News · Meta Details How Closed-Loop Cooling Underpins Its 2030 Water Positive Goal
Meta Details How Closed-Loop Cooling Underpins Its 2030 Water Positive Goal
The company's water announcement ties a specific cooling architecture to a measurable restoration target — and commits to publishing the design through Open Compute.
The cooling architecture doing the heavy lifting
The most concrete engineering claim in Meta's announcement is its typical data center design: direct-to-chip liquid cooling in a closed piped loop, paired with dry coolers that blow air over the pipes to shed heat. Because the coolant recirculates and the heat is rejected to air rather than evaporated, Meta states there is no operational water use in the cooling system.
That distinction matters. Evaporative cooling, the traditional approach for large facilities, consumes water precisely to reject heat. By moving to dry coolers, Meta shifts the water burden off cooling entirely, leaving only domestic, janitorial, equipment-cleaning, and fire-protection needs on site.
The headline figure attached to this is Beaver Dam, Wisconsin: once operational, its total estimated annual water use is anticipated to be less than that of two full-service restaurants in a year. For a facility built to support Meta's apps and AI workloads, that is a deliberately small number, and it only holds because the cooling loop itself is dry.
Restoration as a separate ledger from consumption
Meta separates efficiency from offset. The water positive goal — restoring more than it consumes in the watersheds where it operates — is backed by a specific ratio: 200% restoration of consumed water in high water-stress regions and 100% in medium water-stress regions.
The restoration portfolio is presented with numbers that can be checked against future reports. Since 2017, Meta says it has funded or supported more than 40 projects across nine watersheds. In 2024, operational projects returned more than 1.59 billion gallons; once all projects are fully implemented, the expected annual figure rises to 2.9 to 3.4 billion gallons.
Individual projects carry their own estimates: 64.9 million gallons per year from converting flood irrigation to N-Drip drip irrigation on the Colorado River Indian Reservation, 44 million gallons from restoring 2,000 acres of longleaf pine in the Trinity River Watershed, and 81.5 million gallons from flow restoration on the Isleta Reach in New Mexico. Meta states these projects have a hydrological connection to its source water and are verified by independent third parties.
Community infrastructure Meta builds and gives away
A less discussed part of the announcement is capital infrastructure that leaves Meta's balance sheet. In Kuna, Idaho, Meta says it invested over $70 million to build a water and wastewater treatment facility that has since been gifted to the city. In Richland Parish, Louisiana, it reports over $200 million in local infrastructure, including new wastewater capacity, to be handed to the community.
This reframes the data center as a piece of municipal utility investment rather than purely a private facility. The water treatment capacity Meta funds does not only serve its own site — it becomes public infrastructure that outlasts the company's specific need for it.
Why the Open Compute commitment is the operative detail
For engineering teams, the load-bearing sentence is Meta's pledge to continue sharing its cooling advancements through the Open Compute Project. Efficiency claims and restoration ledgers are Meta's alone; a published dry-cooler and direct-to-chip design is reusable by anyone building at scale.
The practical implication is that the water case for closed-loop cooling is being documented alongside its thermal engineering. If Beaver Dam's near-zero cooling water figure holds in operation, the design becomes a reference point others can cite when local utilities and communities ask how much water a new facility will draw.
The open question the announcement leaves is verification cadence. Meta commits to disclosing consumption and restoration in its annual sustainability report and Volumetric Water Benefits Report. The value of the 2.9-to-3.4-billion-gallon projection depends entirely on whether those future reports show projects reaching full implementation — which is exactly what the transparency pillar promises to make checkable.
Find this story relevant to you?
Contact us to find a unique solution
Need an AI engineering partner that can actually build?
We help businesses integrate AI, build AI-powered products, automate high-value workflows, and modernize the software systems behind them.
Related reading
More analysis around product delivery, operational AI, and the systems work that makes deployment hold up in reality.