Permian Gas Has Been Worth Less Than Nothing for 118 Days This Year

What I Cover My focus spans the global energy landscape and the technologies reshaping it, specifically focusing on oil and gas, renewables, and tech-driven market…

Outside Pecos, Texas, Chevron is getting ready to build a power plant to solve a problem the Permian Basin has been stuck with for years, which is that it produces far more natural gas than anyone nearby is willing to pay for.

Chevron's subsidiary, Energy Forge One, signed a 20-year agreement with Microsoft in June to build a gas-fired plant beside a data center campus, ramping up to roughly 2.67 gigawatts. The final investment decision is due by the end of this year, with power flowing in 2028.

One deal in one county doesn't make a trend, but it does show how three separate energy problems in West Texas, all of them years in the making, are starting to find the same customer.

Most of the talk out of Midland this year has been about restraint, with fewer rigs, tighter budgets and what one trade publication called a transition to slower, more disciplined growth . That describes the oil side well enough, but it misses what's happening with gas.

WTI averaged $65 a barrel in 2025 and roughly $84 through July of this year, well above the $63 to $69 breakeven that Midland and Delaware basin operators reported to the Dallas Fed , so producers kept drilling.

The catch is that every barrel now brings up more gas than it did a few years ago, a rising gas-to-oil ratio that tends to come with age in almost any oil field.

The EIA expects Permian gas production to average 29.2 Bcf/d this year , up 6% and a record for the region.

Waha, the pricing hub for Permian gas, traded negative on 118 of the first 131 trading days this year and is on pace to beat 2024's record for negative days.

Negative means producers pay someone to take the gas off their hands, which they do because the alternative is shutting in the oil well that's actually making them money. So the Permian's trouble isn't production; it's that one of its two products spends much of the year worth less than nothing.

Nobody in the midstream business is acting as if this ends soon, either. Enterprise Products Partners is expanding its Bahia NGL pipeline out of the Midland and Delaware basins, with ExxonMobil taking a 40% stake and contracts running into 2027.

East Daley Analytics figures LNG terminals and data centers together will pull more than 20 Bcf/d of new demand, and that the pipelines now being built toward the Gulf Coast won't sit empty waiting for it.

Their read is that producers may end up drilling gassier acreage on purpose, just to keep both products moving through infrastructure that was designed mostly around crude.

A behind-the-meter data center changes that without anyone laying new pipe...

RBC Capital Markets counts about 38 gigawatts of announced behind-the-meter gas capacity in Texas, more than any other state.

One project alone, a build called GW Ranch , is targeting 7.5 gigawatts of on-site generation in West Texas by 2031, which is roughly what a mid-sized American city draws.

One claim about these projects deserves more skepticism than it tends to get, which is that they're soaking up gas that would otherwise be flared. Nobody has demonstrated that for Project Kilby.

An environmental compliance firm working in the sector made the point in a client note this summer, arguing that sitting near a stranded well doesn't prove a data center's fuel is displacing flared volumes, and that most developers haven't produced the project-level evidence to back it up.

Flaring in the Permian has been falling anyway, with intensity down nearly 10% between 2023 and 2024 , largely because pipelines like Matterhorn Express finally gave the gas somewhere to go.

The simple fact is that the Permian makes more gas than the market can absorb even after the pipelines get built, and data centers are the first buyers in years with enough money and enough impatience to put a power plant where the gas already is.

Nolan County, with Sweetwater at its center, has been called the wind turbine capital of Texas since the mid-2000s, when hundreds of turbines went up on the plateaus between Abilene and Big Spring.

Moving that power east required new transmission, which the state paid for through a roughly $7 billion program called CREZ . Those lines added about 18,000 megawatts of capacity , they've been paid down for years, and they're still carrying electricity today.

That's a large part of why data center developers keep landing along this stretch of I-20, because the wires went in the ground long before anyone was planning gigawatt campuses.

I wrote about what that's done to land values and ranch economics in the corridor in West Texas Cowboys Are Cashing In on the AI Land Rush .

The piece of it that matters here is that wind already solved the region's east-west transmission problem, and everyone building now is spending capital somebody else committed two decades ago.

A new high-voltage line called the Permian Reliability Project is working through siting now, and it's expected to do for far West Texas solar roughly what CREZ did for wind around Abilene, turning land that has plenty of sun and no way to sell the power into land a developer can actually finance.

Where the next round of renewable buildout lands may look different from the last one, because turbines need open ground and distance from flight paths.

One large project near Abilene switched from wind to solar after developers ran into airspace conflicts with Dyess Air Force Base. Panels have fewer of those constraints, and once the Permian Reliability Project's lines are up, solar sits much closer to the data center campuses now scouting land in the basin itself rather than in the Abilene corridor to the northeast.

Putting data centers next to stranded gas isn't new in West Texas, it just hasn't been done at this size before.

Miners have spent years hauling shipping containers full of rigs and small generators out to remote well sites and buying gas straight from producers who had no other way to sell it.

ConocoPhillips ran an early version in the Bakken , ExxonMobil piloted its own , and the model spread across the Permian from there.

Mining works for this because it doesn't care where the power comes from or whether the connection is reliable, so it can take energy nobody else wants, right where it's produced, without waiting on a pipeline permit.

Ryan Dusek, a commodity risk advisor at the consultancy Opportune, has called it one of the few genuinely symbiotic arrangements in energy, since producers get rid of gas they can't move and miners get power at wholesale with no exposure to grid outages.

Pecos runs on that same trade at a completely different scale.

A container of mining rigs might burn a few million cubic feet of gas in a year, while a 2.67-gigawatt power plant is another business entirely, which is why what miners were doing out here a few years ago now looks like a rehearsal for something much bigger.

Governor Greg Abbott ordered a pause on new data center approvals on Aug. 3 , telling ERCOT and the Public Utility Commission of Texas to audit every project sitting in the grid's interconnection queue, which had grown to roughly 474 gigawatts of requests against a record peak demand of about 85 gigawatts.

The audit covers around 300 large projects in ERCOT's Batch Zero process and looks at water use, ownership, public subsidies and a good deal else.

None of it applies to projects that skip the grid connection .

A developer building its own generation on site, the way Chevron and Microsoft are doing at Pecos, can keep going while the queue sits frozen.

That fits what the directive was written for, since Abbott's stated concern is grid reliability for Texans who depend on the grid, and a data center making its own power was never drawing from it.