Is the yttrium control leaking? The direct collapse the customs code CAN see — and the transit hub where genuine ex-China supply confounds the test

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" 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 (direct collapse traceable at element level; transit hub named but confounded by genuine ex-China supply)

Yttrium is the sibling of samarium under the same control — but its circumvention story runs the opposite way to its [samarium companion](2025-samarium-licensing-circumvention-transshipment.md). Samarium's leak is invisible because the customs taxonomy dissolves it into a rare-earth basket. Yttrium's direct-collapse leg is one of the cleanest, element-level traces in the whole corpus — yet the phantom-flag leg still fails GATE 1, for a reason unique to yttrium: the named transit hub is fed mostly by genuine ex-China separators, so a real relabel would be indistinguishable from legitimate diversification.

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 — yttrium (Y) plus Sm, Gd, Tb, Dy, Lu, Sc — under individual, non-automatic export licence covering metals, oxides, alloys and compounds. Licences are granted at MOFCOM's discretion citing end-user and end-use; approvals through 2025 skewed heavily away from US-bound shipments. The Oct-2025 extraterritorial package (Ann. 61/62) and the Nov-2025 US suspension (Ann. 70, to 10 Nov 2026) followed; the April licensing regime itself was never suspended.

2. The dependency is near-total — and it is aerospace + semiconductor, not SmCo defense. The US is 100% net-import-reliant for yttrium, and ~93% of US yttrium-compound imports in 2020–2023 came from China (USGS Mineral Commodity Summaries). China holds ~85% of world rare-earth processing and historically ~99% of heavy/medium-REE separation — yttrium is separated from the same southern-China ion-adsorption clay circuits as the true heavies. Yttrium's value concentrates in yttria-stabilized zirconia (YSZ) — the thermal-barrier coating on jet-engine and gas-turbine blades — plus plasma-facing coatings / chamber components in semiconductor fab equipment, YAG lasers, and phosphors. This is a different blind spot from samarium's SmCo missile magnets: the exposure sits in aerospace coatings and chip-tool consumables, and at least two North-American aerospace-coating suppliers reportedly curtailed output or moved to allocation (prioritising Boeing/Airbus) after the control.

3. The control bites hard — the price wedge screams it. Yttrium oxide ran from under ~$8/kg at end-2024 to ~$126/kg by Nov 2025 (~1,475%), with European spot/short-term quotes reaching a few hundred €/kg (~$270/kg, reported as ~4,400% off the early-2025 base) while Chinese domestic prices stayed at single-digit dollars — an order-of-magnitude ex-China/in-China wedge. Origin-laundering only pays when the material is genuinely scarce ex-China; this wedge is the scarcity signal that makes a leak worth looking for.

The direct-collapse leg — element-level, and it closes

Unlike samarium (basket-blind), yttrium's banned-origin collapse is visible at the element level in Chinese customs statistics:

Source: Chinese customs (GACC) statistics as reported by Reuters and Asia Financial (summarised at materialsdispatch.com, 2026). China's yttrium exports to the rest of the world fell ~30% over the same window, and yttrium/dysprosium/ terbium ran ~50% below their pre-restriction baseline — but ex-China buyers are reluctant to redirect scarce yttrium to the US for fear of Chinese retaliation, a "de-facto embargo layer" on top of the licence.

So the controlled origin's direct channel to the US is gone — a textbook banned-origin collapse. The question is where the material went.

The transit leg — Thailand is the hub, but the phantom-flag test is confounded

The same transit hub that carried the antimony relabel — Thailand — is the node where yttrium's onward flow is visible. Thailand has no domestic rare-earth mine and no chemical-separation plant; its only relevant plant, Neo Magnequench (Korat), is downstream magnet-powder processing, not primary separation. By the volume-implausibility rule, a country with ~zero yttrium separation capacity re-exporting hundreds of tonnes is exactly the phantom-flag signature. And Thailand does re-export at scale:

Source: Thai customs data for HS 28053000, compiled by Climate Connectors (2025). Free national-customs / UN-Comtrade-class dataset.

The structure is real — import from many origins, re-export overwhelmingly to Japan — but two corrections stop it short of a quantified GATE-1 relabel, and both are the honest part of the story:

• The dominant inflows are GENUINE ex-China separators, not phantom flags.

Malaysia = Lynas (LAMP, Kuantan) — the only heavy-REE separation line outside China, which began separated heavy-REE production (Dy oxide) on 15 May 2025 with yttrium among its target elements. Estonia = Neo Performance Materials' Silmet (Sillamäe) — a long-standing Western rare-earth separator. Both are legitimate mode-B supply; material routed Lynas→Thailand→Neo Magnequench→Japan or Silmet→Thailand is a Western/Australian processing web, the opposite of a Chinese shell. A yttrium relabel of Chinese-origin material would be statistically camouflaged inside these genuine ex-China flows — the co-present-genuine-supply confounder, the inverse of samarium's basket problem.

• HS 28053000 mixes scandium + yttrium + other RE metals. It is narrower than

the samarium Eurostat basket (it excludes compounds/oxides, the main yttrium trade form), but it still cannot isolate yttrium metal from scandium or from other rare-earth metals — so a yttrium-specific surge cannot be cleanly sized.

The one suggestive Chinese leg: China's direct shipments into the Thai hub did not collapse the way its US channel did — Chinese-origin material still entered Thailand (48.5 t in 2024; 62.5 t by the Aug-2025 reading) even as China→US fell ~95%. Material China would not license to the US directly still flowed into a processing hub whose output goes to Japan (and onward). Held as inference: the period bases differ (full-year 2024 vs an Aug-2025 reading), so this is a direction, not a sized YoY rise.

The US backfill — detector-NEGATIVE (and that is itself a finding)

For the US buyer specifically, there is no traced third-country surge backfilling yttrium imports — unlike antimony, where Thailand (+110%) and Mexico (+539%) visibly replaced the banned Chinese oxide. US yttrium imports simply went dark (333→17 t). Two readings, both alternative-track:

1. The control is genuinely holding for the US — the "de-facto embargo layer" (ex-China holders won't redirect to the US) plus a truly monopolised element mean no one can backfill at scale in the window. The price spike is consistent with real, unrelieved scarcity. 2. Or the relabel re-enters below the raw-material HS lines — as finished YSZ powder, pre-coated turbine parts, or semiconductor chamber components under machinery/parts codes, exactly the finished-good disguise flagged in the [magnet case](2025-rare-earth-magnets-licensing-circumvention-transshipment.md). A yttrium coating on an imported blade is invisible to HS 2805/2846.

Both keep this GATE 0: the US-side phantom-flag table cannot be built, and the Thai-hub table cannot attribute origin.

The transmission chain

`` China (~99% heavy-REE separation; yttrium from southern ion-adsorption clay) → [Apr-2025 licence; approvals skewed off US-bound shipments] → China→US direct collapses 333→17 t (−95%): the leg the customs code CAN see → onward flow visible only at a processing hub: China (still 62.5 t) + Lynas/Malaysia + Neo-Silmet/Estonia + Vietnam → Thailand (0% domestic Y capacity; Neo Magnequench downstream) → Japan (519 t) + onward — genuine ex-China supply and any Chinese relabel are CO-MINGLED here → US buyer: no traced backfill (embargo-holds OR finished-good disguise) → origin label ≠ economic origin wherever Chinese material rides the hub ``

Common-ownership tell

N — not traced; and the hub inverts the usual pattern. The antimony case had a Chinese-namesake shell chain (Youngsun → Thai Unipet → Youngsun & Essen). Here the Thai hub's processor (Neo Magnequench) and a top inflow origin (Neo Silmet, Estonia) belong to Neo Performance Materials — a Western (Canada-listed) processor, not a Chinese affiliate. No Chinese common-ownership pipe is traceable to a surging yttrium exporter; if anything the hub is a legitimate ex-China processing web. Left as an open question rather than a fabricated hop.

Why it matters for the buyer

1. A visible direct collapse can still hide the real exposure. A risk team can see China→US yttrium fall 95% — cleaner than for any other element under this control — and conclude the dependency is being cut. But 100% import reliance plus a monopolised separation step means the exposure did not diversify; it went latent (scarcity + possible finished-good re-entry). The clean number is the trap. 2. Genuine diversification and circumvention look identical at the hub. Lynas and Neo/Silmet material and any relabelled Chinese material flow through the same Thai node to the same buyer (Japan). Only element-level, capacity-aware, ownership-aware reading separates them — which is exactly why this is an alternative-track signal beside, never inside, the Tier-1 score. 3. The wedge confirms the bind. Yttrium oxide up ~1,475% (Europe far more) with Chinese domestic prices flat is independent confirmation the control bites in the same direction the price wedge prices scarcity.

Caveats

• Inference, not proof. No figure here isolates yttrium at the transit hub

(HS 28053000 bundles scandium + RE metals); the China→US 333→17 t collapse is element-level but is the departure leg, not a relabel. "Consistent with," "fingerprint of," never "proves smuggling."

• Genuine alternatives explicitly labelled. Malaysia/Lynas and Estonia/Neo-Silmet

are real ex-China separation — do not read their flow through Thailand as circumvention. That co-presence is precisely what defeats the phantom-flag test.

• Period-basis mismatch. The China→Thailand "still live" read compares a

full-year 2024 figure with an Aug-2025 reading; treat as directional only.

• US-side detector-negative. No third-country surge into the US is traced;

either the control holds for the US or the relabel rides finished-good HS lines. Element-level compound/oxide (HS 2846) data by origin, or coated-part provenance, 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 and can mute the very relabel we watch for in the window it covers.

Anchor action: `2025-04-04-china-mofcom-heavy-rare-earths-export-licensing`. Sources — control: MOFCOM/GACC Announcement No. 18 (4 Apr 2025), Ann. 61/62 (9 Oct 2025) & No. 70 (7 Nov 2025), via Pillsbury, Andersen Institute, CSIS, Global Trade Alert. Dependency/monopoly: USGS Mineral Commodity Summaries (yttrium: 100% net-import-reliant, ~93% from China 2020–2023); CSIS / Rare Earth Exchanges (separation share; aerospace/semiconductor end-use). Direct collapse: Chinese customs (GACC) as reported by Reuters / Asia Financial, summarised at materialsdispatch.com (333→17 t). Prices: Discovery Alert, NAI 500, Yahoo Finance (~1,475% to ~$126/kg; European ~$270/kg). Transit hub: Thai customs HS 28053000 compiled by Climate Connectors (Malaysia/Vietnam/China/Estonia inflows; Japan re-export; no domestic separation; Neo Magnequench downstream-only). Ex-China separators: Lynas Rare Earths / Bloomberg (LAMP, Dy oxide 15 May 2025, yttrium a target element); Neo Performance Materials (Silmet, Estonia). Presented as the publicly-reported mechanism — NOT a per-shipment accusation beyond what the public record states.