A Critical Examination of Intersecting Policy, Socioeconomic, Energy, and Environmental Health Concerns About Data Centers

Session Submission Type: Panel

Abstract

The rapid expansion of artificial intelligence and cloud computing has fueled an unprecedented boom in data center construction across the United States. Although AI offers major potential benefits for economic productivity, infrastructure supporting it is far from immaterial. Data centers consume large amounts of electricity and water, generate heat and noise, and can impose substantial burdens on host communities. Yet these local consequences and voices remain poorly understood, especially as facilities are increasingly sited based on land availability and affordability rather than community readiness or resilience. This panel brings together four original studies that use quantitative, qualitative, and multiscale approaches to examine the policy, environmental, health, socioeconomic, and energy justice implications of data center expansion.The first paper asks whether data center growth redistributes power reliability within utility service territories. Using a novel dataset that combines data center locations with high-resolution outage data, the authors find that communities near data centers experience improved grid reliability, while more distant communities in the same utility territory face greater outage risk, especially during extreme weather. These distributional effects are often hidden when reliability is measured only at the utility or county level, with important implications for energy equity, grid resilience, and infrastructure planning.The second paper examines the air quality impacts of the xAI data center, powered by 15 natural gas turbines, on low-income, majority African American communities in Southwest Memphis. Drawing on EPA dispersion modeling, satellite aerosol data, and ground-based monitoring, the authors find that although xAI’s operations have not statistically significantly worsened air quality since its September 2024 launch, modeled PM2.5 concentrations exceed daily and annual national standards. The study argues that the central issue is not xAI alone, but the community’s already elevated pollution burden, underscoring the need for sustained community-based monitoring and broader regional mitigation.The third paper extends environmental justice frameworks to data center infrastructure by asking whether data centers function as a new form of locally unwanted land use. Using county-level regression analyses of data center locations, Census indicators, voter turnout, and partisan composition, the authors find that facilities are more likely to be located in lower-income, less-educated, and more racially diverse counties. Counties with lower voter turnout and stronger Republican vote shares also host disproportionately more facilities.The fourth paper examines the policy trade-offs of rapid data center expansion across the Southeast and central Appalachia, where growth is increasingly linked to fossil fuel-based energy development. Using national case studies, stakeholder interviews in Georgia, Tennessee, and South Carolina, and a national survey with regression analyses, the study explores how data center development affects community health, environmental quality, energy costs, and pollution burdens. It also analyzes how institutional trust, procedural justice, place attachment, and regulatory frameworks shape public acceptance, resilience, and views of microgrids and nuclear energy. These papers establish data centers as a consequential and understudied frontier in public policy research and highlight gaps in regulation and monitoring, and offer timely insights for researchers, policymakers, utility regulators, and community advocates confronting the hidden costs of the data center economy.

Policy Area

• Innovations in Science and Technology

Chair

• Sanya R Carley, University of Pennsylvania

Discussants

• Gang He, Baruch College (City University of New York)

• Anthony Rydell Harding, Georgia Institute of Technology

Organizer

• Chien-fei Chen, Clemson University

Individual Presentations

• Do Data Centers Redistribute Power Reliability? - Presenting Author: Brian Y. An, Georgia Institute of Technology

• Impact of the xAI Data Center on Air Quality in Southwest Memphis - Presenting Author: Chunrong Jia, The University of Memphis

• Not the Path of Least Resistance: Ideological Climate and the Siting of U.S. Data Centers - Non-Presenting Co-Author: Yookyung Chin, Korea Advanced Institute of Science and Technology; Presenting Author: Dasom Lee, Korea Advanced Institute of Science and Technology (KAIST)

• Understanding Risk-Risk Trade-offs of Data Centers: A Mixed-Methods Approach - Presenting Author: Chien-fei Chen, Clemson University; Non-Presenting Co-Author: Mary D Willis, Boston University; Non-Presenting Co-Author: Yu Wang, Iowa State University