Latest / Business Strategy Talks with Fexingo: Competitive Advantage, Market Positioning, and Growth Planning / How ASML Built a Moat on Lithography Monopoly
Transcript
- Lucas: There is a company in the Netherlands that builds the single most complicated machine ever made by humans. It weighs two hundred tons, costs about three hundred and fifty million dollars, and there is exactly one firm on earth that can make it. Luna: You're talking about ASML and the EUV lithography machine. The extreme ultraviolet machine. Lucas: Exactly. ASML — Advanced Semiconductor Materials Lithography — is the sole supplier of extreme ultraviolet lithography systems. Without those machines, you cannot make any advanced chip below about seven nanometers. Every smartphone, every AI accelerator, every advanced GPU. ASML is the bottleneck. Luna: And that bottleneck has been twenty-five years in the making. They started the EUV research program in the late 1990s, right? Lucas: More or less. In 1997, the U.S. Department of Energy launched a consortium to figure out whether extreme ultraviolet light could even work for lithography. ASML was a small Dutch company at the time — not the dominant player. Canon and Nikon were the leaders in lithography equipment. Luna: And now Canon and Nikon have essentially no presence in advanced logic lithography. What happened? Lucas: ASML made a bet that EUV was the future, and they committed to it at a scale no competitor matched. But the real genius isn't just the technology. It's how they funded it. ASML essentially made their biggest customers — Intel, Samsung, TSMC — co-invest in the R&D. Luna: Right, the 'customer co-investment model.' They sold equity stakes and also got prepayments for future machines. Lucas: Exactly. In 2012, ASML acquired Cymer, the company making the EUV light source, and they raised capital from Intel, Samsung, and TSMC in exchange for research commitments and guaranteed access to early machines. So the three largest chipmakers in the world had a financial incentive to make EUV succeed. They weren't just customers — they were partners. Luna: That's a brilliant moat. Because not only does ASML have a technological lead that takes decades to replicate, but they also have a financial lock on the very companies that might want to fund a competitor. Lucas: And the switching costs are astronomical. A single fab costs ten to twenty billion dollars. The entire manufacturing process is tuned to ASML's machines. If a competitor appeared tomorrow, it would take at least a decade and tens of billions of dollars to even have a prototype. Luna: But there's a vulnerability here. ASML is a single point of failure for the entire global semiconductor supply chain. That's a geopolitical risk. Lucas: It's the biggest risk. The Dutch government, under pressure from the U.S., has already restricted exports of some ASML machines to China. That creates political tension. And if the EU or Netherlands ever decided to restrict exports further, it could disrupt the whole industry. But it also reinforces ASML's moat — because no government wants to kill the golden goose. Luna: And there's a question of whether the moat is truly unbreakable or just very expensive to breach. Could a competitor emerge from China? Lucas: China is trying. SMEE, Shanghai Micro Electronics Equipment, has been working on lithography for years. But they're probably at least a generation behind. EUV requires a vacuum environment, mirrors with atomic-level precision, and a light source that generates plasma by zapping tin droplets with a laser fifty thousand times per second. It's not something you reverse-engineer quickly. Luna: So the moat is built on multiple layers: proprietary physics, co-investment partnerships, scale, and a decades-long head start. Lucas: And a culture of secrecy. ASML doesn't publish detailed specs. Their machines are installed in clean rooms under non-disclosure agreements. Customers aren't allowed to even open the covers. If something breaks, ASML sends its own engineers. Luna: That's like the ultimate lock-in. You own the machine but you don't really own it. Lucas: Right. And the service revenue is enormous. ASML's gross margins are over 50 percent. They sell the machine, and then they sell the consumables — the pellicles, the optics maintenance, the upgrades. It's a razor and blades model on a three-hundred-million-dollar razor. Luna: And the blade is recurring. Every EUV machine requires regular mirror coatings and light source maintenance. That's high-margin, sticky revenue. Lucas: We should probably pause here for a second and mention something that makes this show possible. It's a quick thing. Luna: Yeah. We don't run ads on this podcast, and we want to keep it that way. Lucas: A small group of listeners already supports us through buy me a coffee dot com slash fexingo, and that's literally what funds making this many episodes. If these conversations have moved your work forward, joining them would mean a lot. Luna: Agreed. And now back to ASML — you mentioned the service revenue. How does that compare to the machine sale itself? Lucas: Service is about 20 to 25 percent of total revenue, but it's growing faster than machine sales. Because the installed base of EUV machines keeps increasing — they've shipped over two hundred EUV systems now. Each one needs ongoing support. Luna: And the next generation — high na EUV — is even more expensive. Those machines reportedly cost around four hundred million dollars each. Lucas: high na EUV is the next step, and again, ASML is the only company making them. They've already shipped early units to Intel and TSMC. The development cost for high na was shared by the same customer consortium. So the moat is self-reinforcing. Luna: Is there any scenario where the moat breaks? I mean, what if a new technology makes lithography obsolete? Lucas: That's the existential question. There are potential alternatives — like directed self-assembly, or nanoimprint lithography, or even quantum tunneling. But none are commercially viable for high-volume manufacturing at single-digit nanometer scales. The industry has bet on EUV for the next decade at least. ASML has a ten-year order backlog. Luna: So the moat is not just technological, but also temporal. Even if someone invented a better mousetrap tomorrow, the industry can't switch quickly. Lucas: Exactly. And ASML knows this. They invest about two billion euros a year in R&D — about 15 percent of revenue. They're not standing still. They're already working on beyond euv technologies. Luna: One more thing — the geographic concentration. ASML is based in Veldhoven, a small town in the Netherlands. That's a risk if the region faces a natural disaster or political instability. Lucas: It's a risk, but the Netherlands is politically stable. And the Dutch government has designated ASML as a strategic national asset. They're not going to let it fail. Plus, ASML has facilities in the U.S., Germany, and Taiwan for service and some manufacturing. Luna: So the takeaway for someone thinking about moats: ASML shows that the deepest moats are often built on a combination of technical difficulty, customer co-option, and time. You can't just spend money to catch up. Lucas: Right. Money helps, but you also need the accumulated learning from decades of iteration. ASML has made mistakes, but they've corrected them. Their machine failure rates are incredibly low now. That kind of reliability is another moat—trust. Luna: And the lesson for smaller companies: you don't need to be a monopoly to have a moat. But if you can find a bottleneck where your product is essential and irreplaceable, you've built something durable. Lucas: Exactly. ASML's moat is extreme, but the principles apply broadly. The question for any business is: what is your EUV? What is the one thing you do that, if you disappeared, the industry would have to rebuild from scratch? Luna: That's a good note to end on. For now, ASML's moat looks unbreachable. But we'll keep watching. Lucas: And we'll keep you posted. That's it for this episode of Business Strategy Talks.