Original Source
Water Scarcity Emerges as New AI Data Center Bottleneck
AI Data Centers Face Enormous Water Demands
New artificial intelligence (AI) data centers are not only consuming vast amounts of electricity but also requiring significant quantities of water for cooling, posing a new challenge. Researchers from UC Riverside, the California Institute of Technology, and the Rochester Institute of Technology point out that data centers typically rely on water-evaporation cooling systems to manage the heat generated by their computers. These systems can consume millions of gallons of water daily.
Increased Strain on Local Water Supplies
Such high water demand from data centers often exceeds the available local water supply in proposed or built locations, necessitating the construction of new water infrastructure, operational delays, or a shift to alternative cooling methods that use less water but consume more electricity. Shaolei Ren, an associate professor of electrical and computer engineering at UC Riverside and a co-author of the study, highlighted that many communities are currently unable to support these facilities' water capacity needs. This challenge could be crucial for the AI industry, particularly in areas like San Francisco and the wider Bay Area.
*Source: San Francisco Examiner (2026-03-15)*
