Subscribe to The Y’all — a weekly dispatch about the people, places and policies defining Texas, produced by Texas Tribune journalists living in communities across the state.
ODESSA — Underneath the Permian Basin, the state’s largest oil field, lies an ocean of toxic, unusable wastewater that bursts out of rock formations when oil companies extract fossil fuels from the ground. For years, companies have struggled with how to dispose of it.
Now, many are turning to an ubiquitous, albeit controversial, technology to solve the problem — artificial intelligence.
In the race to keep up with skyrocketing demand for crude, oil and gas companies are increasingly turning to AI to gain an edge over their competition.
AI, experts and analysts said, is reshaping how oil companies handle the saltwater slush, also known as produced water, by giving operators more information about the region’s geography. It’s also providing access to data more quickly and allowing operators to file permits faster.
Despite its growing use in the industry, few details are available to the public — or to the industry’s regulatory agency.
“It’s a competitive advantage,” said Ramanan Krishnamoorti, professor of petroleum engineering at the University of Houston. “That’s the reason why nobody’s really talking about it.”
The Texas Railroad Commission, which regulates the state’s oil and gas, said in a statement that it is not aware of any use of AI by oil and gas companies.
‘It’s a huge time saver”
Shooting produced water back underground has been a longstanding practice for oil and gas operators. For every barrel of oil, at least three to 10 times as many barrels of produced water rush to the surface.
AI collects information about surrounding wells, including details about the geology and the materials used to build them, such as casing and cement, said Akash Sharma, vice president of product at Enverus, an energy analytics company. Operators are also using AI to determine whether the wells are active. In other cases, AI can assess whether a seismic event is connected to a specific well. The AI can also pull information from multiple datasets and study the relationship between each.
Last summer, the railroad commission made its permitting standards stricter. Companies have to study two miles of the surrounding land where they plan to inject, up from half a mile. There are more limits on how much pressure and water companies can inject. Permits also require information that calls for more intensive engineering and geological work.
Hundreds of other wells can be found within two miles, Sharma said. Companies gather this information before filing for a needed permit, Sharma said. And they can do it faster.
“(AI) allows us to file more permits, allows us to actually think through strategy a lot more effectively, and just overall streamlining the process,” Sharma said. “It’s a huge time saver, especially on things that saltwater disposal teams should be focused more on, what the disposal strategy is, how to do it safely and securely.”
AI can also help determine the best-case scenarios for injection, said Yoshi Pradhan, a former production engineer at Chevron and founder of IronLady Energy Advisors. Pradhan said the condition of the wells can also be examined using the technology.
“You look at historical data,” she said. “You look at historical data to identify where there have been risks, what has happened, and you have AI, gather statistics and extract all that data.”
Running out of space
As fossil fuel production soars in the Permian Basin, gargantuan volumes of produced water rise up to the surface with it. Previously, operators injected the water into disposal wells deep underground, which led to earthquakes. As a result, in 2023, the railroad commission suspended the practice. In 2025, it imposed more limits on injection. Underground, operators are running out of space.
“Texas has never had more produced water every day than it does now,” said Adam Peltz, senior director and lead counsel of the Environmental Defense Fund. “The super boom in the Permian Basin has led to billions of barrels of this toxic salt water that needs to be dealt with.”
The oil and gas industry has begun exploring its options. Some are treating the water for possible discharge into public waterways if they meet state and federal standards. Others are reusing the water to extract more oil.
Permit requirements for saltwater injections have also become increasingly complex, said Kelly Bennett, co-founder and CEO of B3 Insight. Applying for a permit requires additional engineering and geologic work, which AI has helped streamline, Bennett said.
“The investment required to file a permit is significantly higher today than it was before, and a lot of it is because the (railroad commission) is requiring significantly more technical analysis before permitting really begins,” he said.
He said AI can ultimately help reduce disruptions to oil and gas production.
“It really helps contextualize issues that we see today against future market conditions,” Bennett said. “I think that’s really important for this market in which we have to plan differently, we have to develop infrastructure differently. It costs more money.
“The reality is that if you can’t dispose of barrels, you can’t produce barrels.”