The Evolution of Data Center Siting Governance in the U.S.: Zoning, Compensation, and Collaboration

• In Event: The Infrastructure of Intelligence: Navigating Siting Conflicts, Community Equity, and Grid Resilience in the AI Era

Abstract

The demand for hyperscale data centers has surged, and data centers have become critical infrastructure supporting the rapid expansion of generative AI since 2023. This rapid growth in energy use has raised concerns about rising electricity demand and electricity prices (Chen et al. 2025; Karimi et al. 2022; Mytton and Ashtine 2022; Libertson et al. 2021). Data centers may also impose local burdens related to water scarcity, noise pollution, limited local job creation, and gentrification (Privette et al. 2026; Fang and Greenstein 2025; Monstadt and Saltzman 2025; Siddik et al. 2021). In response to these issues, local governments, developers, and residents have clashed, collaborated, and developed a range of solutions to solve these siting challenges.

Previous not-in-my-backyard literature often discusses hazardous facility siting solutions in silos, such as direct (re)zoning regulation, compensation, communication, and collaboration. However, existing data show that different cities and states have often combined these solutions in creative ways for data center siting, and that these package deals have evolved quickly since 2023. Drawing on theories of policy diffusion and risk perception and management, this research asks: How have city-level data center siting solutions evolved and diffused since 2023? What factors drive these changes?

To answer these questions, we compile a national corpus of data center siting documents through web scraping. Using a large-language-model-fine-tuned BERTopic for topic modeling, we computationally identify patterns in siting solutions and construct a panel dataset of data center siting cases in the U.S. In this novel author-constructed dataset, we also combine residents’ engagement records, national and local media coverage, pre-existing data center presence, neighboring city’s siting package, and local political and economic variables. This integrated data enables a comprehensive analysis of longitudinal and cross-sector decision-making.

Methodologically, we use event history analysis based on multistate survival models to examine how data center siting solutions have evolved over time and what factors drive these changes. Our preliminary result shows that, under the influence of residents, media coverage, and political factors, data center siting packages gradually shift from the establishment of zoning standards alone toward broader community benefit assurances and community collaboration.

This research is among the first to analyze data center governance at the national level. It provides a large-scale empirical setting to advance generalizable theories of policy evolution and diffusion, specifically regarding the complex tradeoffs among technological advances, economic development, and environmental and energy concerns. Furthermore, it contributes to the environmental policy and urban planning literature through a longitudinal and cross-sectional analysis of data center siting in the United States.

Author

• Ximing Dong, Arizona State University
  Presenting Author