R
Roel Huis in ‘t Veld
Director, Kriya Materials · Week of October 19, 2026
What ASML’s capacity problem tells a thin-film supplier one hour down the road
Europe’s most valuable company cannot build machines fast enough. The signal travels upstream, and it arrives at companies like ours in a very specific form.
ASML reported third-quarter results on 14 October: sales of €11.7 billion for the quarter, up more than 50% on the €7.5 billion of a year earlier, pre-tax profit of €5.7 billion, and full-year guidance of €44 to €45 billion. For perspective, the company sold €11.8 billion in the whole of 2019. The line that interested me most was not a number. CEO Christophe Fouquet said that planned supply for 2027 and 2028 “will not fully meet the demand of our customers,” and that ASML is now studying how to add capacity for 2028 and beyond, on top of a 30% increase already announced for 2027 and an aspiration of another 30% for 2028. (optics.org)
I run a specialty materials company about an hour’s drive from Veldhoven that originated in the same Philips laboratories in the 1980s and 90s. We do not supply ASML, and this is not a post about lithography. It is a post about what happens to the rest of the optics and thin-film world when the largest buyer of precision optics in Europe decides it needs to build a third more of everything, and then a third more again.
The first thing that happens is that scarcity moves upstream. ASML’s machines are built from mirrors, lenses, coatings, sensors and precision mechanics made by a long chain of suppliers, many of them small and many of them within a few hundred kilometers of here. When the integrator at the top says demand exceeds supply for two more years, every link in that chain hears the same thing: your capacity is now the constraint. Vacuum deposition time, cleanroom space, metrology hours and qualified coating engineers do not appear in a quarter. They are booked, and they are booked by the customer with the deepest pockets. If you make optical coatings for anyone else, you will feel this as longer lead times from your own suppliers and as a quiet drain of people.
The second thing is that the signal comes with a time stamp. The results also mentioned that the next-generation EUV tool is expected to raise throughput from 230 to 260 wafers per hour, and that Samsung and TSMC plan to move high-NA tools into volume production in 2028 and 2030. These are dates. A company my size cannot plan on ASML’s demand, but it can plan on the industry’s calendar: if the big capacity arrives in 2028, the qualification windows for the materials inside it are open now and will be closed by then. That is true whether you sell into the lithography chain directly or into the sensors, displays and packaging that ride on the same AI wave.
The third thing is the one I find hardest. An exponential at the top of the chain creates the temptation to chase it. Every optics company in the Netherlands is being asked, in one form or another, whether it can do something for the semiconductor boom. Some can, and should. But a specialty coatings or optics business that reorganises itself around one customer’s capacity cycle has handed its own planning to someone else’s board. The 2019 to 2026 chart of ASML’s sales is extraordinary. The chart of its suppliers’ margins over the same period would be far less tidy.
So what do I take from a quarter like this? Three practical things. Lock in the inputs now: substrates, targets, precursors, equipment slots, because the queue is forming. We did, and as we did I remembered my former chairman at Polyscope Polymers, Uwe Wascher, who told us to tell our automotive OEM customers: “we plan capacity ahead of demand.” A lesson missed at ASML, but not at Kriya. Capacity is the cheapest thing you can buy; missed sales are the most expensive.
Decide deliberately how much semiconductor-adjacent exposure the company wants, and write the number down before the next sales call rather than after it. And treat the people question as the real constraint: thin-film expertise is scarce, it is about to get scarcer, and the companies that keep their engineers through the boom will be the ones still standing when the cycle turns, as it always does.
Converging exponentials do not arrive as one wave. They arrive as a bid for your capacity, your people and your attention. The discipline is in deciding how much of each to sell, staying disciplined while executing, and thinking about what comes next, because there always is a next.
Source: ASML looks to add more capacity as chip boom continues, optics.org, 14 October 2026