Investigation — Helium supply chain for semiconductor fabs

Investigation Ticker Tape
6 min read

Investigation — Helium supply chain for semiconductor fabs

8 tickers referenced · published Jul 1, 2026

Article published Jul 1, 2026. Prices below use latest available snapshots.

LIN $474.33 -7.6% 30d APD $300.86 +1.8% 30d ENTG $163.09 +17.6% 30d MKSI $320.74 -1.2% 30d ASML $1,883.12 +7.9% 30d XOM $161.46 +9.6% 30d TSM $430.97 +8.2% 30d
no live data (1) — unresolved, delisted, or non-US symbols
$GTLS

Question: Helium supply chain for semiconductor fabs Verdict: fold-into-gulf-infrastructure-strike

Chain map — physics first

Price truth: all price / RSI / trend figures below are read from the per-watchlist scan summaries at summaries (chemicals, wfe-test-metrology, semis), as-of 2026-06-26 close (~3 trading days stale). Code computes the math; nothing is derived in prose.

Helium is not manufactured — it is a byproduct of natural gas with high (>0.3%) native He concentration. That physics fact governs the whole chain: you cannot build a helium plant where the gas isn't, so supply is geologically captive to a handful of fields, and every "producer" is really a gas/LNG operator stripping He as a side stream. Data below as-of 2026-06-26 close (~3 trading days stale; fine).

Layer 1 — Source rock / raw high-He gas (the real chokepoint)

  • Qatar North Field → Ras Laffan — QatarEnergy / RasGas [foreign, state-owned, not investable]. ~30% of Qatar's He ≈ ~11% of global (perspective already corrected the earlier "30% of global" overstatement). Ras Laffan LNG trains carry equity from ExxonMobil (XOM) [diversified — no torque], Shell, TotalEnergies, ConocoPhillips — so "exposure" exists but helium is a rounding error on any of them.
  • US — ExxonMobil Shute Creek / LaBarge (Wyoming) = the single largest US helium source. XOM [gap-but-no-torque]. The old BLM Cliffside Federal Helium Reserve (Amarillo) was privatized/auctioned — no longer a government buffer; a swing-supply cushion that is now gone.
  • Russia — Gazprom Amur GPP [foreign, sanctioned] — gray-market only, structurally offline for Western fabs.
  • Algeria — Sonatrach (Arzew/Skikda) [foreign, state] — Air Liquide/Linde hold the offtake.
  • Tanzania / East Africa Rift — Helium One (HE1.L) [foreign, micro-cap, pre-production]; South Africa Renergen (Virginia Gas Project) [foreign; JSE/ASX primary, RENGF OTC] — the highest-concentration "pure" helium field, but pre-meaningful-production.

Layer 2 — Extraction / purification / liquefaction

  • Majors: LIN [tracked], APD [tracked], Air Liquide (AIQUY) [foreign ADR].
  • Cryogenic liquefier equipment (the capacity bottleneck for any new source): Chart Industries (GTLS) [tracked — trace-lng]; Linde Engineering / Cryostar [inside LIN]; Atlas Copco [foreign].

Layer 3 — Distribution / bulk molecule (the oligopoly)

  • LIN [tracked], APD [tracked], Air Liquide (AIQUY) [foreign], Nippon Sanso / Matheson (TYO:4091) [foreign], Messer [private]. This is a genuine 3–4 firm global oligopoly — helium is priced and allocated here.

Layer 4 — Fab-gate handling (helium-adjacent, NOT helium-specific)

  • Gas purification/filtration: Entegris (ENTG) [tracked — wfe]. Mass-flow control / gas analytics: MKS Instruments (MKSI) [tracked — the sole (mis-tagged) member of trace-helium]. Also UCTT, Ichor [gaps]. These benefit from all fab gas throughput; they are semicap-consumable plays, not a helium-scarcity bet.

Layer 5 — Fab demand (helium intensity rises down the node)

  • TSMC (TSM) [tracked] — sub-7nm / N3 / N2 most He-intensive (chamber cooling, purge, leak-detection). Samsung (005930.KS) [foreign], SK Hynix (000660.KS) [foreign] — Korea carried ~64% Qatar dependency. Intel (INTC) [gap], Micron (MU) [tracked elsewhere]. ASML [tracked] EUV/High-NA uses He for purge/cooling — High-NA reportedly ~doubles He per layer — but ASML is a lithography monopoly bet, not a helium bet.

Layer 6 — Recycling / recovery (the real forward bottleneck, no pure-play)

  • On-site He recovery is a division inside LIN / APD / Air Liquide Advanced Materials (Taiwan), plus ENTG purifiers. Standalone recovery-system vendors (Quantum Design, Cryomech) are [private]. There is no listed recycling pure-play — this is an M&A/embedded-margin story, not an entry.

Investable picks — pure-play vs diversified vs no-clean-play

Hard finding: there is NO clean, US-listed helium pure-play that survives the trend-hold mandate. Every "pure-play" is a foreign micro-cap explorer (Renergen, Helium One, Desert Mountain, Royal Helium, First Helium) — pre-revenue or pre-meaningful-production, TSX-V/AIM/ASX/OTC, exactly the falling-knife / lottery-ticket profile the mandate rules out. Helium exposure for a trend-hold book therefore lives inside diversified industrial-gas compounders, where He is a small but high-margin, pricing-power-rich slice.

Name Layer Class Tape (as-of 06-26) Trend-hold read
LIN distribution/majors diversified oligopoly $519.62, RSI 56.7, strong-up, +11.5% vs SMA200, −1.6% from 52wk high; only chemicals name w/ positive 5y alpha (+12.2) The quality compounder. Cleanest trend-hold vehicle for helium and the broader industrial-gas oligopoly. Helium is diluted torque, but you're buying the best balance sheet in the space in a real uptrend. Verify current technical state before entry.
APD US helium (Doe Canyon moat) diversified, highest He torque $277.79, RSI 41, weak-down / regime "basing", +2.2% vs SMA200, −9.8% from 52wk high; negative 1/3/5y alpha Highest helium leverage, worst tape. Doe Canyon (only non-Gulf semi-grade He source) is the moat, but the tape has refused to price the narrative reversal for 4+ months. This is a turnaround/base, NOT a secular uptrend — against the mandate as a chase. Only interesting on a confirmed base→breakout; do not catch the knife.
GTLS liquefaction equipment diversified cryo/LNG $208.50, RSI 58, strong-up, +0.4% 30d Cleanest "shovels" exposure to new-helium-supply buildout (liquefiers are the capacity gate). But it's a broad cryo/LNG cap-equipment name; helium is a minor end-market.
ENTG / MKSI / ASML fab-gate / litho semicap, helium-adjacent ENTG $161.43 RSI 54.4 strong-up; MKSI $388.61 RSI 59.7 strong-up (+74% 3m); ASML $1,794.62 RSI 53.8 strong-up Strong tapes, but these are WFE/semicap bets that rise on all fab activity. Tagging them as "helium plays" is a category error (see Reconciliation). Already tracked in the right buckets.
AIQUY / Nippon Sanso / Messer distribution foreign / private n/a The rest of the oligopoly. AIQUY is the only way to add helium concentration beyond LIN/APD, but it's a foreign ADR → GATED.

Bottom line: LIN is the trend-hold helium proxy; APD is the high-torque-but-broken option trade; everything else is either foreign, private, semicap-mislabeled, or a speculative explorer.

Reconciliation

vs the trace-helium-supply-chain-for-semico watchlist ([MKSI] only): the prior trace was wrong at the pick level. It correctly identified LIN/APD/ENTG but dropped them as "already tracked," then crowned MKSI as the sole 'NEW' member — and MKSI is a mass-flow-controller / semicap-instrument maker whose revenue keys off total fab gas throughput, not helium scarcity. A helium supply-disruption thesis does not route through MKSI; it routes through the Layer-1 source concentration (Qatar) and the Layer-3 oligopoly (LIN/APD/AIQUY). The trace's own Coverage-Gaps section actually got it right ("no pure-play helium producer is publicly traded in the US… alpha requires tracking He as % of revenue for LIN/APD") — but that correct conclusion never made it into the watchlist, which kept the wrong single name. Correction: the [MKSI] trace-helium watchlist is a mis-tag and should be retired/redirected.

vs gulf-infrastructure-strike (monitoring, priority low): the perspective already owns the macro chain — Ras Laffan chokepoint, the ~11%-of-global magnitude correction, the Doe Canyon moat (via the Revan APD conviction doc, verdict FIZZLED +10.7%), the SK-fab 64%-dependency, and the late-Aug/early-Sep depletion binary (~55-60 days out). Its key_tickers already carry LIN + APD. What this trace adds that the perspective lacks: the full named physical chain by layer, the explicit producer roster (XOM Shute Creek, Sonatrach, Amur, the privatized-away BLM reserve), the liquefaction-equipment layer (GTLS), and the hard verdict that no clean investable pure-play exists. Nothing here contradicts the perspective; it fills the "we haven't traced WHO supplies WHICH fabs" gap the brief named.

Genuinely new (highest-value): the coverage-gap conclusion is the deliverable — the chain is already adequately covered by LIN/APD/ENTG/GTLS/ASML sitting in their correct watchlists, and the "missing pure-play" is missing because it doesn't exist as an investable trend-hold name, not because we failed to find it.

Verdict + reasoning

fold-into-gulf-infrastructure-strike (+ correct the mis-tagged trace watchlist).

Reasoning: (1) No clean US-listed pure-play exists → nothing new to hold, so this does not earn a standalone perspective. (2) The macro thesis and the two real investable names (LIN, APD) are already inside gulf-infrastructure-strike, which is the natural home for the chain map. (3) The one concrete engine action is corrective, not additive: the trace-helium watchlist's sole member [MKSI] is a wrong tag. (4) The perspective is already monitoring/low waiting on the depletion binary — the right posture; this trace doesn't change the trade, it documents the chain so the next agent doesn't re-derive it.

Proposed mutations / follow-ups

Clean US-listed watchlist adds: none. Every clean US name in the chain (LIN, APD, ENTG, MKSI, ASML, GTLS, XOM) is already tracked in an appropriate watchlist. Adding more would be coverage theater.

Corrective mutation (recommended):

  • Retire or redirect the trace-helium-supply-chain-for-semico watchlist — its sole member [MKSI] is a helium-thesis mis-tag. Either delete it or replace its contents with the actual chain core LIN APD GTLS and re-describe as "helium chain (industrial-gas oligopoly + liquefaction), no pure-play exists." (Gated: watchlist deletion/rewrite is a user call.)

Perspective edit (recommended):

  • Append this layer-by-layer chain map + the "no investable pure-play" verdict to gulf-infrastructure-strike/README.md (or its log) so the WHO-supplies-WHICH-fab question the brief flagged is answered in-place. No key_tickers change needed (LIN/APD already present).

GATED coverage decisions (foreign/OTC — need user sign-off; all fail trend-hold as-is, flagged for completeness only):

  • AIQUY (Air Liquide ADR) — only way to add helium concentration beyond LIN/APD; foreign ADR.
  • Speculative explorers — RENGF/Renergen (JSE/ASX), HE1.L/Helium One (AIM), DMEHF/Desert Mountain, RHCCF/Royal Helium, HELIF/First Helium — pre-revenue micro-caps; recommend NOT adding (violates no-lottery-ticket mandate). Breadcrumb: considered, dropped — pre-production He explorers are exactly the falling-knife profile the mandate excludes.

Deep-dive follow-ups:

  • If APD tape ever bases→breaks out, verify current technical state before entry — it's the only name with real helium torque, but only after the tape confirms.
  • ASP Isotopes (ASPI, NASDAQ) — reportedly linked to Renergen; verify whether it is now a US-listed proxy for the highest-concentration helium field before treating as anything but speculative.

Sources

Price, RSI and trend figures read from the desk's validated daily scan summaries (summaries). No number in this note was computed in prose.