Chinese semiconductor industry

Status
Not open for further replies.

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
South Korea is a rival with SEMES. There's also the risk of being banned. It is always better to open new markets than try to squeeze into a crowded existing market because there's literally no competition when you create a new market. That is also how Chinese companies have historically expanded: take domestic market share, create new developing country markets, then finally, take over developed markets or get banned from them. And if banned, then start eroding adversary market shares everywhere outside their home market.

Example: Africa was a small, nonexistent market for smartphones in 2010. US would've deemed them only worthy of using dumb phones on 2G and demanded they be thankful for buying dumb phones for $300 each. Chinese companies put a smartphone in every African's hand for $225 by 2019. Now there's $25 billion worth of Huawei telecom networks in Africa. A 100's billion USD level market was created out of thin air because Chinese businesses empowered Africans to buy modern technology.

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

The exact same thing can be done for Russia and Brazil. They've already signed MOU with China. Finance some initial orders, then they're hooked just like Africans on smartphones and telecom equipment, and a new market has been created.

This is also the logic behind why Intel invested in places that otherwise don't make much sense like Ireland, Is.rael, etc. Are these particularly huge markets? Do they make up a huge portion of tool spend or electronics customers...? Exactly. This is to create markets.
That would still be peanuts compared to just how much money Chinese fab industry is likely to spend on equipment over the next decade. I'm expecting huge expansion in both mature nodes (which we have heard plans for) and advanced nodes (as domestic tools become available). There is not issue with cultivating new market, but the most immediate gains are with advanced fabs at Samsung & TSMC that many Chinese suppliers are already supplying. We've heard about wafers, gases, etching tools, EDAs and such. I'm sure there are more.

My feeling is that NAURA was seeing more like around 50% growth until this past quarter. Which makes sense, since more advanced tools = dis proportionally higher revenues (so we saw almost 80% jump in Q1)
 

gelgoog

Lieutenant General
Registered Member
The amount of Chinese fab investment is depressed by the fact that China cannot order modern equipment due to Western sanctions. The more modern equipment is a lot more expensive per wafer produced. When you consider that even with this trade limitation China is the leader in orders by far, it only makes this more obvious. China is massively expanding capacity to produce legacy nodes.
 

USTBasisRollCarry

New Member
Registered Member

Internal and external troubles TSMC production capacity expansion "sudden brake"​


Please, Log in or Register to view URLs content!

"It is reported that TSMC's 6/7nm process capacity utilization rate was less than 40% by the end of March; 4/5nm process has dropped to 75%"

"TSMC originally planned to build two factories in Kaohsiung, including 7nm and 28nm. However, following the suspension of plans for the 7nm plant of the Kaohsiung plant due to weak demand in the smartphone and PC markets, it was reported that the 28nm plan was also postponed. Delayed, and the list of 28nm equipment that the factory plans to purchase is also cancelled."

"Although TSMC has pressed the "pause button" for the construction of many factories in Taiwan, contrary to this, the construction of new factories in Arizona, USA and Kumamoto, Japan will continue to advance. Among them, the first phase of the Arizona fab is expected to in 2024 start producing 4nm chips. The second phase of the project has also started. It is expected in 2026. to start producing 3nm "


TSMC is clearly digging his own grave here. Once US fabs will be open, they are no more TSMC's, they are 100% made in USA...and owned and ruled by US: TSMC (including its people) are just a tool to produce US chips in US.

And their government does not help them because it is geopolitically controlled by US. The best thing for TSMC is that Taiwan reunites with China in the future, maybe after 2024 elections. No, is not a joke. It is an objective analysis.
Very recent phenomea. In January 2023, TSMC revenue was +16% Y/Y, February 2023, revenue was +11% Y/Y and March 2023 revenue was -15.4% Y/Y and this was after a breakneck 2022 year see below (
Please, Log in or Register to view URLs content!
)

Jan 2022 (y/y change)+36%
Feb 2022 (y/y change)+38%
Mar 2022 (y/y change)+33%
Apr 2022 (y/y change)+55%
May 2022 (y/y change)+65%
Jun 2022 (y/y change)+18%
Jul 2022 (y/y change)+50%
Aug 2022 (y/y change)+59%
Sept 2022 (y/y change)+36%
Oct 2022 (y/y change)+56%
Nov 2022 (y/y change) +50%
Dec 2022 (y/y change) +24%
 

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
The infamous Huawei Whisper wrote this today. I must say that he is normally very accurate and does not exaggerate Chinese progress, so what he said here confirms a lot of things we've been saying.
都看着吧,芯片产业从今年开始要慢慢变天了

我为什么这么说,因为当我们的国产芯片,解决了14nm,12nm和10nm的时候,这意味着我们可以摆脱三星,台积电的过度依赖。

大家觉得是不是夸张了?其实不然,高端芯片目前只有手机等一些设备使用,而新能源汽车,服务器,家用电器等等其他行业使用的芯片无需特别高端。加上我们还有中芯国际的N+1等等技术,可以进一步提升芯片性能,华为在芯片设计上面也有突破,所以把所有综合起来看,我们国产芯片大批量出货之后,导致的必然结果就是三星和台积电订单减少。因为美国的制裁限制,曾经三星和台积电配合美国控制我们芯片。现在随着时间推移,我们已经解决了70%-80%的绝大部分芯片自我供应问题,长江存储也通过技术积累,让存储芯片也有了一席之地。

所以用不了多久,大家就可以看到三星,台积电这些依靠代工的芯片工厂会被反噬,配合美国限制也等于断送自己的庞大收入来源。

等我们突破更高的工艺之后,那么依靠代工的台积电和三星,能否还如前几年一样风光无限呢?

最近有消息称台积电和三星已经宣布先后减产,是芯片真的卖不动了吗?还是说有人已经开始替代了一些领域了呢?

I tweeted about this, but just generally China can already produce 80% of its own chip needs (for logic probably?) and that will only increase as the 7nm process gets improved in both yield and performance. YMTC is also making big in-road. CXMT still sucks, but what can you do?

Anyways, he has started asking the question of why TSMC/Samsung reducing production. Sure, SMIC is experiencing lower sales numbers too, but I don't think its Finfet fab is having any issues attracting orders. So as we get to 2024 and 2025, will Chinese tech firms still be reliant just on Intel & Nvidia chips or will they start buying Huawei or T-Head or Loongson or Biren/Moore thread/Cambrian designed chips that are produced at SMIC? Clearly, Huawei itself has already moved off TSMC produced chip due to sanctions. And Inspur & Sugon need to be buying domestically produced chips too given their appearance on entity list.
 

gelgoog

Lieutenant General
Registered Member
I already said here. The whole semiconductor market is in a major downturn. After massive increase in purchases of consumer PCs when people were working or studying at home because of the lockdowns, pretty much anyone who wanted a PC or consumer hardware got it. There will be a dry spell for at least a couple of years. Might even be half a decade or worse. Consumer demand for hardware is down, so the vendors are cutting production.
 

tinrobert

Junior Member
Registered Member
I already said here. The whole semiconductor market is in a major downturn. After massive increase in purchases of consumer PCs when people were working or studying at home because of the lockdowns, pretty much anyone who wanted a PC or consumer hardware got it. There will be a dry spell for at least a couple of years. Might even be half a decade or worse. Consumer demand for hardware is down, so the vendors are cutting production.
That's correct, but is only one side of the equation. PCs and smartphones have been in a downturn post-covid lockdowns, and will start recovering in 3Q 2023. But the DRAM and NAND memory chip suppliers have been significantly impacted more than logic or foundry. Memory is down because of the large amount of WFE and capex spend by Samsung, hynix, and Micron, which resulted in more chip capacity at a time when supply was already greater than demand post covid.
Combine that with poor monetary and fiscal action by the U.S. Federal Reserve, the shuttering of fossil fuels, and trillions of dollars of handouts and you have a poor economy where people use their money to buy food, petrol, and vehicles at inflated prices rather than electronic gadgets needing semiconductors.
I've written numerous articles and provided data and metrics showing this problem. And I remind readers I forecast the current downturn because of excess capex spend back in mid-2021, 20 months ago. You can access my articles on my web site at
Please, Log in or Register to view URLs content!
and click on the "online blog" button.
 
Last edited:

gelgoog

Lieutenant General
Registered Member
In the 1970s oil crisis there was also a collapse of demand for computers. Tech stocks collapsed. At one point you even had oil companies trying to acquire their way into the electronics business as they were flush with cash from oil sales and didn't know what to do with it.

As for the PC and smartphone markets, I think the differences in features and performance in hardware over the last 8 years are so vanishingly small that people have little reason to upgrade. That was one of the reasons for the 1980s game console crash. So I think it might take quite a while until demand picks back up. In the case of the 1980s crash it took two years to recover. Only recovered with the introduction of the NES basically.

I do not think there will be any major difference in hardware performance in logic until GAA transistors start being used.
 

AETHER

New Member
Registered Member
saw an interesting comment in Bilibili

EUV光刻机这边国内做出来了相差小于1000分之一波长的六镜式极紫外光刻机反射镜样机与转换率大于2%的LPP-EUV光源样机,而且广智院还在与华中科技大学联合研制多束分时光纤激光器驱动极紫外光源,其中已经完成了40Khz液态锡滴打靶装置并完成打靶验证产生极紫外光,除此之外还有清华大学提出的微聚束(SSMB-EUV)光源方案,已经在雄安开工建设光源样机(周长百米级别)不仅如此,哈工大胡组也完成了分辨率优于5皮米的真空激光干涉仪在目标用户处的验证,所以不要把极紫外光刻机当成什么不可能的东西,很多玩意中国已经开始研发甚至样机都出来了。你觉得不可能只是你不知道罢了
translation by chatgpt:

"Domestically in China, the EUV lithography machine has been developed with a six-mirror design for extreme ultraviolet lithography mirrors, with a difference of less than one-thousandth of a wavelength, and a prototype of the LPP-EUV light source with a conversion rate greater than 2%. Additionally, the Guanzhi Institute is collaborating with Huazhong University of Science and Technology to develop a multi-beam time-division fiber laser-driven extreme ultraviolet light source, with a 40 kHz liquid tin droplet target device already completed and verified for generating extreme ultraviolet light. In addition, Tsinghua University has proposed a scheme for a micro-focusing beam (SSMB-EUV) light source, and a prototype has been constructed with a circumference of about 100 meters in Xiong'an. Furthermore, Hu Zu's group at Harbin Institute of Technology has verified the resolution of a vacuum laser interferometer better than 5 picometers at the target user location. Therefore, don't consider extreme ultraviolet lithography machines as something impossible; many things are already being researched and developed in China, and even prototypes have been produced. If you think it's impossible, it's just because you don't know."
 

antiterror13

Brigadier
saw an interesting comment in Bilibili


translation by chatgpt:

"Domestically in China, the EUV lithography machine has been developed with a six-mirror design for extreme ultraviolet lithography mirrors, with a difference of less than one-thousandth of a wavelength, and a prototype of the LPP-EUV light source with a conversion rate greater than 2%. Additionally, the Guanzhi Institute is collaborating with Huazhong University of Science and Technology to develop a multi-beam time-division fiber laser-driven extreme ultraviolet light source, with a 40 kHz liquid tin droplet target device already completed and verified for generating extreme ultraviolet light. In addition, Tsinghua University has proposed a scheme for a micro-focusing beam (SSMB-EUV) light source, and a prototype has been constructed with a circumference of about 100 meters in Xiong'an. Furthermore, Hu Zu's group at Harbin Institute of Technology has verified the resolution of a vacuum laser interferometer better than 5 picometers at the target user location. Therefore, don't consider extreme ultraviolet lithography machines as something impossible; many things are already being researched and developed in China, and even prototypes have been produced. If you think it's impossible, it's just because you don't know."

Same thing most people in the world thought China couldn't build atomic bomb ... guess what China did that in 1964 when China was the most backward and extreme poor .. all done by Chinese without help as the USSR decided to withdraw all technicians and engineers

And guess what that China managed to detonate Hydrogen bomb in 1967, before France ... again when China was the most backward country and poor

Imagine China now .. ohhh boy, China now 1000x or more advanced and much much richer. In term technology advancement, perhaps only the USA is more advanced than China (in general) .. but it is closing very fast
 

ansy1968

Brigadier
Registered Member
Same thing most people in the world thought China couldn't build atomic bomb ... guess what China did that in 1964 when China was the most backward and extreme poor .. all done by Chinese without help as the USSR decided to withdraw all technicians and engineers

And guess what that China managed to detonate Hydrogen bomb in 1967, before France ... again when China was the most backward country and poor

Imagine China now .. ohhh boy, China now 1000x or more advanced and much much richer. In term technology advancement, perhaps only the USA is more advanced than China (in general) .. but it is closing very fast
Then they will accused the Chinese of stealing, they can't accept the fact the Chinaman can innovate or for a better word there is a white man behind every Chinese scientific breakthrough....lol
 
Status
Not open for further replies.
Top