BEWARE A TRADE whose condition has already been met, twice over. Six months ago the question occupying energy desks was whether the ICE gasoil crack — the margin on turning a barrel of crude into diesel — would push above $35 a barrel. It has since touched $74.66. The NYMEX 3-2-1 spread, the American refiner's blended margin, has reached an all-time high of $72, eclipsing even the 2022 energy-crisis peak. The trade is no longer about whether spreads will widen. It is about whether they will persist, and where — if anywhere — the asymmetric upside has not yet been priced in.
The supply-side answer is unusually clean. Ukrainian drone strikes have pushed Russian refinery runs to a 21-year low of around 3.8m barrels a day. Kpler estimates that 4.3m b/d of capacity — some 58% of Russia's refining base — is now in downtime or has been attacked, with 1.5m-2.0m b/d effectively offline. The Financial Times' own count puts more than 30% of operating capacity and 45% of nominal capacity out of action. Moscow's largest refinery, hit in June, is unlikely to restart before 2027. These are not the four-to-six-week outages a refiner can wait out. Crude distillation units, hydrocrackers and catalytic reformers — the equipment that turns a barrel of oil into something useful — take months to repair.
The destruction of Russian product exports is the empirical signature. In July 2026 seaborne product shipments fell by a third month-on-month to about 3.9m tonnes. Diesel and gasoil, the highest-value products, fell 60% to 0.75m tonnes; naphtha dropped 35% to 0.8m; even dark products — the residual fuel oil and vacuum gasoil that make up 60% of volumes — were down 21% to 2.3m tonnes.
Russian July 2026 seaborne oil product exports fell 33% m/m to 3.9 MMt; diesel/gasoil dropped 60% to 0.75 MMt, naphtha 35% to 0.8 MMt, and fuel oil + vacuum gasoil (dark products) 21% to 2.3 MMt.
| category | July 2026 volume (MMt) | M/m change (%) |
|---|---|---|
| Total oil products | 3.9 | -33 |
| Diesel and gasoil | 0.75 | -60 |
| Naphtha | 0.8 | -35 |
| Fuel oil and vacuum gasoil (dark products) | 2.3 | -21 |
Russia's response has been to lock in the tightness rather than relieve it. A gasoline export ban runs through end-2026; from August 1st the ban was widened to cover diesel, marine fuel and gas oil through January 31st 2027. By the KSE Institute's reckoning, that puts up to 36% of total Russian product export volumes at risk. Russia, which supplied roughly 0.9m-1.0m b/d of diesel to the global market before the ban, has unilaterally withdrawn most of that supply.
The result is a market with no spare wall of refining capacity left to switch on. Non-Russian refiners are at their "practical operating ceiling", according to Rapidan Energy, deferring maintenance to keep runs maxed. The IEA reckons global refinery output fell by 4.5m b/d, or 5.4%, in the second quarter of 2026. Seven American refinery closures since 2019 have removed another 1.2m b/d of capacity that will not come back. Middle East export refineries damaged in the Iran war remain shut. China's recent decision to allow its refiners to resume product exports could cushion the gap — but, as TD Securities notes, only by re-tightening the crude market. Refiners are at the ceiling. The pricing of that ceiling is visible in the milestones of the 3-2-1 itself.
The US 3-2-1 crack spread climbed from a $51/bbl Q2 2026 average to a record $72/bbl in early August, surpassing the 2022 energy-crisis peak and more than doubling the thesis's $35/bbl ICE gasoil threshold.
| period | 3-2-1 crack spread (/bbl) |
|---|---|
| Q2 2026 (average) | 51 |
| July 9, 2026 | 60 |
| July 16, 2026 | 70 |
| Early August 2026 | 72 |
The milestones tell the story of how that physical tightness has been priced. The 3-2-1 — three barrels of crude cracked into two of gasoline and one of diesel — averaged just over $51 a barrel in the second quarter, within 20 cents of the 2022 quarterly record. It then climbed through $60 on July 9th, $70 on July 16th, and $72 in early August. The American 3-2-1, the European ICE gasoil crack and the US ULSD-versus-WTI diesel crack (in the high $90s in late July, above $100 in 2022) are three different metrics of the same underlying tightness. All three point the same way: the refiner's margin has decoupled from the crude price.
The concession is real. Retail diesel above $5 a gallon in mid-July and gasoline near $4.55 a gallon in the spring pass directly into trucking and logistics costs, and from there into consumer-goods inflation. Hold policy rates higher for longer and the freight demand that justifies a $70 crack begins to soften. The 2022 precedent — when the US diesel crack eclipsed $100 a barrel — is a reminder that demand destruction and substitution can cap spreads at extremes. Yet the IEA's July Oil Market Report currently sees demand recovering from its May 2026 nadir, with global consumption set to rise by more than 8m b/d from the low by October. By that read, demand destruction is not yet biting. The threshold to watch is whether subsequent IEA monthly reports begin cutting distillate demand estimates for the second half of 2026 — the moment the persistence thesis weakens.
For an energy portfolio, that distinction matters more than the headline number. The persistence case for elevated cracks is structurally intact: Russian capacity is offline for months, the export ban runs into 2027, non-Russian refiners cannot run harder, and demand is recovering rather than collapsing. The trade, however, has already been put on in the obvious place. Marathon PetroleumMPC-- and ValeroVLO-- have nearly doubled in 2026; Phillips 66PSX-- is up 66%. Roughly a third of that move has come in the past month. The pure-play American refiners — the cleanest expression of the 3-2-1 crack — have already done much of the work.
Where the persistence case still leaves room is in the secondary expressions. Integrated majors with refining exposure, such as ExxonMobil, capture the margin uplift but diluted by upstream earnings — less upside, less asymmetric risk. Complex overseas refiners — Reliance, ENI, Total, Repsol — offer exposure to the same product tightness from a less crowded starting point, though currency, tax and crude-sourcing complications make them a less clean expression. None of these is a hidden gem. The honest assessment is that the asymmetric upside available six months ago has narrowed, and what remains is a persistence bet: that Russian refining does not come back faster than expected, that the export ban holds, and that the IEA does not start cutting distillate demand estimates.
The danger is not that the original thesis was wrong. It was right, sooner and bigger than expected. The danger is that the trade has already been put on by everyone else — and that the only investors still finding "asymmetric upside" are the ones arriving last.
Commentaires
Pas encore de commentaires