ASML has confirmed that its high numerical aperture extreme ultraviolet lithography tools have reached production-ready status, marking a pivotal step for the next generation of Artificial Intelligence chips. The Dutch manufacturer, which holds a global monopoly on commercial extreme ultraviolet lithography systems, disclosed the milestone ahead of a technical conference in San Jose. Current-generation extreme ultraviolet machines are nearing their physical limits for advanced Artificial Intelligence chip production, as semiconductors that power large language models and Artificial Intelligence accelerators run up against scaling constraints. High numerical aperture extreme ultraviolet tools are engineered to surpass that barrier by enabling finer and denser circuit patterns in fewer manufacturing steps, directly supporting more powerful and energy efficient processors for demanding Artificial Intelligence workloads.
ASML’s readiness claim is anchored in three key metrics collected from extensive customer testing. The company’s high numerical aperture extreme ultraviolet tools have now processed 500,000 silicon wafers, achieved roughly 80% uptime with a stated target of 90% by year-end, and delivered imaging performance that can replace multiple conventional patterning passes with a single high numerical aperture exposure. According to ASML’s chief technology officer Marco Pieters, those figures signal that chipmakers can begin formal qualification of the machines in their own fabs. The systems are described as among the most expensive pieces of industrial capital equipment ever built, with each unit priced at approximately US$400 million per unit, which is double the cost of the previous extreme ultraviolet generation. TSMC and Intel are cited as among the early adopters preparing to bring the new tools into their advanced process roadmaps.
Despite the technical milestone, ASML stresses that readiness for production and full integration into high volume manufacturing are distinct phases. Pieters notes that integration into large scale semiconductor production lines is expected to take two to three years as customers work through qualification and process development. He expressed confidence that chipmakers already possess the expertise required to qualify the tools, framing the transition as challenging but manageable for the industry. The latest development effectively starts the race among leading foundries and integrated device manufacturers to industrialize high numerical aperture extreme ultraviolet technology, with the next wave of Artificial Intelligence chip performance improvements now moving from theory to a defined execution timeline.
