The Gulf Is Exporting Crude While Its Refineries Burn

Gulf crude is moving again; its refineries are not. Pipeline arithmetic, Kpler's own figures and Saudi Arabia's own plans show the crude recovery is narrower, more exposed and more temporary than the export totals suggest. The refining shortfall it leaves is shifting margins and fuel-security leverage to Asia.

The Middle East has found ways to move crude oil again. It has not found an equally fast way to rebuild the refineries that turn that crude into diesel, jet fuel and petrol.

The headline figures are now widely reported. Kpler estimates that regional crude and condensate exports nearly regained their pre-war volume in September, while more than 2.3m b/d of refining capacity remains offline. On 12 October, Reuters columnist Gavin Maguire argued that this is moving the oil market's centre of gravity east, towards Asian refiners.

OilWatch tested the crude side of that story, and it is weaker than the totals suggest. Saudi Arabia's main bypass, the East-West pipeline, runs about 1.2m b/d below its ceiling, through pumping stations that keep being hit. Part of its planned new capacity must be handed back to the domestic gas system in May. Kpler's own figures for the recovery do not reconcile. And new fires, including one at the Riyadh refinery, are landing on a regional refining system that has not recovered.

The barrels are moving, but through routes that are narrower, more exposed and more temporary than the export totals imply. Washington's new diesel deal with Moscow runs into the same problem: an announcement cannot restart a damaged refinery.

The pipeline is the recovery

Saudi Arabia's share of the recovery runs through one pipe. The East-West line reached its maximum capacity of 7.0m b/d in the first quarter. The energy minister said it carried 5.8m b/d on 6 October, after operations resumed following a September drone attack. That is about 1.2m b/d below its ceiling.

Throughput is not the same as exports. About 2m b/d of the line feeds west-coast refineries, leaving about 5m b/d for export. Vortexa put Yanbu loadings at roughly 3.7m b/d in early September, up from 3.2m in August.

On 12 October, Prince Abdulaziz bin Salman told the WPC Energy Congress in Riyadh that the kingdom could reach 10m b/d of pipeline capacity outside the Gulf within two to three years. The plan rests on upgraded pumping stations and a repurposed gas pipeline. That line could carry about 1.2m b/d of crude within seven to eight weeks, and up to 3m b/d with more investment.

None of this capacity is flowing yet. The conversion would add a separate route, not repair the existing one, and the announcement says nothing about whether the existing line sustains 5.8m b/d or whether its damaged stations are fixed. Whether it lifts sustained exports by 1.2m b/d depends on crude availability, East-West utilisation, terminal capacity at Yanbu and the state of the pumping stations.

The binding nodes are the pumping stations, and they are being hit. An April strike cut the line's capacity by about 700,000 b/d before it was restored days later. The September attack stopped it again. Expanding to 10m b/d through the same stations concentrates more export capacity in a target set that has already been struck.

The bypass moves the exposure rather than removing it. Yanbu loadings were suspended in mid-September after the pipeline was hit, and Red Sea cargoes still sail within range of the Houthis now striking Saudi energy sites.

Every element of the kingdom's stated resilience plan is about moving crude: pipes, pumps and dual-use compressors. None of it concerns refining.

How damaged refineries inflate crude exports

The two halves of the headline are linked, not merely simultaneous. East-West crude is split between Yanbu's export berths and the west-coast refineries. When a refinery goes down, the crude it would have run is freed for export.

Refinery damage therefore raises crude export volumes mechanically. Part of the crude recovery is the refining failure, counted as exports.

The size of the effect is not established. It depends on how much west-coast refining is actually offline: Jazan was confirmed down this summer, Rabigh is unclear. The test is arithmetic: Yanbu loadings should roughly equal pipeline throughput minus west-coast refinery runs, less any build in storage.

The same logic applies across the region wherever a damaged refinery's crude has another way out.

The May problem

The new export capacity is borrowed from next summer. The minister said the repurposed gas line could carry oil only until May. It would then be cleaned and returned to domestic gas supply for the summer.

Saudi Arabia's peak power demand comes in summer. If Hormuz is still constrained in June 2027, the kingdom faces a choice: about 1.2m b/d of western exports, or gas for air-conditioning load. Burning more crude for power is the likely compromise.

That window coincides with Kpler's earliest date for full Gulf refining recovery, the second quarter of 2027. The crude workaround expires just as the refining repair is meant to finish. Dual-use compressors make the trade-off recur every year the strait stays contested.

New fires, unconfirmed damage

Fresh strikes are landing on this system. Reuters photographed smoke rising from the Aramco refinery south of Riyadh on 3 October, and the Houthis claimed missile and drone strikes on Aramco sites in Riyadh and Khurais. Saudi-led coalition officials called the claims misleading.

Aramco's chief executive said on 5 October that "our system is intact", and reports of a fire at Ghawar remain unverified. What the fires have cost is not known: at US government request Planet moved Gulf imagery to case-by-case release, and Vantor imposed its own restrictions, so no public high-resolution imagery can identify damaged equipment. Lost output has been confirmed at none of the sites.

The crude number, and what it measures

The headline recovery is regional, not a recovery of the Strait of Hormuz, and Kpler's own published figures for it do not reconcile.

Kpler figure	Scope	Period	Value
Crude and condensate exports	Middle East excl. Iran	September, as a weekly rate (as reported by Reuters)	18.11m t/week vs 18.22m pre-war (≈19m b/d at 7.3 bbl/t)
Crude and condensate leaving the Gulf	Non-Iranian Gulf	1–28 September	At least 16.5m b/d, stated to match the pre-war average
Major producers' exports	Major Middle Eastern producers	September, stated 29 September	12.8m b/d vs 18.8m in February
Crude crossing Hormuz	Physical strait passage	Week to 6 October	At least 10.1m b/d
Exports from Gulf of Oman coast and Red Sea	Outside the strait	Week to 6 October	6.7m b/d, more than twice pre-war

The route split is internally consistent. Kpler says about 60% of September's 16.5m b/d crossed Hormuz, 23% loaded on the Gulf of Oman coast, mainly at Fujairah, and 17% left through the Red Sea. That partitions exports by route, so a cargo shuttled through the strait and transferred off Oman counts once.

The totals are another matter. For the same month, the tonnage series implies about 19m b/d against at least 16.5m b/d in the barrel series, and Kpler's pre-war baseline appears as 16.5m, about 18m and 18.8m b/d in different releases. Scope and condensate treatment may explain the differences, but Kpler has not published the bridge.

The largest gap is in timing. On 29 September a Kpler analyst put September exports from the major producers at 12.8m b/d. Within days Kpler published at least 16.5m b/d for the same month. The scopes differ, but 3.7m b/d is too large to leave unexplained.

That gap matters because Kpler's figures have themselves become contested. On 2 October Michelle Brouhard, Kpler's head of policy and geopolitical risk, discussed the debate over its Hormuz tanker data in an Oil Ground Up interview. Kpler has not published how its September series reconcile.

The US government figures at the centre of that debate were themselves disputed: in August, trackers put Hormuz flows 3–5m b/d below the Energy Secretary's estimate. Until Kpler publishes a bridge between its series, outlets repeating the same dataset add no corroboration.

Part of the Hormuz volume now runs through one transfer area. Saudi Arabia sold about 60m barrels from Ras Tanura for September and October loading by ship-to-ship transfer off Sohar, roughly 1m b/d carried through the strait on shuttle tankers first. On 11 October a tanker was struck by a drone near Sohar anchorage; UKMTO reclassified the incident as an attack, and India's shipping minister said two Indian crew remained missing. Attribution has not been established, and no suspension of transfers or fall in Saudi loadings has yet been reported.

The thesis does not rest on Kpler's precision. The refining shortfall is measured separately, in plant outages and product exports, and it is large on any of Kpler's crude numbers.

The refining number

The refining side has not recovered. More than 2.3m b/d of regional refining capacity remains offline, Reuters' Gavin Maguire reported. Kpler estimates the Middle East lost around 4m b/d of refined product supply between March and August, as damaged plants shut and undamaged ones cut runs for lack of export routes.

Kpler puts regional refinery runs at about 7.3m b/d, against 9.9m in February, and does not expect pre-war throughput before the second quarter of 2027.

Jazan shows how slowly a single plant returns. Aramco shut the 400,000 b/d refinery on 27 July after a Houthi attack damaged its IGCC complex and tank farm, according to an IIR note seen by Reuters. A further fire broke out there on 9 August, and the restart was pushed to 30 August. Its current operating rate has not been confirmed.

Why crude recovers first

Crude can be rerouted; a refinery cannot. Crude has pipelines to the Red Sea and the Gulf of Oman, protected shuttle tankers running the strait, dark voyages and offshore transfers. Each is a workaround for a blocked route.

A damaged refinery has no workaround. Utilities, compressors, process vessels, control systems and specialist equipment must all return together, and the long-lead parts come from a small number of suppliers. Until the whole chain is back, the plant does not run.

The result is the systemic transition this piece is about. The Gulf has rebuilt a way to export crude faster than it has rebuilt the machinery that turns crude into usable fuel.

Washington's diesel deal faces the same test

The diesel shortage has two engines: damaged Gulf refineries and damaged Russian ones. On 9 October Donald Trump said Russia would supply more than 300,000 tonnes of diesel immediately, 500,000 tonnes in November and 1m tonnes after that, with 3m more to follow. The same day the US Treasury issued General License 135, authorising the sale, delivery and import of Russian-origin diesel until 7 April 2027.

The question is the one this piece asks of Saudi pipelines: does an announcement add physical supply? A US licence does not override Russian export controls, but Moscow moved the next day. On 10 October the government partially lifted its diesel export ban, citing the Putin–Trump arrangement, to allow 500,000 tonnes onto the world market "at this stage", an initial permitted volume rather than a shipment on top of the announced tranches, with oil companies contracting foreign sales under government coordination.

The wider ban otherwise stood: for producers until 31 October, for non-producers until 31 January 2027. The decree also allows producers to export under intergovernmental agreements, a further route, but the 500,000-tonne authorisation is the one actually announced. Whether cargoes have loaded under it is not yet known.

Supply is the harder constraint. Deputy Prime Minister Alexander Novak said production covers both domestic demand and these exports, and set out a path of 300,000 tonnes in October, 500,000 in November, 1m in December and eventually 3m tonnes a month. Ten days earlier, the government had extended the export ban to keep the domestic market stable through the harvest.

Ukrainian drones have shut or crippled three of Russia's six largest diesel refineries, and Volgograd stopped processing on 2 October. Novak's 3m tonnes a month is an aspiration, not a schedule; it would exceed Russia's diesel-only exports a year earlier, around 2.5m tonnes a month, though not the 3.3–3.4m including lower-grade gasoil. Trump's own announcement tied the larger volumes to the condition of Russia's refineries.

The scale is small. The first three tranches total 1.8m tonnes, about 13m barrels. That is less than one month of Russia's diesel exports a year earlier, around 2.5m tonnes, and under half of China's approved October fuel exports.

The ceasefire that would protect Russian refineries is unconfirmed. Trump announced an energy truce on 14 September; Moscow did not confirm it, Kyiv said it was still awaiting Russia's list of facilities, and strikes on refineries continued, including the one that halted Volgograd.

The deal may move prices before it moves fuel. Until cargoes are seen loading, it belongs in the same category as Saudi Arabia's 10m b/d: a promise made against damaged plant.

Where the margin goes

The consequence lands in Asia, not Saudi Arabia. When the Gulf exports crude instead of diesel and jet fuel, the refining margin moves to whoever processes that crude. The natural beneficiaries are refiners in India, China, South Korea and Singapore with spare capacity and access to Gulf barrels.

The shift is already visible in buying. Chinese refiners bought at least 12m barrels of Saudi crude for November on a delivered basis, with further sales to India and Taiwan, Reuters reported. Some cargoes carried premiums of about $20 a barrel over official selling prices. These are reported sales, not deliveries or refinery runs.

The shift brings leverage with it. Importers in Europe, East Africa and South Asia that bought Gulf diesel and jet fuel now depend more on Asian refiners, and on their governments' export policies. Product voyages also lengthen, tightening clean-tanker demand.

Capture is not automatic. Earlier in the war, Asian refiners cut runs for lack of crude, and China has restricted fuel exports since March. After a holiday pause, Beijing approved about 3.7m tonnes of gasoline, diesel and jet exports for October, but now sets them month by month. The margin goes east only if the crude arrives and Beijing keeps letting the products leave.

What would change this assessment

Yanbu loadings well below 4m b/d while the pipeline reports 5.8m would mean the crude recovery is overstated, or crude is going into storage.
Gulf refinery runs back near their pre-war 9.9m b/d would mean refining is recovering faster than Kpler's second-quarter 2027 estimate.
Imagery or Aramco statements confirming lost output at Khurais or the Riyadh refinery would show the new fires are cutting supply, not just producing smoke.
A per-vessel audit of the 6.7m b/d leaving outside Hormuz would test Kpler's regional total. Priorities: Yanbu loadings, single counting of ship-to-ship cargoes, and loaded volume rather than tanker capacity.
Operators suspending ship-to-ship transfers off Sohar, or war-risk premiums there jumping, would cut the crude half of the recovery directly and push more Saudi volume back onto a damaged pipeline.
Kpler publishing a bridge between its 12.8m and 16.5m b/d September figures, and between its tonnage and barrel series, would show how much of the crude recovery is measurement.
An agreement to reopen Hormuz would ease the shuttle, Sohar and May constraints but not the refining damage: the crude half of the argument would weaken, the refining half would not.
Russian diesel cargoes seen loading under Moscow's 10 October authorisation would turn the Trump–Putin deal from a price signal into supply.