The most complicated machine in the world: why nobody can copy ASML

To understand why export controls on lithography equipment work so well, you have to look at the machine itself. EUV systems made by ASML are considered the most complicated industrial machines ever built, and they have become one of the main points of the chip dispute between the United States and China. Industry accounts say many companies were involved in their development, and a single machine is technology worth hundreds of millions of dollars.

Extreme ultraviolet and a network of subcontractors

Wavelength is the key. To create structures below 5 nanometres you need light in the extreme ultraviolet range, and ASML is the only company able to produce such systems at scale. The company could not do it alone. Its advantage is the ecosystem: Zeiss with optics of extreme purity and Trumpf with high-power lasers. Rebuilding that network takes years and billions, not one successful design.

Washington suspects a leak

These machines matter so much that any information about a possible transfer of technology triggers a political reaction. The American side was said to have raised the charge that an advanced EUV machine had reached China. Earlier, Zeiss and Trumpf were named as the companies building the unique components. That shows how narrow the bottleneck of the whole chain is.

Europe's problem is the reverse

Frank Heemskerk, who handles public affairs at ASML, put an uncomfortable thesis to the Dutch debate: the company sells practically nothing in Europe, because there are no factories here that would buy its machines. Demand is concentrated in the United States and Asia. Heemskerk admitted that the machines are wanted in the US and in China, and at the same time he appealed to Brussels to use the market power that Europe does have.

That reverses the usual understanding of technological sovereignty. The Union holds a unique tool for making chips, but it has no buyer for it. Export controls therefore work as leverage over others, and as a lens showing what Europe lacks. For Poland, which aspires to a place in the European semiconductor chain, the conclusion is not romantic: without its own demand for advanced manufacturing, Poland's presence in the sector will stay the role of a service supplier, not a technology one.

The scale of precision explains why the export of a single machine has become a matter of state. EUV is needed from the 5-nanometre generation upwards; China currently has access only to older lithography technologies, and attempts to build its own chain, undertaken by SMIC and Huawei, are costly and drawn out. In the background remains the charge that a modern EUV machine reached China despite sanctions; the American side was said to have reported this publicly.

For Europe the trouble is twofold: the key technology is created on its territory but sold almost entirely outside it, and decisions about who may buy it are largely made in Washington. Technological sovereignty, then, does not come down to owning a factory. It also means having a say over its customers.