ASML Just Cleared the Way for Next-Gen AI Chip Production — Here’s What It Means for Asia

• ASML’s High-NA EUV tools are production-ready.
• 500,000 wafers processed and 80% uptime.
• TSMC and Intel are named adopters, full manufacturing integration 2-3 years away.

The machine that could determine who wins the AI chip production race has just been handed a green light.

ASML, the Dutch company with a global monopoly on extreme ultraviolet lithography equipment, confirmed this week that its next-generation High-NA EUV tools have cleared the bar for high-volume manufacturing. The announcement – made exclusively to Reuters by ASML’s chief technology officer Marco Pieters ahead of a technical conference in San Jose – signals the start of a new time in semiconductor manufacturing, with consequences that ripple well beyond the Netherlands.

Why Asia is paying close attention

For chipmakers in the region, this is not a distant development. TSMC, Asia’s most consequential foundry and one of ASML’s named early adopters, sits at the centre of the High-NA EUV story – though perhaps not in the way you’d expect.

While TSMC has received High-NA EUV systems for testing, it has indicated it may defer mass adoption, with senior vice president Kevin Zhang previously noting that the company’s A16 node could proceed without it.

Analysts at Jefferies, however, expect TSMC to deploy High-NA EUV for its A14 node around 2028. The world’s most dominant foundry is playing it cautiously – and strategically. Intel, meanwhile, has taken the opposite approach.

The US chipmaker is already the furthest along in integrating High-NA EUV into its 14A process roadmap, betting that mastering the technology early could help it claw back ground from TSMC in the AI chip production contest.

What “ready” actually means

Pieters was precise about what production readiness does and doesn’t mean. The tools have now processed 500,000 silicon wafers, achieved roughly 80% uptime – with a target of 90% by year-end – and demonstrated imaging precision sufficient to replace multiple conventional patterning steps with a single High-NA pass. “I think that it’s at an important point to look at the amount of learning cycles that have happened,” he told Reuters.

But technical readiness and manufacturing integration are not the same thing. Despite the milestone, Pieters put the timeline for full deployment into production lines at two to three years. “Chipmakers have all the knowledge to qualify these tools,” he said – a measured vote of confidence in the industry’s readiness to move.

The economics of the next leap

At approximately US$400 million per unit – double the cost of the previous EUV generation – High-NA EUV is not a tool that gets adopted lightly. The calculus is especially pointed in Asia, where foundry economics and customer pricing power shape every major capital decision.

TSMC already prices its 2nm wafers at around US$30,000, a over its 3nm generation. Layering in High-NA EUV adoption adds billions more per fab.

For the AI chip production pipeline – and by extension, the AI industry that depends on it – the message from ASML is that the ceiling on current-gen chipmaking is approaching, and the next floor has just been poured. How quickly Asia’s foundries move to build on it will perhaps define the next chapter of the global semiconductor market.

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.