Cooling compute in dry country.

Yes — you can cool an AI data center in the West Texas desert, because water use is a cooling-design decision, not a consequence of geography. Data centers earned their thirsty reputation from one specific choice: evaporative cooling, which spends water to save electricity. Modern AI campuses can be engineered around dry and closed-loop systems instead, cutting consumptive water use to a small fraction of the evaporative baseline. The right question is not whether the desert has water. It is which cooling architecture the developer chose.

Where the water actually goes

Servers do not consume water; certain cooling systems do. Evaporative designs — cooling towers, evaporative economizers — reject heat by evaporating water into the air. It is thermodynamically elegant and cheap on electricity, which is why it became the industry default in an era when power was the expensive input and water was assumed. But every gallon evaporated is consumed: it leaves as vapor and does not come back. Scale that across a large campus, year-round, and the reputation writes itself.

The dry alternative

Dry cooling rejects heat to air without evaporation — air-cooled chillers and dry coolers, essentially industrial radiators. Closed-loop liquid systems circulate the same fluid indefinitely; the loop is filled once and topped up rarely. The trade-off is electrical: moving heat into desert air takes somewhat more fan and compressor energy than evaporating water does. In most places, that penalty argues for evaporative. In West Texas, where the entire point of siting is abundant low-cost energy, it is the easiest trade in the building.

Why AI hardware makes it easier

AI racks are pushing the industry toward direct-to-chip liquid cooling, because air cannot pull heat off dense accelerators fast enough. Those loops are closed: coolant circulates between the chips and the heat-rejection plant and is not consumed. Pair a closed loop at the rack with dry rejection outside, and the campus's consumptive cooling water falls to a trickle. The desert stops being a liability the moment the design stops asking it for water.

What diligence should ask

A buyer evaluating a West Texas site should ask which cooling architecture is specified — evaporative, hybrid, or dry — and what consumptive use looks like on a peak summer day, not an annual average. Ask where any water comes from, under what rights, and what happens in a drought year. And ask whether the power plan carries the extra electrical load that dry cooling implies. Sites built on abundant behind-the-meter power can afford water-light cooling; the two design choices reinforce each other. Cooling belongs on the same checklist as fuel and redundancy — the rest of that checklist is in how to evaluate a BTM power provider.

About Corley Energy

Corley Energy is a behind-the-meter independent power producer, founded in 2024 by Jake Corley, Tim Bozeman, and Mark Meyer. We convert stranded Permian Basin natural gas into firm, contracted electricity for AI data centers at Power Foundry, our ~1,000-acre development in Upton County, Texas. Start with what a power foundry is, see the company facts, or check current capacity on the Sites page.

Keep reading

Can AI data centers be flexible loads? · ERCOT scarcity pricing and what it means for data centers · How much power do AI data centers need? · Browse the full library