The strongest case for the US bill
On 14 October 2026, ASML reported that shipments to China jumped 109% quarter on quarter in Q3, lifting China's share of its machine sales from 14% to 23%, according to AI Weekly's summary of SCMP reporting. The jump may reflect Chinese memory makers pulling forward purchases of lithography tools, though that explanation is not confirmed. For advocates of the US MATCH Act, this is the exhibit they need. If Chinese buyers can pull forward years of capacity while Washington and The Hague argue, the licensing system looks like a speed bump rather than a barrier. A bill that bans immersion DUV sales and tightens servicing would close that window by statute rather than by diplomacy.
That is a serious argument, and it deserves a serious answer. Allies that do not align their controls hand an advantage to whichever supplier ships fastest, and national security officials are right to worry about it.
What the bill would actually do
As Bits&Chips reported on 24 June 2026, the bipartisan MATCH Act would ban exports of all immersion DUV equipment to China and further restrict servicing there. It would also create mechanisms to pressure partner countries into aligning with US export policy. DutchNews reported that the bill had cleared a committee but not the full Congress, and that the affected machines account for about a fifth of ASML's expected revenue this year.
The Dutch cabinet's own reading is the most precise account available. In a written answer to MP Hoogeveen dated 12 May 2026, Trade Minister Sjoerd Sjoerdsma said the bill instructs the US government to coordinate with allies on country-wide controls with a "presumption of denial." It also contemplates licences for servicing equipment that has already been delivered. If allies fail to harmonise within set periods, the US could impose export restrictions on their companies by its own law. The cabinet's conclusion was blunt: it opposes the extraterritorial effect, because each country is responsible for its own export control legislation.
Why the Dutch objection is principled, not just commercial
It would be easy to dismiss Dutch resistance as lobbying for a national champion. The cabinet's argument is more structural. Under the EU Dual-Use Regulation, Regulation (EU) 2021/821, the legal framework the cabinet's answer calls leading, licence applications are judged case by case rather than through blanket country bans. The cabinet describes its approach as surgical and based on specific security risks, aimed at non-proliferation and preventing unwanted end use.
The servicing provision matters most. Cutting off support for machines already installed would not simply restrict future sales. It would reach into contracts that are already signed and fulfilled, with legal and reputational consequences for a Dutch firm that followed Dutch and EU law. Sjoerdsma put the principle plainly in Washington: "If cooperation becomes cooperation by force, that's undesirable from our point of view." He made the point the same week the Netherlands joined the US-led Pax Silica initiative, which shows the objection is to the mechanism, not to alignment itself.
The stockpiling problem cuts both ways
The Q3 numbers do not settle the argument, and each side will over-read them. A few cautions are in order.
- The 23% figure is a share of system sales in one quarter. CFO Roger Dassen's guidance that China will be about 20% of 2026 business is a full-year measure. The two are not directly comparable.
- If the jump is stockpiling, it is, by definition, front-loaded demand. A surge in one quarter implies weaker shipments later, which is consistent with the guidance that China's full-year share stays near 20%.
- Chinese demand held up against the Dutch government's own regime. The cabinet's answer notes that China exceeded 33% of ASML sales in 2025, as the MP's question put it, and that ASML itself expected the share to fall to about 20% in 2026. The existing licensing system has therefore been compressing the China share, even if it has not stopped it.
The honest reading is that the stockpiling shows a real leakage risk and also shows that the current controls are biting at the margin. Neither supports a statute written in Washington and enforced against Dutch exports.
A proportionate path forward
The pro-innovation position is neither "no new limits" nor "accept whatever Congress drafts." Three principles follow from the evidence.
- Fix timing gaps in the licensing system, not the sovereignty question. If memory makers can pull forward years of purchases, the Netherlands and the EU can apply licence-volume scrutiny and end-use checks through the framework they already control. That keeps decisions case by case, as the Dual-Use Regulation requires.
- Coordinate through plurilateral channels. The Wassenaar Arrangement and the Pax Silica commitments are venues where Japan, South Korea, the Netherlands and the US can set common rules. Coordination of that kind is more durable than coercion, and it spreads the competitive cost across all toolmakers rather than loading it onto one.
- Protect legal certainty on servicing. Any servicing restriction should be narrowly tailored, targeted at named end users, and phased so that existing contracts can be wound down lawfully. Blanket retroactive bans on maintenance invite litigation and push Chinese buyers toward domestic substitutes sooner.
The long-term risk of over-correction is real. ASML funds a large share of the research that keeps European lithography at the frontier. Rules that cut off a fifth of revenue without allied agreement would weaken the company and the ecosystem the controls are meant to protect.
The Q3 surge is a warning to the Dutch government to demonstrate that its own controls are credible. It is not a reason to hand the design of those controls to another legislature.