Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
"Do not press a desperate foe too hard" -SunTzu.
The irony is that U.S. politicians managed to do what the Chinese government failed for years: Force Chinese technological companies to view local tech as an viable alternative, force the collaboration between companies and China research institutions to narrow technology gaps and opened the eyes of many pundits in the industry who though that their supply chain was safe. The U.S sanctions proved to everyone in China that the Chinese government was right.
Huawei could have gotten into the semiconductor manufacturing business years ago and I can assure anyone that they would have been very good at it. But they were sold on the idea that TSMC and Qualcomm are safe suppliers, they were wrong.
Last week I sent an article I wrote in Seeking Alpha on ASML and China, and promised a new article this week. See it attached as it was published today. It's entitled
ASML's Current Battleground Now In China On Several Fronts
I welcome comments, which will help in future articles, as I am in the US and most readers are closer to what's going on in China.@tinrobert Sir I really love this part (in bold), please let us know and share with us news from your sources, we're tracking it in China publication and rumor mills and it seems that this Loch Ness monster do exist.
Back in 2020, there were media reports that a SMEE model SSA800 capable of 28nm resolution would be on the market by the end of 2021, which would be followed by the SSA900 capable of 22nm. Keep in mind that using multiple patterning, resolution could get to 7nm.
Like discussed in the article sanctions on immersion lithography sales to SMIC and other major Chinese semiconductor industry players could have a major impact on ASML's sales. If, for whatever reason, they sanction sales of immersion lithography to China this will provide SMEE a major impetus to not only produce their own machines but also to massively ramp up their production.
@tinrobert mentioned how SMEE had ordered EUV machines from ASML and they were never delivered. CXMT, which is the leading Chinese DRAM manufacturer, also had EUV on its product development roadmap at one point. CXMT's CEO at one point visited ASML to discuss the purchase of equipment. I would not be surprised if sales of EUV equipment were discussed at one point. CXMT, as well as the other Chinese DRAM vendors, have all had friction with Micron. CXMT might end up in a US sanctions list as well. In the case of YTMC I think any sanctions they apply to them will have much less impact since YMTC manufacture V-NAND which uses lower resolution processes with multiple layers.
I doubt the sanctions on SMEE will have any major impact. SMEE is an integrator much like ASML. They assemble products and add the secret sauce to get it all to work together. Their supply chain AFAIK uses no US origin hardware components. In fact all their major components should have Chinese origin.
The numbers SMIC has released on wafer revenue by technology are more obfuscated than they used to be on the past. FinFET/28nm include the 14nm and lower resolution processes i.e. 10nm. IIRC FinFET used to be specifically enumerated in SMIC's reports reports but not any more.
I also agree with one of the commenters that a supply glut over the next two years is definitively possible. Not only from all the new Chinese and other fab construction but also that if, for whatever reason, COVID-19 subsides and people get back to work and children back to school worldwide then hardware sales will drop off a cliff.
I have some doubts about the future success of the Nikon machine. Few companies bothered with the transition towards FinFET. Those which did all have EUV roadmaps. Even Intel now claims they will use EUV. How easy would it be to integrate Nikon and ASML lithography machines in the same production pipeline? The market leaders are transitioning towards 3nm production in the future and that will only happen with EUV. Nikon might at best get some of the companies which did not bother transitioning to EUV yet. In a market like China, Nikon could also find some success given they can't even buy EUV machines. So vendor lock-in by ASML would be much weaker in China than other markets. However if the US does indeed sanction immersion DUV sales this would basically kill any future prospects for Nikon in my opinion.
In the medium term I do not think China will have an impact on EUV sales for ASML. Since China cannot get EUV machines for their own foundries they will just have to purchase services or chips abroad. So the machines will still get sold just not to China. What would cause a major impact on EUV machine sales would be a blanket ban on sales of EUV produced chips to China. This cannot possibly happen, I think, given all the manufacture industry that is in China for all sorts of electronics components. This isn't like putting sanctions on Iraq in the 1990s where back then they couldn't even import a Sony PlayStation 2 console because that had too much compute performance and was classed as a "supercomputer". The US has struggled with this dilemma all along. They use blacklists to sanction some Chinese entities but they can never do a blanket ban because it would impact the world supply market and especially their own companies. Among others you would decimate world smartphone, tablet, personal computer, and server production. This means the sanctions will be weak.
@tinrobert mentioned how SMEE had ordered EUV machines from ASML and they were never delivered. CXMT, which is the leading Chinese DRAM manufacturer, also had EUV on its product development roadmap at one point. CXMT's CEO at one point visited ASML to discuss the purchase of equipment. I would not be surprised if sales of EUV equipment were discussed at one point. CXMT, as well as the other Chinese DRAM vendors, have all had friction with Micron. CXMT might end up in a US sanctions list as well. In the case of YTMC I think any sanctions they apply to them will have much less impact since YMTC manufacture V-NAND which uses lower resolution processes with multiple layers.
I doubt the sanctions on SMEE will have any major impact. SMEE is an integrator much like ASML. They assemble products and add the secret sauce to get it all to work together. Their supply chain AFAIK uses no US origin hardware components. In fact all their major components should have Chinese origin.
The numbers SMIC has released on wafer revenue by technology are more obfuscated than they used to be on the past. FinFET/28nm include the 14nm and lower resolution processes i.e. 10nm. IIRC FinFET used to be specifically enumerated in SMIC's reports reports but not any more.
I also agree with one of the commenters that a supply glut over the next two years is definitively possible. Not only from all the new Chinese and other fab construction but also that if, for whatever reason, COVID-19 subsides and people get back to work and children back to school worldwide then hardware sales will drop off a cliff.
I have some doubts about the future success of the Nikon machine. Few companies bothered with the transition towards FinFET. Those which did all have EUV roadmaps. Even Intel now claims they will use EUV. How easy would it be to integrate Nikon and ASML lithography machines in the same production pipeline? The market leaders are transitioning towards 3nm production in the future and that will only happen with EUV. Nikon might at best get some of the companies which did not bother transitioning to EUV yet. In a market like China, Nikon could also find some success given they can't even buy EUV machines. So vendor lock-in by ASML would be much weaker in China than other markets. However if the US does indeed sanction immersion DUV sales this would basically kill any future prospects for Nikon in my opinion.
In the medium term I do not think China will have an impact on EUV sales for ASML. Since China cannot get EUV machines for their own foundries they will just have to purchase services or chips abroad. So the machines will still get sold just not to China. What would cause a major impact on EUV machine sales would be a blanket ban on sales of EUV produced chips to China. This cannot possibly happen, I think, given all the manufacture industry that is in China for all sorts of electronics components. This isn't like putting sanctions on Iraq in the 1990s where back then they couldn't even import a Sony PlayStation 2 console because that had too much compute performance and was classed as a "supercomputer". The US has struggled with this dilemma all along. They use blacklists to sanction some Chinese entities but they can never do a blanket ban because it would impact the world supply market and especially their own companies. Among others you would decimate world smartphone, tablet, personal computer, and server production. This means the sanctions will be weak.
Last edited:
I'd assume that particle accelerators such as sychrontrons are easily capable of producing EM photons as energy intensive as the lowest wavelength gamma rays in large quantities, as such they should be able to produce EUV photons much more easily, right? Given the gamma photons are much more energy intensive than EUV photons, more energy must be dedicated into producing the former in comparison to the latter in a particle accelerators, right? Steady state microbunching is just about applying ways in which produce and then concentrate coherently a large beam of photons of any wavelength or range or wavelengths, and with regards to EUV lithography that would be EUV photons, right?
Chips and foundry services are different things.
I agree it makes no sense to ban advanced chip sales to China, actually the aim is to maximize the chips sold to China.
Instead stopping foundries from accepting to manufacture Chinese advanced designs it is a much more realistic and possible outcome, and also easier to enforce. This would mean to keep China buying foreign chips, but also stop it from developing independent IP and growing Chinese design companies. The best of both worlds.
The next 5 years will be critical. I would be surprised if they just pass without something unexpected happens.
Skywatcher
Captain
It's odd how all the Chinese related EUV presentations at EUV Litho (apart from the SSMB) seemed to have stopped as of 2016.
ZTE was placed on the Entity List in 2016.
I imagine ZTE was screaming for alternative (100% Chinese) semiconductor fabs at that point.
Just speculation.
Appearance can be deceiving their research in EUV seem quite active, the shift from immersion lithography to EUV is pretty noticeable as they move to commercialize the former. Hopefully they will commercialize EUV as soon as possible.
- Status