Chinese semiconductor thread II

tokenanalyst

Brigadier
Registered Member

keep an eye on this.

Infineon is now talking about getting commodity chips fabb'd in China. This is after ST made the decision to do so at HH fab.

American chipmakers like TI will lose out.
Not only Infineon but also STMicroelectronics

The second phase of Hua Hong Semiconductor's 12-inch wafer fab expansion in Wuxi held a production ceremony on December 10. This event garnered significant attention following the recent announcement of collaboration between STMicroelectronics (STM) and Hua Hong to produce 40nm microcontrollers (MCUs) locally in China by 2025.

Hua Hong specializes in specialty technologies and analog IC manufacturing, aligning well with STM's MCU products. The 40nm process is also favored by STM for local foundry services, focusing on domestic manufacturing and supply within China. This partnership is expected to help STM reduce overseas costs while closer to the Chinese market.

According to Hua Hong, the second phase of the 12-inch production line has been completed 100 days ahead of schedule, entering the operational stage. This facility in Wuxi serves as an important base for Hua Hong, expanding beyond its main manufacturing hub in Shanghai into the Yangtze River Delta region.​


Please, Log in or Register to view URLs content!

Looks like Chines OEMs are demanding local production because US export controls.
 

jli88

Junior Member
Registered Member
Techinsights has just come out with some analysis of Kirin 9020. They are saying no change in process at all. Still the same 7nm. The chip is bigger and incrementally better.

Needless to say, this is a miss from the forum consensus. Would be interesting to find out what memory was used.

Bloomberg report:
Please, Log in or Register to view URLs content!


Can anyone with access give out some more details? @curiouscat
 

Hyper

Junior Member
Registered Member
Techinsights has just come out with some analysis of Kirin 9020. They are saying no change in process at all. Still the same 7nm. The chip is bigger and incrementally better.

Needless to say, this is a miss from the forum consensus. Would be interesting to find out what memory was used.

Bloomberg report:
Please, Log in or Register to view URLs content!


Can anyone with access give out some more details? @curiouscat
Modern cpu are basically bottlenecked by memory. Logic scaled faster than memory. Consequence of von Neumann architecture.
 

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
Techinsights has just come out with some analysis of Kirin 9020. They are saying no change in process at all. Still the same 7nm. The chip is bigger and incrementally better.

Needless to say, this is a miss from the forum consensus. Would be interesting to find out what memory was used.

Bloomberg report:
Please, Log in or Register to view URLs content!


Can anyone with access give out some more details? @curiouscat
well, I don't think people actually expected this chip to use 5nm process, but an improved process vs what they had been using for K9000S.

Going through the Bloomberg article, they are just saying this is 7nm. It wasn't really clear if they are saying transistor density has changed. It's entirely possible that SMIC hasn't made enough progress on yield of N+3 process in time for Huawei to design 9020 around that.

Keep in mind that TechInsights are good for analyzing floor map of the chips, not for analyzing the performances of the chip itself. So, I would ignore all their comments about that.
 

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
Then it means the design optimizations they made are so much more impressive, to use a 7nm process and have similar or better performance than SD8+ gen1, which is fabbed on TSMC 4nm.
the expectation all along was that they got MediaTek 9200 with a 7nm process, the expectation for most people here is that they use something more akin to N7+ this time around. We will have to wait for more info, but it's entirely possible that they did not use a process with higher density at all.
 

tokenanalyst

Brigadier
Registered Member
Techinsights has just come out with some analysis of Kirin 9020. They are saying no change in process at all. Still the same 7nm. The chip is bigger and incrementally better.

Needless to say, this is a miss from the forum consensus. Would be interesting to find out what memory was used.

Bloomberg report:
Please, Log in or Register to view URLs content!


Can anyone with access give out some more details? @curiouscat
That could mean that their yields is getting much better, bigger dies usually impact yield, so Huawei is confident of selling millions of phones with this chip.
 

tokenanalyst

Brigadier
Registered Member

Zhiling Semiconductor: Low-cost 6&8-inch silicon carbide polishing processing solution.​


Zhiling Semiconductor has carried out in-depth development of efficient polishing processes and equipment for silicon carbide wafers, and launched a solution for a fully automatic polishing line for silicon carbide wafers . It has made significant breakthroughs in edge chipping control, scratch control, improvement of precision and efficiency, and cost control, and is expected to promote innovation in polishing processes within the industry.

In 2023, Zhiling Semiconductor completed the development of the heavy-duty chemical mechanical polisher HCP-1280R2-SIA , which is dedicated to high-efficiency and high-precision polishing of 6-inch and 8-inch silicon carbide wafers, and can achieve a polishing pressure of more than 10PSI. In March of this year, Zhiling Semiconductor launched an 8-inch silicon carbide wafer fully automatic polishing line, which attracted the attention of many industry professionals.

1733955826067.png

Please, Log in or Register to view URLs content!
Please, Log in or Register to view URLs content!
 
Top