The trilateral chip-equipment perimeter: parallel national instruments, authority-transfer, and the catch-all migration
The trigger
Between June 2023 and October 2025, the Netherlands and Japan adopted four national export-control measures and one statutory overhaul (NL 2023-06-30, JP 2023-03-31, NL 2024-09-07, NL 2025-01-15, JP 2025-10-09) that, taken together, do something the press coverage did not name: they assemble a supply-side perimeter for chip manufacturing equipment by way of parallel national instruments rather than a treaty, and then quietly transfer control authority away from US extraterritorial reach to allied-state co-sovereignty. This is the structurally distinct pair to the demand-side architecture mapped in case #4 (US chip-control architecture export) — that case's Caveats section flagged the trilateral as a future case; this is it.
What the structured layer shows
The conventional framing is "the US pressured the Netherlands and Japan into joining its chip export controls." The responds_to chain reveals something more specific: a three-pillar split-by-specialisation architecture, deliberately built as separate national legal instruments, in which control authority is migrated from US extraterritoriality to allied-state national licensing as each ally's instrument matures.
Pillar 1 — Exposure / lithography (Netherlands)
The 2022-10-07 BIS rule on advanced AI chips and chipmaking tools extended via the Foreign Direct Product (FDP) rule to ASML's TWINSCAN immersion-DUV systems — but the reach of the FDP rule on a Dutch-made, Dutch-supplied tool was always legally awkward. The Dutch resolution came in June 2023 via the Strategic Goods Decree amendment: ASML TWINSCAN NXT:2000i+ moved under a Dutch national licence (severity 5 — the only sev-5 in the equipment-perimeter chain). The Dutch instrument was country-neutral on its face but, by the press release's own admission, "country of concern" applications face a presumption of denial. China was, in practice, the entire affected market.
The architecturally novel move came in September 2024. The Dutch government extended the national licence to two older ASML DUV models — TWINSCAN NXT:1970i and NXT:1980i — that were already restricted, but through the US Foreign Direct Product rule rather than Dutch national licensing. This was not a tightening — it was an authority transfer. As ASML's own statement put it, the substantive perimeter was unchanged; what changed was which government issues the licence. The Dutch instrument absorbed the US extraterritorial reach: ASML now applies to BHOS rather than BIS for those models' shipments. The press read this as "technical." The structural read is sovereignty consolidation — the trilateral perimeter graduating from US-extraterritorial enforcement to allied-state co-sovereignty.
Pillar 2 — Yield management / metrology (Netherlands + Japan)
The January 2025 Dutch metrology expansion added optical and e-beam inspection systems capable of detecting defects at ≤21 nm — ASML's Hermes Microvision (HMI) e-beam tools and ASMI process-monitoring equipment. This was a different chokepoint: lithography is the exposure step, metrology/inspection is the yield-feedback step. Sub-5 nm process development cannot iterate without scatterometry and e-beam inspection feedback loops, which is why this expansion — a sev-3 by frontmatter — is arguably the most chokepoint-y of the three NL measures. Japan's matching coverage of inspection tools came through the March 2023 23-category list (Lasertec wafer/mask inspection, Hitachi High-Tech) and was reinforced by the October 2025 catch-all overhaul below.
Pillar 3 — Deposition / etch + horizontal catch-all (Japan)
Japan's March 2023 23-category amendment covered the four equipment categories where Japanese tool-makers have global market share: lithography (Nikon, Canon, ArF/KrF/EUV-mask), deposition (Tokyo Electron, Kokusai Electric — CVD/ALD/PVD), etch + cleaning (TEL, SCREEN), and inspection/metrology (Lasertec, Hitachi High-Tech). This brought the Japanese share of the trilateral perimeter to its proper weight — Japanese tool-makers are roughly 25-30 % of global wafer-fab equipment by revenue, the second-largest national cluster after the US.
The structurally newer move is the October 2025 FEFTA catch-all overhaul — Japan's first comprehensive review of its catch-all framework in twelve years. Three structural shifts: (a) two-tier classification splitting "core items" (high dual-use risk, explicitly naming semiconductors and machine tools) from a general tier; (b) extending the end-use requirement and adding a new end-user requirement, both reaching beyond UN-arms-embargo countries to all non-Group-A destinations (covering China); and (c) introducing an "informed" condition for Group-A countries — once METI notifies an exporter of a risk pattern, even Group-A exports become licence-required, plugging transit-circumvention to Russia. This is Japan's catch-all framework migrating from item-list architecture toward US BIS EAR-style horizontal catch-all coverage. It is the long-form codification of what the 2023 item-specific list could only do narrowly.
Who acted, how, when
| Date | Jurisdiction | Instrument | Equipment scope | Authority pattern |
|---|---|---|---|---|
| 2022-10-07 | US BIS | EAR amendment + FDP rule | All advanced-node chipmaking tools, FDP extension to foreign-made | US extraterritorial via FDP |
| 2023-03-31 | JP METI | Foreign Exchange Order amendment, 23-category list | Lithography (ArF/KrF/EUV mask), deposition, etch, cleaning, inspection | Japanese national list, narrow item-level |
| 2023-06-30 | NL BHOS | Strategic Goods Decree amendment | ASML TWINSCAN NXT:2000i+ DUV immersion | Dutch national licence — perimeter close |
| 2024-09-07 | NL BHOS | Ministerial reg. (Staatscourant 2024-29008) | ASML TWINSCAN NXT:1970i / 1980i (older models) | Authority transfer US FDP → Dutch national licence |
| 2025-01-15 | NL BHOS | Besluit Strategische Goederen expansion | ASML HMI e-beam, ASMI ALD/CVD, metrology ≤21 nm defect detection | Dutch national licence, vertical climb (exposure → yield mgmt) |
| 2025-10-09 | JP METI | FEFTA Cabinet Order overhaul | Horizontal catch-all + core-item tier + end-user req + Group A informed | Migration item-list → horizontal catch-all (US BIS-parallel) |
The cumulative perimeter: roughly 85 % of global wafer-fab equipment by revenue (Applied Materials + Lam Research + KLA on the US side; ASML + ASMI on the Dutch side; Tokyo Electron + Lasertec + Nikon + Canon + SCREEN + Hitachi High-Tech on the Japanese side) sits under a national licensing regime that, in practice, blocks shipments to SMIC, CXMT, YMTC, Huawei HiSilicon's foundry partners, and any Chinese sub-5 nm process development. The architecturally important point is that this closure was achieved without a treaty.
What this would have told you in real time
Three falsifiable claims a reader of the register could have written down at specific moments:
1. At June 2023, with the Dutch DUV measure on the table and the Japanese 23-category list already in force, the IPTM responds_to chain made it visible that the trilateral architecture is being deliberately built as parallel national instruments, not a treaty — and therefore that it would prove more durable through political transitions than a US-only or treaty-based regime. The November 2024 US political transition did not break it; the architecture continued to extend (Dutch metrology Jan 2025, Japan FEFTA catch-all Oct 2025) under a different US administration. A reader who had filed the trilateral as "US-led pressure" would have priced political risk to ASML and TEL incorrectly across the transition. The structural read priced it correctly.
2. At September 2024, the Dutch NXT:1970i/1980i amendment's authority-transfer pattern — the same items, restricted by different government — was an early signal that the trilateral was consolidating sovereignty over its perimeter rather than relying on US extraterritoriality. A reader of the IPTM action's Open questions section (which named the January 2025 Klever metrology follow-up explicitly, before it was announced) could have predicted the Dutch metrology expansion four months before it landed. The same responds_to-chain reading would have flagged that the next-most-likely Japanese move was a catch-all overhaul rather than another item-list expansion — Japan had run the item-list architecture as far as it would go.
3. At each chain step, the structural read says the ASML / Tokyo Electron / Lasertec / Nikon / ASMI revenue trajectory in China is on a multi-year secular path-down, not a single-event shock. ASML's China share was ~50 % at its late-2023 peak (a backlog-clearing spike, on tools already ordered before the Dutch DUV measure landed); the post-2024 downshift was already legible from the responds_to chain by Q1 2024. Tokyo Electron's China revenue share moved similarly — China was ~30 % of TEL's revenue in FY22 and declined across FY23-25 per TEL's quarterly disclosures, tracking the layered Japanese restrictions. A reader reading the chain in real time would have priced these trajectories as structural rather than cyclical — and would have been right.
Caveats and what would falsify this
- The trilateral does not include Korea. Samsung Electronics + SK Hynix are major chip producers but not chip-equipment manufacturers; they sit on the demand side of the trilateral perimeter, not the supply side. Inclusion of Korea would shift the regime architecture (different industrial composition, different domestic-politics constraints). If a future Korean instrument matches the Dutch/Japanese pattern, this case's "trilateral" framing needs updating to "quadrilateral."
- Taiwan's role is asymmetric. TSMC is a customer of the trilateral perimeter, not a co-licensor. The asymmetry is structurally important: Taiwan's chip-equipment dependency is on Dutch lithography and Japanese deposition, but Taiwan itself does not legislate equivalent controls. The perimeter passes through Taiwan as a customer node, not as an additional licensing layer.
- The authority-transfer reading of September 2024 is contested. An alternative read is that the Dutch government merely accepted enforcement responsibility for a perimeter the US had set, without any sovereignty significance. The case for the authority-transfer reading rests on (a) ASML's own framing as "technical" being consistent with the licensing-pipeline shift being substantively similar in scope, and (b) the January 2025 Klever metrology expansion using the same Dutch instrument independently — i.e., the Dutch instrument was being actively extended under Dutch initiative, not US pressure alone. If subsequent Dutch coverage stops expanding in 2026-2027, the authority-transfer reading weakens.
- The October 2025 Japan FEFTA overhaul effective date is still recent. Implementation guidance is still being issued by METI; the practical reach of the "informed" condition for Group A countries (does it stick in practice? how often will METI use it?) has not yet generated enough case-by-case licence data to test. The structural claim is that the legal infrastructure now matches US BIS EAR catch-all architecture; the enforcement intensity is a 12-24 month forward read.
- Chinese substitution at the legacy node remains real. SMEE, Naura, AMEC, ACM Research have continued to gain share at lower-end tool categories (Naura is now the largest Chinese semi-equipment company by revenue). The trilateral perimeter does not prevent legacy-node Chinese fab expansion — it caps advanced-node development. SMIC's 7 nm and limited 5 nm parts via multi-patterning on older ASML equipment are the canonical demonstration that the perimeter is porous at the boundary; the question is whether it holds at sub-5 nm yield/throughput economics.
Pairing with case #4 (US demand-side architecture export)
These two architectures are the dual of each other:
- Case #4 (demand-side): US BIS extraterritorial perimeter over chips/compute/model-weights, Tier-1/Tier-2 country tiering, exportable as a compliance template (Malaysia adopted ECCN-mirror codes; UAE adopted the RTE bilateral template). The architecture is a single-jurisdiction extraterritorial regime projected through partner-implementation.
- Case #6 (supply-side): Trilateral chip-equipment perimeter, split-by-specialisation national instruments, authority migration from US extraterritoriality to allied co-sovereignty. The architecture is a multi-jurisdictional national-instrument regime coordinated by technical alignment.
Together, these two regimes implement the US chip-decoupling strategy along its two distinct axes (demand and supply). The fact that they have different architectures — extraterritorial vs. allied-co-sovereignty, single-jurisdiction vs. multi-jurisdictional, template-export vs. instrument-coordination — is itself revealing of the underlying constraint: the demand side can be controlled extraterritorially because the buyers' market structure tolerates it; the supply side has to be controlled through allied co-sovereignty because no single jurisdiction holds the full equipment stack.
Sources
- IPTM actions: 2022-10-07, 2023-03-31, 2023-06-30, 2024-09-07, 2025-01-15, 2025-10-09
- METI press releases (Cabinet Decisions 25 Mar 2025, 4 Apr 2025): meti.go.jp/english/press/2025/0325_005.html, meti.go.jp/english/press/2025/0404_001.html
- Netherlands government press releases: government.nl 2023-06-30, government.nl 2024-09-06, government.nl 2025-01-15
- Staatscourant ministerial regulations (Dutch Government Gazette): 2024 nr. 29008; 2025 nr. 1894 (linked from the 2023-06-30 IPTM action amendment record)
- ASML press releases on Dutch export-control updates: asml.com 2023, asml.com 2024
- Pairing reference: case #4 — US chip-control architecture export (
Caveats and what would falsify thissection flags the trilateral as a future case)