China has built a prototype EUV (extreme ultraviolent) lithography machine in a high-security laboratory in Shenzhen. They had a semiconductor Manhattan Project”. The prototype was completed in early 2025 and is now undergoing testing. It occupies nearly an entire factory floor and was assembled using reverse-engineered components.
Ramp-up timeline
Functional for chip production around 2028.
More realistic full mass-production capability is around 2030 or later.
In 2030, ASML/TSMC are expected to be on high-NA EUV or next-gen nodes. China will be about 1–2 generations behind initially.
Competitiveness assessment
Significant milestone — China has overcome initial barriers faster than many Western analysts predicted, showing the power of state-directed resources, talent recruitment (high salaries/bonuses for PhDs/ex-ASML experts), and reverse engineering.
Still far from commercial parity — The prototype lags in throughput, yield, reliability, and full integration compared to ASML’s mature systems (which took ~20 years to perfect). Yields and costs would likely be much worse initially, and China still depends on imports for some subsystems (e.g., advanced mirrors, resists).
It was built by a team of former engineers from Dutch firm ASML, the world’s leading supplier of photolithography machines that etch circuit patterns onto silicon wafers to create chips. They reverse-engineered ASML’s EUV machine using parts from older ASML machines obtained in secondary markets, the report said, citing two people with knowledge of the project.
China’s prototype EUV machine was yet to produce working chips, the government has set a goal to make advanced integrated circuits on that equipment by 2028.
EUV systems are extremely complex, built from hundreds of thousands of parts across multiple geographies, with no single blueprint.
“Even if China were able to produce some EUV light [from its prototype], turning this into an economically viable manufacturing process could take years, if not decades,” Morningstar analyst Javier Correonero wrote in a research note on Thursday.
China’s EUV efforts are up against “arguably the best supply chain integration ever built”, Correonero said, pointing out ASML’s advantage.
Beyond high-end optics, he said ASML’s EUV machines contain “ultraprecise mechatronics” – the integration of mechanical engineering and computer science, alongside advanced vacuum systems, metrology and software.
According to a research paper published in March, former ASML head scientist Lin Nan led a group of engineers in developing an EUV light source platform that operates at internationally competitive parameters, which offered hope to domestic production of advanced integrated circuits. Lin was among the engineers involved in building the prototype EUV machine in Shenzhen, according to the Reuters report.
Chinese lithography systems maker Shanghai Micro Electronics Equipment, according to a patent filed in March 2023, had also made progress in EUV radiation generators and lithography equipment.
China’s market for lithography machines was 99 per cent controlled by ASML and Japanese suppliers Nikon and Canon.
China manages to meet around 2 per cent of the domestic demand for process control tools in chipmaking from local suppliers, according to TechInsights data. That percentage on the lithography side was about 1.5 per cent, and about 10 per cent on the entire wafer fab equipment sector.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Doesn’t reverse engineering imply that their IP has legal issues and that they can’t ever sell anything overseas?
Making high precision machinery and processes work requires relentless obsessional attention to detail that Chinese industry do not generally reflect with their normal ‘cha bu duo’ (almost good enough) approach to everything. Their CNC machines for example are junk compared to western manufacturers.
Achieving sub-nm level positioning accuracy through 10 or more EUV exposures needed to make a cutting edge chip is a ludicrously difficult process to reverse engineer. It will take them another 5-10yrs to catch up to ASML, though they might be able to short-circuit the UV source with a free electron laser.
According to colleagues who have been in Chinese factories, you should not mistake the garbage they ship abroad as a matter of national policy with what they manufacture for domestic use.
They’re perfectly capable of good work if they’re not out to cheat you.
It’s not about cheating. Chinese will gladly build you what you are willing to pay for. If you want cheap, then that’s what you get. If you are willing to pay a premium, then a premium product is what you get. It’s a pretty simple concept. The problem is 99% of buyers aren’t willing to pay the premium, and just keep pushing on price. Then it’s a race to the bottom…
Right on. Everywhere.
“China has built a prototype EUV (extreme ultraviolent) lithography”
Straight from a Kubrick movie.
Humans are trying to play God, but with non-organic chemicals. The next step is to use organics, such as another COVID-2030. China might lead the way again, as it did with non-metallic buildings, the hardcopy printing press, explosive chemistry, the recycling of organic waste, and high-density farming.
I am waiting for the day Javier Cordoned to eat his word about China taking a decade to come up with its EUV machine.
If you think China isn’t already ahead of this… you are delusional. They certainly aren’t showing you or telling you what they are capable of.. probably already have smaller prototypes running.. they are ahead with riscv memory and light. It will be wild to see the new technology coming from China.
You are right. Chinese is famous of using strategy Sun Tzu Art of War. You will never know their real progress. They will announce to the world they have done it when time come to the surprise of many.
They will not close any gap…. They have only light source…. From this stage to print it on wafer took ASML 20 years….
I know light can be destructive, but violent, even ultraviolent?
Extreme Ultraviolence usually solves the problem whatever it is.
This isn’t as huge of a revelation to me as it would’ve been a few years ago, but it also doesn’t read [at me] as if China is going to suddenly make a leap forward in the market (even in their own market) in terms of production. Going to do some more homework on it. This type of thing has always been of interest to me. Not mostly because of geopolitical stuff, but just because I like seeing who can do what with the tech they get their hands on.
I’m also still interested to see what might be coming out of Japan in terms of semiconductors over the next couple of years.
This post only scratches the surface of what is needed
for an EUV based ecosystem.
I would suggest that anyone really interested should watch the
various EUV related videos posted on the Asianometry youtube channel.
Great channel.
I can’t help but think that when you need machines the size of a large building to manufacture products the size of a postal stamp with features the size of molecules, that you’ve gone down a fundamentally wrong road.
Why?
These are the most useful tools(computers) ever made.
No, I’m not suggesting the products aren’t worth employing them, just that the huge disconnect between the size of the product and the size of the machinery used to manufacture it suggests to me that we’re manufacturing them in the wrong way. As far as I understand it, it IS possible to manufacture state of the art ICs using desktop machinery, just one-off, not mass production, using processes like direct write, instead of extreme ultraviolet lithography.
Maybe a factory full of desktop machines would be better than one colossus. Certainly the capital barrier to getting into the business would be enormously smaller.
Nanoimprint lithography machines are smaller and cheaper than EUV and almost ready for prime time… but the defect rate makes them a better match for memory than logic chips.
That said I think they should scale EUV machines bigger lol. A free electron laser could efficiently power a cluster of a dozen EUV scanners
Direct write lithography gets exponentially slower as feature sizes decrease and number of features increase. If you try to use multiple beams to increase throughput, then pattern fidelity gets much worse as you decrease feature size and increase number of features.
Have you seen the inner workings of an EUV?
They are the most complicated machines ever produced and for good reason
I understand your point of view but if you think about it: we need an object the size of a solar system to have at least one planet in the right place to have macroscopic biological life, developed from microscopic cells constructed from nanoscale proteins. Maybe having very ordered stuff requires very big systems. Obviously a solar system and biological life are not engineered, but maybe there is some underlying principle about entropy. We know that once developed life started to self replicate reducing the “manufacturing scale” from the whole ecosystem of the primordial oceans, to something the size of a cell. That’s a shrinkage of 10 orders of magnitude.
We might think of developing nano machines with self replicating circuitry but I am not sure we want to go that way.
Terrifyingly brilliant
Its at the absolute limit of our technological capabilities the machines are huge and a single spec of dust can ruin the entire yield. I don’t doubt China will figure it out eventually but 2030 is laughably quick even the most generous estimates say 2035.