A Chokepoint Closed and the Forecast Reversed

In December of last year the Energy Information Administration expected American crude oil production to fall. Its outlook put 2026 output at 13.5 million barrels a day, roughly a hundred thousand below 2025, which would have been the first annual decline after four consecutive years of growth. The reasoning was ordinary and sound. The agency assumed West Texas Intermediate averaging about fifty one dollars, and at that price it expected drilling and completion activity to pull back by more than productivity gains could offset.

By June the same agency was forecasting growth. The disruption to the Strait of Hormuz removed a very large volume of Middle Eastern supply from the market, Brent averaged around ninety five dollars for the year with a summer spike higher, and the American production forecast was revised up rather than down. The August outlook now has domestic output rising from 13.6 million barrels a day in 2025 to 13.8 million this year and 14.2 million in 2027. Nothing about American geology changed in those six months. What changed was the price, and what changed the price was a waterway.

The Historical Echo

American oil has done this before on a longer timescale, and the record is a useful corrective to confident forecasting in either direction. United States crude production peaked at 9.6 million barrels a day in 1970. It then declined, with interruptions, for thirty eight years, reaching 5.0 million barrels a day in 2008. That is a fall of roughly half, sustained across four decades, and by the end of it the depletion of American oil was treated in most serious analysis as a settled geological fact rather than a forecast.

It was not settled. Hydraulic fracturing and horizontal drilling reversed it beginning in 2009, and by 2023 output had reached 12.9 million barrels a day, well above the 1970 peak that was supposed to be permanent. Two generations of energy policy had been built on the assumption that the first number was structural, and the assumption was wrong, not because the geology had been misread but because the technology for reaching it had not been invented yet.

The instructive part is not that forecasts fail, which everyone already knows. It is the direction of the failures. The decline from 1970 was underestimated in its early years and then, once it became consensus, extrapolated far past its actual end. The same pattern is visible in the last nine months at compressed scale, with an expected decline replaced by expected growth on the strength of a single geopolitical event.

Where Patient Capital Is Positioning

There is a second detail in the data that deserves more attention than it gets. Between December 2022 and October 2025 the total American rig count fell from 750 to 517, with oil directed rigs down about a third. In the Permian, rigs fell 29 percent over that period while Permian production rose 18 percent, an increase of roughly a million barrels a day. The agency itself notes that the traditional link between rig activity and output has weakened. Fewer machines, drilling better placed wells, are producing more oil. Watching rig counts as a proxy for supply has stopped working.

This is why the ownership structure matters more than the activity headlines. A mineral or royalty interest is a claim on a percentage of gross production from a defined piece of ground, paid before costs and unburdened by them. The holder does not fund the drilling, the completion, the workovers or the eventual plugging, and does not absorb a cost overrun. What the holder owns is a share of whatever comes out, for as long as anything comes out. When operators get better at extraction, as they demonstrably have, the interest benefits without having financed the improvement.

That is a very old arrangement, and its appeal has little to do with any particular price forecast. Someone who bought a royalty on producing acreage in 1975 lived through the 1986 collapse, the long decline, the shale reversal and everything since, and was paid in barrels throughout. The barrels were real whether the forecast was right or not. For a family thinking in generations rather than quarters, that is the durable point. The analysis will keep being revised, sometimes by half a million barrels a day in a single season. The ground does not revise.

The Capital Memo