ASML: Competing with Moore's Law

Executive overview

ASML controls the machinery that makes the world's fastest chips possible. Extreme ultraviolet (EUV) lithography—projecting impossibly small patterns onto silicon to create transistors—is the bottleneck that gates Moore's Law itself. By investing 10+ billion euros over 25 years into technology most competitors abandoned, ASML captured 100% of the leading-edge market and 90%+ of next-gen systems.

Core insight: ASML doesn't compete on price; competitors can't catch up because the company owns the entire ecosystem.

The long path from failed spinout to market monopoly

ASML spun out from Philips in 1984 as a failed division—the tenth-place player with no product, no offices, and skeptical staff. Survival required scraping through 1980s recessions while competitors fell away. By the 1990s, it was one of three leaders (Nikon, Canon, ASML). In 2002, it overtook Nikon. The real inflection came when the U.S. Department of Energy invited ASML into an EUV consortium in the 1990s after domestic players couldn't deliver. Where Japanese rivals Canon and Nikon eventually quit EUV (too costly, too hard), ASML had the funding, determination, and engineering talent to push through 25 years of development and bring the first production EUV systems to market in 2019—a 13-year delay from original targets.

How lithography works and why it's physics-limited

Semiconductors are made by projecting light patterns through a mask onto silicon coated with photoresist, etching away material to form circuits. Early light (visible, then UV) worked fine when transistor patterns were huge. But as transistors shrunk, visible light's wave nature became a blocker—diffraction smeared the pattern. Deep ultraviolet at 193 nanometers could only go so far before shrinking more transistors would require writing your signature with a snow shovel. EUV at 13.5 nanometers wavelength halves the feature size, doubling transistors per chip.

The catch: EUV light is so fragile it gets absorbed by air. Everything inside an EUV machine runs in a vacuum, and the light source doesn't exist on shelves. ASML shoots a laser at tin droplets smaller than dust particles, striking each one twice per 50,000-times-per-second cycle, turning them into plasma 40 times hotter than the sun.

Why the supply chain is ASML's biggest advantage and vulnerability

Rather than vertically integrate (Nikon/Canon approach), ASML became an architect and integrator—80% of cost of goods sold comes from specialized suppliers. This was born from necessity (no capital) but became a strength. The company has unmatched relationships with suppliers and ensures they stay ahead of technology curve. It has acquired or taken stakes in key suppliers (Cymer for light sources, ZEISS for lenses) when they couldn't keep pace independently. If suppliers can't innovate fast enough, entire roadmaps freeze.

Market size and machine economics

ASML sold 345 lithography machines in 2022 at revenues of 21 billion euros and gross margins around 50%. The most expensive EUV systems cost 150+ million euros each. Machines are as large as double-decker buses and ship in pieces requiring three jumbo jets. Once installed, 90% of all machines ASML sold in the past 30 years remain operational. Customers pay down payments upfront, creating favorable working capital. Revenue splits roughly 75% from new machines, 25% from service and upgrades.

Pricing power and customer codependency

ASML's top two customers are ~60% of revenues. Economic theory says ASML could double prices as the sole EUV supplier. It doesn't. The company believes price-gouging would incentivize customers to fund alternatives, disrupting industry progress. Instead, it prices on the basis of productivity improvements delivered and targets a 50-50 split of benefits with customers, then shares gains down the supply chain. This collaborative model keeps the industry moving.

Gross margins and capital deployment

Gross margins improved from ~30% in the 2000s to ~50% today, with operating margins ~30%. The company spends 15–16% of revenues on R&D annually, invests heavily in manufacturing capacity, pays dividends, and executes share buybacks—unusual for a growth company. Management has been remarkably stable; the current CEO took the helm in 2013 after nearly a decade as CFO. Martin Vandenbrink, CTO since the 1983 spinout, is considered a near-genius engineer and has been central to every major breakthrough.

Capacity expansion and roadmap

ASML plans to produce 90 EUV machines per year (vs. 40 sold recently), 600 deep UV systems, and 20 high-NA EUV systems by 2027–2028. Management wants to eliminate itself as a bottleneck. New-generation EUV (High-NA) will be even larger, since smaller transistor patterns require bigger lenses.

Why Moore's Law runs on ASML's roadmap

Moore's Law—doubling transistors every two years—has been gated by photolithography since the 1970s. When lithography stalls, Moore's Law stalls. When EUV first struggled 10 years ago, the industry didn't wait; memory makers pivoted to 3D stacking (layering transistors vertically instead of shrinking laterally), reducing lithography capex intensity. If ASML can't deliver fast enough in any segment, markets will find workarounds. This is why technology risk (competitors, disruption) is acute even with 100% market share.

Key risks to the business

Supply chain velocity. Suppliers must keep pace with ASML's innovation or become a drag on entire roadmaps. ASML has handled this well but faces ongoing capital demands on partners.

Disruptive tech alternatives. ASML's real competitor is the physics of Moore's Law, not other companies. If a cheaper, faster way to make advanced transistors exists, the industry will adopt it.

Geopolitics. 40% of sales are Taiwan, 30% South Korea, 15% China. Export controls or escalation could fracture the most advanced node ecosystem overnight.

Investor takeaways

Perfect entry/exit timing in cycles can blind you to structural 20-year opportunity. ASML's early survival relied partly on luck (1986 recession killed rivals, U.S. government picked ASML over Japanese alternatives). But the core lesson is human ingenuity: Martin Vandenbrink couldn't predict 15 years ahead whether Moore's Law would survive, yet relentless innovation kept extending it. Don't underestimate what unknown unknowns will be solved.