Wait, are we actually cooling AI data centres with pee? Sort of, yes
A cheeky beer advert starring Jason Kelce accidentally landed on a real solution to one of AI's biggest environmental problems. Here is what the science actually says.
Key points
- Loudoun County, Virginia, home to more than 250 data centres, used about 200 million gallons of recycled water per day for cooling as of 2025, covering 43% of daily data centre water needs.
- The remaining 57% of daily data centre water use in Loudoun, around 260 million gallons, still comes from drinking-water supplies.
- Meta has committed at least $270 million to wastewater infrastructure projects near its data centres.
- Advocates are pushing for a 30% tax credit to help industries build out recycled-water systems faster.
- A recent Gallup poll found about seven in ten Americans oppose data centres in their communities.
Jason Kelce did not expect to stumble into environmental engineering. But when the former Philadelphia Eagles star appeared in a Liquid Death and Garage Beer campaign joking that people should "pee on computers together to save humanity", he accidentally described something engineers are genuinely doing, minus one step.
So, could you actually cool a data centre with pee?
Not directly. Pure urine would clog the system fast. But treated wastewater, which includes human urine as one ingredient, is already doing the job at scale.
AI data centres need enormous amounts of water. Servers generate serious heat, and the most common fix is evaporative cooling, a process where hot air passes through water so the heat escapes as steam. The problem is that this eats through drinking water at a rate that concerns local communities.
The alternative is recycled water: wastewater and sewage run through treatment facilities that use membrane bioreactors (filters that use biology and pressure to strip out contaminants), reverse osmosis (a process that pushes water through an ultra-fine membrane to remove dissolved salts and chemicals), and ultraviolet light to kill bacteria. What comes out is clean enough for industrial cooling, and sometimes clean enough to drink.
"We use recycled water for cooling for all kinds of industries, and we have for decades," Dr. Greta Zornes, practice leader for water reuse at engineering firm CDM Smith, told TechCrunch. "Definitely, there's been a boom in recycled water for data centre cooling."
Bruno Pigott, executive director of the WateReuse Association, put it plainly: "You wouldn't just use pee, but you can clean it and make it into useful water."
Skip the cleaning step, though, and things go wrong quickly. Raw urine contains salts, urea, bacteria and organic matter that would leave mineral deposits inside cooling equipment and require constant maintenance. If you pumped it through a hot-air evaporative system, the smell alone would settle the argument.
Why isn't every data centre using recycled water already?
Infrastructure. A data centre needs to sit close to a wastewater treatment plant big enough to supply millions of gallons per day. Many do not.
"When data centres go out into rural areas, a lot of times the wastewater treatment plants just aren't big enough," Zornes said. Building that infrastructure takes years, and the demand from AI is arriving now.
Loudoun County in Virginia shows both the progress and the gap. With more than 250 data centres, it used 200 million gallons of recycled water daily as of 2025. Impressive. But that still only covered 43% of the area's data centre water use. The other 260 million gallons came from the drinking-water supply.
Meta is putting money into closing that gap, committing at least $270 million to wastewater infrastructure near its data centres. On the policy side, Pigott is backing proposed legislation for a 30% tax credit to help industries build recycled-water systems faster.
What does this mean for ordinary people?
If you live near a data centre, this directly affects your water supply. Seven in ten Americans told Gallup they oppose data centres in their communities, and water use is a big part of why.
The good news is that the engineering exists. The race is to build enough treatment infrastructure before AI's water appetite outpaces what local supplies can handle. A beer advert made that conversation go mainstream. That, at least, is useful.
