OilWatch Network Analysis — European diesel inventories held up through September and the European Commission says there is no supply problem. Both are true. But inventory measures the barrels already inside the system — not how easily the next barrel replaces them. On that second measure, the evidence has deteriorated sharply.
On 8 September the Commission's Oil Coordination Group stated that "there is no supply problem in the EU at the present time." Demand for diesel and jet fuel, it said, continues to be met through higher EU refinery production and alternative supplies from global markets, while warning that autumn and winter demand could further tighten the markets. That assessment was correct when made and remains the baseline any claim of shortage has to clear.
Independent gasoil stocks in the Amsterdam–Rotterdam–Antwerp hub support it. Having sat near a four-year low of 12.08 million barrels on 4 September, they held broadly unchanged through the following fortnight. Europe is not emptying out, and anyone writing that it is has the facts against them.
The question this article asks is different. Not how much diesel Europe holds, but what condition the system that replaces it is in.
Crude is recovering. Products are not.
The IEA's September Oil Market Report contains the asymmetry that explains almost everything readers find confusing about current prices. Gulf crude losses have "narrowed to just below 45 per cent" as flows bypass the Strait of Hormuz. Gulf refined product and LPG exports remain "nearly 60 per cent, or 3.7 million b/d, less than in February."
Crude is coming back. Products are not. A crude recovery is not a middle-distillate recovery, and the two have been routinely conflated in coverage of this crisis.
This is why Brent front-month settling at $105.68 on 14 September and $103.87 on the 18th is entirely compatible with Mediterranean ULSD setting an all-time record of $1,664.75 a tonne on the 15th, and Northwest European 10ppm reaching $1,642.25 — both the highest since Platts began the assessments. The physical product market and the crude benchmark have detached, and the product market is the one that matters for European hauliers, farmers and heating-oil buyers.
The concentration of the loss
The IEA's 17 September commentary on Russian refining gives the two numbers that should frame this story, and they are best read together.
Global seaborne gasoil and diesel exports averaged 4.7 million b/d over the first eight months of 2026, down 10 per cent year-on-year, with declines accelerating through the second and third quarters. Over the same period, combined Middle East and Russian diesel exports fell to 520,000 b/d in August — 75 per cent below the same month a year earlier. Middle East exports alone were 390,000 b/d, just over a quarter of pre-war levels.
Global trade down ten per cent; two suppliers down seventy-five. The loss is not distributed across the market. It is concentrated in the Gulf and Russia, and everyone else — principally US, Indian and European refiners — has absorbed it by running harder.
A note on a figure that has caused confusion. The IEA's August report put the year-on-year decline from Russia, the Middle East and Asia at 1.3 million b/d, "about 20 per cent of global seaborne trade"; the September report put Gulf and Russian net exports at 1.6 million b/d below February, when they were "almost 45 per cent of global seaborne trade." These do not conflict and should not be differenced. The first is gross and includes Asia; the second is net and covers two regions. Against the IEA's stated net figure of 4.7 million b/d, Gulf plus Russia at roughly 2.1 million b/d is 45 per cent exactly as described.
What the inventory number cannot tell you
ENGINE's ARA bunker availability outlook has advised buyers to allow five to seven days to secure good coverage — on 12 August, again on 26 August, and again on 16 September.
That reading did not change while independent gasoil stocks fell nine per cent in the first week of August, and it did not change while they held broadly unchanged at around 12.1 million barrels through September. Prompt availability and inventory level have moved independently for five weeks.
Two limits on how far that can be pushed. These are bunker-market figures covering marine fuel grades in one hub; they are not a measurement of wholesale road-diesel cargo availability across Europe. And they predate the Petroline strike by a month, so nothing in them can be attributed to the pipeline. ARA's bunker market provides one visible sign of persistent prompt tightness — not proof of a Europe-wide logistics failure.
But it does establish the principle the headline rests on. Stock level describes what is in storage. Lead time describes how readily the market delivers what a buyer wants now. They are different measurements and they have been telling different stories since mid-August.
The replacement chain has lost redundancy in four places at once
Export capacity. Kuwait's Mina Abdullah, 454,000 b/d, has been shut since March. Al-Zour, 615,000 b/d and purpose-built as a distillate exporter, has run reduced at an unpublished rate. Bahrain's Sitra took severe damage specifically to its diesel hydrotreater and is not expected back before November. Saudi Arabia's Jazan, 400,000 b/d, has been down since 27 July with a restart date that has slipped twice. On 15 September Reuters reported that three of Russia's six largest diesel refineries — which together produce about half the country's diesel — had cut sharply or halted: Kinef shut, Volgograd and NORSI at around a quarter of capacity. Russia was reported on 16 September to have decided to extend its producer diesel export ban through 31 October.
Refining headroom. The IEA describes the global refining system as "stretched to the limit," with August throughput at 81.4 million b/d, 4.2 million below a year earlier. US refiners have been running at 97 to 98 per cent. European runs are forecast at 12.2 million b/d for 2026 with the fourth quarter at 12.45 million — the strongest Q4 since 2019. That is not resilience. That is the buffer working at maximum, before the autumn maintenance that the Commission asked refiners to defer on 31 March.
Route access. Petroline, Saudi Arabia's principal Hormuz bypass at roughly 5.4 million b/d, was shut on 10 September after a drone strike traced to Maysan in Iraq. Saudi Arabia said on the 16th it expects to restore half of capacity within days; Kpler assesses full restoration at up to six weeks, and Bloomberg reported that Aramco itself expects the same. Bloomberg reported on the 18th, citing people familiar with the decision, that Aramco's October suspension applied to all European term buyers. Reuters said it could not independently verify that broader scope, although at least two European customers had been told they would receive no October crude. Separately, Argus reported on the 15th that at least three European refiners had lost late-September cargoes, some postponed into November. Orlen — which takes a substantial share of its crude from Aramco across its Polish, Lithuanian and Czech refineries — bought sixteen extra cargoes to cover through November.
Shipping and insurance. In July, Marsh put additional war-risk premiums for the Strait of Hormuz at 7.5 to 10 per cent of hull value, against one to three per cent shortly before. On a tanker insured at $100 million, that implies $7.5 to 10 million of additional premium for the voyage. Bab el-Mandeb, on the same July assessment, sat at 0.5 per cent — the toll is on the Gulf leg specifically. Marsh's more consequential observation was about willingness rather than price: underwriters were increasingly reluctant to write spot cover at all. A market that stops quoting does not send a price signal. We have found no September update to these figures.
And the fleet itself has shifted. Vortexa records approximately 52 LR2 tankers moving from clean products into dirty crude trading since March, with LR2 clean employment down about 35 per cent against the 2025 average. Smaller MR2s, working the Atlantic basin, held steady. The LR2 is approximately the vessel required for long-haul clean-product arbitrage — the Gulf, India, Korea to Europe.
Europe's substitution strategy therefore requires more long-haul clean-product tonne-miles at precisely the moment part of the fleet capable of supplying them has moved into crude. That does not prove tanker availability is responsible for any particular prompt delay in ARA. It removes another layer of flexibility from an already constrained replacement system.
A note on the price figures
The "$100 diesel crack" figures in circulation are not necessarily the same measurement. Platts assessed the physical ARA diesel crack at $85.16/bbl on 16 September, while the IEA said on the 17th that diesel crack spreads had exceeded $100/bbl during September in both the US Gulf Coast and Northwest Europe. The gap is most likely benchmark, crude reference and assessment period — Platts prices a CIF Northwest Europe cargo against Dated Brent on a named day; the IEA characterises the month. Neither is wrong, and they should not be treated as interchangeable without specifying which is which. Separately, the record reported on 18 September — around $1,565 a tonne on ICE gasoil futures, the equivalent of roughly $210 a barrel — is a flat price, not a crack at all.
What would show this analysis is wrong
ECB experts told Euronews on 19 September, using LSEG futures as at 16 September, that diesel refining margins are expected to peak in October — petrol margins, they note, peaked in August. Refining margin currently accounts for €0.41 a litre, about 19 per cent of the euro-area retail diesel price. The ECB's own September projections make the broader point that margins historically fall as supply and demand adjust, while warning this dislocation may take longer than usual to correct.
That is an institutional forecast published this week and it is the strongest argument against everything above. It deserves a fair test rather than a dismissal.
But a margin peak does not by itself tell us why the market cleared. Margins can fall because supply recovers; they can also fall because demand weakens. The accompanying physical indicators therefore matter.
If margins peak in October alongside stable or rising European diesel consumption, recovering Gulf exports and LR2s returning to clean trade, the replacement system is healing and this analysis is wrong. If margins peak alongside falling consumption while export capacity, shipping availability and refinery headroom remain constrained, that would be consistent with the market clearing through demand destruction rather than supply recovery.
Five indicators, with current readings so a move is measurable rather than impressionistic: ARA gasoil stocks and lead times, currently around 12.1 million barrels and a constant five to seven days; the physical ARA diesel crack, requiring sustained $100-plus with concurrent stock draws to signal deterioration; any significant European refinery outage, given there is no spare replacement capacity; an actual US export restriction, as opposed to the political discussion now under way, which Interior Secretary Doug Burgum has publicly opposed on the grounds it would not lower prices; and whether combined Russian and Gulf diesel exports hold near August's 520,000 b/d into November.
Governments, meanwhile, have stopped debating and started spending. Germany agreed on 18 September to cut energy tax on petrol and diesel by 14 cents a litre — around 17 cents including VAT — from 1 October through the end of 2026. The package is worth roughly €2.5 billion, split evenly between the federal government and the Länder. Berlin also plans talks with the oil industry with the aim of temporarily introducing a fuel-price cap based on the Luxembourg or Belgian model by 1 January 2027. Elsewhere in Europe, several temporary fuel-relief measures are due to expire or be reassessed at the end of September, adding another moving variable to the October price outlook.
The formulation
In July our position was that the diesel is still arriving and the safety margin isn't. The September evidence supports something more specific, and more serious.
Europe is not running out of diesel. But the system that replaces every barrel Europe consumes has simultaneously lost Gulf and Russian supply, refining headroom, shipping flexibility and route resilience. Current inventories conceal that deterioration, because they measure the buffer accumulated yesterday rather than the reliability of tomorrow's replacement barrel.
Its ability to replace the diesel it consumes has degraded faster than its stock level reveals.