"This project introduces a new line of inquiry by shifting the study of data center energy impact from high-level, top-down modeling to a bottom-up, empirical analysis of actual residential economic harm," Garimella said.
Most New Jerseyans have noticed their home electricity bills have skyrocketed during the last couple of years. To discover the cause, a new Rutgers study will examine the impact AI data centers are having on rising household energy costs.
"The widespread deployment of Artificial Intelligence, particularly generative AI, has triggered an 'arms race' for computing power, leading to a massive, unprecedented build-out of energy-intensive data centers," said project lead SC&I Assistant Professor of Library and Information Science Kiran Garimella. "The International Energy Agency (IEA) projects this sector's electricity demand will more than double globally by 2026. This colossal, concentrated, and near-instantaneous new load presents a direct and novel threat to the stability of regional power grids and, critically, to the affordability of electricity for millions of Americans."
The project, titled "Measuring How AI Data Centers Affect New Jersey Household Electricity Bills," has received a seed grant from the New Jersey State Policy Lab (NJSPL), housed in the Rutgers Bloustein School of Planning and Public Policy.
While some academic and industry research has modeled this future grid strain, Garimella said a critical policy gap currently exists. "No one is measuring the current, real-world impact of this build-out on the monthly electricity bills of residents. This study is the first to fill this gap. The most important part of this study will be to quantify this cost, so residents can go to their local governments (who are subsidizing the cost of the data centers) and demand transparency about the financial impacts of data center construction. It's in the public interest that we, as researchers, do this work and demand accountability."
While some academic and industry research has modeled this future grid strain, Garimella said a critical policy gap currently exists. "No one is measuring the current, real-world impact of this build-out on the monthly electricity bills of residents. This study is the first to fill this gap."
Looking beyond New Jersey's borders, Garimella's team will also study the impact of data centers on household electricity costs in Pennsylvania, Delaware, Maryland, and Connecticut. These five states, Garimella said, are all key members of the PJM Interconnection, the largest regional transmission organization (RTO) in North America.
The study's methodology is also innovative. "Technically and methodologically, our project will move beyond traditional utility bill research, which relies on manual, self-reported data, a process that is labor-intensive, expensive, and prone to high error rates," Garimella said. "We will instead develop and deploy a suite of automated data collection tools (e.g., browser extensions). With user consent, these tools will securely log into a participant's utility portal and automatically export their past 24 months of billing data."
His team's plan to employ this "one-click" automation feature, Garimella explained, will radically lower the barrier to participation, reduce error, and create a secure, standardized data pipeline. He also added that their novel methodology is what will make a study of this scale feasible and, more importantly, scalable for future work.
Planning to share their conclusions widely, Garimella said, "The findings are essential for policymakers who are currently 'flying blind,' attempting to balance economic development (attracting data centers) with their primary mandate to protect consumers," Garimella said.
"The findings are essential for policymakers who are currently 'flying blind,' attempting to balance economic development (attracting data centers) with their primary mandate to protect consumers," Garimella said.
To share their findings, Garimella's team will write a policy brief for non-experts and deliver it directly to the NJ Board of Public Utilities, the NJ Governor's Office, and key committee members in the NJ State Legislature. They will also share a non-technical white paper with regional environmental justice organizations, consumer advocacy groups (e.g., AARP), and the press, and present their findings at major national AI/web academic conferences and prepare at least one peer-reviewed manuscript for a high-impact journal.
"If we do find out that we, as New Jersey residents, are paying extra for electricity because a data center has been built in our backyard, this study's findings should help us put pressure on these companies and the local governments to not put the burden on us – the public – to fund these data centers."
Learn more about the Library and Information Science Department at the Rutgers School of Communication and Information on the website.
You May Also Like
Rutgers researchers find most state residents have used at least one artificial intelligence tool and see the technology as helpful, but many say it will eliminate more jobs than it creates.