Did the chokepoint leak, or move? Semiconductor-grade neon after 2022

Trade-flow companion to the price wedge (R72). Every other row in the circumvention corpus asks "is the banned molecule re-entering the buyer under a laundered flag?" Neon asks the inverted question, and the answer is the reason it belongs here: the control did not leak through a phantom origin — the chokepoint physically relocated, and it relocated into the primary strategic rival. This is a DUAL-SCORE / alternative-track signal, never folded into any Tier-1 exposure score. The divergence that IS the signal: a fab that "diversified away from Russian/Ukrainian neon" to a Chinese supplier did not de-risk — it deepened its adversary exposure behind a genuine, non-Chinese-looking supply story. Research, not investment advice; the Russian-feedstock leg below is INFERRED from capacity structure and reported diversion, never asserted as a traced shipment.

Verdict

Neon is not mined. It is a trace by-product of the air-separation units (ASUs) bolted onto large steelworks to make oxygen; the crude gas is then purified to 99.999% "lithography grade" by a handful of specialist refiners, and that purified gas feeds the deep-UV excimer lasers of chip lithography (semiconductors consume up to ~90% of neon demand). For decades the world's purification chokepoint sat in Ukraine — two refiners, Ingas (Mariupol) and Cryoin (Odesa), upgrading crude neon into ~45–54% of the world's semiconductor-grade supply. When Russia widened its invasion in February 2022, both plants halted, removing roughly half of global lithography-grade neon overnight, and China's neon price spiked ~10× in March 2022. Then, on ~30 May 2022, Russia — the dominant crude feedstock source — made noble-gas exports (neon, argon, krypton, xenon) subject to special government clearance through end-2022, claiming ~30% of the world market. Two hard hits, weeks apart, on the same chokepoint.

Two years on there is no phantom-flag surge to catch — and that absence is the finding:

Sources: Ukraine share + Ingas/Cryoin volumes and shutdown — Reuters (via CNBC/AOL), Mar-2022; USITC Executive Briefing "Ukraine, Neon, and Semiconductors." Russia noble-gas export decree (30-May-2022, clearance to 31-Dec-2022; ~30% claim) — Reuters/TASS/Global Times, Jun-2022. Crude rare-gas shares (RU ~15%, UA ~10%) — EU JRC130349 "Rare Gases" impact assessment + industry (SFA Oxford). China ASU/steel dominance, ~14 M m³/yr, foreign-major share — TechNode (Mar-2022), CSIS, gasworld, industry commentary. Korea POSCO/TEMC plants — POSCO Newsroom / gasworld / KED Global (2023, 2026). Prices — China industry press via CNBC/Yahoo Finance / market trackers. No single customs dataset underlies this table — it is assembled from public press + the EU JRC and USITC briefs; the HS 280429 (rare gases) Comtrade cross-check is the GATE-1 gap (see Method).

The classic detector is rejected by construction: no ~0%-capacity country posts an implausible neon surge, because the substitute (China) holds genuine, massive capacity — >40% of the world's ASU crude-neon potential and the steel base to feed it. Mode-A (relabelling) does not apply. What happened instead is closest to the nickel/bauxite mode-B result — a genuine origin masks a deepened dependency — but with a war-severance rather than a resource-nationalism ban as the trigger, and with the concentration landing in China not by industrial policy alone but by who already owned the feedstock when the old node burned down.

The transmission chain (and the inferred wash)

`` Ukraine purification node destroyed (Feb–Mar 2022) + Russia crude-neon export curb (30-May-2022) → world scrambles for lithography-grade neon (price ~10×) → China: already >40% of world ASU crude neon, bundles new purification trains onto existing steel/petrochemical ASUs → China becomes the dominant commercial supplier → China re-exports purified neon to the world's fabs ``

Two legs of this chain are the circumvention question, and both are inferred, not traced:

1. The Russian-feedstock wash (mode-C hypothesis). Russia restricted noble-gas exports in May-2022 — but reporting through 2023 describes Russia diverting rare-gas exports to China, which then ran a surplus and re-exported. If Chinese lithography-grade neon sold onward to Western/Asian fabs is purified from Russian crude neon, then Russia's own export curb is softened by a Chinese purification step (a substantial-transformation origin-wash, the Russian-crude→Indian-diesel / Russian-rough→Indian-diamond pattern on the semiconductor axis) and re-routed to buyers Russia could not sell to directly. The honest caveat is decisive: China has its own enormous ASU fleet and plausibly self-supplies most of its crude neon, so the Russian fraction of any given Chinese export is unquantifiable from the public record — a labelled gap, exactly the uranium-SWU fungibility problem (molecules mix; the customs line cannot resolve provenance).

2. Buyer-diversion to China (the concentration leg). Independent of any wash, the world's lithography-grade neon supply re-concentrated from a Russia-feedstock / Ukraine-purification split into a single Chinese stack. Reliance simply migrated from one geopolitical risk to a larger one — the palladium "dependency-migrates-to-China" leg, here as the primary outcome rather than a side channel.

Against that, the genuine, non-adversary diversification is real and must not be scored as circumvention: Korea (POSCO Air Solution's Gwangyang rare-gas plant on its own steel ASUs; TEMC + SK hynix neon recycling), Taiwan (TSMC's post-2022 push to build a domestic neon supply chain), and Western majors (Linde, Messer, Air Liquide's 2023 China and other neon trains) rebuilt purification onshore or in allied jurisdictions. This is the dual-track split the signal exists to draw: the same shock produced both a healthy ally-onshoring response and a rival-concentration response, and only the customs label distinguishes them.

The ownership tell

N — no namesake shell. Unlike antimony's Youngsun → Thai Unipet → Youngsun & Essen pipe, there is no traced common-ownership laundering entity here. The Russia→China leg is a state-commercial diversion / fungibility integration (uranium-class), not a shell. The one disclosed cross-border ownership crossover — a POSCO Air Solution × Hangzhou Zhongtai rare-gas JV in Korea — sits on the ally side (building Korean capacity), and is a disclosed JV, not a pipe. So the strongest single tell in the corpus (a controlled producer routing through a namesake abroad) is absent, which is itself consistent with a concentration-migration story rather than a relabelling one.

Why it matters for the buyer

1. "De-risking away from Russia/Ukraine neon" can increase adversary exposure. A risk team that reads "we no longer depend on Ukrainian or Russian neon" has answered the wrong question if the new supplier is Chinese. The genuine Chinese origin is not the reassurance it looks like — it is a larger single-country dependency on the strategic rival, wearing a clean label. 2. The chokepoint is upstream of the fab, in a place no chip audit looks. Neon is a consumable, not a component; it rarely appears on a bill of materials. Yet it gates deep-UV lithography, and its supply is welded to steel ASU capacity — which is exactly where China is structurally dominant. Concentration in a hidden consumable is still concentration. 3. A future Chinese neon export curb would fire on a supply base it now owns. The 2022 episode showed a ~10× price move from losing one node. If China — now the dominant commercial supplier — ever restricted lithography-grade neon (as it has for antimony, gallium, germanium, graphite), there is no Ukraine to fall back on; the fallback is the ally-onshoring already underway (Korea, Taiwan, Western majors), whose pace is the real forward signal.

Method & honesty rails

• This is GATE 0. Three honest reasons: (a) no single customs table —

the shares are assembled from public press, the EU JRC rare-gas assessment, and the USITC brief, not one dataset; the HS 280429 (rare gases nes) Comtrade cross-check — China exports up, Ukraine to ~0, Russia→China diversion — is the work that would lift this toward GATE 1. (b) The mode-C Russian-feedstock wash is a hypothesis, not a traced flow (crude-neon provenance is fungible and unresolvable in customs data). (c) The dominant real story is genuine capacity migration, not relabelling, so the volume-implausibility detector is structurally N/A here.

• Control-anchor gap. The underlying controls — the Feb-2022 war severance of

Ingas/Cryoin and Russia's 30-May-2022 noble-gas export decree — are not yet filed IPTM actions; this case is anchored to the nearest filed touchpoint (JBIC's Dec-2025 loan backing Nippon Sanso/Coregas industrial-gas supply-chain resilience). Discovery lead: file the Russia noble-gas export decree (May-2022) and, if traceable, the Ukraine-plant severance as register actions.

• Alternative-track only: never touches buyerRelativeScore or the base

exposure — it sits beside them, like the China–West price wedge.

• Inference, not accusation: the Russian-feedstock and China-re-export legs

are inferred from capacity structure and reported diversion; no individual shipment is asserted to be Russian-origin beyond what the public record states.

• New signature for the detector: *a control (or war-severance) that destroys

a mid-chain processing node does not necessarily leak through a phantom flag — it can re-concentrate legitimately in whoever already owns the upstream feedstock, so the honest read is the capacity map (who holds the ASUs / the purification trains), not a customs origin table. When the feedstock owner is the strategic rival, "diversification" and "deepened dependency" are the same event under two different labels.*