ASML cuts EUV build time by 30% as AI demand surges
ASML is overhauling its production lines to build critical chipmaking machines a third faster, a bottleneck-breaking move that determines the global pace of artificial intelligence expansion.
ASML is cutting the time it takes to build and test its extreme ultraviolet lithography machines from 22 weeks to between 15 and 16 weeks. The Dutch company announced the roughly 30% reduction in cycle time on Wednesday as it races to satisfy a global hunger for advanced chips.
The operational overhaul comes with a significant upgrade to the company's financial outlook. ASML raised its full-year sales guidance to between €43 billion and €45 billion. The firm is close to being fully booked for EUV machines in 2027, with a substantial number of orders already placed for 2028.
Having orders secured this far in advance is rare, according to CFO Roger Dassen. It underscores how aggressively semiconductor manufacturers are locking in capacity to support artificial intelligence infrastructure. “We’re now looking at bringing that down to 15 to 16 weeks,” Dassen told reporters, noting the immense pressure to accelerate output.
ASML plans to produce about 65 of its low-NA EUV machines this year. To hit its target of a 30% capacity boost for 2027, the company is reorganising its clean rooms to prioritise manufacturing over research and development. It is also working closely with suppliers to eliminate supply chain bottlenecks.
A key part of the strategy involves streamlining the final stages of production. “We see opportunities to reduce the testing protocol and still maintain quality,” Dassen said. “We can crank out more tools, and customers are open to that.”
The urgency stems from ASML's monopoly position in the global semiconductor supply chain. It is the sole manufacturer of EUV machines, the essential equipment required by TSMC, Samsung, and Intel to produce any advanced microchip.
Because these complex systems take more than a year to deliver after an order is placed, ASML’s shipping rate acts as a hard limit on how quickly global AI data centres can be built. This production bottleneck sits one step further upstream than the current AI memory shortage, which is already driving up consumer electronics prices.
ASML is investigating a further 30% capacity boost for 2028 to keep pace. How successfully the Dutch firm untangles its own manufacturing constraints will dictate the trajectory of the broader tech economy.
