Conference Agenda

This study examines the relationship between data center expansion, environmental sustainability, and inclusive economic growth across major U.S. digital infrastructure hubs. We integrate EPA Greenhouse Gas Reporting Program (GHGRP) emissions data with Mastercard Inclusive Growth Scores for Northern Virginia and national comparison regions (Santa Clara, Maricopa, and Douglas counties) from 2017 to 2024. Analysis reveals that Loudoun County, processing 70\% of global internet traffic, experienced CO₂ emissions growth from near-zero to 160 million metric tons annually, with power consumption exceeding 50,000 MW by 2024. Despite substantial economic returns (\$700M+ in tax revenue), inclusive growth scores remained modest and volatile, indicating unevenly distributed benefits. Cross-regional analysis exposes distinct sustainability profiles: Santa Clara maintains high inclusive growth (62-66) with moderate emissions; Maricopa dominates national emissions (2+ billion metric tons) with lower inclusive growth. Machine learning models (Random Forest) achieved strong predictive performance (test R² > 0.84) using census tract socioeconomic indicators. Business-as-usual projections forecast Northern Virginia's CO₂ emissions reaching 308,600 metric tons by 2030. However, AI-driven optimization scenarios demonstrate substantial mitigation potential: combined efficiency improvements (40\% reduction) could save 123,440 metric tons regionally—equivalent to removing 26.8 million vehicles annually. Nationally, similar interventions could eliminate emissions equivalent to 250.3 million vehicles. Results reveal persistent decoupling between environmental intensity and inclusive growth, indicating current models fail to distribute benefits equitably. Findings provide evidence-based pathways for sustainable data center expansion integrating renewable energy, advanced cooling, AI optimization, and explicit socioeconomic equity requirements.