The Netherlands Didn't Ask For This Power

The city of Veldhoven is not a place that appears in geopolitical histories. It is a mid-sized municipality in the south of the Netherlands, unremarkable by most measures, known locally for its proximity to Eindhoven and to Philips, from whose corporate spinoffs ASML emerged in the 1980s. It is not where you would expect to find a chokepoint in the most consequential technology competition of the twenty-first century. But that is exactly what it is.

ASML makes the only machines on earth capable of printing the most advanced semiconductor circuits. Extreme ultraviolet lithography — EUV — uses light with a wavelength of 13.5 nanometers, generated by firing a laser at tiny droplets of tin, to etch patterns onto silicon with a precision that has no industrial analogue. The technical barriers to building an EUV machine are so severe that only ASML has ever done it commercially, and the company took nearly two decades and several near-death experiences to get there. The result is a monopoly so complete that “uncontestable” is not hyperbole — it is a technical description.

Every advanced chip made anywhere in the world — every iPhone processor, every data center GPU, every AI accelerator — passes through ASML’s machines or through machines that use ASML’s techniques and components. This is not a vendor relationship. It is a structural dependency of a kind that the technology industry, which prides itself on competition and optionality, has almost never produced.

When the US government began pressuring allies to restrict semiconductor technology exports to China in 2022 and 2023, the Netherlands found itself in a position for which its foreign policy establishment had no preparation. The Dutch had spent decades positioning themselves as a trusted trading nation, a reliable member of NATO, a country whose prosperity depended on open global markets and whose international posture was defined by preference for multilateral institutions over bilateral leverage. Now Washington was asking — with the kind of persistent, escalating firmness that American pressure campaigns have mastered — for the Netherlands to restrict the sale of ASML machines to one of its largest export customers.

The Dutch government did not move quickly. The initial response was to seek a European Union framework that would distribute the decision across the bloc, reducing the bilateral pressure and avoiding the perception of being an American proxy. The EU framework did not arrive on any timeline that was useful. Brussels moves slowly on trade policy under ordinary circumstances; on trade policy that requires alienating China, it moves glacially.

The Netherlands eventually moved unilaterally — first restricting the most advanced EUV systems in early 2023, then progressively tightening the controls to cover older DUV (deep ultraviolet) machines that China had been buying in large volumes. Each tightening followed a similar pattern: the Dutch government announced a new restriction, Chinese officials protested formally, the Dutch government expressed regret for the necessity while maintaining the decision, and ASML’s quarterly earnings call disclosed the revenue impact.

The revenue impact was real. ASML had been selling substantial quantities of DUV machines to Chinese customers including SMIC, and the restriction of those sales cost the company hundreds of millions of euros in annual revenue. ASML’s CEO Peter Wennink, who retired in 2024, was publicly more ambivalent about the restrictions than his government’s position required. He repeatedly pointed out that Chinese customers represented a significant and growing portion of the global semiconductor market, that SMIC was buying DUV machines that were generations behind the frontier, and that the restrictions were delivering pain to ASML and its Dutch shareholders without necessarily achieving the semiconductor supremacy objectives they claimed.

Wennink was right about the economics. Whether he was right about the geopolitics is a different question. The logic of the restrictions is not that DUV machines give China frontier AI capability tomorrow. It is that DUV machines are the path by which China builds the manufacturing experience, yield improvements, and process knowledge that would eventually allow domestic chipmakers to develop competitive frontier production. Restricting older equipment is about restricting the ladder, not the top rung.

What the Netherlands’ experience reveals about the new geopolitics of technology is something that the traditional frameworks of trade policy and national security don’t handle well: the emergence of single-company chokepoints with global strategic significance.

During the Cold War, the US maintained export control regimes (COCOM, its successor organizations) that covered entire technology categories. The logic was industrial: control the class of technology, and you control the capability. The semiconductor era has produced a different structure, where the relevant chokepoints are specific processes at specific companies — ASML’s EUV machines, TSMC’s N3 fabrication process, Cadence and Synopsys’s electronic design automation software. These are not categories. They are singular entities. And the countries that host them have acquired leverage that was never part of their strategic calculation.

South Korea faces an analogous version of this problem with Samsung and SK Hynix. The two companies together produce a dominant share of the world’s DRAM and high-bandwidth memory. US pressure on Korea to restrict HBM sales to China created a genuine policy dilemma: comply and harm major domestic companies with significant Chinese exposure, or resist and damage the US-Korea security relationship. Seoul’s response — partial compliance with substantial delays and carve-outs — is the diplomatically honest version of what happened in the Netherlands, where the scale of the chokepoint made resistance more costly.

Taiwan’s TSMC sits at the extreme end of this spectrum. The company is not merely a chokepoint — it is the entire bridge. But Taiwan’s political situation is uniquely constrained in ways that make the Netherlands’ situation look comfortable. The Dutch government can restrict ASML sales and suffer economic consequences. Taiwan cannot restrict TSMC sales to China without triggering economic consequences that would destabilize its own government, and cannot be seen to align completely with US policy without provoking the cross-strait tensions that its entire security strategy is designed to manage.

The peculiar irony of ASML’s position is that the company’s monopoly was not designed as leverage. It was an accident of industrial history — the result of a bet made by a consortium of European chip companies in the 1980s on a technology that seemed impossibly ambitious, sustained through losses that would have killed a less committed program, eventually vindicated by the physics of what EUV could do that nothing else could. ASML did not set out to become indispensable. It became indispensable by being the only company willing and able to solve a very hard problem.

The lesson that policymakers have drawn — that strategically critical technologies should not be concentrated in single-company monopolies — is correct and almost entirely unactionable. You cannot mandate competitors into existence for a technology that requires decades of accumulated process knowledge. China has been trying to build an EUV alternative since at least 2019, through the Shanghai Institute of Optics and Fine Mechanics among other institutions, and has made genuine progress in understanding the physics while remaining far from producing a commercially viable machine. The gap is not primarily about national will or investment capital. It is about the cumulative engineering knowledge that ASML built over thirty years of doing almost nothing else.

What the Netherlands now lives with is the permanent condition of a country that hosts a piece of industrial infrastructure whose strategic significance exceeds its own. That is a strange position for a country whose self-understanding is as a trading nation and multilateral institution-builder. It is also, probably, a permanent one. The machines in Veldhoven will remain the chokepoint in semiconductor manufacturing for at least the next decade, possibly longer. The Dutch government will continue to navigate between its American alliance and its European commercial interests with the careful ambivalence of someone who holds more cards than they want.

There are worse problems to have. They are still, recognizably, problems.