Is the scandium control leaking? The byproduct monopoly with no customs code of its own

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, triple-blind, byproduct-ubiquity defeats the test)

Scandium is the same 4-April-2025 licensing family as [samarium](2025-samarium-licensing-circumvention-transshipment.md) and [yttrium](2025-yttrium-licensing-circumvention-transshipment.md) — but it is the member the volume-implausibility detector can police least of the three. Four facts sit on top of each other:

1. The control names it directly, alloy included. On 4 April 2025 China's MOFCOM + GACC (Announcement No. 18) placed seven medium/heavy rare earths — Sm, Gd, Tb, Dy, Lu, Sc, Y — under individual, non-automatic export licence, classifying them dual-use. The scandium line explicitly covers metallic scandium, scandium oxide, scandium compounds *and scandium-aluminium (Al-Sc) master alloys — the control drafters reached past the oxide to the alloy, because the alloy is how scandium actually moves. The 9 October 2025 Announcements 61/62 asserted extraterritorial reach (0.1% de-minimis, 50%-affiliate, foreign-direct-product); 7 November 2025 Announcement No. 70 suspended the October package for the US to 10 Nov 2026 (Busan truce). The April licensing regime itself was never suspended; as of December 2025 it remained in force, with China issuing general* licences only to selected exporters.

2. The monopoly is near-total, and defence-critical. China controls ~85% of refined chemical scandium and ~100% of metallized scandium exports (Discovery Alert / Crux Investor, 2025). The GAO confirmed zero US scandium suppliers in 2025, extending aerospace lead times up to six months; the US has not mined scandium since 1969 (Mining Engineering / NioCorp). The global market is tiny — USGS puts Sc₂O₃ consumption at ~30–60 t/yr against ~80 t/yr capacity — which makes even a few tonnes of laundered flow market-moving. Scandium's value concentrates in Al-Sc master alloys (lightweight, weldable aerospace/defence structures — armour, missile components, airframes) and solid-oxide fuel-cell electrolytes; a scaled ex-China chain is years out (see mode-B builds below).

3. Scandium is 100% a *byproduct* — so it has no clean "zero-capacity flag." There is no primary scandium mine anywhere on earth. It is recovered as a trace byproduct of unrelated operations: China (Bayan Obo REE/iron), Russia (Kola Peninsula / Solikamsk), Ukraine (Zhovti Vody uranium & iron), the Philippines (nickel-laterite HPAL), Kazakhstan (uranium). This quietly breaks the premise the antimony template rests on. There, a Thailand surge is damning because Thailand mines ~0% of world antimony. For scandium, almost any large nickel-laterite, uranium, titanium-pigment or bauxite-residue operation could in principle recover trace scandium — so a surge from an off-list flag is harder to call "implausible." The zero-capacity denominator that powers layer 1 is itself blurred.

4. The customs taxonomy gives scandium no code of its own — it is *triple*-blind. This is the sharpest reason for GATE 0, and it is worse than samarium's two-way blindness: - Metal: HS/HTS 2805.30 is literally "Rare-earth metals, scandium and yttrium, whether or not intermixed or interalloyed" — scandium shares one basket line with yttrium and the lanthanide metals. - Oxide/compounds: HS 2846 is "Compounds… of rare-earth metals, of yttrium or of scandium, or of mixtures" — scandium oxide is basketed with yttrium and all lanthanide compounds. - The alloy escapes rare earths entirely. Al-Sc master alloy is aluminium by weight and moves under aluminium HS lines (72/76-series), wholly outside the rare-earth chapters — a legal code-hop the April control text anticipated but the trade taxonomy does not reflect.

So the one material whose phantom-flag test is cleanest in principle (essentially no legitimate ex-China refined supply) is the one the customs system resolves least — across three layers at once.

Why the table can't close

For antimony we could print China-collapses / Thailand-surges to the tonne because HS 282580 is antimony-specific. For scandium there is no equivalent row to print. A scandium-specific relabel would appear as, at most, a wobble inside the HS 2805.30 / 2846 rare-earth aggregate — indistinguishable from any yttrium or lanthanide movement — or it would not appear in the rare-earth chapters at all, having left as Al-Sc master alloy under an aluminium code. Publishing a China-vs-transit tonnage table here would be fabricating resolution the data does not have, so we do not. The honest artifact is the absence of one.

The likely disguises (what a scandium leak would wear)

`` China (~85% refined / ~100% metallized scandium; Bayan Obo byproduct) → [Apr-2025 dual-use licence on metal + oxide + Al-Sc alloy; never suspended] → leak paths, none of which the customs code can photograph: · (A) trans-shipment inside the HS 2805.30 / 2846 rare-earth basket — a scandium surge is invisible in a mixed-REE aggregate · (A′) byproduct-origin ambiguity — a Philippines/Kazakhstan/Ukraine flag is *plausible* (they host Ni-laterite / U byproduct), so the "zero-capacity" alarm never trips · (C) transformation-wash / code-hop — exported as Al-Sc master alloy under aluminium HS lines, outside the rare-earth chapters entirely → buyer (US/EU aerospace & defence), stated origin ≠ economic origin ``

Scandium's leak, if it is flowing, most likely wears the aluminium disguise: the highest-value form (Al-Sc master alloy) is also the form that leaves the rare-earth customs universe. That is the corpus's cleanest example of a control-anticipated code-hop the trade-reporting system has not caught up to.

Genuine ex-China builds — label them as supply, not laundering

Two real mode-B (capacity-relocation) projects are standing up ex-China scandium and must not be mistaken for relabelling:

• NioCorp / Elk Creek Resources (Nebraska). US DoD awarded **up to $10 M

under the Defense Production Act (Title III) to advance domestic scandium; NioCorp acquired Al-Sc master-alloy assets to aim at *America's first vertically-integrated scandium mine-to-market* (mining → Sc₂O₃ → Al-Sc alloy). Not yet in production** — contingent on final project financing.

• Rio Tinto / Canada Growth Fund, Sorel-Tracy (Québec). ~C$25 M to expand

scandium-oxide output at the titanium-dioxide plant to ~9 t/yr nameplate — a genuine byproduct-recovery build, Western-owned.

These are the answer to the control, not evidence of a leak. As with Lynas in the samarium/yttrium cases, a rising "non-China" scandium share driven by NioCorp or Rio Tinto is real diversification; scoring it as circumvention would be exactly the error this alternative-track is built to avoid.

Common-ownership tell

N — not traced to a named entity. Chinese producers are scaling (e.g. Hunan Oriental Scandium Co. publicly guiding oxide-capacity growth), but the public record I can cite does not name an offshore shell, namesake, or address linking a surging transit-country scandium exporter to a controlled Chinese producer — unlike the antimony Guangxi 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 rare-earth import stats will never see "scandium" resolve as a line — and if it leaks as Al-Sc master alloy it leaves the rare-earth chapters altogether. Any exposure model keyed on stated rare-earth origin is blind to scandium in both directions at once. 2. The control is binding — the evidence is the vacuum, not a surge. With GAO confirming zero US suppliers, +6-month aerospace lead times, and a US DPA response, the April licence is demonstrably biting; origin-laundering only pays when a material is genuinely scarce ex-China, the same scarcity the price wedge prices. 3. Byproduct-ubiquity is a permanent blind spot, not a data-lag one. Because scandium has no primary mine, the "implausible origin" test will never be as crisp as it is for antimony — a Philippines or Kazakhstan flag is always at least arguable. This is structural: element-level and alloy-level customs granularity would be required to move scandium toward GATE 1, and neither exists today.

Caveats

• Inference, not proof. No figure here isolates scandium; no clean

scandium-specific customs series exists to isolate. "Consistent with" and "fingerprint of," never "proves smuggling."

• Genuine alternatives explicitly labelled. NioCorp/Elk Creek and Rio

Tinto/Sorel-Tracy are real ex-China builds (mode B), not phantom flags; do not read their output as circumvention.

• Byproduct premise. The volume-implausibility test's zero-capacity

denominator is blurred for scandium; an off-list origin can be plausibly argued, so the layer-1 detector is structurally weak here regardless of data quality.

• Suspension noise. The Oct-package suspension (to Nov-2026) does not touch

the April licensing regime that controls scandium, but the broader US-China truce can mute observable relabelling 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) incl. scandium metal/oxide/compounds/Al-Sc alloys as dual-use, per Holland & Knight and Discovery Alert ("China's Scandium Export Controls"); Announcements 61/62 (9 Oct 2025) & No. 70 (7 Nov 2025). Monopoly/defence: Discovery Alert (~85% refined / 100% metallized), Crux Investor ("Filling the Scandium Vacuum"), US GAO (zero US suppliers 2025, +6-mo lead times). Byproduct sources & market size: USGS Mineral Commodity Summaries 2025/2026 (Scandium); byproduct sites (Bayan Obo / Kola / Zhovti Vody / Philippines Ni-laterite). Trade taxonomy: HTS/HS 2805.30 (rare-earth metals, scandium and yttrium) and HS 2846 (compounds of yttrium or scandium), via Datamyne/Flexport/UNSD; Al-Sc master alloy under aluminium HS lines. Ex-China builds: NioCorp/Elk Creek DoD DPA (Mining Engineering, NioCorp release), Rio Tinto/Canada Growth Fund Sorel-Tracy (Mining.com). Presented as the publicly-reported mechanism — NOT a per-shipment accusation beyond what the public record states.