Author(s) :

Katie Chung.

Volume/Issue :

Abstract :

The rapid growth of Artificial Intelligence (AI) has sharply increased demand for data center infrastructure, raising concerns about energy use, greenhouse gas emissions, and environmental justice. This study examines AI-driven data center development in New Jersey, estimating effects on electricity demand, air pollution, and public health using a transparent screening framework. Facility-specific capacity data are compiled from publicly accessible sources, EPA eGRID 2022 average emission rates are applied under two utilization scenarios, public health effects are assessed using the EPA COBRA screening model, and environmental justice implications are evaluated using EPA EJScreen and NJDEP overburdened-community mapping tools. Screening results indicate that New Jersey’s disclosed data center capacity could expand 4.16-fold, increasing estimated electricity demand from approximately 2.06–2.74 TWh per year (2.8–3.7% of statewide retail sales) to 8.55–11.40 TWh per year (11.6–15.5% of statewide retail sales). Associated CO₂ emissions under full build-out are estimated at approximately 2.55–3.40 million metric tons per year, with modeled health damages reaching approximately $271–$450 million annually under high utilization. Several existing and planned facilities are located in or near communities already identified as socioeconomically vulnerable or facing elevated pollution burdens. Comparative analysis with Northern Virginia and Central Ohio illustrates how rapid, large-scale data center growth creates grid reliability, cost allocation, and environmental equity challenges in the absence of proactive policy frameworks. These findings support improved facility-level disclosure, targeted tariff design, cleaner incremental electricity supply, and rigorous tract-level environmental justice review.