ASML and the West’s technology dilemma

Veldhoven, a quiet and comfortable town in the southern Netherlands, does not seem like the kind of place that should be on the frontlines of geo-economic competition between the United States and China. But, alongside the many workers who commute to Eindhoven, it is also home to one of the world’s most important companies.

ASML has a decent claim to be the most significant European company that the fewest people have heard of. It manufactures advanced photolithography machines, the tool which allows chipmakers to etch circuit patterns into silicon wafers. Every leading global chipmaker, from Taiwan’s TSMC, to Korea’s Samsung, to Intel, relies on ASML’s extreme ultraviolet (EUV) lithographic systems to produce the most advanced chips. While there are other sources of deep ultraviolet (DUV) lithographic machines – capable of producing less advanced chips – when it comes to EUV machinery, ASML is the world’s sole supplier. Anyone producing the chips required to power the AI race finds themselves reliant on the Dutch firm.

Beginning during President Trump’s first term, continuing throughout the Biden Administration and now picked up once again by Trump’s team, the Dutch government has been pressured into gradually tightening ASML’s sales of machinery to China. The flow of not only EUV machinery but also of the more advanced models of DUV tools have been gradually squeezed off. Reports last month that China has begun producing its own DUV machines – heavily based on Dutch technology – were enough to both seriously dent ASML’s share price and to cause a round of worry among US geopolitical strategists. China’s lithographic industry is perhaps a decade behind ASML, but, given the Dutch firm had a four-decade head start, it is catching up rapidly.

To observers of economic and technological history, this is nothing new. The same kind of concerns over what a modern economist would term ‘technology transfer’ are at least as old as modern economic growth. During the Industrial Revolution, as Britain build up a commanding – if temporary – economic lead over its peers and rivals, a whole series of parliamentary statutes sought to not only prevent the export of key machinery but went as far as banning skilled artisans from taking their knowledge outside the country.

The overall efficiency of such a knowledge blockade was mixed. Take, as one prominent example, the case of Samuel Slater. Born in Derbyshire in 1768, he went to work in a cotton mill at the age of 10 and later became an apprentice, learning how to use the still novel technology of the water frame, which was revolutionising British textile production. At the age of 21 and declaring himself to be a farmer to evade restrictions, he emigrated to New York. By 1793 he was in New England operating a mill that not only borrowed heavily from British technology but also copied and adapted British organisational and management techniques. On a visit to one of his establishments in the 1830s, then President Andrew Jackson, hailed him as ‘the father of the American factory system’. Back in Britain he was known as ‘Slater the traitor’.

The technology blockade was always porous, never absolute. The water frame, which Slater took to North America, for example, was first employed in England in 1769 and such frames, despite the restrictions, were to be found in France by 1779 and the Netherlands by 1785. At least 1,000 British artisans are known to have made the journey to France between 1710 and 1800.

Eventually, the British government changed track. The ban on emigration was dropped in the mid-1820s and controls on exporting machinery were drastically scaled back before finally being repealed altogether in the 1840s. Rather than seeking to safeguard close secrets, British manufacturers became increasingly keen to sell their technology overseas. By the 1840s, with Britain’s swing towards free trade, such restrictions were increasingly hard to justify.

The consensus view of economic historians is that, even if technology and skilled workers regularly slipped through the net, the overall approach raised the costs of replicating Britain’s approach. The British government could not stop technology from being used in France, the Netherlands or the United States but they could materially delay its introduction.

Cold War efforts to block technology transfer were, if anything, more successful. From the late 1940s onwards, the leading Western economies not only sought to prevent the export of military technologies to their ideological enemies, but also applied controls to the sale of many industrial goods and processes, especially in the fields of computing and semi-conductors. Again, workarounds were found – as they always are. Sometimes they involved third – or even fourth party – transaction chains and, at other times, more old-fashioned industrial espionage. Modern researchers have found good evidence that this was one of many factors holding back Soviet and Eastern Bloc productivity throughout the 1960s, 1970s and 1980s.

It is, in all likelihood, impossible to stop a technology crossing borders forever, though forever is a long time. The history books are full of examples of what trade economists might describe as ‘frictions’ working to delay such technology transfers for years, even decades.

The more interesting question, especially in the case of a competitor such as China, is how the embargoed party responds. Recent work looking at US restrictions on exports to China over the course of the 2000s and 2010s has found that, in the short run, such geo-economic tools tend to work well; Chinese imports of the restricted items fall materially. In the medium term, however, such restrictions have spurred innovation. When Western supplies of a key component or machine are cut off, Chinese firms typically respond by increasing their own R&D spend and the number of patents they hold. Such spurs to innovation are, in the findings of a recent paper, ‘economically meaningful in magnitude and persistence’.

In other words, policymakers have a careful balance to strike. Denying ASML the ability to export to China has almost certainly undercut Chinese efforts in the AI race, but this may come at the price of spurring China to develop its own replacements, creating a stronger, rather than a weaker, Chinese AI and chipmaking sector.