Summary: The Information reported on July 27 that a state-supported Shanghai operator moved immersion DUV lithography tools into serial manufacturing. First deliveries are expected for SMIC, Hua Hong, and CXMT at roughly five systems in 2026 and twenty in 2027. Analysts caution that yield parity, fleet reliability, and global support—not just working tools—remain the hard tests.
The Information reported on July 27, citing two people familiar with the effort, that a state-supported operator in Shanghai has moved immersion deep ultraviolet (DUV) lithography tools into serial manufacturing. The market reacted the same session: ASML shares fell as much as 8%. First deliveries go to China’s leading fabs later this year — SMIC, Hua Hong Semiconductor, and Changxin Memory Technologies (CXMT) — on a cadence of about five systems in 2026 and roughly twenty in 2027.
Engineering Teams Pulled Together to Run a Line
The piece did not name the builder. What sources did describe was how the program was staffed: DUV engineering groups were drawn from several Chinese companies, Shanghai Yuliangsheng Technology among them. China’s first homegrown immersion DUV scanners, a separate Tom’s Hardware report says, belong to Shanghai Aishengna. SMIC has been testing a Yuliangsheng immersion unit since September 2025.
The customer roster is revealing by itself. SMIC, Hua Hong, and CXMT are exactly the names a U.S. House measure (H.R. 8170) would put on a restricted list. Building domestically hands the fabs a supply route the export controls had narrowed.
One Pass Covers 28nm; 7nm Requires Multipatterning
A single immersion DUV exposure reaches 28nm-class geometry. Reaching 7nm takes multipatterning, and each extra pass buys overlay error and yield loss — more steps, thinner margins. Parts in the Chinese stack come mostly from local suppliers, though certain critical items still arrive from Japan; when domestic vendors slip, the 2026 output goal slips with them.
ASML is the yardstick: roughly 130 immersion shipments expected in 2026, a 98.7% share of the immersion market, with another 30% of immersion capacity planned for 2027 and a further 30% under review for 2028. Five against one hundred thirty is a gap counted in multiples, not in percentage points.
Beyond a Working Demo, Production Reality
Nick Patience, AI lead at Futurum Group, set the bar bluntly: Chinese chipmakers must reach “at least yield parity, not just have a working tool.” Yields from SMIC’s imported DUV fleet, he added, still trail TSMC by a wide margin, and reliability is something fabs earn through years of field iteration. SemiAnalysis called production scale “the most underestimated part,” pointing to single-tool performance, fleet-wide performance, ecosystem holes, and poor economics against ASML machines that are already fully depreciated.
Analysts elsewhere counsel even more restraint. Stephane Houri of ODDO BHF said the announcement should be taken “with a pinch of salt,” warning that Chinese output “could be limited to the very low end.” Closing the distance to ASML, said Paul Triolo of DGA Albright Stonebridge Group, would mean building “reliable machines serially” with worldwide fab support — “quite a stretch.” A June read from the AI Futures Project was darker still: commercially meaningful Chinese immersion DUV arrives in the mid-2030s.
Controls, Supply Chains, and What Changes
Over the near term, ASML absorbs limited damage. SemiAnalysis argues the local tools “displace revenue ASML already lost to export controls” — “a tool ASML cannot legally or physically supply being built locally does not subtract from a sold-out order book.” Shares remain up more than 123% year to date. China accounted for 14% of ASML’s €6.6 billion in Q2 net system sales and about 20% across the full year, down from 33% in 2025.
Mostly, the movement is legislative. Beyond H.R. 8170, the MATCH Act — filed in April, reported by the House Foreign Affairs Committee on April 22, with a Senate companion in S.4281 — would stretch restrictions to servicing and technical support for equipment already installed. On EUV, Reuters has documented a working Chinese prototype, yet ASML spent roughly twenty years and about $10 billion on R&D, co-funded by Intel, TSMC, and Samsung, before EUV turned profitable at scale around 2018–2019. Triolo’s bottom line: “I think EUV is out of reach. Never say never… but it’s a completely different technology.”
What counts is less 2026’s five tools than whether the twenty in 2027 actually ship. That is where serial production — not one-offs — gets proven, and where yield stops being an analyst’s footnote and becomes a fab’s own numbers.