Autoresearch: DOE/EIA primary on 2022–2026 US SPR exchange loans, injection rates, and refill-as-price-support

Generated by /autoresearch on 2026-09-19. Synthesized across 3 rounds from government HTML/PDF plus a small set of secondary recaps of named-bank interviews. Anchored by Grokipedia Strategic_Petroleum_Reserve_(United_States) after the un-qualified slug resolved to India. See Provenance. Treat as raw material — review before promoting. No buy/sell/size. Specific avoidance of a refill-bull is in scope; this note does not recommend a short. Context: vault/projects/stock-market (hypothesis spr-exchange-loans-to-oil-refill-floor)

Summary

DOE primary treats SPR exchanges as time-loans repaid in similar-quality crude plus premium barrels, not as sales. Cash appears only as a close-out/quality residual, not as the ordinary repayment instrument (DOE SPR FAQs; 10 CFR 626.7; FY22 Winter Exchange RFP; FY26 Oil Release No. 2 RFP). The 2026 Hormuz/IEA action is the same legal vehicle: Secretary Wright’s 11 March 2026 DOE statement authorized 172 million barrels to be replaced by ~200 million barrels “within the next year… 20% more… at no cost to the taxpayer” (energy.gov, 2026-03-11). Subsequent RFPs put minimum premiums at 8–24% with return windows into 2027–2029, so the “within a year / 20%” headline is not the contract calendar (FY26 No. 2 RFP; FY26 No. 3 RFP).

EIA weekly stocks have not yet turned to net injection. From a 2026 peak of 415.441 million barrels (week ending 20 Feb) the reserve was 284.957 million barrels in the week ending 11 Sep 2026 (EIA release 16 Sep) — still a draw, slowed to ~58 kb/d that week, not a refill (EIA WCSSTUS1 weekly table). Alhajji’s ~400 kb/d injection ceiling is directionally near GAO’s December 2025 effective fill 0.440 Mb/d, but it is not DOE’s design fill (0.785 Mb/d) and is below FY26 RFP anticipated receipt (225 + 225 + 125 + 225 = 800 kb/d). The 400 figure is therefore not a demonstrated technical hard cap, and observed post-slowdown flow has not tested it because injection has not started (GAO-26-106918; FY26 No. 2 RFP B.6(f)).

On the refill-bull: DOE’s own acquisition rule tells the Department to refuse or suspend an exchange if it would add significant upward pressure to prices (10 CFR 626.7(c)(4)). Wright’s primary is “no cost to the taxpayer” / premium-in-kind, not a cash bid. A July 2026 Goldman recap still treats global SPR rebuilding (~1 Mb/d) as a floor under prices that nonetheless leaves a ~2 Mb/d 2027 surplus (OilPrice recap of Samantha Dart / Bloomberg TV, 2026-07-01). No fetched sell-side note names XLE or USO, or states that those multiples still embed a US government refill-rally. Step 6 of the hypothesis stays unverified on instruments.

Findings

1. Legal structure: exchanges are in-kind loans; cash is residual, not the loan

DOE’s FAQ is explicit that crude-oil exchanges acquire oil for the SPR at no cost, that “loans” are a form of time exchange after a disruption outside the company’s control, and that exchange authority requires oil of similar quality repaid “along with premium barrels (similar to interest)” within a negotiated time; drawdown and transportation costs are folded into the premium (SPR FAQs). Broad authority is EPCA §159: the Secretary may acquire oil by “purchase, exchange or otherwise” (same). The history page repeats that EPCA lets DOE negotiate exchanges so the SPR “ultimately receives more oil than it released” (History of SPR Releases).

The regulation matches the FAQ and then adds a cash-shaped escape hatch that must not be over-read. 10 CFR 626.2 defines exchange as SPR petroleum provided “in return for petroleum products of comparable quality plus a premium quantity of petroleum products (in barrels)—or another form of premium as permitted by law—delivered to the SPR in the future,” and premium as “the additional amount of petroleum product (in barrels)—or another form of payment as permitted by law” (CFR 2025 title 10 part 626). §626.7(b)(6) (emergency request) and §626.7(c)(3) (solicitation) both say repayment “shall be in the form of barrels of petroleum products, or another form of repayment as permitted by law,” and shall include (i) principal barrels, (ii) DOE’s costs, (iii) a premium (same).

The FY22 and FY26 RFPs operationalize that as molecules first. Premium barrels are returned in addition to awarded volume, calculated on NSV barrels delivered from the SPR. Example from FY22: 1,000,000 Bryan Mound sour barrels returned by 28 Feb 2023 require +39,000 premium barrels (3.9%) (FY22 Winter Exchange RFP B.5). Cash shows up in two residual places only: B.7 close-out invoices undelivered volume (or pays over-delivery) at the cost value of Exchange Oil on the last DD250, drawable on the letter of credit; B.8 quality (API/sulfur outside ±0.5° / ±0.10 wt%) is settled in U.S. dollars (FY22 RFP B.7–B.8). That is not an optional cash repurchase of the loan principal.

Hypothesis implication: Alhajji/Doomberg “loans not sales / repaid in molecules plus interest in molecules” is supported by DOE primary. A reading that 2026 releases were ordinary sales with an optional repurchase, or that repayment is cash as a matter of course, is not supported by the FAQ, the CFR, or the RFPs fetched here.

2. 2022 winter exchange: 32 Mbbl, 2.3–9.1% premiums, returns 2022–2024

The 23–24 Nov 2021 50-million-barrel package was 32 million barrels exchange + 18 million barrels Congressionally mandated sale (Bipartisan Budget Act of 2018 / FY2022) (DOE summary; DOE solicitation notice). Exchange oil moved 16 Dec 2021 – 30 Apr 2022; return oil went back to the same SPR site in FY2022, 2023, and 2024, with higher premium for later return (same summary).

FY22 RFP fixed premiums (selected rows from B.5(a)):

(FY22 Winter Exchange RFP)

Anticipated FY22 receipt (injection) capability in that RFP: Bryan Mound 140 kb/d, Big Hill 225 kb/d, West Hackberry 120 kb/d, Bayou Choctaw 110 kb/d (sum 595 kb/d), with a note that receipt may decline as caverns fill (FY22 RFP B.6(f)).

Do not collapse 2022 into “all loans.” DOE’s own history page books the 2022 IEA coordinated action as an Operational Sale of 180.0 MMbbl, distinct from the Exchange Agreements section (History of SPR Releases). FY2022 mandated sales delivered 38.0 MMbbl; FY2023 mandated sales 26.0 MMbbl (same). The loan-not-sale claim is load-bearing for 2026 emergency exchanges, and for the 32 Mbbl 2021–22 winter exchange, not for every barrel that left the SPR in 2022.

The history page’s Exchange list, as fetched, still ends at a 2025 ExxonMobil emergency exchange of 0.5 Mbbl and does not yet enumerate the 2026 Hormuz RFPs (History of SPR Releases).

3. 2026 Hormuz/IEA exchanges: 172 Mbbl authorization, in-kind premiums 8–28%, returns 2027–2029

Authorization (DOE primary, Wright): On 11 March 2026 Wright said IEA members agreed to a coordinated 400 million barrel release and that the President authorized DOE to release 172 million barrels from the SPR, “approximately 120 days to deliver based on planned discharge rates.” He added: the United States has arranged to “more than replace these strategic reserves with approximately 200 million barrels within the next year—20% more barrels than will be drawn down—and at no cost to the taxpayer” (energy.gov, 2026-03-11). 172 × 1.20 = 206.4; “~200 million” is that 20% arithmetic, not an independent volume audit.

Press premiums (DOE, not RFP tables): On 30 April 2026 DOE issued an RFP for an emergency exchange of up to 92.5 million barrels, building on “three previous emergency exchange RFPs, which together quickly awarded approximately 80 million barrels across two completed exchanges” and “securing a 24 percent premium in returned crude oil barrels” (energy.gov, 2026-04-30). A later awards release said DOE would exchange more than 53.3 million barrels while securing an “approximately 28 percent return premium—representing 15.1 million barrels,” and that “to date, approximately 35 million barrels have been delivered to the market, while … [the] effort has generated approximately 35 million barrels of additional volume for the SPR at no additional cost to taxpayers” (energy.gov awards page). 15.1 / 53.3 ≈ 28.3%. The “35 million additional volume” line is DOE’s own aggregation; it is not broken out as barrels already injected versus contracted future premium, and it is not the same number as Wright’s March “~28 million extra on 172.”

RFP minimums (primary contract math):

FY26 Oil Release No. 2 (DE-RP96-26PO00004), up to 92.5 MMbbl — selected minimum premiums plus extra for later return periods (FY26 No. 2 RFP B.1):

• Bryan Mound sour Jun/Jul-26: 18.0% min; additional 0.0 / 0.5 / 1.5 / 3.0 pp for Jan–Mar 2027 through Oct–Dec 2028 windows.
• Bryan Mound sweet Aug-26: 18.0% min; additional 3.5 / 4.0 / 4.5 pp for 2029 windows.
• Bayou Choctaw sour Jun–Aug-26: 18.5% min.
• West Hackberry sweet Jul-26: 22.0% min.
• West Hackberry sour Aug-26: 24.0% min (+2.5 pp additional in the listed window).
• Big Hill sour Aug-26: 23.0% min.

Nominal receipt rate per full month in that table is 3,000,000 bbl/month at Bryan Mound and Big Hill (~100 kb/d if spread over a 30-day month) and 1,200,000 bbl/month at Bayou Choctaw — a scheduling number, not the site’s physical max. B.6(f) anticipated receipt capability: Bryan Mound 225,000 b/d, Big Hill 225,000 b/d, Bayou Choctaw 125,000 b/d, West Hackberry 225,000 b/d (sum 800 kb/d), “may decline as SPR caverns are filled” (FY26 No. 2 RFP B.6(f)).

FY26 Oil Release No. 3 (DE-RP96-26PO00005), up to 40 MMbbl, Bryan Mound / Big Hill sour, lower minimums: BM Jul-26 9.0%, BM Sep-26 8.0%, BH Aug-26 8.5%, BH Sep-26 8.0%, with extra pp for later 2027–2028 windows; award note as of 22 Jun 2026 shows Vitol 500,000 barrels only (FY26 No. 3 RFP; No. 3 award PDF).

FY26 Release No. 1.b (DE-RP96-26PO00003) award sheet (as of 17 Apr 2026): 26.030 million barrels to Alon, BP, Energy Transfer, ExxonMobil, Macquarie, Marathon, Shell, Trafigura (10.030), Vitol — premium table for this RFP was not in the award PDF fetched (No. 1.b award PDF).

Doomberg “up to 24%” is on a DOE RFP line (West Hackberry sour 24.0% minimum in No. 2) and in the 30 Apr press release (“24 percent premium”). It is a contractual ceiling/minimum on listed streams, not a uniform rate on every barrel. Wright “35 to 40 million extra as premiums” is not a single DOE table. Closest primaries: Wright ~28 million extra on 172 at 20% (11 Mar); awards page 15.1 million extra on 53.3 at ~28% plus a separate “35 million additional volume generated” clause whose accounting identity is unspecified. A Reuters 5 Jun 2026 Fox Business interview (not fetched; search snippet only) attributes to Wright “extra 40 million barrels,” “premiums of up to 24%,” “1.25 barrels back,” and “loaning about 133 million barrels.” Those Reuters figures are not verified against a DOE transcript in this pass and should not be treated as primary.

Calendar contradiction (material to “refill is a 2026 bid”): Wright’s 11 Mar phrase is “within the next year.” The RFPs schedule principal+premium returns into 2027, 2028, and 2029. Even RP1 for some Bryan Mound sour is Jan–Mar 2027. The physical refill cannot be a 2026 cash bid on this paperwork; it is a dated in-kind return.

Grokipedia’s US SPR page (encyclopedic, not primary) correctly flags the 2026 action as an emergency exchange with ≥18% premium and returns as late as 2028, and records an initial 20 Mar 2026 award of 45.2 million barrels (Bayou Choctaw 10 / Bryan Mound 15.7 / West Hackberry 19.5) (Grokipedia, Strategic Petroleum Reserve (United States)). Its ~243 million barrel post-draw projection is already stale versus EIA 285 million.

4. Observed SPR stocks: still drawing as of 11 Sep 2026; 400 kb/d ceiling not yet tested

EIA weekly ending stocks of crude in SPR (WCSSTUS1), thousand barrels (history through 28 Aug; latest six weeks, release 16 Sep 2026):

Peak-to-11-Sep draw: 415.441 − 284.957 = 130.484 million barrels. Wright’s 172 million in ~120 days would be ~1.43 Mb/d if fully executed on that clock; the fastest EIA week averaged 1.42 Mb/d (week ending 15 May). Alhajji’s 1.9 Mb/d peak release is not visible as a weekly-average EIA print. It may be an intra-week or design-rate figure; this pass found no DOE primary stating 1.9 Mb/d for 2026. Design drawdown remains the longstanding ~4.4 Mb/d hydraulic capability when full (DOE FY2026 SPR budget vol. 3; GAO-26-106918).

Construction Physics’ ~120 Mbbl Hormuz draw and below 300M the week of 7 Aug match EIA (298.694 kbbl week ending 7 Aug; 130.5 Mbbl from the Feb peak by 11 Sep, with the draw not finished). The 252.4 million barrel figure in the FAQs is the EPCA limited-drawdown floor (“if there are fewer than 252,400,000 barrels stored in the Reserve”) (SPR FAQs). At 284.957 million, remaining room to that legal floor is ~32.5 million barrels, not the ~73 million named when inventory was higher. 73 Mbbl was a level-dependent overhang, not a constant.

Injection vs 400 kb/d:

• GAO, using DOE estimates as of December 2025: effective fill 0.440 Mb/d vs design 0.785 Mb/d (56%). Site split in that table: Bayou Choctaw 0.105 / 0.110, one site 0.225 / 0.225, Big Hill 0.000 / 0.225 (construction outage), West Hackberry 0.110 / 0.225. Notes: effective rates are point-in-time December 2025; volume column is January 2026 (GAO-26-106918).
• FY22 RFP anticipated receipt 595 kb/d (FY22 RFP B.6(f)).
• FY26 No. 2 RFP anticipated receipt 800 kb/d (FY26 No. 2 RFP B.6(f)) — Big Hill is back in the receipt table at 225 kb/d, so the Dec 2025 Big Hill-zero snapshot is not the April 2026 RFP snapshot.
• EIA has not printed a net-injection week in the 2026 Hormuz window through 11 Sep. The slowdown is a smaller draw (58 kb/d), not a 400 kb/d fill.

Hypothesis implication: “Peak release ≫ injection, so refill is a floor not a rally” is structurally intact (draw weeks at 1.0–1.4 Mb/d versus fill design 0.785 and effective 0.44 in late 2025). The specific 400 kb/d technical ceiling is one expert’s round number next to GAO’s 440, not a DOE-published cavern law, and has not been observed because returns have not started. If FY26 B.6(f) 800 kb/d is achievable once returns land, Alhajji’s 400 understates physical fill.

5. Is a government refill bid 2026–2027 crude-price support? DOE says avoid lifting prices; sell-side talks a global floor, not XLE/USO

DOE regulation, opposite sign to a support bid. 10 CFR 626.7(c)(4): “Based on the market analysis … DOE may refuse offers, decrease the rate of acquisition, or suspend the exchange process if DOE determines acquisition will add significant upward pressure to prices either regionally or on a worldwide basis.” Factors include recent prices, private inventories, other stockpilers, production outlook, logistical problems, macro (CFR part 626). §626.6(b)(4) says the same for purchases. §626.4(a)(4) lists “possible effect on consumer and market prices of any SPR acquisition” among criteria before commencing fill. This is an anti-spike acquisition rule, not a mandate to bid crude higher.

Wright/DOE political line is stabilize via emergency release and restock via in-kind premium at no taxpayer cost, not “we will buy barrels in 2026–27 to put a floor under WTI.” The 11 Mar statement’s purpose clause is “to lower energy prices” with the IEA release (energy.gov, 2026-03-11). Haustveit’s RFP quotes are “move oil quickly into the market” and “return of premium barrels” (30 Apr RFP release; awards release).

Secondary (fetched): Goldman’s Samantha Dart, via Bloomberg TV as recapped 1 Jul 2026, said depleted inventories “will now have to be rebuilt — a process that’s likely to put a floor under oil prices,” while still seeing ~3 Mb/d global surplus next year and “a little over 1 million barrels a day just of SPR rebuilding globally,” leaving ~2 Mb/d surplus (OilPrice, 2026-07-01). That is global SPR + commercial rebuild, not a US DOE cash tender, and Goldman’s punchline is glut anyway. Morgan Stanley is cited there as cutting the 18-month oil forecast on Hormuz reopening (same).

Not fetched (do not treat as verified quotes): Reuters 5 Jun 2026 Wright/Fox Business; Washington Examiner 14 Sep 2026 Wright/Bloomberg TV “It’ll start being refilled in the next few months”; EnergyNow 403 on a July 2026 piece that search-attributed to Energy Aspects “SPR restocking will lead to a higher price floor in 2027” and to Kpler 506–664 kb/d of global reserve demand. Those remain gaps.

XLE / USO: this pass found no sell-side note, positioning print, or implied-vol write-up that names XLE or USO as embedding a US SPR refill-rally. Absence of evidence in a 15-URL crawl is not proof the tape never priced it; it is enough to keep hypothesis step 6 ⚠ unverified.

Hypothesis implication (avoidance, not a short): DOE primary strengthens “the 2026 US refill is repayment in kind, not a government market bid.” It does not by itself show that energy-equity or crude-ETF multiples still contain that bid. A global “inventory rebuild floor” argument still circulates at Goldman; it is weaker than a US cash-bid story and is paired with a 2027 glut call.

Contradictions and open questions

• Wright “~200 Mbbl back within a year / 20%” vs RFPs (returns 2027–2029, minimums 8–24%, one awards print at ~28%). The political 20% is not the contract grid. Which booked premium barrels DOE will actually receive, and when, is not in a public cumulative ledger fetched here.
• 35 vs ~28 vs 15.1 vs (unverified) 40 million “extra” barrels. Do not collapse these. 15.1 is 53.3 × ~28% on one award; 35 is an unspecified “generated” total on the same press page; 28 is 20% of 172; 40 is Reuters-only until a transcript is fetched.
• 172 Mbbl authorized vs 130.5 Mbbl EIA draw through 11 Sep. Either the 172 is incomplete, stretched past 120 days, partially unawarded (No. 3 shows only 0.5 Mbbl as of 22 Jun), or EIA stocks net out concurrent small returns. Unreconciled.
• 1.9 Mb/d peak (Alhajji) vs 1.42 Mb/d fastest EIA week vs 1.43 Mb/d implied by 172/120 days. No DOE primary for 1.9 found.
• 400 kb/d “max injection” (Alhajji) vs GAO 440 effective (Dec 2025) vs design 785 vs FY26 RFP 800 receipt. 400 is not a published DOE technical law. Observed injection is still ~0 (net draw).
• Grokipedia ~243 Mbbl post-172 projection vs EIA 285. Encyclopedic lag; use EIA.
• ~73 Mbbl below the 252.4 legal floor was inventory-relative; at 285 it is ~32.5 Mbbl. Still waivable for a full emergency drawdown; the limited-drawdown statute is what cites 252.4 million (SPR FAQs).
• Does XLE/USO still embed a refill-rally? Unanswered. Goldman’s fetched recap is global floor + 2027 glut, not those tickers.
• Japan/China refill blocks (Alhajji) were out of scope for this DOE/EIA pass and remain untested here.

Provenance

Rounds run: 3 of 3 (full)

Sub-questions by round:

Round 1 (broad survey):

1. DOE legal/FAQ/CFR structure of SPR exchanges vs sales (molecules vs cash)
2. 2022–2026 volumes, premiums, return schedules in primary documents
3. EIA weekly SPR stocks through the 2026 Hormuz window
4. DOE/Wright primary on refill as (non-)price support

Round 2 (drill-down):

1. FY26 RFP premium tables vs DOE 24%/28%/35-Mbbl press math — targeting the Wright/Doomberg premium claim
2. GAO and RFP fill/receipt rates vs Alhajji ~400 kb/d — targeting the rate-split gap
3. Whether EIA had turned from draw to injection after the Hormuz peak — targeting “observed refill”

Round 3 (resolve remaining uncertainty):

1. EIA weeks of 4 Sep and 11 Sep 2026 — targeting whether injection had begun
2. Wright 11 Mar 2026 DOE statement vs RFP calendars — targeting “within a year” vs 2027–29 returns
3. Sell-side “SPR bid / price floor” language, and any XLE/USO note — targeting hypothesis step 6

Anchor source (Grokipedia, fetched before round 1):

• First helper call on "Strategic Petroleum Reserve" → Strategic Petroleum Reserve (India) — wrong country; not used as US primer.
• Search + fetch: Strategic Petroleum Reserve (United States) — 92,251 chars, 127 citations, fetched 2026-09-19. Useful for EPCA/loan vocabulary and the 2026 Iran-exchange narrative (172 Mbbl, ≥18% premium, returns to 2028, 45.2 Mbbl first awards). Not used as the source of contract premiums, EIA weeks, or fill rates.

X sources: not run (--include-x not passed).

URLs fetched (successful unless noted):

Anchor:

• Grokipedia US SPR — encyclopedic primer — 2026 exchange framing
• Grokipedia India SPR (mis-resolved slug) — discarded for this topic

Round 1:

• DOE SPR FAQs — official — loans = time exchanges; premium barrels; 252,400,000 limited-drawdown floor (WebFetch timed out; curl extraction used)
• History of SPR Releases — official — exchanges vs 2022 operational sale 180 MMbbl; 2025 Exxon 0.5 Mbbl; mandated sales (WebFetch timed out; search-extracted HTML used)
• 10 CFR Part 626 (govinfo PDF) — federal primary, PDF source — exchange/premium definitions; §626.7 repayment in barrels; §626.7(c)(4) no-upward-price-pressure rule
• FY22 Winter Exchange RFP — official PDF — 32 MMbbl, 2.3–9.1% premiums, 2022–24 returns, cash only for imbalance/quality; B.6(f) 595 kb/d receipt
• Summary of 50 Million Barrel Release — official — 32 exchange + 18 mandated sale
• DOE announces 50 million barrel release — official — returns CY2022–2024
• EIA weekly SPR stocks history — official — weekly series through 28 Aug 2026
• DOE 92.5 MMbbl RFP press — official — 24% premium on earlier ~80 MMbbl awards
• DOE 53.3 MMbbl awards press — official — ~28% / 15.1 MMbbl; “35 million additional volume”

Round 2:

• FY26 SPR Oil Release No. 2 RFP — official PDF — up to 92.5 MMbbl; 18–24% minimums; B.6(f) 800 kb/d receipt; returns 2027–2029
• FY26 No. 2 Amendment 0001 — official PDF — same premium grid
• FY26 SPR Oil Release No. 3 RFP — official PDF — up to 40 MMbbl; 8–9% minimums
• FY26 No. 3 award information — official PDF — Vitol 500,000 bbl as of 22 Jun 2026
• FY26 No. 1.b award information — official PDF — 26.03 MMbbl awarded 17 Apr 2026; no premium table in this PDF
• GAO-26-106918 — official — Dec 2025 effective fill 0.440 / design 0.785 Mb/d; 172 MMbbl via emergency exchanges as of May 2026
• DOE FY2026 Volume 3 SPR — official PDF — site fill KPPs 225 / 225 / 110 MBD; drawdown 4.4 Mb/d design

Round 3:

• Wright 172-million-barrel statement, 2026-03-11 — official — 172 Mbbl / ~200 back / 20% / no taxpayer cost / ~120 days
• EIA WCSSTUS1 latest weekly table — official — 285,360 (4 Sep) and 284,957 (11 Sep); release 16 Sep 2026
• EIA Weekly Petroleum Status Report landing — official — week ending 11 Sep 2026, next release 23 Sep
• OilPrice / Goldman Sachs glut recap, 2026-07-01 — news recap of Bloomberg TV — global SPR rebuild ~1 Mb/d as a floor; ~2 Mb/d leftover surplus
• EnergyNow “SPR buying through 2028” — fetch failed (403)
• Reuters 5 Jun 2026 Wright/Fox Business and Washington Examiner 14 Sep 2026 Wright/Bloomberg TV — search snippets only; not fetched

Tools used: WebSearch, WebFetch, curl/HTML extract for energy.gov timeouts, grokipedia.py fetch/search, Read on govinfo/spr.doe.gov/GAO PDFs (whitelisted *.gov).
Generated: 2026-09-19 (UTC).