Opinion: Chevron Just Changed the Economics of Permian Gas

Chevron Just Changed the Economics of Permian Gas Waha hub gas, the stranded byproduct of West Texas oil drilling, has spent much of the past three years as a disposal problem rather than an asset. In 2023, spot prices at the hub turned negative on multiple occasions: producers actually paid buyers to take the gas away, rather than invest in gathering infrastructure or absorb the regulatory costs associated with flaring. The Permian Basin currently flares roughly 500 million cubic feet per day, losing around $1 billion in annual value that has nowhere economic to go. Microsoft's 20-year power purchase agreement with Chevron changes that calculation directly. The deal, a 2.67 gigawatt gas plant built on Chevron's Reeves County acreage, purpose-built for a dedicated AI data centre with no grid connection, creates a long-duration off-taker for gas that would otherwise compete for scarce pipeline takeaway capacity or be burned as waste. First power is expected in 2028. Coverage of the deal has focused almost entirely on Microsoft's energy strategy: the tension between a 20-year fossil fuel commitment and a 2030 carbon-negative pledge, and whether AI data centres can plausibly run on renewables. Those are legitimate questions. But the more consequential market effect runs in the opposite direction, what this deal does to the economics of Permian oil production. Associated gas in the Permian is a byproduct of crude oil drilling, not a primary commodity. When an operator sinks a well targeting the Wolfcamp or the Delaware Basin, they get oil, which is what they want, and they get gas, which they must either sell, flare, or reinject. For most of the past several years, the gas has been a drag. Pipeline takeaway capacity has lagged production growth; the Matterhorn Express pipeline, adding 2.5 billion cubic feet per day of capacity in 2024, only partially relieved the constraint. Waha prices averaged around $1.50 per million British thermal units in 2024 and have printed negative in summer months when pipeline congestion peaks. A 20-year off-take at any positive price transforms that drag into a revenue line. Chevron has converted previously stranded gas into a long-duration contracted income stream. Once that income stream exists, the breakeven economics of every Permian well sending gas to the Kilby project shift lower. An operator who previously drilled an oil well while treating the gas as a cost now drills the same well while treating the gas as a partial revenue source. Multiplied across Chevron's West Texas acreage, and, as further deals follow, across the basin more broadly, the aggregate effect on drilling economics is material. This mechanism has precedent at smaller scale. Bitcoin miners began co-locating with Permian flare sites precisely because stranded gas had a monetisation problem, and modular compute loads solved it. The Chevron-Microsoft deal applies the same logic at a scale that actually affects basin-wide supply dynamics. The 2.67 GW facility, running at standard heat rates, could consume somewhere in the range of 300 to 500 million cubic feet of gas per day, a substantial fraction of current basin-wide flaring volumes. The direction of the effect cuts against the deal's surface narrative. Publicly, it frames as a solution to AI's power problem and, with creative accounting, as a climate improvement over flaring. On the first point, the arithmetic is straightforward: data centres require 24/7 baseload power, interconnection queues for new grid-connected capacity now run five to seven years, and stranded Permian gas is cheap and physically available. Henry Hub settled at $3.23 per million British thermal units at Friday's close; Waha has historically traded at steep discounts to that benchmark. The maths works for Microsoft. The climate framing is shakier. Flared gas is methane being combusted rather than vented, which reduces its near-term warming impact compared to raw methane leakage. But compared to leaving the oil in the ground, the accounting looks different. More to the point: by improving Permian drilling economics, the deal may increase total gas production volume, some portion of which will still be flared or vented by operators without a Microsoft anchor contract elsewhere in the basin. The duration of the commitment deserves scrutiny in its own right. Corporate power purchase agreements typically run 10 to 12 years, long enough to support project finance, short enough to preserve flexibility as energy markets evolve. Twenty years is a categorically different proposition. It signals that Microsoft's internal planning horizon for AI data centre load does not include a scenario where renewable capacity costs fall far enough, fast enough, to displace gas-fired baseload before the mid-2040s. The company's sustainability team may hold a different view. The contracts team, evidently, does not. The second-order effect on Atlantic Basin gas markets is less immediate but traceable. TTF settled at $41.00 per megawatt-hour at Friday's close, against JKM at $15.52, a gulf reflecting asymmetric demand pressure following the Ras Laffan outage, which knocked an estimated 17% of Qatar's LNG export capacity offline. European buyers have been absorbing whatever Atlantic Basin swing supply is available. Every molecule of Permian gas locked into a 20-year domestic US power deal is a molecule that will not be liquefied and exported westward. The Chevron-Microsoft deal is one contract. Cleanview data showing 59 self-supply projects totalling 90 gigawatts globally suggests it is the leading edge of a structural trend. For oil market analysts, the Kilby project warrants more attention than it has received. Chevron has solved a gas disposal problem in the most profitable way currently available. The consequence, a floor under Waha pricing and a corresponding improvement in Permian drilling economics, is a supply variable that does not yet appear in consensus crude production forecasts. It should.