Why do the Low Lands produce world-class IC designers, but no world-class IC design companies, Bram Nauta wonders.
I once sold ASML shares at five dollars. Five dollars! I thought I’d made a great investment, because I’d bought them for only about three dollars. I bought a used car with the proceeds. Had I been more patient (or smart), I could now buy a garage full of brand-new Maseratis. Oh well …
As great a company as ASML is, an investment in semiconductor firms further upstream in the supply chain would have paid off even better: Most value is in design, not in manufacturing equipment. Look at the market caps. Nvidia: over five trillion dollars, Broadcom: 2 trillion, AMD: 700 billion.
These companies sell ideas cast in silicon. They own no factories, touch no lithography machines and sometimes never handle a single wafer. They design. They think about architectures, about algorithms that run in hardware, about circuit topologies that make a billion decisions per second. And the world pays handsomely for it. It’s like best-selling books: Not the printing company but the authors stash the cash.
Now for the remarkable part: The Netherlands and Belgium train world-class IC designers in Delft, Eindhoven, Enschede and Leuven. So why has no serious Dutch chip design startup scaled to a billion-dollar company?
The answer is as simple as it is frustrating: money. Or rather, the lack of it. A chip takes three to five years to design, another year or two to reach the market. The investor who wants returns by year three doesn’t understand the industry, or won’t accept the risk. So, the engineers join an American or Taiwanese firm that thinks in decades, or get acquired before they can grow.
What would actually help? Deep-tech funds with a ten- to fifteen-year horizon, leveraging a mix of public money and institutional capital from pension funds. Dutch citizens – including me – save hundreds of billions for retirement, and allocating a small strategic slice to domestic deep tech is not a reckless idea. Give our pension funds this nudge! Also sorely needed: a national program that supports chip design startups through their first five years, with easy-to-obtain subsidies and introductions to customers alike.
We have everything we need to build a European Nvidia: the knowledge, the universities and the infrastructure. I already see strong potential for Axelera AI and Euclyd in Eindhoven, and more startups are popping up as you read this.
Technology, especially AI, moves fast, and today’s winners can be tomorrow’s losers. Our pension funds invest in today’s winners; I’d rather have them invest more in tomorrow’s winners. What we lack is the willingness to wait and the courage to risk our money. If I still had my pile of ASML shares, I’d sell some to invest in young startups. That’s more fun and carries less risk of a crash than racing a Maserati with me behind the wheel.
Yes, those ASML shares were an expensive lesson. But the real lesson is this: The greatest returns aren’t found in what you can measure today; they live in what becomes possible after tomorrow. When the Netherlands starts believing that, with all our capital, not just words, we’ll have a real chance.
