Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/will-china-master-14nm-and-what-is-their-future-in-semi-production.14607/
        )

    [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] => 2021370
            [XFI] => 1050270
        )

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

Will China master 14nm and what is their future in semi production?

Arthur Hanson

Well-known member
Are there any observations on China and its efforts to have homegrown production of semis and the possible timelines involved? Will China be able to compete in what areas of semi-production and when? I don't see them competing at the high end if ever because of patent and copyright laws that would be costly to buy if available at all and the restricted access to the tools needed. Will China resort to the outright theft of IP and tools to meet their ends? Any thoughts or opinions on this would be appreciated. My thoughts on this are that it will be a long time for China to catch up to the state of the art if ever, unless they earn a level of trust they don't currently have.

 
I believe China will master 14nm but not sure it's 2022.
Another key is the competitiveness of Chinese fabless companies. China needs more fabless companies like Hisilicon to support SMIC.
when shortage is over in 2023, SMIC will face big challenge if China fabless does not grow up.
 
I believe China will master 14nm but not sure it's 2022.
Another key is the competitiveness of Chinese fabless companies. China needs more fabless companies like Hisilicon to support SMIC.
when shortage is over in 2023, SMIC will face big challenge if China fabless does not grow up.

From what I am told yield is getting better on SMIC 14nm so you will see more of it in 2022 but maybe not considered HVM. I doubt that 14nm will hit double digit revenue for SMIC until 2025. Right now SMIC is focusing on 28nm design wins to fill the extra capacity planned for 2024/25:

 
SMIC 14nm is a clone of TSMC N16FFC based on the Kirin 710A teardown work that TechInsights did. Fab tools are from the same suppliers, though there seem to be domestic players in the periphery like FOUPs (https://www.digitimes.com/news/a20170914PD209.html). SMICs revenue though shows 28nm "+FinFET" as one bucket, and that is not nearly the majority of their production, so not clear if their yields are all that great yet, but sufficient for Huawei to ship production silicon already.
 
SMIC 14nm is a clone of TSMC N16FFC based on the Kirin 710A teardown work that TechInsights did. Fab tools are from the same suppliers, though there seem to be domestic players in the periphery like FOUPs (https://www.digitimes.com/news/a20170914PD209.html). SMICs revenue though shows 28nm "+FinFET" as one bucket, and that is not nearly the majority of their production, so not clear if their yields are all that great yet, but sufficient for Huawei to ship production silicon already.

TSMC built a 16nm fab in China so the SMIC 14nm clone is not a surprise. UMC is having the same problem with 14nm, never made it into HVM. GF had to license Samsung's 14nm so SMIC is not alone. FinFETs are complicated to design and manufacture. Of course the outside media disagrees:




FinFETs are more than 60% of TSMC's revenue and climbing quickly. What future does a fab have if they cannot make a profit on FinFET wafers?
 
TSMC built a 16nm fab in China so the SMIC 14nm clone is not a surprise. UMC is having the same problem with 14nm, never made it into HVM. GF had to license Samsung's 14nm so SMIC is not alone. FinFETs are complicated to design and manufacture. Of course the outside media disagrees:




FinFETs are more than 60% of TSMC's revenue and climbing quickly. What future does a fab have if they cannot make a profit on FinFET wafers?

It's not hard to believe SMIC has achieved 95% yield on 14nm process considering that they have aggressively recruited ex-TSMC employees at all costs. From what I've heard, they regularly offered minimum 3x salary or more.
 
  • Like
Reactions: VCT
China has the will and financial wherewithal to eventually catch up. I think it's more of a question of how long it could take. My guess is that the leading foundries will slow down after 3nm and that will provide an opening for China to close the gap.
 
You're always wrong so shrink will continue at 3nm.
They certainly have the will and likely the money (thank CCP for both!), but these aren't the reasons they can or will succeed. Not enough tools (EUV) nor legit IP - if they outright steal it they maybe have a 50/50 shot. More likely is takeover of Taiwan & TSMC, so whichever works for them...
 
It's not hard to believe SMIC has achieved 95% yield on 14nm process considering that they have aggressively recruited ex-TSMC employees at all costs. From what I've heard, they regularly offered minimum 3x salary or more.
14nm was set up with Qualcomm/IMEC/Huawei collaboration initially.
 
Shrink will continue, but at a slower pace. The pace of node shrinks has already slowed down a fair bit, and will likely slow down further at 3nm.
Other technologies will take the place of shrink in increasing performance and China will stay behind until they establish a fair and open market without massive IP theft or forced partnerships.
 
Back
Top