Researchers at the University of Texas at Austin have proposed an innovative solution for managing water consumption in AI data centers. As the AI boom places significant pressure on global water and energy resources, using contaminated water from the oil and gas industry could serve as a sustainable source for supplying water to these centers.
Water Needs of AI Data Centers
With the rapid growth of large data centers, the demand for water resources has significantly increased. According to statistics, a large data center can consume up to 5 million gallons of water per day. Meanwhile, actual water consumption may even exceed this amount. Research from the University of Michigan indicates that the water needed for generating the energy required for these centers must also be considered. For example, in three large data centers in Michigan, the total water consumption, including the water needed for energy production, rises to 14 billion gallons.
Read more: GT Wings Unveils AirWing™S as an Innovation in Small Vessels
Innovative Solutions for Cooling Data Centers
In a new article published in the scientific journal Water Research, researchers have examined various methods for treating contaminated water with the aim of using it for cooling data centers. Reverse osmosis is recognized as the most efficient method, but it faces challenges with water that is saltier than seawater. In cases where this method is not feasible, converting water to steam and collecting the resulting freshwater is another option, although this method is more energy-intensive.
Greg Stenton, a professor at the University of California, Riverside, emphasizes the importance of considering indirect water consumption, stating, "This indirect consumption is sometimes much greater than direct consumption and has serious impacts on regional water resources."
Researchers have also pointed out that using contaminated water from the oil and gas industry can help reduce the water needs of data centers while simultaneously addressing a larger problem for the fossil fuel industry. According to Vaibhav Bahadur, one of the researchers in this study, the Permian Basin alone produces about 790 million gallons of contaminated water daily, which is more than 150 times the water consumed by a large data center.
Thus, the Permian Basin, producing the equivalent of 1200 Olympic-sized swimming pools per day, poses serious challenges for storing this water. Currently, this water is transported over long distances and injected deep into the ground, which carries risks including seismic activity.
Bahadur also highlighted the urgent need to find solutions for utilizing this water, stating, "We will soon face a shortage of space for burying this water, so much so that by 2030 this problem will become more serious. Cleaning this water for use in data centers is a great way to solve this issue."
Read more: The Role of Expanding Iranian Catalysts in Enhancing Steel Industry Productivity · The VLCC Market Faces Challenges in Week 42 of 2026
