Why a joke about peeing on computers points to a real data center water crisis

AI data centers are consuming huge volumes of water to keep servers from overheating, and the viral Liquid Death ad about using urine to cool computers is based on a real industry trend: operators are increasingly turning to recycled wastewater instead of drinking water. The joke landed because it taps into a serious infrastructure problem that is now shaping where data centers can be built and how communities respond to them.

The campaign, which featured former Philadelphia Eagles lineman Jason Kelce, framed the issue in absurdist terms — “let’s pee on computers together” — but water experts say the underlying point is not as far-fetched as it sounds. Across the U.S., data center developers and utilities are expanding the use of reclaimed water, while policymakers and engineers are trying to keep the artificial-intelligence boom from putting even more pressure on municipal drinking supplies.

What the Liquid Death ad got right

The commercial is a joke. The problem behind it is not. Data centers generate intense heat and rely on cooling systems that can require vast amounts of water, especially in evaporative cooling setups. That demand has become a growing concern as AI workloads expand and more facilities are built to support them.

In the ad, Kelce riffs on the idea that “AI data centers waste millions of gallons of water” and suggests that people should contribute their own urine to the cause. The punchline is crude, but the campaign accidentally highlights a legitimate industry workaround: using treated wastewater and other non-potable sources to cool equipment.

“The commercial is funny and tongue-in-cheek,” said Michael Obradovitch, vice president of Data Center Global Accounts at Ecolab. “But in reality, there is a fair amount of alternative water sources already being used to a similar extent to cool these data centers.”

In other words, the real story is not whether anyone should literally supply urine to a cooling tower. It is that the tech sector is being pushed to rethink how it gets water in the first place.

How data centers cool servers without drinking water

Data centers do not need potable water specifically; they need reliable water that can move heat away from racks of servers. That opens the door to sources that are not safe to drink but are suitable for industrial use once treated properly.

One of the most common alternatives is recycled water, also called reclaimed or reused water. This is produced by treating wastewater, sewage, and other effluent so that it can be used again for industrial systems, irrigation, and in some cases even potable applications.

Why recycled water matters

Recycled water reduces dependence on municipal drinking supplies, which is especially important in regions facing growth, drought, or limits on freshwater availability. For data centers, it can be the difference between a project that fits local water capacity and one that strains it.

Bruno Pigott, executive director of the WateReuse Association and a former acting assistant administrator for water at the U.S. Environmental Protection Agency, said the concept is straightforward: sewage contains water that can be cleaned and reused, but it must be treated before it goes anywhere near sensitive cooling equipment.

“You wouldn’t just use pee, but you can clean it and make it into useful water, and that’s what we advocate,” Pigott said.

He also noted that untreated urine would be a bad fit for cooling systems because it carries salts, urea, bacteria, and organic material that can cause scaling and deposits. In an evaporative cooling tower, those contaminants would quickly create maintenance issues and reduce efficiency.

What treatment looks like in practice

Modern water-reuse systems can involve multiple stages, including membrane bioreactors, reverse osmosis, ultraviolet disinfection, and other filtration and purification methods. Depending on the end use, that treatment can make water clean enough for industrial systems or, in some cases, safe to drink.

The same basic idea is already used in space, where astronauts recycle urine and other waste streams because every drop of water must be recovered and reused. The difference is scale: turning wastewater into usable water for a spacecraft is a highly controlled process; turning municipal sewage into a daily industrial supply for huge data campuses requires large facilities, large budgets, and dependable infrastructure.

Why data centers are driving a recycled-water boom

According to water engineers, the artificial-intelligence buildout has accelerated demand for recycled-water projects. Where these systems were once part of a broader industrial toolkit, they are now becoming a central part of planning for new data center clusters.

“We use recycled water for cooling for all kinds of industries, and we have for decades,” said Dr. Greta Zornes, practice leader for water reuse at engineering firm CDM Smith. “So this is only one application, but definitely, there’s been a boom in recycled water for data center cooling.”

Zornes said her own workload has shifted as demand from data center clients has surged.

“Every day right now, I’m working on recycled water for data centers,” she said.

That reflects a broader trend: as cloud computing and AI training expand, so does the need for cooling. In many markets, utilities and municipalities are now treating water strategy as part of site selection, permitting, and public engagement for new data infrastructure.

Where recycled water works best

Recycled water tends to make the most sense when a data center is close to a large enough wastewater treatment plant. If the treatment plant cannot supply enough volume, the economics and logistics break down quickly.

That is why the solution is easier to deploy in denser metro areas with mature infrastructure than in remote or fast-growing rural zones.

Zornes said proximity matters because the wastewater plant has to be large enough to provide enough treated water. In rural areas, she added, nearby facilities often are not built to handle the scale required for major data center campuses.

This is one of the main constraints in the current AI infrastructure race: the bottleneck is not just electricity and land, but also water delivery and treatment capacity.

What the numbers show in Loudoun County

Loudoun County, Virginia, the northern Virginia hub just outside Washington, D.C., shows how fast this issue is growing. The county is home to more than 250 data centers, and at least two dozen more are planned.

Even with extensive use of recycled water, the region’s overall water demand remains enormous. In 2025, Loudoun data centers were using about 200 million gallons of recycled water per day, but that still covered only part of the total demand. The remaining 260 million gallons per day, or 57% of daily water usage, came from potable water supplies, according to Loudoun Water.

That means recycled water accounted for about 43% of daily data center water use in the county, while drinking water sources still made up the majority. The figures underscore both the promise and the limits of reuse strategies when data center density reaches extreme levels.

How can the AI industry scale water reuse?

The answer is by paying for the infrastructure that makes reuse possible. That means building or expanding treatment plants, laying pipes, creating storage systems, and coordinating with local water authorities so reclaimed water can reach facilities in high volumes.

Obradovitch argued that data centers could become important anchor tenants for that kind of public infrastructure investment, especially if companies are willing to fund projects that also benefit surrounding communities.

He said there are examples of data centers committing capital to municipal water projects as part of their community engagement, helping to close the gap between industrial demand and local treatment capacity.

Meta has already signaled how expensive this kind of support can be. The company plans to invest at least $270 million in wastewater infrastructure projects near its data centers, a sign that major tech firms may need to help pay for the systems their facilities depend on.

Those investments are not necessarily charity; they are also risk management. Without dependable water access, approvals can stall, community opposition can rise, and some sites may become impossible to develop.

What communities are asking for

Communities increasingly want stronger guarantees that data centers will not drain local water supplies. Residents often view large server farms as noisy, power-hungry, and opaque. Water has now joined electricity and land use as one of the defining flashpoints in siting fights.

• More transparency about daily water use
• Preference for non-potable or recycled supplies
• Funding for local treatment and distribution systems
• Limits on drinking-water dependence in stressed regions

For utilities and local governments, the challenge is balancing economic development against long-term water security. For companies, the challenge is proving that AI infrastructure can scale without overwhelming neighborhoods that already have limited resources.

Why policy could determine the next phase

Industry solutions may not scale fast enough without government support. Pigott said recycled-water adoption could accelerate meaningfully if lawmakers created a federal tax incentive for water-reuse infrastructure.

He is advocating for a 30% tax credit that would help industries build out recycled-water systems more quickly.

Such a policy could lower the upfront cost of treatment plants, pipelines, and related equipment, making reuse more attractive to data center operators and municipalities. It could also push more projects toward reclaimed water before drinking-water systems are forced to absorb the full burden of AI growth.

The timing matters because infrastructure takes years to design, permit, and build. Data center demand, by contrast, is moving much faster. That mismatch is creating pressure on utilities, regulators, and companies to solve problems before they become crises.

What the viral moment says about public sentiment

The ad’s popularity also points to a deeper unease about how AI is reshaping everyday life. Public concern about data centers is no longer limited to engineers or local planners; it is spilling into the mainstream, often through satire and backlash.

Recent polling cited by TechCrunch suggests that around seven in 10 Americans oppose data centers in their communities. That resistance reflects worries about land use, water demand, noise, emissions, and whether the benefits of AI are worth the local tradeoffs.

For the tech industry, the challenge is not just technical. It is also reputational. AI products may be useful to companies and investors, but they can still face pushback from consumers who feel they are being imposed on society without enough accountability.

Pigott said he welcomes any attention that helps raise awareness of water, even if the method is crude or comedic, because it creates an opening to educate the public about reuse and conservation.

That may be the best outcome of the joke: it makes an invisible infrastructure issue visible. A crude punchline about urine turns out to be an entry point into one of the most practical questions in the AI era — how to power digital growth without draining the water systems people depend on.

The bottom line

No, data centers should not literally be cooled with human urine. But the ad’s joke lands on a real truth: the AI industry is already leaning more heavily on recycled wastewater, and that shift will likely shape where new data centers go, how they are financed, and which communities are willing to host them.

As AI infrastructure expands, water reuse is moving from niche sustainability concept to core planning issue. The companies that can secure non-potable supplies may gain a crucial advantage, while policymakers and utilities face growing pressure to build the systems that make those supplies available at scale.