The Cushion Ran Out — Iran Is Won, Hormuz Is Shut, and the Bill Is Landing on the Treasury

Hormuz has been closed since February 28 — the SPR cushion is exhausted and Brent is at $92. The second-order effects have now reached the Treasury: foreign UST holdings fell three of four months, net interest is running 15% of federal spending, and the Treasury is doubling its buybacks to fight a sell-off. Iran clears its oil in yuan through CIPS. Russia and China ran the war in real time. And the war was lost, in part, because Palantir's AI targeting failed.

The most important sentence of this cycle is not a headline. It is the one Larry Johnson wrote about the Strategic Petroleum Reserve: the cushion ran out. Five months of Hormuz closure were masked by reserve drawdowns. The masking is over. And when a market stops being able to hide a shortage, it stops being able to hide the consequences of the shortage — which this week arrived not in the oil complex but in the Treasury market.

The War Was Already Lost

Start with what Johnson has been building toward since February: the Strait of Hormuz has been shut since February 28, 2026, removing an estimated 11–16 million barrels per day of Gulf supply. Shipping through the strait is running at roughly 17% of its pre-conflict average, per UKMTO. Brent is pushing $92, up about 25% from pre-war levels. And Scott Bessent — the man who announced “the toughest sanctions in history” — is, in Johnson’s telling, bewildered that oil keeps rising.

The bewilderment is the confession. A Treasury Secretary does not express surprise at a price move he engineered unless the mechanism he believed in has stopped working. The mechanism was: sanctions isolate Iran, isolation collapses the war effort, the strait reopens. Five months in, the strait is shut, Iran’s war economy is functioning, and the sanctions have instead isolated the dollar.

Ritter supplies the military verdict. The United States lost its war of choice with Iran, and one principal reason is that Palantir’s Maven AI targeting system failed. The human OODA loop — Boyd’s, the one trained operators actually execute — beat the algorithmic kill chain. This is the same argument Ritter makes about Israel’s Gaza apparatus: Lavender, Gospel, Fire Factory, “Where’s Daddy” — tens of thousands of casualties and a strategic result that looks like defeat. The technology that is supposed to be the future of war is losing wars now. (Hold that thought; it matters for the AI-capex trade far more than it matters for Iran.)

The Bill Lands on the Treasury

Here is where this cycle stops being a Middle East story and becomes a rates story.

Johnson’s reporting connects three facts that the market has been treating as separate:

  1. Foreign holdings of US debt fell to $9.299 trillion in June — down $72.1 billion in a month, the third decline in four months. China cut to roughly $765 billion.
  2. Net interest on US debt ran ~$963 billion in the first ten months of FY2026 — about 15% of all federal spending.
  3. The Treasury announced it will at least double its long-dated buybacks, from $2 billion to $4 billion per operation, to stem the sell-off.

Read those together. The buyer’s strike is no longer a forecast — it is a measured fact, printed in the TIC data. The foreigners are walking. The interest bill is a structural fiscal drag. And the Treasury’s response is to do more of the thing that is a maturity swap, not debt reduction — buying back off-the-run bonds and issuing fresh benchmarks to keep the long end from seizing.

This site has written at length about what the buyback actually is ([[US Treasury Buybacks — Maturity Swap Mechanics and the Buyer’s Strike]]): the state fighting its own rollover against a $40 trillion stock, where the marginal buyer is no longer the foreign official account but the levered shadow bank — the same shadow bank now pulled toward financing AI capex. The doubling of buybacks is that thesis moving from theory to policy.

De-Dollarization Is Now Operational

The deepest claim this cycle is not about oil at all. It is that Iran sells its oil to China — China buys 80–90% of Iran’s seaborne exports — settled in yuan through CIPS, entirely outside the dollar clearing system. CIPS hit a single-day record of roughly $178.5 billion in March 2026. This follows the January 29 Iran-China-Russia trilateral pact whose central plank is unified opposition to reimposed sanctions.

Johnson’s analytical kicker is the reason the whole thing works this time when it didn’t before. The 2015 JCPOA sanctions regime succeeded because isolation was universal — enforced by the US, Europe, and the UN Security Council including Russia and China. That is gone. Russia supplied Iran with real-time intelligence on US warship and aircraft positions. China supplied satellite imagery and BeiDou navigation. The United States is fighting a coalition while aiming its sanctions weapon at half of that coalition — and the weapon it chose (dollar exclusion) is the one that corrodes the currency it needs to fund the war.

Bessent doesn’t understand that he can’t weaponize the dollar against Iran and threaten Chinese banks with dollar exclusion and expect the rest of the world to keep buying the debt. The contradiction is the story.

The Covert-Action Record

Johnson’s other contribution this cycle is a reassembly, not a scoop: the documentary chain that makes the current war the terminal phase of a policy running since 2007. Bush’s classified finding authorizing up to $400 million for covert operations against Iran. Obama’s Timber Sycamore — over $1 billion, Saudi co-funded — arming Syrian rebels whose weapons reached al-Nusra. The declassified 2012 DIA memo that predicted a “Salafist principality” in eastern Syria, which was “exactly what the supporting powers to the opposition want.” HTS, the al-Qaeda descendant, toppling Assad in December 2024 — and the US then revoking its terrorist designation.

The through-line: the Global War on Terror was operationally subordinate to an undeclared war on Iran, with Sunni jihadism as the expendable instrument. The State Department’s own 2004–2023 terrorism data — roughly 90% of global attacks committed by Sunni extremists, not Iran-linked groups — contradicts the “Iran is the #1 sponsor of terrorism” narrative the escalation is built on.

This is not a new trade. It is the scaffolding that raises the ceiling on the de-dollarization and escalation threads, and it explains why the war has no clean political exit: the people running it have been running it since before the current president.

The Materials Ceiling

Ritter’s and Johnson’s military reporting converges on a hard fact: the US cannot out-produce the adversary it chose to fight. A 2025 Govini report found more than 80,000 parts across roughly 1,900 US weapon systems incorporate five minerals — antimony, gallium, germanium, tungsten, tellurium — whose processed supply runs through China. The gallium/germanium ban alone is estimated to cost $3.4 billion. The DFARS rules meant to bar Chinese-origin minerals keep slipping.

The war that was supposed to demonstrate American industrial supremacy is instead demonstrating American industrial captivity. This is the materialist version of Ritter’s AI point: the arsenal was built for a different war (counterinsurgency, precision strikes against non-state actors), the refinery complex was built for a globalized oil market where sour crude was always available, and the chip-mineral supply chain was built on the assumption the adversary would never be the supplier. All three assumptions are now false, and none can be corrected inside the timeframe the conflict imposes.

What This Means for Rates

The configuration is what we flagged in July and August, now with the physical data to match: a two-front energy war the US cannot resource, a buyer’s strike that is measured rather than predicted, and a de-dollarization loop that accelerates with every sanctions escalation.

Long end: structurally under supply pressure. The Treasury’s $4-billion buyback is demand-side support that one or two large IG/AI syndications can swamp — the exact point we have been making about the maturity-swap architecture. Term premium is repricing, and the buyback is a teaspoon against a tap.

Oil↔rates correlation: this is now positive, not negative. A supply-driven oil shock (Brent $92 and rising) passes into inflation breakevens and the front end, while the fiscal/de-dollarization story hits the long end. The curve steers bear-flattening, with breakevens structurally bid.

The dollar: the mechanism is the currency regime. Weaponized dollar → yuan settlement via CIPS → declining foreign appetite for UST → higher term premium → larger interest bill → more issuance. Each turn of the loop is slower than the trade, which is why it keeps being underpriced.

The AI read-through (our distinctive call): Ritter’s Maven-failure thesis is the mirror image of the AI-capex financialization we track in [[AI as Fictitious Capital]] and [[The Opportunity Cost of Survival — The Green-to-AI Capital Reallocation]]. The entire “AI is additive, excludable, national-security-essential” justification rests on the technology winning wars. If its warfighting value collapses on first contact with a real adversary, the capex narrative loses its hardest empirical prop — and the vendor-financing loop (NVIDIA, Apollo, Blackstone, KKR) that is structurally identical to the Treasury’s own maturity swap starts to look less like a frontier and more like the next buyer’s strike.

Confidence: HIGH on the oil persistence (the closure is five months old and the cushion is gone). MEDIUM on the pace of dollar erosion (the direction is measured, the slope is not). LOW on Johnson’s aggregate damage percentages — directionally useful, numerically unaudited.

What would falsify: a Hormuz selective reopening. The whole one-directional oil read — and the correlation flip it drives — rests on the closure holding. If the Navy’s logistics strain (Johnson documents the Bahrain hub destroyed, crews without food or functioning toilets, a Combat Logistics Force frozen at ~34 ships for a decade) forces a partial reopening, oil becomes two-sided and the term premium story decouples from the war story.

Sources: Larry C. Johnson (Sonar21), Scott Ritter (substack), Aug 11–23 2026. Cross-referenced against prior dispatches: Hormuz Arithmetic, Trump Blinks, Oil Revision, Munitions Depletion. NewsRoom digest at sources/NewsRoom Digest — 2026-08-23 supports all factual claims. Helmer Bar applied per item: single-source intelligence flagged as directional, document-sourced claims (Govini, DIA memo, TIC data) treated as verifiable.