For years the bottleneck was obvious. China could design ambitious chips, but the machines that print them remained locked behind export controls. Now the country is trying something harder than designing around the rules: building the tools itself.
The machine everyone is watching
The tool at the center of the story is the deep-ultraviolet (DUV) lithography machine — the workhorse that patterns circuits onto wafers. Export rules have long barred the most advanced EUV systems from China and tightened even on older DUV gear. So a domestic alternative has moved from research curiosity to national priority.
Reuters reported in late July 2026 that a state-backed group, Shanghai Aishengna Electronic Technology Group (上海爱晟纳电子技术集团), is leading China's effort to produce its own immersion DUV systems. The reported plan targets about five machines in 2026 and roughly 20 in 2027, with the first units destined for major Chinese foundries including SMIC.
The Financial Times has separately reported that Yuliangsheng (宇量昇), another Shanghai equipment developer, is building DUV prototypes and aiming for around a dozen systems by the end of 2026, with SMIC already testing them on simpler chips.
Why a handful of machines rattled markets
The headline numbers are tiny next to global demand. China imported about 95 DUV systems in 2025 alone, according to Hong Kong China News Agency reporting cited by TrendForce — so a planned output of five to twenty homegrown tools barely dents that. Yet the symbolic weight was enough to move ASML's stock, which fell about 5% in premarket trading on the Reuters report.
The reason is the trendline, not the volume. ASML's own results show China accounted for 14% of its net system sales by shipment location in the second quarter of 2026, down from 19% in the first quarter. ASML simultaneously raised its full-year net sales forecast to between €43 billion and €45 billion, signalling that demand elsewhere is more than making up the difference. But the direction of travel — China buying less, making some itself — is exactly what export controls were meant to slow.
Huawei's quiet role in the toolkit
Huawei is not building the光刻 machines alone, but it is knitting the supply chain together. Through its investment arms, the company has backed specialist suppliers attacking the hardest sub-components. Reported investments include Fujian Zhiqi Photonics (福建智奇光子), working on the ultra-precise projection lenses that are among the toughest parts to localize, and Beijing RSLaser (北京瑞信光电), a developer of DUV light sources. These are the exact bottlenecks — optics and light — that determine whether a homegrown tool can ever match a foreign one.
This fits a broader Huawei doctrine. In May 2026 the company unveiled what it calls the "Taoding law" (韬定律) at an international circuits conference, arguing that chip performance can advance through "time scaling" — stacking and folding logic layers — rather than only shrinking transistors. The idea is to sidestep the need for the most advanced lithography entirely. How far that claim holds up in production remains unproven, but the strategic intent is clear: own more of the stack.
What homegrown DUV can and cannot do
The realistic ceiling is modest. The first domestic immersion tools are aimed at the 28-nanometre node, and analysts note that multi-patterning tricks could push them toward 7-nanometre-class output — but at much lower yield and higher cost than a mature foreign line. Against ASML's EUV systems, which remain the only route to the leading edge, the gap is still years wide.
There is also the unglamorous reality of chipmaking: a new lithography tool typically needs a year or more of in-fab tuning before it is trustworthy enough for volume production. Buying a machine is not the same as running one at yield. For the fight over compute (算力), tools are necessary but far from sufficient.
The export-control feedback loop
The pressure is not one-directional. TrendForce reported in August 2026 that Washington is weighing a near-total extension of curbs to ASML's remaining DUV sales and even servicing of machines already in China, under a proposed MATCH Act. Whether or not that passes, the mere debate sharpens Beijing's incentive to build its own tools. Every restriction that limits foreign supply is, from Beijing's view, a subsidy for the domestic alternative — which is precisely why a five-machine plan draws outsized attention.
Honest limitations
- The production targets (about five machines in 2026, roughly 20 in 2027) come from Reuters and FT reporting, not from a company-confirmed production schedule.
- ASML's 14% and 19% figures are the company's own reported shares of net system sales by shipment location, not China revenue, and exclude services.
- We have not independently validated the specifications, reliability, or yields of any homegrown DUV system; claims about optics and light-source progress remain vendor- and media-reported.
- Domestic DUV, even if delivered, remains well behind ASML's EUV-class capability and is not a near-term replacement for leading-edge manufacturing.
What readers can do now
- Track ASML's quarterly China share as a live barometer — a falling percentage can mean either weakening demand or rising local substitution.
- Watch the supplier layer, not just chip launches — follow Huawei-backed equipment investments (optics, light sources, packaging) to see where self-reliance is actually progressing.
- Discount any "China replaced ASML" headline — compare unit counts, node coverage, and yield, not press releases.