AI models require massive computing power, leading to an unprecedented demand for water to cool data centers.

By 2030, U.S. data centers could require enough additional water capacity to match New York City's daily supply.

A UC Riverside study reveals that limited public water capacity is quickly becoming a major obstacle to data center expansion.

Tech companies are forced to partner with local communities, spending hundreds of millions on water infrastructure upgrades.

Data centers operate continuously, generating extreme heat. Liquid cooling techniques are efficient but highly water-intensive.

Even closed-loop systems consume huge amounts of water through evaporative cooling towers transferring heat outside.

Peak daily water demand for a large, modern data center can reach between 1 million and 8 million gallons.

Public water systems must engineer for maximum demand, making a data center's peak usage a critical planning factor.

Building new peak water capacity could cost up to $58 billion, with the financial burden often falling on local residents.

When water is unavailable, facilities use inefficient dry cooling, increasing electricity demand and stressing summer power grids.

Experts urge data centers to report peak demand and provide corporate funding for upgrades so residents don't pay the price.