Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
So it's clear now. The reason SMIC only build new 28nm fabs but not 14nm fabs，is that SMIC still needs（or wants）to comply with US regulations.
I remember earlier in this thread，I said that SMIC keep building new 28nm fabs but not 14nm fabs，is a sign that SMIC cannot get US equipment for 14nm process. A lot of people disagree，they say that SMIC don't build 14nm fabs is because there is no demand for 14nm in China，and SMIC isn't interest in building 14nm fab. They also claim that SMIC can totally build new 14nm fab without US equipments right now.
Well now SMIC made it loud and clear，the reason their newly expanded 28nm capacity isn't affected by the new US restriction，is NOT because they don't need US tech for 28nm fab anymore. Rather，it is because their technology is below the threshold set by the US.
See it？SMIC cannot even live without US tech for 28nm nodes，let alone 14nm
New capacity not affected by new U.S. regulations
A questioner asked about overseas customers' demand for SMIC's production capacity. Zhao Haijun said that overseas companies tend to obtain long-term commitments from Chinese foundries to allocate capacity to ensure supply security. He explained that under the influence of the shortage of supply in various industries last year, today's overseas customers will maintain a certain proportion of products produced by Chinese companies. In order to ensure the security of the industrial chain, overseas customers and SMIC have formed a strategic alliance to obtain long-term commitments in production capacity allocation. He said that SMIC's cooperation with overseas customers is very fast.
In addition to industry cyclical factors, Zhao Haijun also mentioned that, according to preliminary interpretation, the production capacity and technology nodes of SMIC's customers fully comply with the new US regulations. He explained that the new regulations in the United States are mainly aimed at the production of American design companies in China, and the company's new production capacity and technology nodes are in full compliance with the relevant regulations. However, some customers need buffer time to interpret the new US export control regulations. The company maintains close communication with customers and suppliers, clarifies some concepts in the new regulations, and assesses the impact on the company. The work is still in progress.
My feeling is that 28nm to 14nm goes past an important threshold.
5nm and 7nm are gradual improvement to the 14nm chip.
But 14nm and 28nm are kind of like a different chip.
SMIC would have a hole in its head if it focuses on 14nm instead of 5nm or 7nm and the important 28nm node whose chips wind up in most industrial products.
There is no way SMIC should be building a 14nm fab.
It has to be 28nm fabs, and 7nm or 5nm fabs.
In other words, the 5nm or 7nm could eventually make the 14nm chip obsolete inside the computer device.
But the 5nm or 7nm chips will not make the 28nm chip or higher obsolete inside industrial components that require a chip.
So why would SMIC be expected to invest in the 14nm node which is going obsolete, and there is excess capacity worldwide to begin with at the current moment.
SMIC has to go where the demand is and that action is industrial chips in China, their bread and butter market.
That is becoming very clear, the strategy inside China cannot be more obvious.
When presented with a problem, there are always choices on what to do.
Here it seems clearer everyday.
1. Most important point to maintain production. At current nodes and mature nodes, and stockpile chips and equipment, all in the name of sustaining production.
2. When production is stable, and in China's case production in IC is growing the fastest in the world, that will provide the needed capital for R&D and when the time comes, to upgrade to more advanced nodes.
The old one-two punches in IC boxing.
But the 5nm or 7nm chips will not make the 28nm chip or higher obsolete inside industrial components that require a chip.
So why would SMIC be expected to invest in the 14nm node which is going obsolete, and there is excess capacity worldwide to begin with at the current moment.
SMIC has to go where the demand is and that action is industrial chips in China, their bread and butter market.
That is becoming very clear, the strategy inside China cannot be more obvious.
When presented with a problem, there are always choices on what to do.
Here it seems clearer everyday.
1. Most important point to maintain production. At current nodes and mature nodes, and stockpile chips and equipment, all in the name of sustaining production.
2. When production is stable, and in China's case production in IC is growing the fastest in the world, that will provide the needed capital for R&D and when the time comes, to upgrade to more advanced nodes.
The old one-two punches in IC boxing.
"The new regulations in the United States are mainly aimed at the production of American design companies in China, and the company 's new production capacity and technology nodes are in full compliance with the relevant regulations."
If they want to keep expanding in 28nm probably they have no option than use local and non US equipment, they will have to de-Americanize their supply chain whatever they like it or not. There is not options.
You see there is not a lot of difference in the equipment used by a 28nm process node and a 14nm process node or a 7nm. A chips is made of a lot of layers, with different materials, semiconductors, dopants, oxides, metals, dielectrics. Deposited, patterned, etched, Implanted, diffused, polished, cleaned and so on. One in top of the others. Some layers may require more sophisticated equipment but that majority are made with legacy equipment. Is all about the cooking.
The same high ratio etch machines is used across multiple process node, immersion lithography machines are used in single patterning features of 45nm+ but can be use for 7nm nodes using multipatterning techniques, ALD equipment is used across multiple layers and nodes. resist developing equipment used across multiple photoresist, cleaning equipment, metrology, CD-SEM and so on.
Because BIS is too lazy to do research and find out was is unique for advanced nodes and what is not, they make no distinction between mature nodes and advance nodes Just "anything used for 16nm or less".
Basically speaking unless Estevez made some changes in the rules, Americans companies are out of China, except maybe for old equipment.
That is how I see the whole situation that I have been warning for quite some time.
But how can they do this research?
The American BIS could go ask Intel.
What if they are adding Chinese ingredients to the cooking in Asian fabs?
Seems to me everyone is using a different cookbook, lol.
And more bad news.
Firstly，compare to the hotly debated topic "domestic substitution" by the public. Chinese semiconductor firms aren't actually so enthusiastic about it. Some of them still have delusions about the US may lift the ban in the future，as stated in the artical below，and some are just satisfied with status quo. Which is the same thing.
Because as the US semiconductor tech advance further in future，they may update the threshold of the ban. And many Chinese firms may just satisfied with status quo of trailing behind the US，satisfied with taking what ever the US allowable，not interest with catching up.
There are domestic equipment manufactures using imported core components，to quickly push put new products，and claim credit for it. When the ban on these core components comes，these domestic equipment will be just as unreliable as foreign ones.
Also，there are fabs loves to talk about how supportive they are to domestic equipment/materials，but it's mere rhetoric，lack real action.
Prove it that there is no demand for 14nm in China
No industry experts ever said this. Can you provide any source？
Yes, they clearly needed to do two step back to do a jump, A LONG TIME AGO at the time ZTE was added to the blacklist and not become dependent of unreliable companies.
Components is not a big deal as long is not American or made in the US. Locally made components should have the biggest market share, followed by non US components and with the current situation US components should be treated as mercury salts.
- Status