U.S. crude output keeps breaking records even as bearish forecasts miss by nearly a million barrels a day
With WTI front-month at $100.50 and the EIA projecting yet another production record, the case for a U.S. supply slowdown is harder to construct than it was a year ago.
WTI crude front-month traded at $100.50 a barrel on Friday (2026-09-11), down half a percent on the session but sitting roughly 54% above the $65 annual average that U.S. producers navigated through 2025, when they set a national output record anyway. Production grew 2.7% to 13.586 million barrels per day that year, a record, despite prices that bearish analysts had expected to deter growth. The arithmetic of that combination is worth sitting with: operators expanded volumes through softer prices, and now face considerably better margins.2
The EIA's August 2026 Short-Term Energy Outlook, released on August 11 (2026-08-11), projected that annual U.S. crude production will reach another record this year, extending a streak the agency traces to 2018, when the U.S. first overtook Russia as the world's largest crude producer.5 In 2025, U.S. crude and lease condensate averaged 13.586 million bpd, accounting for 15.8% of global supply. Russia averaged 10.161 million bpd. Saudi Arabia managed 9.727 million bpd. U.S. output ran roughly 40% above those two combined.2
The bearish view had genuine foundations in mid-2025. S&P Global, in a June 2025 forecast, projected U.S. crude output would fall 640,000 bpd by year-end 2026, to 12.96 million bpd — the first annual decline since 2020.3 That call is now running close to a million barrels per day behind where EIA data and projections point. The miss is large enough that the structural assumptions behind it deserve scrutiny before the same framework is applied to the rest of 2026 and into 2027.3
The Permian Basin is the core of the discrepancy. It accounted for 48% of U.S. crude output in 2025, averaging 6.6 million bpd after a 4% annual increase, per EIA data.2 Production concentration at that scale means aggregate U.S. output is effectively a single basin's story, and bearish models built on industry-wide drilling economics will miss if the Permian operates well inside typical cost curves rather than near them.
Longer laterals and more productive well completions drove that Permian growth, not a price windfall. Operators there expanded volumes through 2025's $65 WTI average; at $100.50 on Friday (2026-09-11), those same efficiencies are considerably more profitable. An operator who grew production through $65 crude does not slow activity when prices move into triple digits, absent a specific geological or capital constraint that the efficiency data has not yet revealed.2
The EIA's May 2026 STEO projected U.S. crude will average 14.10 million bpd in 2027.1 Annual U.S. field production has never averaged 14 million bpd, nor reached that level even in a single month, according to EIA data going back to January 1920.1 Crossing that threshold would mark a step-change in U.S. supply capacity. Prior to 2025, only one year — 2024, at 13.235 million bpd — had ever averaged above 13 million bpd.1
Natural gas is running the same pattern. EIA projects U.S. marketed gas output will average 122.5 billion cubic feet per day in 2026, above the record 118.5 Bcf/d set in 2025, according to the August 2026 STEO.4 NYMEX Henry Hub front-month held at $2.82 per MMBtu on Friday (2026-09-11), with EIA projecting a production record regardless. Simultaneous crude and gas output records, driven by overlapping basin efficiency gains, point to a growth trajectory that does not depend on any single commodity's price for momentum.4
The monthly EIA production data for the second half of 2026 will either support or complicate the bullish supply thesis. The second-highest monthly U.S. crude output figure on record was 13.828 million bpd in September 2025, and the third-highest was 13.810 million bpd in August 2025, per EIA data.1 If late-2026 readings approach or surpass those marks with WTI front-month in triple digits, the sustained-slowdown thesis becomes very difficult to maintain. If output stalls well short, it would signal that geological limits are beginning to bite in the Permian in ways the efficiency data does not yet show — and that the bears missed the timing but not the direction.1