22/09/2026
Clean water is older than every technology we build, and it remains essential to every community.
As artificial intelligence expands, the data centers that power it are becoming part of local conversations about water, land, energy, and long-term planning. These facilities run servers around the clock, producing heat that must be managed. In many large data centers, evaporative cooling systems use substantial amounts of water to remove that heat.
Water use varies by facility, climate, and cooling design. Some systems rely heavily on evaporation, while newer closed-loop, direct liquid cooling, immersion, and water-free systems can reduce freshwater demand. Those choices matter because water impacts are local. A facility drawing from a river, aquifer, reservoir, or municipal supply affects the watershed where it is built, not the place where an AI tool is being used.
For Indigenous Nations and communities, water is not only a utility. It is tied to food systems, fisheries, medicines, ceremonies, treaty rights, and responsibilities carried across generations. When new infrastructure is proposed, questions about water availability, water quality, discharge, electricity demand, land use, and transparency are all part of understanding the full footprint.
Reports cited in recent research and policy discussions have noted that some large data centers can use millions of gallons of water per day, especially under peak conditions. Other studies point to rising concerns as data center growth expands into water-stressed regions. At the same time, researchers and engineers are identifying ways to reduce water demand through better siting, improved cooling technology, recycled water, and stronger monitoring.
The growth of AI is often described as digital, but its infrastructure is physical. It sits on land, draws power, and can depend on water from specific places.
What questions should communities be asking before major data center projects are approved in their watersheds?
Sources: Lincoln Institute of Land Policy, UN News, CAES Field Report, Sciencedirect, Cornell Chronicle
