A joke about cooling AI data centers with pee landed because it points to a serious problem: servers need cooling, cooling can require massive quantities of water, and communities are increasingly paying attention to where that water comes from.
Liquid Death and Garage Beer turned the issue into a marketing campaign featuring former Philadelphia Eagles star Jason Kelce. The joke was crude by design. The underlying question is not: should people send urine to data centers? It is: can recycled wastewater help reduce the pressure that data centers place on drinking water supplies?
The joke rests on a real water reuse idea
The campaign suggests, facetiously, that people could contribute pee to cool data centers. Experts in the source article make clear that direct use of urine is not the answer. Urine contains salts, urea, bacteria, organic matter and material that could leave mineral deposits, creating cleaning problems in cooling systems.
The useful version of the idea is different. Wastewater and sewage water can be treated, cleaned and reused. Because wastewater includes human urine, the campaign accidentally gestures toward a real category of water reuse, even if the literal proposal is not practical.
Michael Obradovitch, vice president of Data Center Global Accounts at Ecolab, told TechCrunch that alternative water sources are already being used to cool data centers. Bruno Pigott, executive director of the WateReuse Association and former acting assistant administrator in water for the U.S. Environmental Protection Agency (EPA), also framed the answer around cleaning wastewater rather than using urine directly.
The point is straightforward: data center cooling does not always have to draw from potable drinking water. In the right place, with the right treatment systems, recycled water can take on part of that job.
Why untreated urine would fail as a cooling source
Cooling systems are not designed for a steady stream of raw waste. One method mentioned in the source article is evaporative cooling, where hot air passes through water and heat is removed through evaporation. Running untreated urine through that kind of system would create obvious operational and odor problems.
The source article also notes that humans can treat urine into potable drinking water in space, where astronauts have limited water supplies. But that approach is not described as efficient at large scale for data center use. The practical route is to rely on wastewater and sewage systems that already collect large volumes of water.
Water treatment facilities can use membrane bioreactors, reverse osmosis, ultraviolet light and other processes to make water clean enough for industrial use. In some cases, the water can be treated until it is drinkable. For data center cooling, the key value is that recycled water can partially offset demand for potable water.
Dr. Greta Zornes, practice leader for water reuse at engineering firm CDM Smith, told TechCrunch that recycled water has been used for cooling across industries for decades. What is changing now is the pace of demand from data centers.
Infrastructure decides where recycled water can work
Recycled water is not available everywhere a data center might be built. Zornes said data centers need to be somewhat near a wastewater treatment facility that is large enough to provide enough water. That creates a practical constraint for facilities located in rural areas, where wastewater treatment plants may not process enough water to support large cooling needs.
This makes the water question less about a single technology and more about local infrastructure. A data center cannot simply choose recycled water if there is no nearby treatment capacity, no distribution system and not enough wastewater volume to reuse.
That is why buildout time matters. Treatment plants, pipes and related systems are physical infrastructure, and the source article emphasizes that the needed capacity often does not already exist. Even when data center operators want to use more recycled water, the switch can be limited by what the local water system can handle.
The article gives Loudoun County, Virginia, located outside of Washington, D.C., as a major example. It is home to more than 250 data centers, with plans to construct at least another two dozen. As of 2025, Loudoun data centers collectively used about 200 million gallons of recycled water each day.
That sounds large, but it still accounted for only 43% of daily data center water usage in the area. The other 260 million gallons, or 57% of daily data center water usage, came from potable water supplies, according to Loudoun Water.
Data centers may become water infrastructure anchors
One implication from the source article is that AI growth could push money toward water reuse infrastructure. Obradovitch said data centers can become anchors of water infrastructure when they support municipalities in addressing treatment and capacity challenges.
Meta is one example cited in the source article. The company will invest at least $270 million in wastewater infrastructure projects near its data centers. The same article notes that Meta also loses about $4 billion each quarter on its Reality Labs division, underscoring the scale at which major technology companies can spend.
Policy could also shape the pace of adoption. Pigott is advocating for legislation that would provide a 30% tax credit to help industries scale recycled water infrastructure. His view, as described in the source article, is that this could accelerate the move into recycled water for data centers and other industries.
The broader issue is not whether recycled water works in theory. It already has industrial uses. The challenge is whether enough treatment capacity can be built where data centers need it, quickly enough to reduce reliance on potable water.
Public concern is becoming part of the story
The Liquid Death campaign became notable because data center water use is no longer an obscure infrastructure topic. The source article cites a recent Gallup poll finding that about seven out of 10 Americans oppose data centers in their communities. It also notes that AI products continue to face backlash from consumers who feel the technology is being forced into their lives.
That public concern changes the stakes for AI companies. Cooling decisions are no longer just engineering choices hidden behind facility walls. They can affect local water supplies, municipal planning and community trust.
The practical answer is not to cool computers with raw urine. It is to expand recycled water systems where they make sense, reduce pressure on potable water and treat data center growth as part of local water planning. The joke may be crude, but the infrastructure question behind it is real.