The Multilateral Alignment of Technology Controls on Hardware Act would write semiconductor-equipment export controls into law, cut off five named Chinese chipmakers — and give U.S. allies 150 days to follow suit or be overruled.
Most of the unilateral U.S. controls the bill codifies already exist. Its real marginal value is the alliance ultimatum — and three moving parts.
The MATCH Act would (1) write semiconductor-equipment controls into statute, stripping Commerce of discretion to quietly loosen them; (2) designate SMIC, Huawei, CXMT, YMTC and Hua Hong (plus all affiliates) as “covered facilities,” cutting off both sales and the servicing lifeline; and (3) give allies ~150 days to adopt equivalent controls or face an extraterritorial Foreign Direct Product Rule. It targets equipment, not chips.
Controls live in BIS regulation today, so any administration can quietly loosen them. The bill makes them permanent — and harder to trade away in a negotiation.
Instead of case-by-case Entity List fights, it names the firms — and every subsidiary, affiliate and JV partner — to defeat the rename-and-restructure game.
The Netherlands, Japan and Germany must adopt deny-by-default controls, or the U.S. reaches over their heads via the extraterritorial Foreign Direct Product Rule.
The whole strategy rests on one fact: U.S. technology is woven through every advanced chipmaking tool on the planet. That embeddedness is the leverage.
MATCH is the latest move in a campaign that began in 2022 — and an attempt to make the whole edifice permanent and multilateral.
Commerce restricts China's ability to both buy and build advanced chips, citing military and AI end-uses — the first attempt to choke advanced-node manufacturing capacity, not just finished chips.
Washington secures an informal agreement from the two key allied toolmaking nations to restrict some equipment — the multilateral seed MATCH wants to harden.
Huawei's Mate 60 Pro ships a SMIC-made 7nm chip — built on restricted DUV tools via multi-patterning, no EUV. Proof controls slow but don't stop a determined rival. (China has stalled at 7nm since.)
Geographic scope broadened to worldwide subsidiaries to curb third-country diversion; advanced-computing chip thresholds tightened.
~140 Entity List additions, controls on 24 types of equipment and 3 software tools, and the first-ever controls on high-bandwidth memory (HBM).
Introduced in both chambers. Within two weeks, industry lobbying drops the country-wide cryo-etch ban and softens the servicing presumption. DUV stays.
Ordered reported as a substitute amid what lawmakers called the largest export-control markup in the history of Congress.
Both chambers share the same operative skeleton, executed against statutory deadlines that climb from mapping to diplomacy to coercion.
Commerce, State (and in the House, Energy & Defense) jointly identify every covered tool and covered facility, and report the list to Congress.
The U.S. immediately presses allied supplier countries to adopt deny-by-default controls and servicing restrictions, then briefs Congress on progress.
Binding regulations on U.S. equipment take effect. Where an ally hasn't aligned, the U.S. unilaterally extends jurisdiction over their tools via the FDPR.
An annual report and recurring certification that every relevant export and servicing action faces a license under a policy of denial. Backsliding re-triggers in 60 days.
Lithography tools aren't fire-and-forget — ASML services its China DUV fleet roughly every six months. CNAS estimates that cutting that off would "significantly degrade" China's installed immersion fleet within about a year— without seizing a single machine. The bill's sweeping definition of servicing covers in-person and remote support.
It also captures every subsidiary, affiliate and JV partner of a named firm — designed to defeat the rename-and-restructure game that has long blunted the Entity List.
The bills are ~90% identical in structure. The House version is broader and more aggressive in scope but more constrained procedurally; the Senate version is narrower and more diplomatically textured.
Three datasets frame the whole debate: China's rising mature-node share, its accelerating self-sufficiency, and the Western revenue on the line.
Backed by $150B+ in state subsidies and ~55% of all planned global capacity expansion. These chips end up in cars, weapons and infrastructure.
Every refused Western sale hands market share to SMEE, Naura, AMEC and ACM. The stubborn gap remains lithography, metrology and EDA.
Figures are approximate and vary by quarter; cutting this revenue is the core of the "less money for R&D" critique.
The core tension is durability versus backfire: sound chokepoint logic and statutory permanence on one side; accelerated indigenization, lost R&D and retaliation on the other.
Chips obsolesce in a few years; restricting manufacturing capacity constrains output for a decade. “The machines matter more.”
China can't replicate DUV immersion lithography before ~2030 — and bulky tools are far harder to smuggle than pocket-sized chips.
Cut off original-supplier maintenance and software updates, and installed Chinese capacity quietly degrades on its own.
Writing controls into law prevents administration whiplash and makes them far harder to bargain away in a trade deal.
A successful 150-day push could lock the Netherlands, Japan and Germany into a leak-proof, deny-by-default coalition.
The big U.S. toolmakers booked ~$19B in China sales in 2025; ASML ~30% of revenue. Cutting it may erode the very lead the bill protects.
The 2022 controls pushed China's domestic toolmaker share from ~13% to ~30%. Every refused sale feeds SMEE, Naura & AMEC.
Beijing banned gallium, germanium & antimony to the U.S. one day after the Dec 2024 crackdown — antimony exports fell ~97%. It controls ~99% of gallium and ~60% of magnets, and only suspended the curbs in the 2025 truce.
FDPR coercion stings most because of the asymmetry: allied firms eat the losses while U.S. AI firms still sell to China. The EU built an Anti-Coercion Instrument partly against this.
MATCH controls equipment sold to China, not Chinese chips imported into the U.S. That problem lives on a different, slower track.
MATCH's name promises multilateral alignment — but its mechanism is unilateral coercion of allies. That sits on two fault lines.
Builds its own national licensing and frames controls as sovereign acts. ASML's CEO openly questions the security rationale. China was still ~36% of 2024 sales.
METI controls 23 tool categories, but Japan fears mineral retaliation — Toyota privately warned Tokyo of a critical-minerals cutoff.
Sole supplier of the optics inside every ASML machine, most auto-exposed, not yet in the trilateral — and shielded by the EU's Anti-Coercion Instrument.
The critique most relevant to anyone worried about Chinese chips already in American cars, weapons and infrastructure: MATCH barely touches the legacy-chip problem.
A December 2024 BIS survey found the embeddedness is already here: more than two-thirds of surveyed companies' products contain Chinese-made chips — and roughly half can't even tell whether they do.China's share of global legacy (~28nm+) production is climbing toward ~39% by 2027, backed by $150B+ in subsidies.
But MATCH restricts equipment sold to China, not Chinese chips imported into the U.S.That fight lives on slower, separate tracks — and they're thinner than advertised. The Section 301 probe concluded in Dec 2025 adding 0% in tariffs today, with a hike only scheduled to be considered from mid-2027; NDAA §5949(federal procurement only) doesn't bite until Dec 2027. You also can't easily tariff what half of importers can't even detect.
It can slow China at the leading edge and buy time — if, and only if, the allies align. The U.S. controls it codifies mostly exist already; the genuinely new thing is the 150-day alliance ultimatum. Lock the Netherlands, Japan and Germany into a durable coalition and the fence finally tightens at its leakiest joints. Provoke them instead, and the bill becomes an expensive way to accelerate Chinese self-sufficiency.
Allies align within 150 days. A durable deny-by-default bloc forms; China's advanced-node progress stalls near 7nm and the U.S. AI-compute lead is preserved for years.
Some allies align, some drag; the FDPR is invoked selectively. China's leading edge is slowed and made costlier, but indigenization accelerates at mature nodes.
Allies balk at extraterritorial pressure; China retaliates on minerals and races to self-sufficiency. U.S. toolmakers lose the market permanently and fund the rivals that replace them.
Probably not as a full-strength standalone. The realistic path is as a rider on the FY2027 NDAA, likely diluted — with the FDPR and 150-day ally-coercion provisions the most likely to be stripped. It is the best-targeted export-control instrument yet proposed: a scalpel aimed precisely at the leading edge of a problem whose mass sits at the trailing edge — written by a Congress legislating against the grain of an administration busy trading the same controls away.