ASML: Competing with Moore's Law

Executive overview

ASML is the only supplier of extreme ultraviolet lithography machines, the critical technology enabling continued miniaturization of semiconductors. Spun out from Philips as a struggling underdog in 1984, the company invested 25+ years and 10 billion euros in R&D to pioneer EUV technology, which 2019 finally enabled the next generation of chip manufacturing. Core insight: Photolithography is the gating technology of Moore's Law; ASML's monopoly on advanced EUV equipment gives it unprecedented competitive advantage.

From startup outsider to industry leader

• Spun out from Philips in 1984 as the number 10 player among 10 lithography competitors with no revenue, no product, and no offices
• Early years were survival mode through brutal industry cycles in the 1980s, kept alive by Philips and ASM capital injections
• Emerged as one of three leading players by the mid-1990s alongside Nikon and Canon, finally overtaking Nikon as number one in 2002
• Martin Vandenbrink joined from Philips in 1983 and became CTO; led development of company's first commercial product and has remained instrumental to success

How photolithography works

• Semiconductors work by embedding microscopic electrical circuits on silicon using light projected through a mask onto chemically-coated surfaces
• Early approach used visible light, but as patterns shrunk to single-digit nanometers (vs. 193nm wavelength of deep ultraviolet), light diffraction became impossible to overcome
• Process conceptually similar to cinema projection but inverted: lens shrinks image down rather than projecting it large
• Dimensions involved are so small they're measured in handfuls of atoms; defects from any manufacturing flaw are catastrophic

The evolution to extreme ultraviolet

• Deep ultraviolet light topped out at 193nm wavelength but leading-edge chips today are 5nm nodes; trying to write signature with snow shovel
• After recognition in 1990s that new technology was essential, the US funded EUV research through Department of Energy and DARPA via consortium
• Americans invited ASML to join because they had no domestic lithography maker strong enough and didn't want to give technology to Japan
• Original timeline was mid-2000s delivery; actually took until 2019—13-14 years late but critical momentum shift

ASML's dominance in EUV

• 100% market share of world's most advanced EUV lithography equipment, 90%+ on next-generation machines
• Massive barriers to entry: 25+ years and 10 billion euros of R&D investment; no competitors attempted to follow after Nikon and Canon gave up
• Light source requires shooting laser at tin droplets smaller than dust, striking each twice to create plasma 40x hotter than sun's surface—50,000 times per second
• Acquired light source provider CYMA in 2013 because no other supplier could innovate fast enough; mirrors and other components equally irreplaceable
• Each machine requires three jumbo jets to transport; size will increase further as transistors shrink because larger lenses needed

The business model and customer relationships

• Only 345 photolithography machines sold in 2022 despite 21 billion euros revenue; EUV machines cost 150+ million euros each
• 75% of revenue from new machine sales, 25% from service, maintenance, and field upgrades
• 90% of machines sold in last 30 years still in operation; machines installed 10 years ago can be upgraded to markedly improve productivity
• Top two customers (TSMC and Samsung/Intel) represent nearly 60% of revenues; only ~3 companies globally have capacity to buy and operate this equipment
• ASML accounts for 20-25% of total wafer fabrication equipment costs depending on chip type

Architectural integration approach

• Describes itself as "architects and integrators" rather than traditional vertically-integrated manufacturer
• 80% of cost of goods sold comes from components purchased from external suppliers; 20% is onsite labor
• Born from necessity in early years: didn't have capital or time for vertical integration like Nikon and Canon pursued
• Strategy has become strategic strength: managing world-class supply chain relationships that enable ecosystem to move forward together without bottlenecks

Pricing and profitability strategy

• Gross margins around 50% (up from mid-40s last decade), operating margins around 30%
• Despite monopoly position, ASML explicitly doesn't price gouge; philosophically committed to 50-50 split of profitability improvements with customers
• Rationale: if ASML doubled prices, customers would invest in alternative technologies or suppliers, destroying long-term collaborative relationships
• Cash generation fantastic due to customer down payments upfront; nearly all net income converts to free cash flow
• Company spends 15-16% of revenues on R&D annually, invests heavily in CapEx expansion, yet still pays dividends and executes substantial share buybacks

Capacity expansion plans

• Currently producing ~40 EUV machines annually; planning to expand to 90 EUV machines per year by 2027-28
• Also targeting 600 deep ultraviolet (DUV) machines and ~20 high-NA EUV machines annually
• Explicitly positioning itself to be supply enabler, not bottleneck; doesn't want customers incentivized to find alternative manufacturing approaches
• Manufacturing capacity primarily means assembly facility expansion since 80% of components sourced externally

Key risks and long-term challenges

• Supply chain vulnerability: Suppliers must keep pace with technology advances or ASML forced to acquire them (Cyma 2013, stake in Zeiss)
• Disruptive technologies: ASML's real competition is Moore's Law itself; if alternative manufacturing approaches become cheaper/faster (e.g., 3D NAND stacking), lithography demand falls
• Geopolitics: 40% of sales to Taiwan, 30% to South Korea, 15% to China; concentrated customer base in geopolitically volatile region makes supply/export restrictions risk

Holistic lithography and next-generation opportunities

• Can't use optical techniques on leading-edge chips anymore; use electron beams to detect defects down to single nanometer resolution (four silicon atoms)
• Feeding better defect data into computer models improves lithography machine adjustments, creating virtuous cycle of improvement
• Will continue incremental improvements on current machines while pursuing next-generation EUV (high NA) and beyond
• Modular design approach enables de-risked deployment: if one module upgrade fails, rest of machine continues operating

Lessons for investors and founders

• Don't fixate on perfect entry/exit timing in cyclical industries; structural opportunity can be obscured by cycle obsession
• Underestimated luck in early years: 1986 recession killed competitors when ASML lacked real product; EUV technology transfer happened only because US had no alternative to Japan threat
• Human ingenuity and long-term innovation persistence matters more than most realize; CTO Martin Vandenbrink has guided company through 40+ years while maintaining that Moore's Law will run 15+ more years (and proven right repeatedly)