ASML
ASML is the sole manufacturer of EUV lithography systems in the world — not the leading supplier, the *only* supplier. Every chip below the 7nm node, from every major foundry on earth, passes through a machine that only one company on the planet currently knows how to build.
The machine itself is almost absurd in scale. A TWINSCAN NXE/EXE system weighs around 180 tons — roughly the size of a school bus — built from over 100,000 individual parts sourced across those 5,000+ suppliers, and it genuinely takes multiple Boeing 747 cargo flights just to ship one unit to a customer's fab. Installation and calibration on-site can take months before the tool is qualified for production.
The product line itself has a clear generational split:
| System | NA | Node range |
|---|---|---|
| TWINSCAN NXE:3400C / 3600D | 0.33 (standard EUV) | 7nm–3nm |
| TWINSCAN EXE:5000 | 0.55 (High-NA EUV) | 2nm and beyond |
| TWINSCAN NXT:2000i (DUV, ArF immersion) | up to 1.35 | still used for less-critical layers even at leading-edge nodes |
Even a 2nm-class chip isn't exposed entirely on EUV tools — only the handful of layers that genuinely need EUV's resolution use it; the rest of the mask set still runs on mature, far cheaper 193nm immersion scanners, because paying for EUV precision on a layer that doesn't need it would be pure waste.
The monopoly isn't a market-share statistic — it's a genuine single point of failure for the entire industry. No competitor currently builds a commercially viable EUV scanner, and credible estimates put a competitor's time-to-market at a decade or more, given how deep and specialized ASML's supplier network has become. That's why EUV scanner export policy is treated as a matter of national industrial strategy rather than ordinary trade regulation — restricting access to these specific machines is one of the few remaining ways to meaningfully slow a rival's access to leading-edge chip manufacturing.