Artificial intelligence often feels invisible. We type a prompt, stream content, or store files in the cloud, and everything works instantly. But behind that convenience is a growing issue that receives far less attention than energy use: AI data center water use.

Data centers rely heavily on water to keep servers from overheating. Cooling systems often use water directly, and in many cases, that water is lost through evaporation during the cooling process. Research shows that data centers also consume water indirectly through electricity generation, especially when power comes from thermoelectric or hydropower systems (ScinceDirect) . This means the environmental impact of AI is not just about energy, it is also about the water required to sustain that energy.

The scale of this demand is significant. One analysis estimates that U.S. data centers consume billions of liters of water annually, with usage increasing alongside the growth of artificial intelligence (ScienceDirect). Earlier research also notes that data centers can rely on potable water for cooling and that water consumption is often underreported, with limited transparency across the industry (nature).

What makes this issue more complex is how uneven water use can be. The amount of water required can vary dramatically depending on location, cooling systems, and energy sources, meaning there is no single solution to reducing its impact (ScienceDirect). As AI continues to expand, these hidden demands will likely grow as well.

AI may feel weightless, but the infrastructure behind it is not. Understanding AI data center water use reveals that the environmental cost of digital technology extends beyond electricity, into the resources we depend on every day.

Alt text: Data center releasing large amounts of steam from cooling processes.

Work Cited:

Lei, Nuoa, Jun Lu, Arman Shehabi, and Eric Masanet. “The Water Use of Data Center Workloads: A Review and Assessment of Key Determinants.” Science of the Total Environment, 2025. https://www.sciencedirect.com/science/article/abs/pii/S0921344925001892

Mytton, David. “Data Centre Water Consumption.” npj Clean Water, vol. 4, Article number: 11, 2021. https://www.nature.com/articles/s41545-021-00101-w