CLARIDAS NEWS The world, seen clearly

WORLD

How Iran's Oil Most Likely Reaches China: A Route the Trade Ledgers Don't Reconcile

By U.S. estimates, Iran shipped about 1.58 million barrels a day of crude in 2025 and roughly 1.57 million of it toward one buyer. No one can watch the paperwork change — but the U.S. government's own report, and an unreconciled gap between two countries' trade ledgers, are consistent with how it would.

Correction — 2026-08-10: An earlier version argued the China–Malaysia trade-ledger gap 'settles' or 'proves' relabeling and said Malaysia 'cannot export crude it did not produce.' Production is not a ceiling on exports — a refining-and-transshipment hub re-exports imported crude — so the unreconciled mirror gap is strong evidence of, not proof of, relabeling; the where and when of the origin edit cannot be determined from aggregate trade data and is now tagged [speculative]. A timeliness note was added: the record described is 2023–2025, and the 2026 war and Gulf-shipping disruption have sharply reduced current physical flows. The central finding was retained then, but is further narrowed in the 2026-08-18 note below. Flagged by an external cross-model review and by our new inference-discipline gate; reviewed by our Publisher-of-Record. · 2026-08-10 (2nd): The EIA/Vortexa export figures — ~99% to China, the ~$48B revenue estimate, and the 1.57M Vortexa restatement — were tagged [verified]; they are MODELED estimates (EIA's own report calls the 99% share 'an assumption, not an observation'), now retagged [modeled]. The reproducible UN Comtrade mirror-gap arithmetic remains reproducible as an operation, but see the 2026-08-18 notes: because Comtrade flags China's quantity as estimated (`isQtyEstimated: true`) while Malaysia's is not, the 4.53× comparison itself is [modeled], not a clean reported-to-reported ratio. Flagged by our cross-LLM (Alden) and adversarial (Vista) review. · 2026-08-18: A belated cross-LLM (Alden) gate held the live piece 4/10 for presenting an INFERRED route as observed. Remediated under the Correction Protocol on a new SHA: reframed the headline, subhed, and thesis as an estimated route CONSISTENT WITH an unreconciled mirror discrepancy, not a directly observed nationality edit; retagged the origin-edit inference [speculative] while keeping the raw Comtrade discrepancy [verified]; scoped the single OFAC-documented LYNN transfer as one example, not a governing mechanism; attributed the sanctions cost-chain link by link; removed the secondary-sourced shadow-fleet (~3,300) and Kpler (1.38M) figures and the unsourced 2026 current-flow comparison; disclosed the mirror-gap confounders (consignment-vs-origin, re-exports, free-zone coverage, timing, revisions, kg-to-barrel density); trimmed Room for Disagreement within the word limit; removed the non-compliant closer. The NARROWED finding: a large, reproducible mirror discrepancy between China's and Malaysia's crude records is consistent with Iranian-origin relabeling — which the public evidence infers but cannot directly observe.

The Facts

Iran exported an estimated 1,576 thousand barrels per day (b/d) of crude oil and condensate in 2025, and an estimated 1,567 thousand b/d of that went to China — about 99% modeled. Both figures are from the U.S. Energy Information Administration's June 2026 "Report on Iranian Petroleum and Petroleum Products Exports," required by the SHIP Act, using Vortexa tanker-tracking data accessed March 2026. The same report states its method plainly: for volumes from 2020 onward, EIA assesses that Iranian crude sent to "unknown" destinations and to Malaysia, Singapore, and Vietnam was "likely destined for China". EIA estimates Iran's crude-and-condensate export revenue at about $48 billion in 2025 modeled. The trade record itself carries a discrepancy consistent with relabeling. EIA reports that China imported 54% more crude oil from Malaysia in 2023 than in 2022 — about 1.1 million b/d — a volume that "exceeded Malaysia's total crude oil production". EIA states that "industry analysts believe that much of the oil shipped from Iran to China was relabeled as originating from countries such as Malaysia, the United Arab Emirates, and Oman to avoid U.S. sanctions". Malaysia's own petroleum-and-other-liquids production had declined to 597,000 b/d by 2023 as its fields matured. The arithmetic is reproducible in one free database, with a caveat the database itself attaches. In UN Comtrade, China's record returns 46,939,955,132 kg of crude (HS 2709) imported "from Malaysia" in 2023 — a quantity Comtrade marks *estimated* (`isQtyEstimated: true`) — while Malaysia's record returns 10,367,873,490 kg exported to the entire world, marked not estimated. The two retrieved values and the division are; but because the numerator is a Comtrade estimate rather than a clean reported weight, the resulting 4.53× comparison is modeled, not a reported-to-reported ratio. Over the same window, China's directly-recorded crude imports from Iran return zero rows in Comtrade for 2023 and 2024 — an absence of records, not a measured zero. On the physical leg, U.S. Treasury sanctions notices name the ships. In its "Economic Fury" action targeting Iran's oil trade and shadow fleet, OFAC states that the Hong Kong-flagged tanker LYNN "engaged in ship-to-ship transfers of Iranian crude oil with a shadow fleet tanker off the coast of Malaysia in May 2025, ultimately delivering the cargo to China". The same release states that "since February 2025, OFAC has sanctioned over 1,000 Iran-related persons, vessels, and aircraft as part of this campaign". That is one documented cargo and a campaign count — not a measured share of the 2025 flow.

The Analysis

The following is analysis, not fact. The striking thing about this trade is not that it is hidden. It is how much of it is written down. Iran's oil leaves the country the way any oil does — loaded at a Gulf terminal onto a tanker — and then, on the best available reading of the trade records, somewhere between the loading port and a Chinese refinery its paperwork appears to change nationality speculative. The border it crosses may be less a coastline than a label. Three mechanisms would fit the traces the records leave. First, a destination edit: cargoes booked to "unknown" ports or Southeast Asian ones that EIA's analysts assess as China-bound modeled. Second, an origin edit: China's customs record returns a "Malaysian" crude quantity — one Comtrade flags estimated — exceeding everything Malaysia reports exporting to the world modeled; reading that gap as a deliberate origin rewrite is the further inference speculative. Third, a physical leg: OFAC documents one ship-to-ship transfer off Malaysia in May 2025 (the tanker LYNN) — a single verified example, not a measured share of the flow. The first two are accounting; the third, in the one documented case, is seamanship. The tell is a mismatch anyone can pull. In 2023 China's customs booked about 1.1 million b/d of "Malaysian" crude — more than Malaysia's own records say it exported to the entire world. Production alone is not a ceiling: a refining-and-transshipment hub re-exports crude it imports, but re-exports sit inside a country's own export ledger, and Malaysia's does not appear to carry them at this scale. The discrepancy reproduces from the retrieved fields, though China's side is a Comtrade estimate — one more reason to read the gap cautiously modeled; what it does not by itself rule out — consignment-versus-origin reporting, free-zone coverage, timing and later revisions, the estimated numerator's construction, and the kg-to-barrel conversion — is why we read it as consistent with, not proof of, Iranian relabeling speculative. Comparing its magnitude to Iran's missing barrels would need a crude-density conversion and aligned time-and-commodity coverage this piece does not attempt. What would route the trade this way is sanctions, not smuggling for its own sake. The U.S. Treasury's "Economic Fury" action shows the rule in force: OFAC's secondary-sanctions campaign penalizes buyers of Iranian oil and the vessels that carry it. The label change would be the cost of doing business under that rule — but the links are not equally established: OFAC documents sanctioned vessels and one at-sea transfer, while the price discount Iran accepts, the systematic use of older tankers, and a destination-wide origin rewrite are plausible inferences the public record does not separately prove speculative. What the estimates do agree on is direction: on the modeled data, a large, one-directional flow routed almost entirely through a single buyer modeled.

Room for Disagreement

The strongest counter is that the "99% to China" figure and the mirror-gap arithmetic both rest on estimates and inference, not on a clean bilateral receipt. That counter has real force, and the U.S. government's own report concedes most of it. EIA's SHIP Act report says directly that "because of limited availability and transparency of information on Iranian oil exports, nearly all" of its figures are estimates, and that its 99%-to-China share depends on an *assessment* that "unknown" and Southeast-Asian-destined cargoes are China-bound — an assumption, not an observation. Remove that assumption and the measured share falls. On the labeling itself, the report is blunt: its section titled "Iran's labeling practices" reads, in full, "We do not have any source for this information". The relabeling is inferred from the trade-ledger gap; it is not something the reporting agency can document at source. What the estimates do not erase is the mirror gap: China's customs record returns about 4.5 times more "Malaysian" crude than Malaysia reports exporting to the entire world — though China's side of that ratio is a Comtrade-flagged estimate, so the comparison is modeled, not a clean reported one. It is a discrepancy re-exports and refining throughput may not close, since those already sit inside Malaysia's own export figure. The estimates argue over the size of the flow; the unreconciled gap is what is consistent with relabeling speculative. A further limit is timing: the record here is 2023–2025 and describes how the trade routed then, not necessarily what moves today.

The View From

From Kuala Lumpur, the story is an accusation the numbers make on Malaysia's behalf, not against it. Malaysia's own Comtrade filing reports exporting about 10.4 billion kg of crude to the whole world in 2023, while China's filing records about 46.9 billion kg "from Malaysia" — a figure Comtrade marks estimated modeled. What is observable is that the discrepancy sits in China's own import record; where, or when, any origin label was actually changed, aggregate trade data cannot say speculative. The transshipment label borrows a neighbor's name, and on these two records — one of them a Comtrade estimate — the figures do not reconcile.

Notable

How this was made. Models: World pod — Opus writer/editor; remediated by Onett per Alden HOLD 4/10 → 6/10 (2026-08-18). [verified] = the retrieved EIA/Treasury/Comtrade fields and the arithmetic operations on them; the 4.53× Comtrade comparison is [modeled] because Comtrade flags China's numerator estimated (`isQtyEstimated: true`) while Malaysia's is not; the relabeling route and the sanctions cost-chain are [speculative] inferences the public evidence is consistent with but does not directly observe. Data vintage: EIA report June 2026, Comtrade 2023 filings, OFAC 2025 campaign.. Publisher of Record: Unruly Labs LP. Published August 10, 2026.

Confidence. Every factual claim here is verified against a cited primary source. A marker appears only where a claim is modeledmodeled, speculativespeculative, or preprintpreprint — the departures from verified worth flagging.

Sources. U.S. EIA, Report on Iranian Petroleum and Petroleum Products Exports (SHIP Act, June 2026): 2025 crude+condensate exports ~1,576 kb/d, ~1,567 kb/d (~99%) to China; Vortexa method assessing Malaysia/Singapore/Vietnam/unknown as China-bound; ~$48B 2025 export revenue; and the 'We do not have any source for this information' admission on labeling practices. (retrieved 2026-08-10) · U.S. EIA, Today in Energy #61843: China imported 54% more crude (~1.1M b/d) 'from Malaysia' in 2023, a volume exceeding Malaysia's total crude production; analysts believe Iranian oil was relabeled as Malaysian, UAE, and Omani origin. (retrieved 2026-08-10) · U.S. EIA, Malaysia Country Analysis Brief: Malaysia's petroleum-and-other-liquids production declined to 597,000 b/d by 2023 — the baseline against which China's 'from Malaysia' crude figure exceeds what Malaysia reports exporting to the world. (retrieved 2026-08-10) · UN Comtrade (HS 2709, crude petroleum, 2023): China imports 'from Malaysia' 46,939,955,132 kg; Malaysia exports to world 10,367,873,490 kg (4.53x); China's directly-recorded crude imports from Iran fall to zero recorded rows in 2023 and 2024. Re-pulled this run. (retrieved 2026-08-10) · U.S. Treasury OFAC, 'Economic Fury Targets Global Network Fueling Iran's Oil Trade and Shadow Fleet' (press release sb0472): the tanker LYNN engaged in ship-to-ship transfers of Iranian crude off Malaysia in May 2025, delivering to China; 'since February 2025, OFAC has sanctioned over 1,000 Iran-related persons, vessels, and aircraft as part of this campaign.' (retrieved 2026-08-10)