Diesel cracked above $100 a barrel on the US Gulf Coast last week while crude itself drifted lower, and the IEA expects global refinery runs to bottom at 81.3 mb/d this month. The binding constraint in the oil market is no longer the volume of crude produced but the capacity available to turn it into middle distillates.
TL;DR
- The crack: front-month NYMEX ULSD against WTI stood at $96.80/bbl on 9 October, having fallen $6.45 on the day; the US Gulf Coast ULSD crack against WTI MEH averaged $104.45/bbl that week.
- Refinery runs: the IEA projects a seasonal low of 81.3 mb/d in October, recovering to 84.6 mb/d by year-end.
- Not enough: the agency states that recovery will be insufficient to offset the middle-distillate deficit already accumulated this year.
- Russian throughput: 3.8 mb/d in June against roughly 6.5 mb/d of installed capacity, down approximately 30% year on year.
- Crude is soft: WTI traded at $94.03 and Brent at $104.30 on Thursday morning, both lower on the session.
- Forward risk: Ukraine has said it will resume strikes on Russian refineries, raising the prospect of further losses.
The Market Snapshot
| Instrument | Latest | Prior |
|---|
| WTI crude, Nov | $94.03 | $94.10 |
| Brent crude, Dec | $104.30 | $105.51 |
| Brent–WTI spread | $10.27 | $11.41 |
| NYMEX ULSD–WTI crack, front month | $96.80/bbl | — |
| US Gulf Coast ULSD–WTI MEH crack | $104.45/bbl | $101.03/bbl |
| US Gulf Coast cracking margin, WTI MEH | $54.61/bbl | — |
WTI and Brent are screen observations on 15 October 2026 against 14 October closes. The NYMEX crack is the front-month contract against WTI at 19:33 GMT on 9 October, as reported by S&P Global, down $6.45/bbl that day. Gulf Coast figures are S&P Global Commodity Insights weekly averages to 9 October, with the September average shown as prior; they reference WTI MEH rather than Cushing, so the two cracks are not directly comparable.
Crude is not the story. WTI has moved seven cents in a session and Brent has given back just over a dollar. The Brent–WTI spread narrowed from $11.41 to $10.27, which bears on regional crude pricing and says nothing by itself about product availability.
The product market is where the pressure sits, and two separately defined benchmarks confirm it. The front-month NYMEX ULSD crack against WTI was $96.80 a barrel on 9 October. The Gulf Coast ULSD crack against WTI MEH averaged $104.45 over the same week, up from $101.03 in September. The difference between them reflects both the crude reference and the regional product market rather than one figure correcting the other. What they establish together is that diesel cracks ran above $100 a barrel, which the IEA's own September analysis also records for both the US Gulf Coast and north-west Europe.
One Stream Is Carrying the Barrel
The margin data show where that premium is going. Against a Gulf Coast diesel crack of $104.45, the Gulf Coast refinery cracking margin on the same WTI MEH basis averaged $54.61 a barrel over the week to 9 October.
A refinery does not produce only diesel. The gap between the two figures is what the rest of the barrel is worth, and it is substantially less than the distillate stream. Refiners running for diesel are earning an extraordinary margin on that cut while gasoline, fuel oil and the lighter ends contribute far less than the headline crack implies.
That matters for how quickly the squeeze resolves. A refiner maximising distillate yield is already doing so; the configuration of existing plant limits how far the barrel can be shifted towards middle distillates regardless of price. A crack of this size is a signal that cannot be fully acted upon, which is why it persists.
Three Different Constraints
The temptation is to treat the refining squeeze as a single event. It is not. Three distinct pressures are operating, with different causes and different durations, and the distinction determines how long the tightness lasts.
Seasonal maintenance. The IEA projects global refinery runs reaching a seasonal low of 81.3 mb/d in October before recovering to 84.6 mb/d by year-end. Autumn maintenance happens every year and reverses every year. On its own this would be unremarkable.
Existing physical damage. Russian refinery throughput ran at 3.8 mb/d in June against installed capacity of approximately 6.5 mb/d, a year-on-year decline of around 30%, with gasoline production down 20% and diesel production down close to 30%. The IEA forecasts throughput of 4.0 mb/d for the remainder of 2026. Russian and Middle Eastern diesel export losses totalled 1.3 mb/d in September, roughly 4.5% of the global diesel and gasoil demand forecast for this year, and global seaborne diesel and gasoil exports averaged 4.7 mb/d between January and August, down 10% year on year. Middle Eastern product shipments remain below half their pre-war level.
Prospective disruption. President Zelenskyy said on 15 October that Ukraine would resume strikes on Russian oil refineries. That raises the probability of further capacity losses without establishing any.
The gap between Russia's 6.5 mb/d of installed capacity and its 3.8 mb/d of actual throughput should not be read as a measure of capacity destroyed. Throughput reflects damage, economics, feedstock and maintenance together. What it does measure is crude actually processed, which is the figure that matters for product supply.
Why Diesel and Not Crude
The mechanism explaining the whole divergence sits in which units have been hit.
The IEA notes that attacks on hydrocrackers and hydrotreaters can reduce diesel production even where crude distillation capacity remains operational, and that serious damage to those secondary units can take six to eight months to repair. A refinery can therefore continue processing crude while producing materially less diesel than before.
That is why the product market is tight when the crude market is not, and why the easing the IEA foresees arrives in the second half of 2027 rather than next quarter. It also explains a point that is frequently misread: the recently announced easing of Russian diesel export restrictions is unlikely to change the balance materially, because the constraint is a lack of operational refining capacity rather than export policy. Permission to export does not create barrels no refinery has produced.
The Recovery Will Not Close the Gap
The IEA's central finding is more consequential than the October trough. The agency states that the expected year-end recovery in refinery operations, while it should begin to replenish stocks, will be insufficient to offset the deficit accumulated in middle distillates this year.
Diesel markets have been under what the agency calls significant strain, with prolonged supply disruptions forcing counter-seasonal stock draws — inventories falling during the period when they would normally build, ahead of autumn maintenance and peak winter demand. US distillate stocks stood at 105.1 million barrels for the week ended 2 October, approximately 12% below the five-year average.
The agency expects the global gasoil balance to remain tight through the first half of 2027, easing only in the second half, and makes that easing conditional on a recovery in Middle Eastern product supply.
The Crude Balance Is Looser
The crude picture explains why benchmark prices are flat while products are not.
The IEA forecasts global oil supply averaging 100.7 mb/d in 2026, a decline of 5.8 mb/d year on year, rebounding 7.7 mb/d in 2027. Demand is expected to fall 2.6 mb/d to 102.4 mb/d this year before recovering 2 mb/d in 2027, which would still leave consumption 620,000 b/d below 2025 levels. Diesel accounts for roughly a quarter of this year's demand decline.
Supply falling faster than demand keeps the market in deficit through year-end despite a sharp demand contraction. But a crude deficit of that construction, driven by outages rather than consumption strength, produces a different price signal from one driven by demand. Crude has somewhere to go when the disruption eases, which is why the 2027 rebound of 7.7 mb/d is expected to tip the market into surplus. Refining capacity damaged in secondary units does not recover on the same schedule.
What the Falling Crack Tells You
The most informative detail in the pricing is that the NYMEX crack was falling when observed. The $96.80 reading on 9 October came after a $6.45 decline that day, following announcements regarding possible additional Russian diesel supplies and a US sanctions waiver.
A crack at this level and rising would indicate a market discovering scarcity. A crack at this level and falling on supply news indicates a market that has already priced severe scarcity and is marking down the margin as the probability of relief improves. The level remains extraordinary; the direction on that day was towards easing.
It also shows how sensitive the product market has become to policy headlines rather than refining fundamentals. A sanctions waiver moved the crack more than six dollars in a session, which is characteristic of a market where physical availability is thin enough that marginal supply news carries disproportionate weight.
Outlook
The immediate question is whether the projected recovery in refinery runs from 81.3 mb/d to 84.6 mb/d arrives on schedule. Maintenance programmes are planned and this is the most predictable element in the current balance.
Beyond that, two things determine whether middle distillates stay tight into 2027. The first is Russian refining throughput, running at 3.8 mb/d in June with the IEA forecasting 4.0 mb/d for the remainder of this year, and where Ukraine has now signalled further strikes against secondary units that take six to eight months to repair. The second is Middle Eastern product supply, which the IEA identifies as the condition for the gasoil balance easing in the second half of 2027.
Crude prices are the least informative variable in this market. With supply and demand both contracting and the deficit driven by outages rather than consumption, the benchmark tells you less about physical conditions than the crack does.
Key Risks
Refinery recovery timing. The move from 81.3 mb/d to 84.6 mb/d is a forecast. Delay would extend the distillate deficit into the winter demand peak.
Secondary-unit damage. Hydrocracker and hydrotreater repairs take six to eight months. Further strikes would extend the diesel constraint well beyond any crude recovery.
Relief is priced in increments. The NYMEX crack fell $6.45 on a sanctions waiver and supply announcements, so marginal good news moves the product market materially.
Crude and product divergence. Benchmark crude prices are a poor proxy for the constraint, and analysis anchored to WTI or Brent alone will misread the physical market.
Forecast dependence. The IEA's 2027 easing is conditional on Middle Eastern product supply recovering, which is itself conditional on conflict resolution.
Intelligence Monitoring Points
- Global refinery runs, against the IEA's 81.3 mb/d October trough and 84.6 mb/d year-end projection.
- Diesel cracks, from $96.80/bbl on NYMEX and $104.45/bbl on the Gulf Coast in the week to 9 October.
- Gulf Coast cracking margins, at $54.61/bbl against the diesel crack of $104.45/bbl.
- Russian refinery throughput, at 3.8 mb/d in June against an IEA forecast of 4.0 mb/d for the remainder of 2026.
- US distillate stocks, at 105.1 million barrels and approximately 12% below the five-year average.
- Middle Eastern product exports, currently below half pre-war levels.
The Bottom Line
The oil market's constraint has moved downstream. Crude is trading within a few cents of yesterday's level while diesel commands a premium over it above $100 a barrel on the Gulf Coast, and the IEA expects middle distillates to remain tight well into 2027 regardless of what crude production does.
The reason is specific rather than general. Russian refinery throughput has fallen about 30% year on year to 3.8 mb/d, and the damage includes hydrocrackers and hydrotreaters — the units that make diesel rather than the ones that process crude. Those take six to eight months to repair. A refinery can run crude and still not make distillate, which is why export permission changes nothing and why the agency's easing sits in the second half of 2027.
The margin data show the constraint from the other side. A Gulf Coast diesel crack of $104.45 against a total cracking margin of $54.61 means one stream is carrying the barrel while the rest earns considerably less. Refiners already maximising distillate yield cannot respond further, which is precisely why a price signal this large has not resolved itself.
Data and source note: WTI and Brent prices are screen observations on 15 October 2026 compared with 14 October closes and are not exchange settlements. The front-month NYMEX ULSD–WTI crack of $96.80/bbl refers to 19:33 GMT on 9 October 2026 as reported by S&P Global Commodity Insights, down $6.45/bbl that day. US Gulf Coast ULSD–WTI MEH cracks and cracking margins are S&P Global Commodity Insights averages for the week to 9 October 2026 and for September 2026; they use WTI MEH as the crude reference rather than WTI at Cushing and are not directly comparable with the NYMEX figure. All crack observations are dated and none is a current reading. Russian refining capacity, throughput, production and global seaborne export figures are from the IEA's assessment of 17 September 2026; installed capacity of approximately 6.5 mb/d against throughput of 3.8 mb/d does not represent a measurement of capacity destroyed. Global supply, demand, refinery run and middle distillate figures are from the IEA Oil Market Report of 14 October 2026; the 2026 annual average and second-half average refer to different periods. US distillate stocks refer to the week ended 2 October 2026. President Zelenskyy's statement on resuming strikes is an announced intention and no resulting capacity loss is established here.
Sources
This document is published by Bloodstone Research for informational and institutional research purposes only. It does not constitute investment advice, an investment recommendation, an offer or solicitation to buy or sell any financial instrument, commodity or security, or a forecast of future performance. Market conditions and data may change without notice. Readers should conduct their own analysis and, where appropriate, seek independent professional advice before making investment decisions. For institutional enquiries contact research@bloodstonecapital.co.uk.
