SPR releases are loans, not sales — so the refill is not the bullish catalyst the bulls expect

One-line summary: The 2026 US SPR releases were exchange loans repaid in kind, not sales — so borrowers return barrels at today's lower prices, and the physical refill rate (~400 kb/d) is a fraction of the release rate (~1.9 Mb/d), which means "the SPR must be refilled, therefore oil goes up" is arithmetically wrong.

The insight

A standard bull argument in mid-2026: the SPR is drained, refilling it is a national priority, therefore a large sustained government bid arrives and lifts crude. anas-alhajji dismantles this in four separate ways, and each is independently sufficient.

1. They are loans, so the repayment flow is a supply, not a demand. Companies borrowed barrels at ~$120 and must return them with interest in kind (his illustration: borrow 10M, return 12M). They will return them when oil is $60–80. Two consequences: the oil companies and traders capture the spread — "making hundreds of millions of dollars of the taxpayer money" — and, critically, the government does not need to buy the barrels back at market. The refill arrives as repayment, not as a bid.

2. The rates are asymmetric by an order of magnitude. Peak release ran ~1.9 Mb/d; maximum technical injection is ~400 kb/d. "we cannot return this amount at the same rate that we released it." A 400 kb/d bid is a price floor, not a rally: "Yes, it supports prices, it creates a floor, but it's not going to raise prices substantially."

3. The inventory-draw statistic is being misread — and this is the sharpest point. Analysts cite falling global inventories as bullish. Alhajji: most of the decline was in strategic reserves, not commercial ones, and the two have opposite price signs. "when commercial inventories decline, oil prices go up. But when the Strategic Petroleum Reserve decline, prices go down. Why? Because they are intended for that." Releasing SPR is the policy response to high prices, so an SPR-driven inventory draw is evidence the mitigation is working, not evidence of scarcity. He calls the aggregate framing "a lie" and notes the decline was concentrated in just two countries (US and Japan) for two unrelated reasons.

4. Nobody is in a hurry to refill. Japan will not refill while the yen is weak — the April oil price in yen was "the highest in history", which is why Japan drained. China will not refill unless prices are below $70, "and we have data for the last 10 years to prove this point." And the constraints on further US draws are legal, not physical: the 252M floor is a Congressional limit the President can waive in an emergency; the US is a net exporter so the IEA's 90-day rule does not bind it; and the caverns are fine ("It's filled literally with water instead of oil to maintain the integrity").

The real constraint is crude quality, not quantity — and this is where the concept connects to the refining bottleneck. Diesel needs medium sour crude; shale is light sweet. The SPR's strategic value was that it held medium sour barrels, exactly what Hormuz stopped supplying. Alhajji: without the releases "diesel prices in the United States probably would be in the range of $12 a gallon." So shale is not a substitute SPR: "we can produce a lot of gasoline from it, but we cannot produce a lot of diesel." See refining-bottleneck-to-refiner-crack-capture.

Evidence

All from anas-alhajji in 2026-07-16-podcast-macro-voices-macrovoices-541-dr-anas-alhajji-bab-el-mandeb-the.

• Loans, not sales: "Those releases are loans, they are not sales. And this is very important to the balances of the market... when companies borrowed the oil at 120, it was extremely attractive for them to borrow because they need to return this oil later on with some interest in kind. So, for example, if ExxonMobil borrowed, let's say 10 million barrels, they have to return them 12 million barrels later on... they borrowed the oil at 120 and later on when they return it, they will return it when oil is 60 or 70 or 80. So even with the interest, the oil companies basically and the traders are making hundreds of millions of dollars of the taxpayer money... So the idea here is when they refill, they are not going to refill at a higher price, they are going to refill at a lower price. So those who are expecting prices will go up, companies are not going to play the game that way because it's a loan, it's not a purchase."
• Rate asymmetry: "we have asymmetry between the releases and the injection. For example, we released at certain days, we release 1.9 million barrels a day. We did that. But we cannot refill that amount because technically speaking, at the current situation, we cannot inject more than 400,000 barrels a day... Yes, it supports prices, it creates a floor, but it's not going to raise prices substantially. We've seen this before."
• Strategic vs commercial inventories have opposite signs: "Most of the decrease was in strategic petroleum reserves, not in commercial inventories. And the impact here is different because when commercial inventories decline, oil prices go up. But when the Strategic Petroleum Reserve decline, prices go down. Why? Because they are intended for that. The government holds them for emergencies... And they kept telling us global oil inventories declined this much. That is a lie. That is not correct. And the reason why? Because most of that decline was only in two countries, the United States and Japan... And if you look at the small decline in commercial inventories, they were only in two countries, the United States and China. So we cannot even say there were a global decline in inventories."
• Volumes: "we already released about 99 [million barrels] and we still have about 73."
• The limit is legal, not technical: "We've seen people basically talking about, oh, we have a technical limit. We cannot go below 250. This is a lie. It's not a technical limit, it's a legal limit... the Congress says that the President, in case of emergency, can go below the 250... The technical limits are way, way lower than that." And on the IEA: "the condition for the IEA is to have 90 days of your net imports and the United States is a net exporter and therefore the IEA rules do not apply to the United States."
• The caverns are fine: "those who are talking about a collapse of those caverns, they need to study geology... some of them basically think that we took that oil out of the caverns and there is nothing but air. That is not correct. It's filled literally with water instead of oil to maintain the integrity of those caverns."
• No refill bid from Japan or China: "Japan is not going to refill as long as the yen is low. Because even by today's prices, the price is too high for Japan." And: "The Chinese are not going to refill their inventories. And we have data for the last 10 years to prove this point unless prices are below $70."
• Quality, not quantity: "you can look at shale literally as the spr, because we can produce oil quickly from shale. Here's the problem. Crude quality. What we need for diesel basically is medium sour crude. What we produce from shale is light sweet crude. So we can produce a lot of gasoline from it, but we cannot produce a lot of diesel."
• What the medium-sour releases prevented: "without the US Strategic Petroleum reserve releases, diesel prices in the United States probably would be in the range of $12 a gallon... those massive releases of spr, especially of the medium sour crude, saved the world."

Implications

• Kills a specific bull argument. "The SPR must be refilled → sustained government bid → higher crude" does not survive: the barrels come back as loan repayments, the injection rate is ~1/5 the release rate, and the two largest other refill candidates (Japan, China) are both price- or FX-blocked. The SPR refill is a floor, not a catalyst.
• Corrects an inventory read that this vault could otherwise get backwards. Global-inventory-draw headlines in 2026 are mostly an SPR artifact concentrated in two countries. Do not treat aggregate inventory declines as a bullish commercial-tightness signal without decomposing strategic vs commercial.
• There is still ~73 Mbbl of politically-available release below the legal floor, waivable by the President. From 2026-07-17-podcast-forward-guidance-the-ai-unwind-is-forcing-a-historic-market, the Forward Guidance panel reads this the same way — Felix: "There's still ample reserves actually to draw down... if you're a pumper in chief looking to manipulate markets and trying to sustain things into midterms, there is still functionally reserves to be drained that would support risk and keep a lid on oil." Independent corroboration from an unrelated source that the SPR remains a bearish overhang into the midterms, not a bullish one.
• Feeds refining-bottleneck-to-refiner-crack-capture — the quality argument (medium sour vs light sweet) is the same constraint seen from the reserve side.

Contradictions / tensions

• ⚠ Rate figures still one-expert; loan structure is now DOE-primary (2026-09-19). The 99/73 Mbbl split, the 1.9 Mb/d vs 400 kb/d rates, and the ExxonMobil illustration remain Alhajji’s. The loan structure is now confirmed against 10 CFR 626.7 + FY22/FY26 RFPs (2026-09-19-autoresearch-doe-eia-spr-exchange-loan-terms). The 400 kb/d ceiling is not DOE design fill and has not been observed.
• His ExxonMobil figures are explicitly hypothetical ("This is just an example"), not a disclosed transaction.
• Interested party: Alhajji sells energy research and is arguing against the "perma bulls" — a position he has staked publicly and repeatedly, which creates a consistency incentive.
• The claim that the President can freely waive the 252M floor is a legal reading offered without citation.

Open questions

• Can the SPR exchange (loan) terms — volumes, return schedules, interest-in-kind rates — be confirmed from DOE primary sources? Everything on this page turns on it.
• What is the actual observed injection rate since releases stopped, and does it corroborate the ~400 kb/d technical ceiling?
• If ~73 Mbbl remains releasable below the legal floor, is there a dated political catalyst (midterms) for using it? Two independent sources now suggest so.

Related

• refining-bottleneck-to-refiner-crack-capture
• war-risk-insurance-as-chokepoint-transmission
• energy-shock-2026-vs-2022
• anas-alhajji
• 2026-08-27-feed-construction-physics-how-the-strategic-petroleum-reserve-works

---

Update (2026-08-11) — independently corroborated, and the premium is now quantified

From 2026-08-09-feed-doomberg-endangered-specious. Until today this concept rested on a single voice (anas-alhajji, on a podcast). Doomberg is a written, independent, different-source-type second leg — which is what makes confirmed mean corroboration rather than two outlets chasing one narrative.

Doomberg describes the same mechanism in the same terms — "molecular loans will be repaid in molecules, plus interest — also in the form of molecules" — and frames it as a physical calendar spread. Three additions beyond what the wiki held:

• The premium is quantified: "premiums of up to 24% in the form of extra oil." The prior source's illustration was a round "borrow 10M, return 12M"; 24% is the reported contractual ceiling.
• Volume: Energy Secretary Chris Wright expects "about 35 to 40 million barrels of extra oil will be returned this year and next in the form of premiums." That is the premium alone, not the principal.
• The political mechanism, which the wiki did not have: "With no cash changing hands, and thus no need for congressional approval." This is why the program is durable — it routes around appropriations entirely. It also makes the refill insensitive to political turnover, strengthening the concept's claim that the refill is not a discretionary bid that can be switched on to support prices.

Second datapoint, filed for the energy chains: Venezuelan oil exports surpassed 1 million bpd in July, the third consecutive month, up from roughly half that level in December — attributed to a US-led revitalization of the Venezuelan oil industry. A supply-side offset to the Iran shock that this book's energy chains should carry; the energy-oil cluster has been the worst-performing in the live feed for five consecutive weeks.

⚠ partial — the free portion is Doomberg's setup; the four-part analysis it promises (pump prices, US/Canada production, refining capacity) is paywalled and unread. Doomberg is also an opinionated, stance-taking publication whose framing here is explicitly favourable to the current Energy Secretary; the mechanism is corroborated, the editorial judgment ("shrewdness and prudence") is not evidence.

Update (2026-08-28) — Construction Physics: caverns are always full of liquid; the 300M "technical limit" is not in official reports

From 2026-08-27-feed-construction-physics-how-the-strategic-petroleum-reserve-works (Brian Potter / Construction Physics, published 2026-08-27). Independent written engineering source — does not re-argue loans-vs-sales. It corroborates Alhajji's cavern-integrity claim and dates the Hormuz draw.

• ~120 million barrels released for Hormuz as of writing. The SPR went below 300 million barrels the week of August 7.
• Caverns are always completely full of liquid; oil sits on brine; drawdown pumps water in, fill pumps oil in and brine out. That is the same geology Alhajji summarized as "filled literally with water instead of oil."
• No official DOE / Sandia / GAO SPR-wide fill level below which cavern damage occurs, other than roof oil (~12 million barrels / ~1.7% of capacity) to keep water off the ceiling. DOE has denied a 300 million barrel limit. Analyst claims (150–300M) are inconsistent and unpublished.
• Every full drawdown grows a cavern by ~15% (100 barrels of water dissolve ~15 barrels of salt). Phase I caverns include single-cycle wells; 11 Phase I caverns still in use, six with one drawdown or fewer remaining. Cavern creep costs up to 2.4 million barrels/year of capacity.
• Potter does not confirm the loan/exchange structure. The load-bearing loans-vs-sales claim still rests on Alhajji + Doomberg. This page's caverns-are-fine / 300M-is-not-a-technical-floor leg is now a third independent source-type.

Update (2026-09-19) — DOE primary confirms loan-not-sale; injection has not started

From 2026-09-19-autoresearch-doe-eia-spr-exchange-loan-terms (/explore-chain on spr-exchange-loans-to-oil-refill-floor). The load-bearing claim is no longer Alhajji + Doomberg only.

• 10 CFR 626.7 + DOE SPR FAQs + FY22/FY26 RFPs: exchanges are time-loans repaid in similar-quality crude plus premium barrels. Cash is a close-out/quality residual. Wright (11 Mar 2026): 172 Mbbl replaced by ~200 Mbbl “at no cost to the taxpayer.” FY26 minimum premiums 8–24%; return windows 2027–2029.
• EIA weekly still drawing through week ending 11 Sep 2026: 284.957 Mbbl, −58 kb/d that week. Injection has not started; Alhajji’s ~400 kb/d is not DOE law and is untested.
• XLE / USO embedding unverified. Do not graduate the hypothesis. The “SPR must refill → government bid → crude up” argument is weaker with DOE primary; the instrument leg is not. Filed as a question: spr-exchange-loans-to-oil-refill-floor.