The Hidden Costs of AI: Data Centers, Water, and the Future of Rural America

Artificial intelligence has ignited a historic expansion of data center infrastructure across the United States. But a growing body of research shows that this buildout comes with hidden costs >>> not just water security, but also pressure on an already stretched-thin power grid. As AI-scale data centers demand enormous amounts of water for cooling and massive amounts of electricity, rural communities in particular are being forced to grapple with resource trade-offs that were rarely considered in traditional technology procurement.

A Midwestern Farm.

According to a recent Data Center Dynamics article by Zachary Skidmore, Senior Reporter – Energy and Sustainability, citing Bloomberg report, covered by Data Center Dynamics, nearly two-thirds of new U.S. data centers since 2022 have been sited in regions of high water stress. Five states, California, Arizona, Texas, Illinois, and Virginia, together account for 72% of these new deployments, even though each of them is battling existing water challenges.

Photo by Enrique Macias

“Every part of Texas is facing this water-energy nexus crisis.”

Amy Bush, a hydrologist at RMBJ Geo Inc, told Bloomberg, “Every part of the state is facing this water-energy nexus crisis,” speaking specifically about Texas, which has seen a huge influx of AI data center projects thanks to its low taxes and abundant energy resources.

Why Data Centers Use So Much Water

One 100MW data center in the U.S. consumes roughly two million liters (528,340 gallons) of water per day, the International Energy Agency reported in April. Globally, the data center industry now uses more than 560 billion liters (147.9 billion U.S. gallons) of water each year, and projections suggest that figure could more than double to 1,200 billion liters by 2030.

The reason? Most data centers still rely on evaporative cooling, a method where roughly 80% of the water used evaporates during cooling, with the rest ending up as wastewater. While major hyperscalers have made public commitments to improve water stewardship — for example, Google’s partnerships in Chile, California, Taiwan, and France — these efforts may be too little, too late in communities already struggling with stressed aquifers and agricultural needs.

Communities Push Back

Data centers have started to face stronger resistance from local residents. Last year, for example, Google was forced to delay a data center project in Chile after an environmental court intervened, citing risks to an overdrawn aquifer serving the capital region. Similar community backlash could grow in the U.S. as well, particularly in rural counties that lack both water security and the resources to negotiate fair trade-offs with data center operators.

For residents of Tucker County, West Virginia, the fight against the coal-powered AI data center is about more than just preserving their views of Blackwater Falls (pictured) or their favorite hiking trails. It’s about protecting their health, their homes, and their future.

Farmers Fighting for Their Way of Life

In May, we published the blog The True Cost of the AI Revolution: Data Centers' Impact on Rural. We wrote about Beverly and Jeff Morris who bought their home in Mansfield, Georgia, in 2016, seeking a peaceful countryside escape just an hour from downtown Atlanta. For them, this piece of land represented more than just a home—it was their sanctuary and a foundation for their farming livelihood. "When we found this place, we decided that this was it. It was perfect," Beverly Morris explains in an interview with Ben Lieberman. However, their rural paradise was disrupted when Meta began construction of a data center in 2018, just 400 yards from their property. The impacts have been devastating.

Beverly Morris explains that they’ve had to save up water just to flush their toilet, highlighting the severity of living next door to a Meta data center.

Beverly and Jeff Morris had their rural paradise disrupted when Meta/Facebook began construction of a data center in 2018, just 400 yards from their property. The impacts have been devastating and costly.

Can Cooling Innovation Solve It?

There is promising innovation on the horizon. Microsoft studies have shown that advanced cooling technologies — including cold plates and two-phase immersion — could cut water usage by 31–52% over their life cycles compared to traditional air cooling. Other solutions include chilled water reuse, rainwater harvesting, river or geothermal cooling, and closed-loop systems that dramatically cut evaporative loss. Nature provides insights into these innovative cooling technologies.

Still, large-scale adoption of these solutions will take time — and policy leadership — to ensure water sustainability becomes a first-class consideration in data center siting decisions.

Platocom’s View: Standards, Transparency, and Rural Support

From our perspective, water should no longer be treated as an afterthought in data center procurement. As AI reshapes the digital economy, data centers are emerging as essential critical infrastructure — but with enormous local impacts. Rural and semi-rural communities in particular often lack the technical capacity, regulatory resources, or negotiating leverage to manage these trade-offs on their own.

Platocom suggests that federal and state authorities develop:

✅ Water stewardship reporting frameworks for data center operators

✅ Minimum water-impact standards for cooling technology

✅ Technical support for rural counties to help them evaluate proposals and weigh community impacts

✅ Neutral reporting bodies for vendors and community members to flag water or resource concerns early

Otherwise, as AI-driven data center sprawl accelerates, many rural communities risk trading one crisis — the digital divide — for another: water insecurity.

In short, the future of AI infrastructure must balance performance with sustainable resource use — because resilience is worthless if it runs dry.