The Problem
AI data centers require massive amounts of water to keep their servers cool. Cooling towers evaporate water to shed heat generated by thousands of processors running around the clock. Unlike electricity, this water is mostly lost — it doesn't return to local rivers or groundwater supplies.
Key Facts
- A large data center can use up to 5 million gallons of water per day.
- That's equivalent to the daily water use of a town of 10,000–50,000 people.
- Training a single large AI model can consume hundreds of thousands of liters.
- Many data centers are built in dry regions, worsening local water scarcity.
Why It Matters
When data centers are placed in already water-stressed areas, they compete directly with farming communities and residential water needs. Overuse of groundwater can lower water tables permanently, affecting entire regions for decades. Warm water discharged back into rivers also raises local temperatures, harming fish and plant life.
What Can Be Done
Some companies are beginning to adopt air-cooled or closed-loop liquid cooling systems that use far less water. Others are exploring the use of recycled or treated wastewater instead of fresh water for cooling. Requiring data centers to publicly report water usage is another important step: transparency helps communities and regulators respond before damage becomes irreversible.