AI data centers are thirsty, and Texas is starting to notice. A new study from the University of Texas at Austin suggests a possible fix: cool the servers with water nobody else wants.
The researchers modeled how desalinated brackish groundwater and oil and gas wastewater could supply data center cooling. Their conclusion is that it may be workable, at a modest cost to the facilities’ energy budgets.
What the researchers found
The paper appeared online in the journal Water Research in September 2026. UT Austin publicly announced the findings on October 8, 2026.
Mechanical engineering professor Vaibhav Bahadur led the work. His team looked at established desalination methods, chiefly reverse osmosis, to see whether they could produce usable cooling water for data centers.
The other water source in the study is produced water. That is the briny water that comes up from wells alongside oil and gas, a byproduct the industry has to deal with one way or another.
According to the study, the energy and cost of treating these water sources would represent only a small fraction of a data center’s overall energy demand.
A new way to measure the strain
The team also introduced a new metric called Saline Water Utilization Intensity, or SWUI. It is designed to gauge local water stress under different treatment scenarios.
The researchers ran several case studies through the metric. A proposed data center project in Longview scored best. That setup paired brackish water with a two-stage reverse osmosis system.
The numbers behind the idea
The Permian Basin produces approximately 8 billion gallons of produced water every day. That is roughly the volume of 1,200 Olympic-sized swimming pools.
Meanwhile, a large data center can need up to 5 million gallons of water per day for cooling.
Why the timing matters
In September 2026, Governor Greg Abbott paused all new data center permits. That move came as the state Attorney General’s office stepped up investigations into how existing facilities use water.
What this means
For data center developers, the study offers a possible argument to bring back to regulators. If a project can show it draws on brackish or produced water instead of freshwater, it may be easier to defend in a state now scrutinizing water use closely.
There are open questions worth watching. The research is a modeling study, not a working installation. Brine disposal is another practical issue: desalination leaves behind concentrated salt waste, and any large-scale rollout would need a plan for it.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.