Is the samarium control leaking? The defense near-monopoly the customs code can't see

Trade-flow companion to the price wedge (R72). The wedge measures scarcity (what a controlled material costs ex-China). This case measures the other half: is the control actually holding, or is controlled material re-entering the buyer through a laundered origin? This is a DUAL-SCORE / alternative-track signal — never folded into any Tier-1 exposure score. The divergence between "who China licensed to" and "who suddenly started shipping to the EU/US" IS the signal. Research, not investment advice; origin-relabelling is INFERRED from statistical implausibility plus public records, never asserted as smuggling on any single shipment.

Live chart + method: [/situation-room → "Is the ban leaking?"](/situation-room).

Verdict — GATE 0 (near-monopoly, basket-blind, but leak state-confirmed)

Samarium is the sharpest single-origin dependency in the entire controlled set, and it is the one the customs taxonomy is least able to police. Three facts sit on top of each other:

1. The control names it directly. On 4 April 2025 China's MOFCOM + GACC (Announcement No. 18) placed seven medium/heavy rare earths — samarium (Sm) plus Gd, Tb, Dy, Lu, Sc, Y — under individual, non-automatic export licence: metals, oxides, alloys, compounds, and samarium-cobalt (SmCo) permanent-magnet materials. On 9 October 2025 Announcements 61/62 asserted extraterritorial reach (0.1% de-minimis, 50%-affiliate, foreign-direct-product) over foreign-made goods containing Chinese-origin controlled REEs; on 7 November 2025 Announcement No. 70 suspended the October package for the US to 10 Nov 2026 (Busan truce). Critically, the Apr-2025 text explicitly restricts re-export "to a third country or region" of SmCo materials — an anti-trans-shipment clause written into the control itself.

2. The monopoly is near-total, and defense-critical. China held ~99% of global heavy/medium rare-earth *separation through 2023; the only ex-China separation line, a refinery in Vietnam, has been idle for over a year on a tax dispute — so for essentially the whole post-control window of 2025, no one outside China could legitimately produce separated samarium at all. Samarium's value is concentrated in SmCo magnets, which hold their magnetism to ~350 °C (specialty grades to ~550 °C) — irreplaceable in high-temperature defense hardware (precision-guided munitions, missile fin actuators, fighter aircraft incl. the F-35). Meaningful ex-China resilience is assessed 3–7 years out; a scaled mine-to-magnet chain 7–12+ years*.

3. The leak is not merely inferred here — the controlling state confirms it. In May 2025 China launched a special campaign against the smuggling and illegal export of strategic minerals, explicitly targeting "third-country laundering" and "misdeclaration / under-declaration," and stated it had uncovered cases of overseas organisations transferring controlled Chinese-origin rare earths for use in sensitive military fields. On 24 June 2026 MOFCOM Announcement No. 26 formalised a violation-reporting mechanism (effective 1 Jul 2026). This is the corpus's first case where the controlling state publicly asserts that trans-shipment of the material is happening and is policing it — corroboration most cases can only infer from mirror data.

So why GATE 0, not GATE 1? Because the two things that would let us quantify the leak both fail for samarium — as detailed below.

The origin shift — and why the table can't close

The EU import structure moved exactly as a leaking control predicts, but only at the rare-earth-basket level:

Source: Eurostat, "EU trade in rare earth elements increased in 2025" (news release 2026-06-29, Combined-Nomenclature rare-earth codes; free public dataset). Total 2024 imports ≈ 12,900 t implied by the +17.1% rebound; the release publishes 2025 partner shares but not a 2024 partner-by-partner split.

On its face this is the fingerprint: the controlling origin's share receding while Russia + Malaysia together take ~49% of EU rare-earth imports. But two corrections stop it short of a quantified GATE-1 relabel — and both are the honest part of the story:

• The customs code cannot isolate samarium. Eurostat's rare-earth lines are a

basket (mixed compounds/metals across all lanthanides). The one material whose phantom-flag test is cleanest in principle — because ex-China separation was ~zero — is the one the trade taxonomy resolves least. This is the same below-basket-resolution blind spot the Myanmar-HREE case hit from the other side (there, intra-origin; here, intra-basket). A samarium-specific surge from a zero-capacity flag would be invisible inside a rising rare-earth aggregate.

• **Both surgers hold genuine — if limited — capacity, so neither is a clean

zero-capacity phantom flag: - Malaysia = Lynas (LAMP, Kuantan). This is a *genuine* mode-B alternative, not a relabel: Australian-owned, fed by Mount Weld concentrate, and in 2025 the first commercial separated-heavy-REE producer outside China (Dy commissioned early-2025, Tb mid-2025, ~1,500 t/yr HREE circuit). But it only began producing samarium oxide "since late 2025" — so for the post-control window Malaysia's rare-earth surge is overwhelmingly *other* elements, and its samarium is genuine but near-nil. Malaysia's rise is real supply diversification; it is not evidence of samarium laundering, and it would be wrong to score it as such. - Russia = Solikamsk (SMZ). Russia's *sole* separator produces "basic carbonates and mixed powders rather than the refined individual elements" (~2,500 t/yr finished), and is only *now* standing up in-house separation of individual compounds incl. samarium/gadolinium/europium concentrates. So a Russia jump to ~26% of EU rare-earth imports is the more implausible leg for *high-value separated* material — and Russia was in talks with China over rare-earth extraction-technology transfer (Sept 2025). Suggestive of a China-Russia rare-earth nexus and a possible relabel/blend vector — but Russia's partial genuine output plus the basket code mean this stays inference, not a traced relabel.**

The transmission chain

`` China (near-100% samarium separation; SmCo magnet chain) → [Apr-2025 licence, incl. explicit re-export ban on SmCo] → leak paths China itself names: third-country laundering + mis-declaration · into the rare-earth basket (samarium invisible inside a rising aggregate) · via partial-capacity flags (Russia's low-separation line; blend/relabel) · as finished SmCo magnets/alloys under other HS lines (see magnet case) → buyer (EU/US defense & industrial), origin label ≠ economic origin ``

The samarium leak, if it is flowing, most likely wears two disguises at once: a basket disguise (aggregated into "rare earths" so no single-element surge shows) and a finished-good disguise (exported as SmCo alloy/magnet, caught only partially by the finished-magnet HS 850511 line in the companion [rare-earth-magnets case](2025-rare-earth-magnets-licensing-circumvention-transshipment.md)). The corpus's cleanest monopoly is therefore its hardest to photograph — which is itself the finding.

Common-ownership tell

N — not traced to a named entity. China's May-2025 campaign disclosed that "overseas organisations and individuals" moved controlled Chinese-origin rare earths for military end-use, but the public record I can cite does not name the shell, namesake, or address linking a surging exporter to a controlled Chinese producer (unlike the antimony Youngsun → Thai Unipet → Youngsun & Essen chain). Left as an open question rather than a fabricated hop.

Why it matters for the buyer

1. The sharpest dependency is the least visible in official data. A risk team reading "our rare-earth imports diversified — China is now under half" would read the samarium exposure as easing. For separated samarium specifically the dependency is essentially intact; it is hidden inside a rising basket and behind finished-good codes. Any exposure model keyed on stated rare-earth origin over-credits diversification for the one element that matters most to defense. 2. The control is binding — the controlling state says so. Origin-laundering only happens when a material is genuinely scarce and valuable ex-China. China's own anti-smuggling campaign is independent confirmation that the samarium/REE control bites, in the same direction the price wedge prices scarcity. 3. Genuine diversification and circumvention look identical in the aggregate. Malaysia's Lynas surge (real) and a hypothetical Russia relabel (inferred) both raise a non-China rare-earth share. Only element-level, capacity-aware reading separates them — which is exactly why we hold this as an alternative-track signal beside, never inside, the Tier-1 score.

Caveats

• Inference, not proof. No figure here isolates samarium; the Eurostat table is

the rare-earth basket. A single Russian or Malaysian shipment is innocent on its face. "Consistent with" and "fingerprint of," never "proves smuggling."

• Genuine alternative explicitly labelled. Malaysia/Lynas is real ex-China

separation (mode B), not a phantom flag; do not read its rise as circumvention.

• Data lag + basket resolution. Eurostat CN rare-earth codes cannot resolve

samarium; a samarium-specific relabel is structurally invisible to the volume-implausibility detector at this resolution. Element-level customs data (or SmCo-specific HS granularity) would be required to move this toward GATE 1.

• Suspension noise. The Nov-2025 US suspension (to Nov-2026) partially reopens

the direct US channel, which can mute the very relabel we are watching for in the window it covers; the EU channel is unaffected.

Anchor action: `2025-04-04-china-mofcom-heavy-rare-earths-export-licensing`. Sources — control: MOFCOM/GACC Announcement No. 18 (4 Apr 2025), Announcements 61/62 (9 Oct 2025) & No. 70 (7 Nov 2025), as reported by Holland & Knight, CIRS, CSIS, White & Case, Mayer Brown, Global Times. Monopoly/defense: CSIS, Arnold Magnetic Technologies, Dura Magnetics, Rare Earth Exchanges. Ex-China separation: Lynas Rare Earths & Magnetics Magazine (Malaysia LAMP, samarium oxide from late 2025); Discovery Alert & S&P Global (Russia/Solikamsk + China tech-transfer talks). Trade: Eurostat news release 2026-06-29. State enforcement: MOFCOM special campaign (May 2025) & Announcement No. 26 (24 Jun 2026) via Global Times / Morgan Lewis. Presented as the publicly-reported mechanism — NOT a per-shipment accusation beyond what the public record states.