Credit: ZBMAUE oopAAM via Wikimedia Commons (public domain) China’s effort to build its own chipmaking machines finally has a coordinator. It is Huawei. And the flagship machine it is coordinating still runs on German lenses. Eleanor Olcott and Zijing Wu reported the arrangement for the Financial Times, citing multiple people familiar with the matter.
TNW has not independently verified it. Huawei is investing across the industry that supplies lithography equipment. It is also helping those suppliers get their machines into Chinese fabs, including SMIC. Lithography is the step that prints circuits onto silicon.
It is the one part of chipmaking China has never managed to replace. What the arrangement actually is Lin Qingyuan, a semiconductor analyst at Bernstein, gave the FT the clearest description of it. “Huawei is playing a project management role in China’s DUV push,” he said. A prototype on its own achieves nothing, he added. The work needs thousands of vendors and customers moving together, and “a central orchestrating power to bring this effort together”.
That describes a role nobody previously held. China has had lithography ambitions and scattered suppliers for years. What it lacked was somebody with enough money and leverage to make a fab commit to an unproven domestic tool. The company at the centre, and a denial Yuliangsheng, a Shanghai equipment maker, co-developed the first Chinese-made advanced deep ultraviolet machine.
SMIC is testing it, and so is Huawei on its own production lines. The company was spun out of SiCarrier. The FT has previously reported that SiCarrier has close technical and business ties to Huawei. Huawei has denied any affiliation with SiCarrier.
That denial sits awkwardly beside the reporting, and both belong in the record. We have met Yuliangsheng before. In July, TNW reported that a state-owned group, Shanghai Aishengna, was absorbing teams from Chinese lithography startups, Yuliangsheng among them. The FT now describes Yuliangsheng as a SiCarrier spinout building its own machine.
Those two accounts are not the same, and nobody has confirmed either on the record. The corporate structure of Chinese lithography is not a settled fact. Anyone writing about it should say so. The Zeiss problem Here is the detail that undercuts the whole project, and the FT put it in plainly.
Yuliangsheng uses projection lenses made by Zeiss, according to two people familiar with the matter. Zeiss said it sells its lithography optics only to ASML. Industry insiders told the FT the lenses could be obtained on the secondary market in China. Nobody in that paragraph alleges anything unlawful, and neither do we.
But look at what it describes. A machine built to escape export controls currently depends on a component its manufacturer does not sell into the country. Huawei is working on that. Its venture arm Habo and a second Huawei-backed fund have both invested in Fujian Zhiqi Photonics.
That company is an ultra-precision optics maker, spun out of the Shenzhen-listed group Castech in 2022. The fund took a 37% stake for about $15m in 2023. Castech is worth a second look. It traces its origins to a Chinese Academy of Sciences institute and makes optical crystals for laser systems.
It has also disclosed that both ASML and Huawei are customers. So a supplier to ASML has spun out a company now funded by Huawei to displace ASML’s lens supplier. The other bottleneck is light A DUV machine has tens of thousands of parts across roughly a dozen major component groups. China lags in two of them, and the second is the light source.
Japan’s Gigaphoton and Cymer, which ASML owns, dominate that market. “Building a light source is easy,” Lin said, “but having one that works very stably at such high power is difficult.” Beijing RSLaser Opto-Electronics is the domestic effort, and Habo has invested there too. It has supplied light sources for older machines from Shanghai Micro Electronics Equipment. Chinese media report it is now working with SMEE on a source for a 28-nanometre tool. How far along this is Yuliangsheng expects to produce 12 DUV machines by the end of this year.
The FT previously reported SMIC testing a 28nm Yuliangsheng tool. SMIC pairs it with multi-patterning, a technique that runs a wafer through the machine repeatedly, to reach 7nm chips. That sounds closer than it is. UBS analysis this month put Chinese lithography at roughly where ASML stood in 2004.
Lin told the FT the machines will take several years to approach ASML productivity. The commercial evidence agrees. Fabs including SMIC have bought the domestic tools and are testing them. They still run production on ASML.
The Dutch company’s China sales are projected at about a fifth of its revenue this year, and CXMT and YMTC have reportedly stockpiled enough ASML machines to cover three years of expansion. Meanwhile the target keeps moving The gap being closed is not stationary. ASML spent this week working with TSMC on a shift to 12-inch photomasks. The point is to let its High NA machines print the largest AI chips in a single exposure.
So China is testing 28nm DUV while ASML redesigns the top of the market. That is the honest shape of it. Domestic tools are real, they are being installed, and they are chasing a moving target with two component groups still unsolved. The argument that should worry Washington Lin ended on a point that cuts against the policy driving all of this.
US restrictions could accelerate China’s progress rather than slow it, he said, by pushing local fabs towards local suppliers. “If the restrictions push local fabs to collaborate with local DUV suppliers, they could increase their technological capabilities much faster than expected.” The pattern is familiar. Export controls already pushed Chinese firms towards custom silicon rather than stopping them. A fab does not lack the will to buy domestic. What it lacks is tolerance for yield loss while an unproven tool matures.
Restrictions remove the alternative that makes that loss intolerable. What to watch Whether the 12 machines are delivered, and to whom. Whether Fujian Zhiqi ships a lens that replaces the Zeiss part. That is the clearest single test of whether this becomes a domestic supply chain or stays an assembly of imported components.
And ASML’s China revenue line. It is a fifth of the business, and the most-watched number in European technology. When it starts falling for reasons other than export policy, this project will have worked.
