Beyond Hormuz Lies an Oil Surplus

On 12 August 2026 OPEC and the International Energy Agency (IEA) published their monthly views of the oil market. OPEC's: demand grows 0.6 million barrels per day (mbpd) this year. The IEA's: demand falls 1.6 mbpd this year, the first annual decline since the pandemic. Two institutions watched the same war for six months and disagree by over 2 mbpd on oil demand with 4 months left in the year.

Hormuz, crisis or mild inconvenience?

Iran closed the Strait of Hormuz at the end of February. Oil moving through it fell from 21.6 mbpd in Q4 2025 to 4.9 mbpd in Q2 2026; liquefied natural gas (LNG) fell from 10.5 billion cubic feet (bcf) a day to less than 1 bcf (which is Qatari gas going offline). Brent crude went from $71 in late February to $112 by the third week of March and touched $126 at the end of April.

Then it went back. On 2 July, with the strait still contested, Brent settled at $72 – a round trip to the pre-war price with the Strait still very much constrained. It has spent August bouncing around $85-95. The world lost its most important oil transport lane, an estimated 6 mbpd of supply¹, and the market treated it as an inconvenience.

First part of the answer is that demand fell almost as sharply as supply: the second quarter ran 4.9 mbpd below the year before, price destroying consumption at a pace no forecaster carried. Second part of it is that the buffers paid for the rest: 410 million barrels drawn from observed global reserves between late February and the end of July, and an American Strategic Petroleum Reserve that fell below 300 million barrels on 10 August, its lowest level since 1983, after starting the year above 415 million barrels.⁵

The third and longest-lasting part of the answer is that the global oil market entered the war in an oil surplus. In 2025 the world produced 106.2 mbpd and burned 104.8: a surplus of roughly 1.4 mbpd for a year. China took almost exactly that much off the market, stockpiling an average of 1.1 mbpd and hitting a record 2.7 mbpd in December, seven weeks before the Iran closed Hormuz. It entered 2026 holding over 1.2 billion barrels onshore, near 100 days of import cover, and has spent the war drawing on it while holding refined product at home under wartime export controls.

That 1.2 billion barrels is a floor rather than an estimate. The figure is assembled from satellite readings of floating tank roofs, so it counts only what can be photographed from above. The American reserve sits in salt caverns that no satellite can see and is counted in full; China's underground storage is not. To believe 1.2 billion barrels is the true number you have to believe Beijing built only the tanks with lids.

The market mistook China's import total as a demand signal, as opposed it what was essentially partially a treasury operation: accumulate cheap, release dear, disclose nothing. The glut it was absorbing into storage has been paying for the war.

Demand is peaking, and the war brought it forward

Ignore the wartime impact for a moment and we see that global oil demand was barely moving. Consumption sat near 105 mbpd in 2025, falling to 103 mbpd in the 2026 wartimes, projecting to rebound back to 105 mbpd, where it stays into 2030.² The bulge through 2027 and 2028, 106 to 107 mbpd, is not people burning more oil; it's governments and traders buying barrels to refill their strategic reserves, and when the refilling stops the line drifts back to 105 mbpd.

The IEA outlines why the demand is flattening: India's fast electrification of 3-wheelers, heavy truck electrification, new energy vehicles predominantly coming out of China, solar with storage displacing diesel generators. The war accelerated these forces.³ In July, new energy vehicles (NEVs) took 60.4% of Chinese domestic vehicle sales, the first month above 60%. NEV exports ran 553,000 units, up 145.5% on the year, with January to July of 2026 up 120%. Chinese solar module exports hit a record 68 gigawatts in March, double February's, with shipments to Africa up 176% in the month. A diesel generator retired in Lagos at $112 Brent does not come back at $70.

In addition to transport electrification, coal-to-olefins, the Chinese practice of making plastics from coal instead of oil is also dampering demand.⁴ Integrated capacity stood at 13.4 million tonnes (Mt) at the end of 2024 and ran at 96%, displacing roughly 0.7 mbpd of naphtha. Another 7.3 Mt are under construction for start-up between 2026 and 2029. Petrochemicals are the last growth sector in every published oil forecast, and this is the technology aimed at exactly that barrel.

The demand suppression forces are not a war story. Electric heavy trucks went from 14% of Chinese sales in 2024 to 29% in 2025, and the coal-to-olefins plants were commissioned years ago. What the war did was move the fence sitters to the electrification side, and not only the buyers: a disruption proven possible rewrites energy policy for good, and energy and finance ministries from Tokyo to Casablanca are redrawing their plans in real time, whatever crude costs in 2028. Once these purchases are made, that oil demand is removed for as long as the panels and cars are still operational. The IEA sees demand for combustible fossil fuels peaking as early as 2027, even without the Hormuz clarion call.

There's more oil to come

While demand flattens, the production capacity rises on capital and construction committed years ago. The IEA's last published medium-term outlook puts world production capacity at 114.7 mbpd by 2030, up 5.1 mbpd from 2024.²

The notable additions are: Guyana goes from 0.75 mbpd in 2025 to 1.3 mbpd by2030; Brazil reached 4.0 mbpd this year; Argentina is at 0.8 mbpd and climbing. None of them are in OPEC, all of them were financed before the war. Their cash costs one extraction runs between $4 and $7 a barrel, so they'll be pumping. Inside what remains of the OPEC, the United Arab Emirates produced 3.4 mbpd in 2025 against an estimated capacity of 4.2 mbpd⁸. In April, they left OPEC altogether, then announced a second pipeline to the Gulf of Oman and signaled they'll be pumping more oil.

Someone has to produce less

Capacity of 114.7 mbpd against demand of 104.8 mbpd leaves 9.9 mbpd with nowhere to go. Most of that is ordinary: a market in balance carries 2-4 mbpd of spare capacity, and Saudi Arabia alone accounts for half of it, pumping 9.3 mbpd against a capacity of 11.6 mbpd. They'll simply continue to run below capacity in the event of a glut. The question is who absorbs the rest: the 5 mbpd or so of capacity beyond what a healthy market keeps idle.

It will not be the low-cost producers, because no plausible price stops a barrel that with a $4 production cost.⁶ Saudi Aramco produces at $4 a barrel, Petrobras pumps pre-salt crude for $4, Hess's Guyana barrels cost $7, and Cenovus runs its oil sands at $9 a barrel of operating cost. Their capital is sunk cost, their platforms last 20 years, and their decades more crude under them.

American shale has the opposite operating dynamics. Shale producers make money at $43 a barrel on wells they have already drilled. But more than 70% of a shale well's production is gone within the first year, so the industry makes incremental re-drilling decisions constantly – and the latest estimates put new-well breakevens in the Permian, Bakken and other basins at $60–70 per barrel.⁶

Put those two facts together and the adjustment mechanism is visible. With crude at or below $60, new American wells are not worth drilling and existing one keeps pumping till depletion. Once they stop drilling, the supply mechanistically declines 2-3 mbpd over two or three years.

OPEC could keep shale alive by keeping prices high, if it cut its own production substantially instead. It has done exactly that before: 2.2 mbpd withheld from 2023, unwound through 2025. Whether it does so again is a question its own members have started answering with their feet. Qatar and the UAE have already exited the bloc, and the members who remain need crude between roughly $80 and $92 to balance their budgets.⁶ The UAE left to gain market share rather than enforce price discipline. An OPEC price rescue seems unlikely this time around.

One giant ad campaign for electrification

Taking 6 mbpd offline for six months, forcing petroleum reserve drawdowns, and keeping crude prices elevated has shown the importance of oil to the global economy. It has also been the greatest marketing campaign for the products China sells. The momentum behind electrification has increased again, most of all in the two products that displace diesel directly.⁹

If any policymaker was still hesitant about sprinting toward electrification, the Hormuz closure has removed the doubt, and China is there to supply the alternative.

What's up with diesel spreads

Diesel prices are the combination of crude prices, which have certainly risen, and the refining margin, called the crack spread. The spread averaged about $26 per barrel through the first nine months of 2025, stepped up to $42 in November when EU sanctions hit Russian oil companies and Kuwait's Al-Zour refinery went down, and doubled when the strait closed. Then between May and July, as Brent fell $23 per barrel, diesel went the other way: the crack rose $21, reaching $81 in July and still rising into August.⁷

Two crises share the chart, and they end on different clocks. The Hormuz crisis is a logistics problem: Gulf refineries are intact but their product tankers are trapped, Japan and Korea cut runs because crude cargoes stopped arriving on schedule, and China implemented a refined-products export ban in March. These factors reverse eventually: cargoes reload, export quotas reopen (as Beijing has already started doing), and the region's refining returns in weeks.

The real crisis is in Russia, where drone strikes have removed a third of refining capacity, taking runs down to 3.6 mbpd in July, the lowest since 2002. Moscow has banned diesel exports from what remains. That capacity rebuilds under sanctions, on another war's timeline.

For the crude thesis the chart is confirmation rather than complication: if oil itself were scarce, the record would be in the oil price. Instead it is in the refining margin. One consequence follows for the wider economy: a falling oil price no longer buys the disinflation it used to, because the fuel that sets producer costs – freight, farming, mining – has decoupled from Brent crude. Diesel's buyers are responding in kind: fleet operators facing once-in-a-generation prices for the second time in five years are not deferring purchases but crossing over, and this time the replacement is in stock, because Chinese electric trucks have reached export maturity. In the same way Hormuz is an advertising campaign for electrification, the Russian refining crisis is an advertising campaign for electric heavy trucks specifically, alongside more solar with storage. The root causes differ, but the same country is selling the solution.

Sources for every load-bearing figure, and the derivations behind every number this publication constructed itself.

1. An estimated 6 mbpd was shut in. Five published measurements describe three different things.

   • Through-strait flows fell from 21.6 mbpd in Q4 2025 to 4.9 mbpd in Q2 2026 on the EIA's tracking, a series counting crude, condensate and products together.
   • In the week to 16 August the US Energy Secretary put 9 mbpd through the strait, against roughly 20 mbpd before the war.
   • Gulf output shut in during July: 8.3 mbpd on the IEA's estimate, 5.5 mbpd on the EIA's, with 6.57 mbpd forecast for the third quarter.
   • World supply in July was 101.5 mbpd, 6.3 mbpd below the previous July. This is the outcome measure and already nets bypass pipelines and non-Gulf growth against the loss.
   • Covert shuttle volumes are estimated near 4 mbpd and reported to be running higher; roughly 150 ships now float off the Omani coast against about 40 in January. A barrel moving without a transponder is recorded as shut in when it is merely unobserved, so any tracked shut-in figure is a ceiling.
   • Gross transit loss, 12–17 mbpd depending on the window, is not supply loss and enters no balance here.

   6 mbpd is the Plain Sight synthesis of the data above: between the two agency estimates, nearer the lower one, because the measurement error runs one way.↩

2. The IEA's demand path plus two small adjustments. Two IEA series exist and they are not interchangeable. The monthly Oil Market Report runs to 2027 and carries the war. The medium-term Oil 2025, published in June 2025, runs to 2030 and predates it.

   Series	Published	2025	2026	2027	2030
   Oil Market Report	Aug 2026	104.8	103.3	105.7	—
   Oil 2025 (pre-war)	Jun 2025	103.7	104.4	104.8	105.5
   Plain Sight	Aug 2026	—	—	104.6	104.8

   • All figures are mbpd. The two IEA series sit about 1.1 mbpd apart on the same year, because the agency has revised the completed year upward repeatedly, including by 0.3 mbpd between its July and August 2026 reports. Levels from the two documents may not be placed on one axis, and none here are.
   • Oil 2025 publishes endpoints and a shape rather than an annual table in its free summary: demand rising 2.5 mbpd from 2024 to a plateau of about 105.5 mbpd, growth of roughly 0.7 mbpd in 2025 and 2026 slowing to a trickle, and a small decline in 2030. The year-by-year path above is interpolated from those.
   • Plain Sight sits 0.7 mbpd below the Oil 2025 path at 2030: 0.2 mbpd for coal-to-olefins and 0.5 mbpd for accelerated electrification.
   • The 114.7 mbpd capacity path is Oil 2025's, and no post-war edition has revised it: the medium-term outlook that would have has not appeared.
   • Exhibit 1 is stated in whole numbers because at this horizon the revisions are larger than the decimals.
   ↩

3. Deriving the electrification adjustment. The adjustment is 0.5 mbpd by 2030. Part of it is purchases pulled forward; part is trajectory – policies redrawn after a proven disruption do not revert. Treating most of it as timing keeps 0.5 mbpd the conservative reading.

   • The IEA's Global EV Outlook has the world electric fleet displacing 1.7 mbpd of oil in 2025, about 1.0 mbpd of it in China; Ember puts the same figure at 1.8 mbpd.
   • Those are stocks, not flows – the cumulative effect of every electric vehicle ever sold. The annual increment is around 0.4 mbpd, with Jefferies putting China's fleet displacement at 1.4 mbpd in the first half of 2026 against roughly 1.0 mbpd a year earlier.
   • New energy vehicles reached 60.4% of Chinese domestic sales in July 2026; exports ran 553,000 units, up 145.5% on the previous July, and 2.91 million units over the first seven months, up 120%.
   • Chinese solar module exports hit a record 68 GW in March, twice February's volume, with Africa up 176% in the month.
   • Basis caution: the monthly export series counts wholesale shipments including plug-in hybrids, while the customs series counts battery-electric vehicles and ran 68.7% higher in the first half. They are different measurements and are not juxtaposed here.
   • Basis caution: electric heavy trucks went from 14% of Chinese sales in 2024 to 29% in 2025, both pre-war. December 2025's reading above half was a subsidy-deadline pull-forward that fell back to roughly 30% by April 2026, while first-half 2026 volumes ran 77.7% above the year before, which is the honest number.
   • Vehicles and panels bought early cap the pull-forward component; the policy response is the open-ended one. The 0.5 mbpd accrues over the decade and counts the open-ended part as zero.
   ↩

4. Coal-to-olefins estimate. Chinese plants make olefins from coal rather than from oil-derived naphtha. The displacement per unit of capacity:

   1 Mt/yr olefins × 2.15 t naphtha/t olefins × 8.75 bbl/t ÷ 365 days = 0.0515 mbpd of naphtha displaced per Mt/yr of capacity
   • Conversion factors: 2.15 tonnes of naphtha per tonne of olefins, credible range 2.08 to 2.33 from Honeywell UOP, ICIS and the Oxford Institute for Energy Studies; 8.75 barrels per tonne of naphtha, where Platts uses 8.9.
   • Integrated capacity was 13.42 Mt/yr at the end of 2024, producing 12.89 Mt, a 96% utilization rate that is the tell on the economics.
   • Standing displacement is therefore 13.42 × 0.0515 = 0.69 mbpd. Propagating the factor ranges gives 0.62 to 0.74 mbpd.
   • An additional 7.3 Mt/yr is under construction for start-up between 2026 and 2029, displacing a further 7.3 × 0.0515 = 0.38 mbpd, with perhaps another 20 Mt/yr at feasibility and permitting.
   • Commissioning schedules are public, so a competent forecaster already carries most of the first tranche and none of the second. The adjustment taken here is 0.2 mbpd by 2030.
   ↩

5. The Strategic Petroleum Reserves will need to be refilled. Refilling is demand with an expiry date, and it is what lifts 2027 and 2028 in Exhibit 1.

   • Observed global reserves fell 410 million barrels between late February and the end of July, a 2.7 mbpd draw – and the war is not over. Another 100–200 million barrels of draws before it ends is well within range, so this profile is a snapshot, not a settled total.
   • The American reserve ran 415.1 million barrels at the end of the first quarter and fell below 300 million on 10 August, its lowest since 1983. Regaining the first-quarter level alone takes 116 million barrels, and a post-war rebuild is unlikely to stop there.
   • China accumulated at 1.13 mbpd through 2025, hit a record 2.67 mbpd in December, and will rebuild at low prices. The closure has also given every importer a reason to hold more than it did before: the refill is drawn barrels restored plus reserves that did not previously exist, China's expansion first among them.
   • Exhibit 1 therefore carries a band rather than a point: 1–2 mbpd of refill across 2027 and 2028, tapering through 2029 – roughly 800 million barrels of restored draws plus new strategic demand on top. A bid of that size is a floor under crude through 2028 and into 2029; strategic buyers may hold the $60 floor and step away when the prices go much higher than that.
   ↩

6. Estimations of breakeven crude prices. Three different categories carry the same word, and no ratio is taken between them.

   • Prices a driller needs. The Dallas Fed Energy Survey collected responses from 135 energy firms, 92 of them exploration and production companies, between 11 and 19 March 2026: $66 a barrel to profitably drill a new well, $67 in the Permian, $62 to $70 across the basins; $59 for large producers and $68 for small ones. Covering operating costs on an existing well takes $43, ranging $34 to $47. Both are West Texas Intermediate prices covering royalties, severance taxes, overheads and interest.
   • Field cash costs. Aramco $3.51 a barrel of oil equivalent, Petrobras pre-salt $4.19, Hess Guyana $6.73 on 2024 accounts, Cenovus oil sands $9. These exclude government take and capital recovery.
   • Company breakevens. Canadian Natural puts its corporate breakeven including dividends just above $40 a barrel, the comparable figure for a business rather than a barrel.
   • Government fiscal breakevens. From the IMF Regional Economic Outlook series as republished by the St Louis Fed, on 2025 projections: Qatar $43.2, UAE $49.95, Kuwait $81.84, Saudi Arabia $90.94, Iraq $92.43. Later vintages of the same series differ by four to five dollars a line, which is why the vintage is named.
   ↩

7. The diesel crack, and what is broken. The series in Exhibit 4 is the New York Harbor ultra-low-sulfur diesel spot price times 42, minus the Europe Brent spot price, monthly averages, both from EIA data. The widely quoted August records above $100/bbl are front-month futures cracks on different bases and are not this series.

   • Pre-war step: cracks rose from a ~$26/bbl 2025 average to $42 in November 2025 on EU sanctions against Rosneft, Lukoil and Gazprom Neft, Ukrainian strikes on Russian refining, and Kuwait's Al-Zour outage.
   • Russia: refinery runs fell to 3.6 mbpd in July 2026, the lowest since May 2002 and roughly a third below the 5.3–5.6 mbpd seasonal norm, after 18 refineries were struck in July alone; Moscow banned diesel exports. At a ~40% distillate yield the lost runs are roughly 0.7 mbpd of diesel – the conversion is an estimate.
   • Hormuz side: Gulf refineries are intact but product exports are trapped behind the strait; Japan drew strategic reserves from 16 March and secured crude month-to-month; China banned product exports on 5 March, then released quota tranches – July diesel exports of 810,000 tons (~0.2 mbpd) were up 88% on June but total product exports remained about half the pre-war rate.
   • The balancing system: US refiners ran at full rates and exported a record 1.9 mbpd of diesel, drawing US diesel inventories to their lowest since 1996, 12% below the five-year average.
   ↩

8. The UAE capacity estimate. The 4.2 mbpd is the EIA's estimate of effective production capacity in 2025. ADNOC's own stated figure is 4.85–5 mbpd, a 2027 target; the two bases differ by 0.65 mbpd and both sit well above the 3.4 mbpd produced.↩

9. Electrification indicators in Exhibit 3. All three bars are export volumes for the first half of 2026 against the first half of 2025, each on its publisher's basis.

   • Electric vehicles +68.7%: China customs, battery-electric vehicles. The industry association's monthly count, which includes plug-in hybrids, ran higher still – exports of 553,000 units in July, up 145.5% on the year, and 2.91 million over January to July, up 120%.
   • Storage batteries +28.5%: 58.6 GWh exported in the first half, on the China Automotive Battery Innovation Alliance's count. Total battery exports were 181.3 GWh, up 42.5%, and domestic storage battery sales of 318.1 GWh were up 83.4%.
   • Solar modules −2.5% by volume, against exports of wafers, cells and modules worth $17.18 billion, up 24.3% by value, with cell volumes up 36.8%. Module shipments spiked to 37.32 GW in March ahead of the removal of VAT export rebates for PV products on 1 April 2026, then fell for three consecutive months, July running 21.4% below the previous July. Chinese domestic installations fell 66% in the first half on a separate pricing reform. The solar contraction of 2026 is a Chinese policy event, not a demand signal, which is why the body claims nothing from it.
   ↩

Disclosure: No positions in oil futures, oil equities, or energy ETFs. Positions are disclosed live at disclosure.plainsightresearch.com.