Presidio: A 3.2% Free Lunch in the Oilfield, Priced Like a Yield Trap
There are two Presidio Production Companies in this transcript. One is a small Oklahoma-and-Arkoma PDP buyer, roughly a year into public life, that lost $46.4 million in the third quarter. The other is a self-styled industrial AI platform that lifted production 3.2% on wells it already owned without spending a dollar of capital — and did it while the rest of the market was busy marking down anything with "AI" in the ticker.
The company's own framing is a Formula 1 parable: Ross Brawn buying Honda's abandoned team for a reported single pound sterling and winning a championship. “You don't win by building everything from scratch. You buy a platform and then you use the skills that are uniquely yours to make it more valuable than anyone else could.” — William Ulrich, Chairman and Co-CEO · 2026-11-13 Translated: Presidio does not drill. It buys producing wells cheap, fires the prior operator's cost structure, and monetizes the undeveloped acreage underneath through carried interests and outright sales.
The AI that isn't a capex story
The credo got its first real test this quarter because Canyon Creek — closed in July, forty-two wells — was in the numbers for a full period for the first time. Company-wide lease operating expense fell to $9.10/BOE from $9.39, production hit a public-company record of 26,133 BOE/d, and management projects about a 32% LOE cut at Canyon Creek after roughly 30% at the IPO assets.
But the number that matters for the multiple is the AI contribution. Field Intelligence, the surveillance agent, ranks every well against its expected rate; a weekend well-check module sends shorthanded crews to the right well first; an AI-driven plunger box sits on 25 wells; and an AI workover adviser reviews every job. The scoreboard went from 2.3% in Q2 to 3.2% in Q3 against a 3%–5% full-year target. “In the third quarter, our AI programs added approximately 3.2%, up from the 2.3% in the second quarter.” — Christopher Hammack, Co-CEO and Director · 2026-11-13
Then the tell. “When we set these production goals for 3% to 5% gains in '26, we were — I think we thought we were being aggressive, and we've come to find out we were actually probably being very conservative.” — Christopher Hammack, Co-CEO and Director · 2026-11-13 Will Ulrich sizes the prize at a 15% gap between actual and theoretical production per well — so the 3%–5% was only ever a quarter to a third of the addressable uplift. Yet when an analyst asked what the 2027 plunger-box rollout looks like, the answer was flat: “I don't think that we have any thoughts on it right now.” — William Ulrich, Chairman and Co-CEO · 2026-11-13 Enthusiasm on the mechanism, silence on the number.
That reticence is worth weighing against the tape. The market's AI trade is de-rating: over the last 30 days the biggest decliners in our theme screens are High Bandwidth Memory, co-packaged optics and data-center capacity names, all off double digits. Meanwhile AI agents have become a consensus talking point across this week's reporters — from job-search agents to agentic media buying. Presidio's version is capital-light and attached to oilfield hardware, which is why it earns a hearing the generic personal AI agents chorus does not. Will even pre-empts the objection with a jab at the industry's Shale 1.0 scoreboard:
AI agents are having their Shale 1.0 moment right now. The technology is just as real, but look at how companies are keeping score. They count how many agents they've deployed, how many tokens they've burned, how many lines of code they've shipped... You spent how much to build what?
The rate hike that isn't their problem
Rates were among the most-discussed themes in the market's Q4 vocabulary for a reason: the CFO notes Polymarket now quotes a 98% probability of at least a 25bp hike at the December FOMC, with more expected in the first half of 2027. For most leveraged small caps that is a headwind.
Presidio turned it into a moat. In June the company fixed the coupon on its $344 million core ABS warehouse debt at 6.38% — roughly 50bp inside where it would price today — and after quarter-end swapped 75% of the $54 million variable-rate warehouse balance to a fixed 7.2%. Blended, the cost of debt is effectively 6.6% and locked. “In June, we fixed the interest rate of our core $344 million of debt at 6.38%. Today, I estimate that facility would be close to 50 basis points more expensive given the rise in rates.” — John Brawley, Executive Vice President and CFO · 2026-11-13 Better still, the same rate hike environment they've immunized against should widen acquisition discount rates for the sellers they buy from. “Higher rates will also have the potential to widen acquisition discount rates, all of which is actually helpful to our strategy.” — John Brawley, Executive Vice President and CFO · 2026-11-13
The catch is that fixed-rate debt is not the same as a fixed-rate business. The quarter's GAAP loss was almost entirely a $66.5 million unrealized non-cash mark on the commodity hedge book as prices rose — the insurance policy paying out in the wrong direction on paper. Adjusted EBITDA also stepped down sequentially, from $33.2 million to $29.1 million, which management attributed to the shape of pricing in that same hedge book. Hedge durability costs money in good quarters; that is the trade.
Unmodeled profit, and a borrowing base that shrinks
The quietest value driver is the one that never made it into guidance. Leasehold sales — selling undeveloped acreage to operators who do want to drill — generated $17 million year to date, including $7.1 million already in Q4, and flows straight to the bottom line outside the EBITDA numbers. Management was candid that this wasn't in the IPO outlook at all. Expect the same playbook on the newer Arkoma assets once the in-house technical read is done.
Against that, two items deserve a raised eyebrow. The RBL borrowing base is expected to fall from $60 million to roughly $45 million at the winter redetermination on hedge cash-flow roll-off — immaterial with the facility undrawn, but a reminder that the collateral is a depleting asset. And when asked why raise the dividend rather than pay down debt, the answer leaned on narrative as much as math: “There's the market perception and there's also the realities of our business.” — William Ulrich, Chairman and Co-CEO · 2026-11-13 The realities, per the CFO's own arithmetic, are $349.4 million of net debt against $124.5 million of last-six-months-annualized adjusted EBITDA — 2.8x — with liquidity near $100 million and a $1 billion ABS warehouse as dry powder against a claimed $17 billion actionable acquisition pipeline across three basins. The quarterly distribution rose to $1.50 annualized from $1.35, funded by the exit rate potential of the reactivated wells rather than by the mark-to-market noise.
What to watch
Three things decide whether this is a compounding acquirer or a yield trap. First, whether Deal Intelligence and the plunger box rollout convert into a quantified 2027 production target — the company has the mechanism and refuses to give the number. Second, whether leasehold and non-producing asset sales stay a recurring, budgeted line rather than a happy accident that flatters a quarter. Third, whether the next Canyon Creek-scale deal clears at the same levered returns now that everyone else's cost of capital has moved against them. Presidio's argument is that its engine, not its balance sheet, produces the return — which is exactly the claim a rate cycle is built to test.