technologies, which can feel extractive despite offering a sense of connection.