Energy and Water Usage from Data Centers
Faculty Mentor Information
Dr. Kendra Kaiser, University of Idaho; and Dr. Mitchell Meads, University of Idaho
Presentation Date
7-15-2026
Abstract
Historically, technological advancements generate widespread economic wealth and productivity but leaves local communities to absorb the environmental costs. Today’s rapid expansion of AI data centers follows this trend, driving unprecedented resource demands, with limited access or knowledge of their associated impacts. Using one of the only publicly accessible datasets on natural resource use of data centers, this project analyzed Meta’s data center resource demands to quantify the impacts data centers have on local energy and water resources. Temporal trend and correlation analyses between energy and water resource withdrawals were evaluated at the site-level for all of Meta’s global, active data centers. Over an 8 year period from 2017-2024, Meta's data center footprint grew from 6 to 18 data centers; roughly 0.43 to about 1.5 billion gallons of water withdrawal per year; and about 2.5 to 18.4 million MegaWatt hours of electricity usage per year. By illustrating these resource consumption trends, this project enables regional planners and policy makers to use a data-driven approach and resource allocation policies to make effective decisions for data center proposals.
Energy and Water Usage from Data Centers
Historically, technological advancements generate widespread economic wealth and productivity but leaves local communities to absorb the environmental costs. Today’s rapid expansion of AI data centers follows this trend, driving unprecedented resource demands, with limited access or knowledge of their associated impacts. Using one of the only publicly accessible datasets on natural resource use of data centers, this project analyzed Meta’s data center resource demands to quantify the impacts data centers have on local energy and water resources. Temporal trend and correlation analyses between energy and water resource withdrawals were evaluated at the site-level for all of Meta’s global, active data centers. Over an 8 year period from 2017-2024, Meta's data center footprint grew from 6 to 18 data centers; roughly 0.43 to about 1.5 billion gallons of water withdrawal per year; and about 2.5 to 18.4 million MegaWatt hours of electricity usage per year. By illustrating these resource consumption trends, this project enables regional planners and policy makers to use a data-driven approach and resource allocation policies to make effective decisions for data center proposals.