thirst for data

Texas data center boom could strain water supply, new report warns

A new report shows that Texas data centers used 25 billion gallons of water in 2025. Photo via HARC report.

As data centers continue to boom throughout Texas, a new report from the Houston Advanced Research Center (HARC) warns that the trend could strain the state’s water supply.

HARC estimates Texas data centers used 25 billion gallons of water in 2025—and that the demand for water will continue to rise to meet the needs of the 464 data centers currently in Texas, as well as 70 additional sites currently under development.

In the report, titled “Thirsty Data and the Lone Star State: The Impact of Data Center Growth on Texas’ Water Supply,” The Woodlands-based nonprofit says that water use for cooling data centers is expected to double or triple by 2028 on the national level. If projections hold, the total annual water use for data centers in Texas will increase by 0.5 percent to 2.7 percent by 2030, or to between 29 billion and 161 billion gallons of water consumed.

Data centers often use water for cooling, though water demand is dependent on the type of cooling used, the size and type of the data center. Although used water can be reused, some new water withdrawals are always needed to replace evaporated water and other systems’ water losses. Water is also used to cool the power plants that generate electricity used by the data centers.

The HARC report offers guidance to address the overall concerns of water demands by data centers, including:

  • Dry cooling methods
  • Increased reliance on wind and solar energy sources
  • Alternative water supplies, like treated wastewater or brackish water for cooling
  • Adjusted operating schedules to accommodate water usage
  • Partnering with local companies to develop projects that reduce water leaks
  • Companies creating their own water infrastructure investments

The report goes on to explain that the Texas State Water Plan, produced by the Texas Water Development Board, projects shortages of 1.6 trillion gallons by 2030 and 2.3 trillion gallons by 2070. HARC posits that the recent surge in water demand from AI data centers is not fully reflected in those projections.

"Texas water plans always look backward, not forward," the report reads. "That means the 2027 water plan, which is in development now, will be based on 2026 regional water plans that do not include forecasted data center water use. Data centers that began operation in 2025 will not be added to the State Water Plan until 2032."

Currently, there are no state regulations that require data centers to report how much water they use. However, the Public Utility Commission of Texas (PUC) plans to survey operators of data centers and cryptocurrency mining facilities on their water consumption, cooling methods and electricity sources this spring. It is expected to release the results by the end of the year. The companies will have six weeks to respond. The Texas Water Development Board will assist the PUCT on the questions.

“I think we all recognize the importance of data centers and the technology they support and what they give to our modern-day life,” PUC Commissioner Courtney Hjaltman said during the last commission meeting. “Texans, regulators and the legislature really need that understanding of data centers, really need to understand the water they’re using so that we can plan and create the Texas we want.”

See the full HARC report here.

Trending News

A View From HETI

SLB and NVIDIA plan to build an "AI factory for energy" to help the industry transform massive amounts of data into actionable information. Photo courtesy SLB

Houston-based energy technology company SLB has expanded its 18-year tech collaboration with chipmaker NVIDIA to include the development of an “AI factory for energy.”

Through their partnership, SLB and NVIDIA will create AI infrastructure and models built around SLB’s existing digital platforms to help energy companies scale AI for data and operations.

In addition to the development of the “AI factory,” SLB will:

  • Provide modular design services to enhance NVIDIA’s blueprint for building, launching and operating gigawatt-scale AI data centers. In this case, modular design involves manufacturing data center components off-site.
  • Use NVIDIA’s AI infrastructure to improve the processing of large datasets and AI models across SLB’s digital platforms.

Energy companies generate vast amounts of operational data, which can slow down and silo decision-making, SLB says. By combining NVIDIA’s Omniverse libraries and its Nemotron open models with SLB’s digital and AI platforms, the companies aim to more rapidly transform data into actionable insights.

Omniverse libraries are sets of prebuilt 3D elements, such as objects, surfaces and interactive features, that make it easier to construct detailed virtual spaces without having to design everything manually. They’re commonly used for building immersive environments, digital replicas of real-world systems and simulation scenarios.

Nemotron open models are AI models that are freely available to download and modify. Instead of relying on a hosted service, you can run them on your own infrastructure and tailor them to fit specific needs.

Vladimir Troy, vice president of AI infrastructure at NVIDIA, says the energy sector is at the forefront of AI driving a “new industrial revolution.”

“The winners in AI will be companies with the best data, the deepest domain expertise, and the ability to scale,” Demos Pafitis, SLB’s chief technology officer, added. “By collaborating with NVIDIA to advance modular data center construction and harness our domain expertise and digital platforms, we’re enabling the energy industry to deploy AI at scale and transform operational data into smarter decisions.”

Trending News