Thirsty Algorithms: Why AI Is Straining the Tap

Research identifies systemic risks as AI infrastructure scales, citing transparency gaps and rising costs for local households

16 Mar 2026

A new study has identified four primary risks to water security posed by the rapid expansion of artificial intelligence data centers across the United States. The research, published in PLOS Water, suggests that while AI tools can improve utility management, the physical infrastructure supporting them creates significant pressure on local resources.

The findings move beyond national consumption estimates to examine the impact on specific municipalities. Researchers from the University of Washington found that the scale of water required for cooling these facilities often rivals the needs of entire towns.

A primary concern involves a lack of transparency. The study notes that data center operators and local governments have used public records exemptions to shield site-level water data from the public. Without this information, residents and utilities cannot accurately assess the impact on local supplies.

Financial equity presents a second challenge. Expanding infrastructure to meet the high demand of a data center requires significant capital investment. Utilities often distribute these costs across their entire customer base. This means residential ratepayers may end up subsidizing the infrastructure for facilities that offer them little direct benefit.

Operational resilience is the third risk identified. The continuous water demand from these sites reduces the flexibility utilities need during droughts or heatwaves. When a facility requires a non-negotiable volume of water to function, the margin for error during environmental stress narrows.

The fourth risk is indirect. Beyond on-site cooling, the electricity generation required to power these centers consumes vast quantities of water at thermal plants. This secondary layer of consumption is rarely included in standard corporate reporting.

The author acknowledges that AI offers genuine benefits, such as predictive leak detection and demand forecasting. However, the study argues that these advantages depend on transparent governance. It calls for mandatory public disclosure of water use and new regulatory frameworks to protect communities during extreme weather events.

As US utilities face aging infrastructure and rising digital demand, the research suggests that the benefits of the AI transition may be undermined by a failure to account for its environmental costs.