Water Use Emerges as Major Issue for AI Data Center Growth
Water use has emerged as a major issue alongside energy for AI data center infrastructure.
Why it matters: Legal and compliance teams need to monitor evolving regulations addressing water consumption from AI data centers. Understanding water trends is crucial for managing operational and regulatory risks.
- Google’s data centers used 8.1 billion gallons of water in 2024, doubling 2021 levels.
- A single Google facility in Iowa consumed 1 billion gallons in 2024.
- By 2028, U.S. AI data centers could use between 38 and 73 billion gallons annually.
- About 57% of water for data centers comes from potable municipal sources, which raises regulatory concern.
The rapid growth of AI workloads is significantly increasing water use by data centers, raising legal and regulatory challenges alongside energy concerns. Google's data centers, for example, consumed 8.1 billion gallons of water in 2024, nearly twice their water use from 2021. One facility in Council Bluffs, Iowa, accounted for 1 billion gallons alone, showing localized strain on water resources (WaterVerse, 2026).
In Loudoun County, Virginia, data centers used 1.6 billion gallons in 2023, about 10% of the county's total water consumption. By 2028, AI data centers across the U.S. could consume between 38 and 73 billion gallons annually, driven largely by AI workloads (WaterVerse, 2026). This rise intensifies pressure on municipal water supplies, especially since 57% of this water comes from potable (drinking) water systems.
The reliance on drinking water for cooling raises public and regulatory concern, signaling potential new compliance requirements. For legal teams, this means early attention to environmental regulations and water use constraints is essential for ongoing data center operations and expansions.
Some tech leaders are responding. Microsoft has developed AI data centers using air-based coolers and reusing water through closed-loop cooling—a system that recycles water rather than consuming it continuously. This reduces water demand compared to traditional cooling towers, which evaporate large volumes of water (Axios, 2026). However, many existing and planned data centers still depend heavily on water.
Community resistance to water and energy use is mounting. In Hillsboro, Oregon, citizens and advocacy groups have opposed data center expansions citing water shortages, energy use, and tax concerns. Legal challenges are targeting tax incentives that support these expansions (Axios, 2026).
Further complicating matters, nearly two-thirds of the 809 U.S. AI data centers planned as of mid-2026 are located in drought-affected areas, intensifying the risk of regulatory scrutiny and legal action (Tom's Hardware, 2026). Despite efforts by Google, Amazon, and Microsoft to improve water recycling and replenishment, public concern remains high. Polls indicate 70% of Americans worry about local environmental impacts from data centers (Axios, 2026).
For legal pros advising AI infrastructure projects, understanding water use and anticipating tighter environmental rules will be critical. Compliance will shape approvals, local community relations, and ongoing operations in this evolving sector.
By the numbers:
- 8.1 billion gallons — Water used by Google data centers in 2024
- 1 billion gallons — Water consumed by Google's Iowa facility in 2024
- 38 to 73 billion gallons — Projected U.S. AI data center water use by 2028
Yes, but: While some AI data centers employ advanced cooling technologies reducing water use, many existing and planned sites remain heavily water-dependent, risking regulatory and community pushback.
What's next: Legal teams should watch for new municipal and state regulations addressing data center water use, as well as ongoing legal challenges to data center expansions targeting environmental and tax policies.