Great, now China can ban purchases of American chips and chip tools in the future without any recourse.U.S. Senate eyes tightened restrictions on Chinese semiconductors
Bro, I've been posting their mature node expansion for weeks. Especially for power chips segment. There is no question they are on a great path here.A post I want to share with you from a mentor I truly respected. Like what I said before, the US can have the 3nm chips, while the Chinese will concentrate in auto and industrial chips. Those fancy chips needs a lot of customer to be profitable and with the current global economic malaise. Just as @tokenanalyst posted above, Apple need to sell their Iphone at $2,000 to maintain their high profit margin or accept a lesser one.
While some of the news I shared may not be what you expect, there are some positive news that i would like to share.
While most on this forum only see the Big 4 from news media (SMIC, Huahong, CXMT, YMTC), they failed to see the progress elsewhere. A little bit here and there do add up to big sum.
For example, in the Shenzen Greater Bay Area, there are many fab expansion in progress. Due to the US sanction, most are now limited to 28nm and up. But if you add these up, I see upwards of 600K wafers per month total 28nm & up capacity to be added between ~2024 to 2030. This in addition to the 340K wafers per month announced by SMIC. And we see Huahong Fab8 & Fab9 being planned as well.
So, while I see China 16nm and below expansion stunted by US BIS action (this is contrary to what SDF members are predicting), but what we see in the industry is quite a bit of 28nm expansion that's much more optimistic than what SDF sees.
So, brother, what I'm saying is we see about one million 12" wafers per month of 28nm fab capacity by 2030. Despite of lack of advanced nodes, the sheer volume of 28nm that we should be able to build up without US interference is cause for celebration (US knows they need China to support the more legacy nodes...since I don't think they are interested in investing in the older nodes themselves).
Not entirely. He thinks this is absolutely damaging in the short term to Chinese semi industry. That clearly hasn't happened.This guy gets it.
It happened to HiSilicon. They basically imploded. And much the same might happen to CXMT and even YMTC. I initially thought the memory companies might not be affected by the ban on litho but turns out they expanded the rules just to cover the memory companies.Not entirely. He thinks this is absolutely damaging in the short term to Chinese semi industry. That clearly hasn't happened.
HiSilicon is old news. I will give you that CXMT/YMTC are likely to be the most affected here. We will have to wait a while to find out if they are able to expand production. We know they did quite a bit of stocking up. It's unclear to us how much not having Lam/KLA staff around will hurt them.It happened to HiSilicon. They basically imploded. And much the same might happen to CXMT and even YMTC. I initially thought the memory companies might not be affected by the ban on litho but turns out they expanded the rules just to cover the memory companies.
Well, SMEE arfi scanner already "landed' last year so it's just a matter of iterative improvements and scaling up production. The big question for me is EUV. We are going to have wait 2 to 3 years to see if the optimistic rumors have any truths to them.However China already surpassed the US in total 300mm wafer production capacity and its capacity keeps increasing. China will find some way to keep the machines operating. The cat is out of the bag I think.
It remains to be seen if the US will manage to keep the West's edge on semiconductors long term or not though. It depends on if the Chinese industry can make its own tools or not.
what was the tweet about?
Ah that was deleted? Apparently aji no moto entering semiconductor industrywhat was the tweet about?
Yes Sir agree on you on all points, what I'm saying, China is playing the long game and is feinting vulnerability. My thesis is that China short term goal is to increase onshore production of mature nodes needed for their Fourth Industrial Revolution, 28nm and 14nm nodes are enough While developing the core competency for 7nm and 5nm Chips. SMEE SSA 800 28NM DUVL can produce chips from 28nm to 7nm using multi patterning, future variant like the mythical SSA 900 22nm DUVL can lowered it to 5nm. But all of these are pure projection with 2024/25 as the timeframe. What matter most for China is having the volume production capacity from 28nm to 14nm plus the Core Competency from 7nm to 5nm. And in 3 years we shall overwhelm the US in the mature legacy chip and partially on par with the US regarding their onshore 5nm. Now I understand why TSMC had cautiously announce that the 3nm maybe produce at their Arizona FAB, the US want to stay at least a generation ahead. 2025 is the inflection point, a parity is possible IF we follow the current trend (onshore capability) BUT the US may think otherwise.Bro, I've been posting their mature node expansion for weeks. Especially for power chips segment. There is no question they are on a great path here.
there is nothing to celebrate for just taking over the mature segment. It's non-Chinese like to be this unambitious.
Again, they are not buying all these latest ASML Arfi scanners to just waste them away. They are buying them because they want to increase their advanced node production. HPC chips have wide usage outside of chip industry. If they really get banned at some point without domestic alternatives, it would be a real problem for China's AI, super computing, cloud and other industries. The effect of that goes far beyond the benefits of just being able to produce industrial chips.