Water-AI Nexus report provides roadmap for water sustainability in the AI era

The report outlines four principles for responsible water use by data centers in our current digital landscape.

Data centers comprise critical infrastructure for key sectors, including utilities, government, and finance. As AI continues to advance, data centers are scaling up to provide the computing power needed to develop the AI tools we increasingly use. Some data center designs use water for cooling to increase power efficiency. This helps to reduce carbon emissions and provides relief for power distribution constraints while infrastructure providers work to upgrade electrical grids. With rising water scarcity and demands for sustainable solutions, efficient use of water and energy is essential for the data industry’s growth.

Supporting environmental and economic goals while also responding to policy needs and the public interest, the Water-AI Nexus™ Center of Excellence, founded by the Water Environment Federation, Amazon, The Water Center at Penn, and the Leading Utilities of the World, aims to guide both utilities and data center developers on responsible practices to minimize impacts on watersheds, starting with this report.

“Water utilities worldwide are facing unprecedented challenges from climate change and aging infrastructure, while simultaneously working to serve communities that depend on reliable, affordable water services,” Water Environment Federation President Howard Carter said. “The Water-AI Nexus Center of Excellence will accelerate innovation by connecting water professionals with AI experts to develop solutions that benefit both sectors as well as the communities they serve.”

The report outlines four principles on data center design: 

1. Improve design and siting practices

Where and how data centers are built matters. There are multiple factors for developers to consider including water availability, climate conditions, and energy access when determining a data center’s location and cooling technology.

The report cites location as a key consideration, as a location’s water and energy availability influences data center design decisions. 

2. Reduce water consumption

Water usage effectiveness (WUE) is a valuable metric for understanding water use in data centers. WUE varies by cooling system design and by location. 

3. Reuse and sustainably source water

The data center industry is increasingly developing and using municipal recycled water for cooling. This eases constraints around potable supplies and helps prioritize potable water for community purposes.

Examples of data center operators investing in this space include AWS, which announced expanding to water reuse in 120 locations by 2030, and Meta, which has invested in water infrastructure that allows spent cooling water to be used as irrigation for non-edible crops in Idaho. 

4. Engage with communities

Data centers should work with region-specific water utilities, government agencies, and community groups to understand local water stresses. Water replenishment serves as a pathway to deliver water back to communities and ecosystems and represents the final critical component of comprehensive data center water stewardship strategy. 

Looking ahead

The report’s authors highlight the crucial relationship between water and technology as we move toward a more water-conscious future. By adopting these principles, they say, data centers can minimize their water footprint while continuing to power the digital services we rely on.

Learn more about the Water-AI Nexus™ Center of Excellence.

Sign up for our monthly newsletter to get Amazon sustainability updates sent directly to your inbox.