Can China build its own ASML?
nikkei.shorthandstories.com
nikkei.shorthandstories.com
I remember when their cars were a joke, and when their cellphones were cheap trash. Now I don't think I'd buy a non-Chinese new car or cellphone, lol.
Producing 10x of 10-years old GPUs is not exactly useful for modern AI codebases for instance.
This matters for AI datacenters in particular, if they want to be autonomous in building them they need to be able to build advanced microprocessors locally.
But I agree that, given enough time, China should be able to. But Volvo en Tesla handing over IP themselves, and phone makers letting China produce and assemble phones (maybe not as extensive these days anymore) is something different.
EDIT: I mentioned Saab when I meant Volvo. Not sure if people downvoted based on that, I have no info on Saab so consider that as a mistake.
China has Money and Smart people. (and very effective corporate / nationstate espionage) so they can most certainly reproduce advanced machines.
They might not have incentive yet to do it because it will not make them popular, and potentially output products would be banned on US/EU. They play a long game and want US and EU consumers to ask tehir governments to please allow the chinese products.. so their market share is safe and stable.
I think once there is enough incentive for them they would do it. They simply do not want to do it currently.
> people overstate complexity of ASmL machines. They are not impossible to make or use, specialist work sure but its possible. The only reason why no one does it is: 1) IP laws, 2) Costs
The real reasons are: enormous complexity, lots of original research involved, deeply specialized supply chains. Recreating that takes lots of time and money. Even if you cut corners and steal IP. So that leaves China with 'time' as the real impediment. And who knows, AI will be a boost to get to that goal?
And they only allocated a small number of employees and resources to bring their first EUV machine to market?
ASML had 7,685 employees prior to their EUV success (Including contracts), of these 7,685 employees only 2500 were engineers. ASML put half of these engineers to develop the NXE:3100 and they only spent 10 billion dollars.
So total resources = 1250 engineers, 10 billion to make EUV a reality and 13 years.
So yes the complexity and difficulty of EUV machines is vastly overstated. The amount of resources spent on it is coins found in couch cushions to huge mega corps.
This isn't to say that automated translation is bad, I use it extensively, but not all interaction worth having is online.
Many cultures will probably much rather interact with a foreigner speaking proficient AI translation, than a language-learner speaking with a poor accent and constant grammatical imperfections. Go to the Netherlands now as a Dutch learner who hasn't reached B2 level, for example, and local people won't like it, it just wastes their time compared to using English.
No, not in this case.
> I remember when their cars were a joke, and when their cellphones were cheap trash.
Any country can make cars and phones now if they REALLY want to, they just don't because they wouldn't be competitive on the global market at performance and at scale. But not anyone can just make an EUV machine if they want to.
Becoming a leader in commodity white goods like a phones and cars is a different beast than EUV machines. The challenges are not even remotely comparable.
China doesn't have a Zeiss, it doesn't have an ASM, it doesn't have a Cymer, it doesn't have a Trumpf, and it doesn't have a dozen other domestic competitors to western suppliers of critical parts that make an EUV machine.
And each of those suppliers, like Zeiss for example, is a leader because it has decades of expertise that can't be speed-run, even with IP theft. So before China can have an ASML it first needs a Zeiss. So ask yourself where is their Zeiss and are they beating the west's Zeiss?
The "moat" we're talking about with private companies is that it's very hard for competitors to get enough funding to compete, and private investors are unwilling to invest in a company that will compete with a big established player. That's completely different when a state has a strategy and decides to invest to achieve a goal.
They were promised upwards mobility, but their leaders need factory workers, more than ever.
Chinese factories are more automated than ever. More and more skilled chinese do skilled white collar work not repetitive factory assembly work.
It might take them a decade, but they'll get there. Working on the basis that you can somehow stop China rather than merely delaying them is a comforting illusion.
Where's the proof in that? China can't even create jet engines that outperform dated soviet designs even when they bought their IP. Some things are just almost impossible to do without the right people at the right time, and you can spend virtually unlimited money and still never get there.
Russia does not have a 5th generation engine core in even the validation phase.
China has a 5th generation engine core the WS-15, it has not only passed validation but is currently in mass production.
A true 5th generation engine core allows true supercruise
You can see the gap between the F-22, J-20 and Su-57
The F-22 supercruises from mach 1.5 -> mach 1.8 (This is because they have 5th gen engines, and is it burns its fuel stores)
The J-20 supercruises at mach 1.4 -> mach 1.8 (This is because the J-20 has a much larger internal fuel load, vs the F-22 as it burns this fuel it speeds up)
The Su-57 super cruises at mach 1.3 -> mach 1.5 (This is because it has a 4th gen engine, the actual 5th gen engine intended to power the SU-57 is stuck in development hell the Izdeliye 30 (or AL-51F1) so Russia put a souped up 4th gen engine inside the SU-57 the AL-41F1 which is the same core as the AL-31F. If they do complete their 5th gen engine the SU-57 is expected to supercruise at mach 1.8-> mach 2.0 as it burns up its fuel. Su-57 accepts more stealth penalties in exchange for superior aerodynamics.
It has the Changchun Institute of Optics, Fine Mechanics and Physics (CIOMP)
Founded 1952. One of its things:
>4-Metre Silicon Carbide (SiC) Mirror: CIOMP successfully manufactured the world's largest single-piece Silicon Carbide mirror blank. It provides the optical backbone for high-resolution space telescope imaging systems
Apparently carbide mirrors are used for EUV lithography.
Nature article + pic of mirror https://www.nature.com/articles/s41377-022-01050-w
So are modern Western cars - and arguably also phones.
If I'm forced to drive in a bug-riddled impossible-to-repair privacy-invading spaceship either way, why would I go for the overpriced outdated Western one rather than the affordable innovative Chinese one?
I would prefer a boring 2000s car with an electric drivetrain, but it's not like anyone is making them...
Vendor-lock is quite high, possibly higher in the US.
Look at iPhone, hardware is behind, restricted, unauditable privacy and questionable software in terms of performance (especially starting iOS 26, where you lost the choice for Liquid Glass).
Impossible to reasonably sync AirPods or watches, Macbook or choose the firmware version.
In comparison, the irony is that China offers the freedom (like with Qwen).
On these supposedly evil Chinese devices (phones, headphones, watches, etc), you have an open platform with great hardware, that you can modify as much as you want, and the only people who restrict you is... Google (making sure you cannot root without losing to important apps, so technically you can root, but practically you can't).
All the claims about "China = bad hardware and bad software" were true 20 years ago, but this is not the case anymore.
That being said, should certainly be quite cautious about default setup... you have US surveillance, PLUS, Chinese surveillance.
btw after one pixel with swollen battery I'm never buying one again, but I have to admit they have great cameras, reasonable dimensions and easily unlockable bootloader, sadly it's wrote pointless with faulty unreliable hardware and shit bare bone software
My in law has a BYD Seagull, fail to see how it lags compared to similarly priced cars.
Chinese cars are incredibly popular in Europe, because they are better cars at the same price tag, even the more expensive ones.
I think it's asinine to think their high end manufacturing is crap, keeps us non competitive, let alone ignoring most of our stuff or large parts of it already comes from China.
And it has nothing to do with wages, modern high end manufacturing is highly automated and skilled engineers are as expensive as in southern or central Europe.
It's like those people thinking Aliexpress/Temu products are subsidized by Chinese gov, not realizing that Ali/Temu is extremely expensive and you can buy products there for like 1/2 - 1/10th of their price in Chinese online shops (especially Pinduoduo), people claiming China is subsidizing this and that don't realize how cheap is to produce products. Been in China last summer, bought 4 big suitcases full of things mostly from Pinduoduo.
What have you tried so far on a daily basis to say this?
Have you tried latest Huawei models?
By the way, Foxconn makes iPhones, does it count as Chinese quality or not?
as for the cars, I'd agree maybe compared to most western brands, but the question is whether is worth paying 30-100% extra for only slightly better western car
This doesn't mean they will always be behind though.
If I was ASML I would have an AI generated honeypot of techniques that are plausible but incorrect for China to go after on this and make sure you get hacked by them.
Washington, via the recent MATCH Act, has requested ASML to not supply or maintain any of the older equipment.
ASML has travelled to Washington to tell US Congress to take a long walk off a short pier.
The only thing to stop them is the will of more powerful states.
If you take Iran as an example, they would have possessed Nuclear weapons long ago if it wasn't for USA/Israel as Iran as a state will throw everything at it.
p.s. And that's a very detailed and well laid out visual article above usual broadsheet standards.
A nation state will fail at most things, because there's no budget, and there's commercial pressure, democratic pressure etc.
One of the few things to make it work is lack of competition. Iran: they have a nuclear program because they can't buy what they want from the market.
So far they have announced road maps and benchmarks for their upcoming products using this. A new Kirin-series phone/laptop chip and an Ascend AI accelerator - stated performance comparable to leading US products made with EUVL.
Products are due in August.
https://www.huawei.com/en/news/2026/5/ieee-iscas-tau-scaling
[Shanghai, China, May 25, 2026] Today, at the 2026 IEEE International Symposium on Circuits and Systems (ISCAS), He Tingbo from HUAWEI delivered a keynote speech titled "New Semiconductor Path in Practice". In her speech, she presented the Tau (τ) Scaling Law, a new principle for guiding the future development of the semiconductor industry. This law proposes replacing geometric scaling with time (τ) scaling as a new guiding principle for the evolution of both semiconductors and electronic systems. Based on this principle, innovative technologies such as LogicFolding can be used to continuously compress signal propagation delay and steadily improve transistor density, which will drive the ongoing evolution of semiconductors and electronic systems.
Looking into it, they are "folding" a logic chip to have shorter paths with their SRAMs, which is the main application of 3D integration these days, chucking memory (low-power) above high power logic die.
stacking logic chips is hard because they have to dissipate a lot of power, and stacking them multiplies your power density (which is already very high for the logic die, generally) by the number dies you stack~effectively. Most of the progress on these sorts of things have just been stacking layer on layer of memory on a single logic die.
It looks like they are folding other components other than memory onto the logic die which I haven't seen before, but this isn't cold fusion and even intel has thrusts doing this type of 3D integration.
R&D is difficult and expensive which is why the US and Europe gives up on many cutting edge technologies and relies on the tried and true and discounting opportunities. This will be the end of their dominance.
US sanctioning GPUs to China, then blocking access to latest SOTA models will impact everyone in the world negatively.
We need competition, Europe is not an option anymore, hopefully Chinese labs will keep open sourcing SOTA models
Based on the trend, if China wins, everyone benefits, if US wins, only Americans will benefit and everyone else will suffer
Protectionism policies are blocking low prices (what about tariffs ? what about monopolies ?) and are benefiting only the ones holding the knowledge and factories at the detriment of the rest of humanity.
Can't wait for ASML / TSMC / Zeiss equivalents so we can have access to memory sticks and GPUs / AI accelerators to run Qwen Super-Large distilled on Claude Zulu model, GTA VIII or whatever will come at that time.
They basically (almost) did. In 2013 ASML bought US Cymer(the maker of EUV light sources) and in 2001 ASML acquired the US-based lithography equipment manufacturer Silicon Valley Group (SVG) after it had encountered liquidity issues.
Basically, the US had everything needed, the EUV light sources and lithography machines, just spread over different companies facing financial issues, so ASML came at the right moment and bought them all and integrated into their own business.
Just because ASML is based in NL doesn't mean all of its EUV secret sauce IP is domestically Dutch. Most of it comes from the US, which is why the US government maintains such a high influence of ASML's trade restrictions.
It is true those companies you referred to were bought and their IP has helped ASML in _EUV_, but they had 20 years of European IP before that, all related other parts of the lithography machine (and not EUV) that made ASML the company it is.
If you think that complexity is only in (light) source for EUV (is it already profitable for ASML, by the way?), then you have no idea the complexity of that machine!
Disclaimer: I worked in their metrology department.
Was any of that 20 year old European DUV lithography machine IP superior to the IP from the other DUV champions, Canon and Nikon?
So it doesn't answer my question if ASML had superior DUV IP.
The main flabbergasting stuff about Coca-Cola is the fact that they choose to use HFCS in some countries, and sugar in others, and this is due to... protectionism again that artificially skews competition like for electronics.
China’s willpower and centralized deployment of that willpower is legendary.
The US is ran by a moronic kleptocracy of conmen and number-go-up techbros that give precisely zero shits about human rights and improving life for the average citizen.
There is no way to defend the moral or political superiority of the west anymore. And who needs to worry about foreign intervention destabilizing internal affairs when we let the extremists, the populists do that for us?
At least if China pulls this off we might get some semi-affordable RAM and SSDs.
Though to be fair, I think everyone knew that China was always going to have their 100% domestic chip manufacturing supply chain. I'd argue that the blocks were mainly a delaying tactic by the USA oligarchy. Simply blocking ASML from doing business with China would in itself motivate China to move faster, but I guess the decision makers and their advisors calculated that it would be slower than letting China buy the machinery and reverse engineer it.
Of course that didn't really work out. The only reason the media is picking up on these stories is that China, did, get their hands on the machinery, but then... of course they did.
Because the light source comes from the US.
the qn is if it will be what china wants and needs
more likely something better or more suited to the ecosystem in china will emerge
Meh. For a year or so, his computers will work a little faster. Unless they power an armageddon-device, it's just a short head-start in a long race.
Can China build its own ASML? NEVER! Western brains and economic greatness cannot be equalled!
Yes, China will catch up. It won't give them anything but a short lead.
I do however wonder how long China can afford to pay the price for this. China is effectively heavily subsidising it's export and research efforts to catch up. But subsidizing with what? They don't have a magic money tree.
To flesh it out a bit more: people in the West complain that the Chinese are unfairly competing, by selling their stuff - cars, batteries, chips etc - at an unfairly low price, either via subsidies or an artificially low exchange rate (or both). There's also IP theft, sure, but there's plenty of home grown IP too.
So this means that when German workers build a car, they get paid a lot more than the Chinese workers. The Chinese AI researchers get paid less than the Americans - for the same work. Sure, it's a mix of patriotic fervor, different purchasing power of the salary (agricultural goods are cheaper in China because farmers get paid less for the same work), etc etc. Subsidies are the same thing scaled up - Chinese government collects taxes, i.e. makes Chinese people work more for the same ultimate outcome.
But this extra economic heft is not coming from nowhere. China is channeling funds that it could spend elsewhere, but instead spends on high IP industries.
Europe and US could do the same. But at a cost! They could tax people more, who would have less money to spend (making them unhappy and affecting GDP growth), or cut investment elsewhere. You could finance it with debt, and we all know how well this tends to go.
I don't know enough about the Chinese economy to even try to have an opinion as to where this money is coming from, and what they aren't spending it on but should. But the unwind is coming sooner or later. You can't subsidise something forever. You can't grow your GDP with subsidies. So what's going to give?
It might be that they are hoping to kill foreign industry, become a monopoly and milk it, like the US under Trump is trying to in smaller scale. But of they were a free economy, it would be a race - can they stay solvent for long enough? Liberal or not, they are still an economy, and you can't beat gravity forever. I'm curious how it plays out.
They do though. The advantage of state capitalism is that at anytime you can switch from a fully capitalist economy, where the market is everywhere save the family cell, to a statist economy where the state decides what goes where, to anything in between. If the CCP says 'retirement won't be paid in RMB but in foodstuffs for everyone, and the funds are reappropriated', they kill their debt.
China is much more free at moving them from one place to another, in particular allocating them differently to the hedonistic preferences of their citizens.
But it's still only moving not creating. For now it seems, the thing that needs to give is the workers' consumption. What else? No idea.
But they are not the only way, you can go back to a more autarkical model for the rural population, or to a more directed economy like during WW2, or a mix of both. Also, if i want to be exact, i don't think China is totally capitalist yet anyway, you still have rural villages that does not allocate all resources through the market.
More than what Intel could do with all their money and access.
Also ASML is not alone. In fact ASML can not exists without its partners, first and foremost American Department of Energy but also s IMEC in Belgium and Zeiss in Germany. Those are all SOTA in either R&D or production. They are not "some of the best" they are literally the best in their very complex expertise. In fact they have to collaborate reach such level.
... and yet, it's "just" that. There is nothing magical about ASML. Yes it might be practically impossible because of IP, economics, etc but still China (or anyone else) can definitely pour a lot of resources to try and make significant process. Will the result be competitive though in light of the moat ASML has, in particular partnerships, that is hard to imagine.
PS edit : I did like Chris Miller's "Chip War: The Fight for the World's Most Critical Technology" and FWIW did have a former colleague working at ASML and been invited to IMEC events.