ASML holds an effective monopoly on EUV lithography, the machines every advanced chip on Earth depends on. Yet the company says its European sales are near zero, with demand concentrated in Taiwan, South Korea, the US, and China. The signal is uncomfortable: owning the most strategic tool in the semiconductor stack does not translate into a domestic industry that uses it. Europe exports the crown jewel and imports the finished chips.
The global implication is that industrial policy built on subsidies alone does not manufacture demand. The EU Chips Act poured money into fab ambitions, but fabs need anchor customers, packaging, materials, skilled labor, and downstream design houses. A single champion supplier cannot compensate for a missing ecosystem. Washington learned a version of this with the CHIPS Act; capacity without a customer base and full supply chain becomes stranded capital. ASML publicly asking the EU to help create demand is a rare admission that the bloc lacks the gravitational pull to keep its own value at home.
There is also a strategic exposure hiding in plain sight. ASML's dependence on Asian and American buyers means European leverage over the tool is real, but European resilience in actual chip supply is thin. In a fragmentation scenario, the region controls a chokepoint it barely uses itself.
For Japan, this is a mirror worth studying closely. Japan sits in a stronger position than Europe here, holding deep strength in materials, chemicals, and equipment through firms across the supply chain, plus the Rapidus and TSMC-Kumamoto bets aimed at rebuilding domestic leading-edge capacity. The ASML lesson is direct: METI's subsidies must be tied to committed demand and a full ecosystem, not just fab construction, or Japan risks the same stranded-asset trap.
For Japanese enterprises, SIers, and development teams, the near-term takeaway is supply-chain concentration risk. Advanced compute for AI workloads still funnels through a handful of geographies and a single lithography vendor. SIers advising clients on AI infrastructure, on-premise GPU procurement, and long-horizon capacity planning should treat hardware availability as a geopolitical variable, not a catalog line item. Build sourcing flexibility and cloud fallback into architecture decisions now, because the fragility exposed at ASML's home market is the same fragility sitting under every enterprise AI roadmap.
