Production buffers across the OPEC+ alliance have eroded steadily through the first quarter of 2026. As of the IEA's May Oil Market Report, 3.8 mb/d of spare capacity remains in the system — the tightest reading since February 2022 and roughly 0.3 mb/d below the five-year average.[1]
The concentration matters as much as the total. Saudi Arabia alone accounts for 2.4 mb/d of the cushion, with the United Arab Emirates contributing a further 0.9 mb/d. The remaining 0.5 mb/d is distributed across five other producers, each carrying less than one hundred and fifty thousand barrels of headroom.[2]
A coordinated supply incident in the Persian Gulf could absorb half the global cushion within seven days.
That arithmetic is what gives the HORMUZ-2026 scenario, wargamed by our analysts and reviewed for Resiplan corporate clients last month, its uncomfortable weight. The scenario assumes a non-state actor disrupts tanker transit for five working days. In our base case, two-thirds of available spare capacity is mobilised by day three; the cushion does not recover to its pre-incident level for fourteen weeks.
A second pathway, which we treat as plausible but lower probability, involves cascading refinery turnarounds in Asia coinciding with seasonal demand. Here the binding constraint is not crude availability but downstream product — and the spare-capacity metric loses some of its explanatory power.[3]
A tighter definition, and what it changes.
Public discussions of "spare capacity" routinely conflate two ideas. The first is the production that producers could bring online if they were willing — the so-called nameplate buffer. The second, and the one analysts actually use, is the production that can be brought online within thirty days and sustained for at least ninety. The gap between the two is rarely material in calm markets, but it grows under stress, when maintenance backlogs and field-decline rates start to bite.
Our spare-capacity series uses the second definition — the one consistent with the IEA, OPEC Secretariat, and Argus methodologies — and computes a country-level breakdown that we publish monthly with the OMR release. The methodology document is here; the underlying time series is downloadable as CSV and queryable through the API.[4]
A buffer that is not what it was
The shaded band is the post-2018 five-year rolling mean ±0.4 mb/d. The series has crossed into the lower decile and stayed there for forty-seven days.
Note — Spare capacity defined as production that can be brought online within thirty days and sustained for ninety. Shaded band: post-2018 five-year rolling mean ±0.4 mb/d. Last reading: 3.8 mb/d, −0.3 mb/d versus mean.
SOURCE — IEA Oil Market Report (monthly), OPEC Secretariat, Argus Media. DEMO DATA, not for citation.
What the wargame revealed.
The HORMUZ-2026 wargame, conducted at the end of April with a panel that included two former insurance underwriters and a former operations director at a Gulf-state producer, produced a result that surprised half the room and confirmed the priors of the other half. The binding constraint was not crude. It was the speed at which the spare capacity could be physically mobilised, the lag in tanker re-routing, and — most pointedly — the price-discovery problem in war-risk insurance premiums.
In the base case, war-risk premiums on hulls transiting the Gulf doubled within thirty-six hours of the disruption. Charter rates followed. The spot crude price moved less than implied by the supply shock alone, because traders priced in the eventual restoration; what moved was the delivered price, including freight and insurance.[5]
The implication for risk managers is uncomfortable. The most-watched headline — the spot price of Brent — is the least informative number in the first seventy-two hours of a Gulf incident. The ones to watch are the war-risk index, the charter rate spread between affected and unaffected lanes, and the change in spare-capacity utilisation reported by the affected producers.
What it means for monitoring.
We have added war-risk premium and lane-spread charter rates to the country-level dashboards for the eight states with direct Hormuz exposure. The composite stress index for Iran, Saudi Arabia, the United Arab Emirates, Kuwait, Bahrain, Qatar, Oman, and Iraq will reflect these inputs starting with the May 16 build (v3.2). The methodology change is documented here.
For agent IO and machine readers, the new fields are exposed under /api/v1/countries/{iso}/risk and listed in the updated agent-card.json at /.well-known/. The MCP server adds two new skills: get-war-risk-premiums and get-lane-spreads.
N. Vermeulen is the energy editor at shortage.life. He previously covered downstream markets at Argus and refinery operations at Reuters. Reach him at nv@shortage.life.