Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/chinese-chip-makers-seek-way-out-of-us-curbs.16984/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Chinese chip makers seek way out of US curbs

Daniel Nenni

Admin
Staff member
Some Chinese firms try to clear their names from sanction list; others try to lure non-US foreign engineers

Some Chinese semiconductor makers are reported to have taken the initiative to contact the United States embassy in Beijing after they were put on a trade watch list of the US Commerce Department last month.

These are companies that are required to pass an end-use check within 60 days or they will be added to the entity list of the US Commerce Department’s Bureau of Industry and Security (BIS), the South China Morning Post (SCMP) reported, citing unnamed sources.

A BIS official from the US embassy in Beijing will visit Wuhan to meet some Chinese firms that are now on the bureau’s unverified list, said the report.

As US citizens are now restricted from working for Chinese chip makers, some of these companies had recently offered huge salary packages to lure talent from Taiwan, South Korea, Japan and Singapore, said some columnists in Taiwan and Singapore.

Mainland Chinese commentators said the US curbs will have only a short-term impact on China, as the country has a strong talent pipeline from its technology schools.
They expect that many of the departed Chinese American engineers will eventually give up their American citizenship and return to China.

On October 7, the BIS added Yangtze Memory Technologies Co (YMTC), China’s top memory chipmaker, and 30 other Chinese “entities” to an unverified list.

The bureau said it was unable to verify the bona fides of the Chinese firms because an end-use check could not be completed satisfactorily for reasons outside the control of US authorities. It said the named companies would face restrictions when trying to purchase US products.

The SCMP reported on October 25 that trade officials from the US embassy in Beijing had held talks with executives of Naura Technology Group, a Chinese integrated micro-electronics firm, as the company’s unit was among those on the unverified list.

It has become urgent for the companies on the list to seek contact with the US trade officials as the officials need to complete the check of end-users within 60 days, according to the report. An unnamed executive at Wuhan Juhere Photonic Technologies said the company had contacted the US export control officials in Beijing and tried to remove itself from the unverified list.

A columnist named Yang Lu wrote in an article published by Singapore-based Initium Media that the US chip export control on China has had a huge impact on China’s semiconductor industry as the depth and breadth of the sanctions are unprecedented.

Yang said the US-China competition has become a zero-sum game due to the export ban unveiled on October 7. He said the ban can hurt China’s ability to develop supercomputers capable of being used to direct hypersonic missiles and develop nuclear weapons.

The BIS also banned US citizens from supporting the development or production of integrated circuits at certain chip fabs in China, without licenses, from October 12.

GigaDevice Semiconductor, a Chinese NOR flash memory designer, said its two Chinese American executives had resigned.

A Taipei-based columnist surnamed Chen wrote in an article on Digitimes.com that the chip export ban unveiled by the US in early October has widened the technology gap between Chinese chip makers and global industry players.

Chen said China’s plan to self-supply 70% of its semiconductors by 2025 and become a hub of global chip foundries by 2030 has basically vanished due to the US curbs. The columnist said China’s SMIC and Hua Hong Semiconductor can now produce only 28-nanometer (nm) chips, instead of the more advanced 14-16 nm chips.

Chen said some Chinese American engineers resigned from their chip firms but kept operating them informally. She added that, due to this round of US sanctions, some Chinese chip firms recently offered huge salary packages to attract and retain foreign engineers.

Most mainland Chinese IT writers admitted that the US curbs are having a big negative impact on China’s chip sector but some of them said the situation remains manageable.

Xiao Fengkan, a Shanxi-based columnist, on October 22 published an article with the title, “It is not a bad thing that American talents are urged to leave China’s chip industry.”

Xiao found it inevitable that some Chinese American engineers and executives will leave China as they have been forced to choose between their careers in the country or their US citizenship. He added that China can replenish its talent pool with its top university students in a matter of years.

A Shenzhen-based writer said people should not be over-worried when some Chinese American engineers leave China. He said many of them will return to China in future as it will be difficult for them to enjoy equivalent job opportunities in the US.

The writer suggested that China should grant Chinese nationality so that such talent will not have to worry about giving up American citizenship.

 
Is SMIC on the broader list? It seems they would be the best target to permanently derail Chinese semiconductor ambitions. One does have to keep in mind however that the CCP has gotten very clever in moving around subsidiaries to get around export controls on technicalities I.E huawei and honor phones. Cheers
 
It's probably too late to rein in SMIC. They can toggle volume between FinFET and non-FinFET nodes.
Can SMIC survive without American and Japanese equipment/IP though? Also is the domestic Chinese market big enough to sustain SMIC as I reckon they will have a hell of a time trying to sell their “7nm” node internationally considering its strongly suspected to be stolen TSMC IP. Thoughts?
 
Can SMIC survive without American and Japanese equipment/IP though? Also is the domestic Chinese market big enough to sustain SMIC as I reckon they will have a hell of a time trying to sell their “7nm” node internationally considering its strongly suspected to be stolen TSMC IP. Thoughts?
The vendor support almost certainly has local Chinese-speaking representatives and parts would be stocked as much as possible, although once run out, US-only parts could be limiting.

For general availability (https://www.smics.com/en/site/design_process), the most advanced logic process is 28nm, and the most advanced NVM process is 24nm NAND and 45nm NOR. The FinFET processes like 14nm and 7nm are extremely rare use, probably reserved only for Huawei and some occasional small companies. However most such companies have used TSMC 7nm in the past, so switching to SMIC would not be automatic, many features may not be matching. In case they needed the FinFET volume suddenly, they would need to use the 28nm resources, hence the need to toggle.
 
Last edited:
The vendor support almost certainly has local Chinese-speaking representatives and parts would be stocked as much as possible, although once run out, US-only parts could be limiting.

For general availability (https://www.smics.com/en/site/design_process), the most advanced logic process is 28nm, and the most advanced NVM process is 24nm NAND and 45nm NOR. The FinFET processes like 14nm and 7nm are extremely rare use, probably reserved only for Huawei and some occasional small companies. However most such companies have used TSMC 7nm in the past, so switching to SMIC would not be automatic, many features may not be matching. In case they needed the FinFET volume suddenly, they would need to use the 28nm resources, hence the need to toggle.
I see. So if I am understanding you correctly SMIC would have to go without it’s primary revenue base to transition to finfet in any sort of meaningful capacity? Hence the toggle? Who owns the patents for finet anyways? I know Intel pioneered it so how does that work?
 
Is SMIC on the broader list? It seems they would be the best target to permanently derail Chinese semiconductor ambitions. One does have to keep in mind however that the CCP has gotten very clever in moving around subsidiaries to get around export controls on technicalities I.E huawei and honor phones. Cheers

SMIC was blacklisted long ago
 
Can SMIC survive without American and Japanese equipment/IP though? Also is the domestic Chinese market big enough to sustain SMIC as I reckon they will have a hell of a time trying to sell their “7nm” node internationally considering its strongly suspected to be stolen TSMC IP. Thoughts?

Cannot speak for the equipment/IP part,but market is certainly not a problem. There is tons of Chinese fabless using TSMC‘s 7nm process,surely they will need SMIC's help,if they are denied by TSMC.
 
Can SMIC survive without American and Japanese equipment/IP though? Also is the domestic Chinese market big enough to sustain SMIC as I reckon they will have a hell of a time trying to sell their “7nm” node internationally considering its strongly suspected to be stolen TSMC IP. Thoughts?

SMIC will thrive. Money will not be a problem:


China is the largest semiconductor consumer of the world and of course the largest SMIC customer.
 
About the "US person" part. Here is my thought.

You have to understand why those Chinese Americans come to China to establish semiconductor companies in the first place. Beacuse of "nationalism" or “patriotism” ?Not really,not to my knowledge.

Most of them come to China because they are facing a "bottleneck" in their career in the US. They have little chance to be promote to CEO position due to "bamboo ceiling".The entrepreneur-path is also difficult,because semiconductor in the US is a very mature industry,big companies already have firm control of the market,it's very difficult for startup company in areas like EDA/material/equipment etc to compete with established giants.

That's why they come to China,because in China the environment of semiconductor sector is totally different. They have the opportunity to accomplish something that they couldn't back in the US. And that opportunity is still there,the recent US ban doesn't change it.

So my conclusion is:US persons working in Chinese semiconductor industry. For those who don't have a very high ambition and prefer a low-profile but stable career path,they may choose to leave China. For those who are ambitious and aspire for achievement,they will stay in China. They may revoke their US green card or citizenship,and get citizenship from some third country such as Canada or Singapore
 
About the "US person" part. Here is my thought.

You have to understand why those Chinese Americans come to China to establish semiconductor companies in the first place. Beacuse of "nationalism" or “patriotism” ?Not really,not to my knowledge.

Most of them come to China because they are facing a "bottleneck" in their career in the US. They have little chance to be promote to CEO position due to "bamboo ceiling".The entrepreneur-path is also difficult,because semiconductor in the US is a very mature industry, big companies already have firm control of the market,it's very difficult for startup company in areas like EDA/material/equipment etc to compete with established giants.

That's why they come to China, because in China the environment of semiconductor sector is totally different. They have the opportunity to accomplish something that they couldn't back in the US. And that opportunity is still there, the recent US ban doesn't change it.

So my conclusion is:US persons working in Chinese semiconductor industry. For those who don't have a very high ambition and prefer a low-profile but stable career path, they may choose to leave China. For those who are ambitious and aspire for achievement, they will stay in China. They may revoke their US green card or citizenship, and get citizenship from some third country such as Canada or Singapore

I have asked many Chinese citizens why they work in Silicon Valley versus China. They all have said they would have to work too much in China for less money.

I work with many start-ups, here in Silicon Valley race is not an issue. When I started in semiconductors in the 1980s race was an issue but not today. It really is a melting pot out here and that is one of the strengths of Silicon Valley, absolutely.
 
I see. So if I am understanding you correctly SMIC would have to go without it’s primary revenue base to transition to finfet in any sort of meaningful capacity? Hence the toggle? Who owns the patents for finet anyways? I know Intel pioneered it so how does that work?
No one. Everyone has their own FinFet patents. As long as your idea isn't identical to something that is already patented you can apply for a patient.
 
What's the big deal?Semiconductor industry already moving into GAA now,FinFET is outdated
A brilliant architecture in 14nm FinFET process may outperform an outdated or poorly conceived chip architecture in 3nm process. You can theoretically design and fab some very impressive chips (or chiplets) in 7nm FinFET process like SMIC has. The best process in the world can't eliminate the disadvantages of inefficient legacy designs. The tools people have at their disposal have a huge influence over the solutions they pursue. Oftentimes advanced, generation ahead fab process has been used to cover for less than brilliant designs. I would argue this was the case with Intel CPUs for years. Teams constrained by a lessor process may be highly incentivized to innovate in design.
 
I see. So if I am understanding you correctly SMIC would have to go without it’s primary revenue base to transition to finfet in any sort of meaningful capacity? Hence the toggle? Who owns the patents for finet anyways? I know Intel pioneered it so how does that work?
US6413802 has already expired.
 
No one. Everyone has their own FinFet patents. As long as your idea isn't identical to something that is already patented you can apply for a patient.

UC Berkeley did the heavy lifting:


The potential of Digh Hisamoto's research on DELTA transistors drew the attention of the Defense Advanced Research Projects Agency (DARPA), which in 1997 awarded a contract to a research group at UC Berkeley to develop a deep sub-micron transistor based on DELTA technology.[13] The group was led by Hisamoto along with TSMC's Chenming Hu. The team made the following breakthroughs between 1998 and 2004.[14]

  • 1998 – N-channel FinFET (17 nm) – Digh Hisamoto, Chenming Hu, Tsu-Jae King Liu, Jeffrey Bokor, Wen-Chin Lee, Jakub Kedzierski, Erik Anderson, Hideki Takeuchi, Kazuya Asano[15]
  • 1999 – P-channel FinFET (sub-50 nm) – Digh Hisamoto, Chenming Hu, Xuejue Huang, Wen-Chin Lee, Charles Kuo, Leland Chang, Jakub Kedzierski, Erik Anderson, Hideki Takeuchi[16]
  • 2001 – 15 nm FinFET – Chenming Hu, Yang-Kyu Choi, Nick Lindert, P. Xuan, S. Tang, D. Ha, Erik Anderson, Tsu-Jae King Liu, Jeffrey Bokor[17]
  • 2002 – 10 nm FinFET – Shibly Ahmed, Scott Bell, Cyrus Tabery, Jeffrey Bokor, David Kyser, Chenming Hu, Tsu-Jae King Liu, Bin Yu, Leland Chang[18]
  • 2004 – High-κ/metal gate FinFET – D. Ha, Hideki Takeuchi, Yang-Kyu Choi, Tsu-Jae King Liu, W. Bai, D.-L. Kwong, A. Agarwal, M. Ameen
They coined the term "FinFET" (fin field-effect transistor) in a December 2000 paper,[19] used to describe a non-planar, double-gate transistor built on an SOI substrate.[20]
 
A brilliant architecture in 14nm FinFET process may outperform an outdated or poorly conceived chip architecture in 3nm process. You can theoretically design and fab some very impressive chips (or chiplets) in 7nm FinFET process like SMIC has. The best process in the world can't eliminate the disadvantages of inefficient legacy designs. The tools people have at their disposal have a huge influence over the solutions they pursue. Oftentimes advanced, generation ahead fab process has been used to cover for less than brilliant designs. I would argue this was the case with Intel CPUs for years. Teams constrained by a lessor process may be highly incentivized to innovate in design.
I don't know if intel is totally applicable here. i7 is an inferior node to N5, no arguments there, but I feel like the big efficiency gain of RPL over ADL can't just be due to a better design. I feel like a large part of this has to be intarnode improvements to 10nm and intel's IDM advantage at work.
 
Back
Top