Chinese semiconductor industry

Status
Not open for further replies.

gelgoog

Brigadier
Registered Member
Please, Log in or Register to view URLs content!
The coercive and unilateral measures issued against Venezuela since August 2017, have prevented the country from assuming its obligations to debt creditors. In this way, the US authorities deliberately placed CITGO at high risk, as a prelude to the asset freeze that came into effect in 2019 simultaneously with the US’ sponsorship of Guaidó.
...
In US courts, CITGO has not had been defended by legal teams of the legitimate authorities of Venezuela. Rather, the company has been “defended” by the fake attorney's office represented by José Ignacio Hernández, who is accused of negligence and is also clearly subordinated to those interested in dismembering the company, as in his previous work Hernández worked for the plaintiff, Crystallex.

The US first prevented Venezuela from being able to pay their bond, then didn't allow them to defend themselves in court, and seized the assets.
 

nlalyst

Junior Member
Registered Member
Please, Log in or Register to view URLs content!



Also "process qualification" as in your slides doesn't mean production.
Can you show me a 180nm product TSMC shipped in 1998?

The earliest evidence of products shipped on 0.18 micron by TSMC was from 1999:
Please, Log in or Register to view URLs content!


However, the actual devices using those chips, afaik, were not launched until 2000. All the Nvidia, 3Dfx and ATI TSMC manufactured graphic chips that came to market in 1999 were produced on the 250nm node. The first 180nm graphics manufactured by TSMC was available only in 2000, the Radeon 7000 PCI.

Therefore, I still stand by my observation that SMIC was 3 years behind TSMC in 2001.
 

gelgoog

Brigadier
Registered Member
Can you show me a 180nm product TSMC shipped in 1998?
The earliest evidence of products shipped on 0.18 micron by TSMC was from 1999
...
However, the actual devices using those chips, afaik, were not launched until 2000. All the Nvidia, 3Dfx and ATI TSMC manufactured graphic chips that came to market in 1999 were produced on the 250nm node. The first 180nm graphics manufactured by TSMC was available only in 2000, the Radeon 7000 PCI.
Therefore, I still stand by my observation that SMIC was 3 years behind TSMC in 2001.

So you are telling me you would rather believe a Wiki's claims of when TSMC started producing chips rather than TSMC themselves? Great.
It is little surprising that TSMC wouldn't have ramped 180nm to mass production on the start. No one does.

If you read the article I posted before, it claims SMIC only had 250nm in mass production in 2002. Not only that, they were using 200mm wafers while TSMC already had 300mm wafer production.

Same deal with SMIC 10nm right now. Only 14nm is in mass production. When they start candidate production of 7nm, 10nm might be in mass production, rise repeat.
 
Last edited:
D

Deleted member 15949

Guest
It is way more complicated than what you think @SleepyStudent

Please, Log in or Register to view URLs content!


AeroComposit (who makes the MC-21 wings), and Aviadvigatel (who makes the PD-14 engine) were put under the Entity List.
Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!


The US can just keep squeezing the Russians more and more.

The United Aircraft Corporation is one of the enterprises sanctioned under CAATSA for example.

With regards to Venezuelan Citgo, the debt is bullshit. Their asset seizure in the US happened around the same time the gold reserves the Venezuelan government had in the Bank of England were frozen. They even intercept Iranian tankers in the open sea headed to Venezuela.
Yes, US sanctions capture all items with a US-origin. Euro (or even Chinese) factories in the United States falls under that remit.
CAATSA are primary and anything produced in foreign countries with <25% content by value can be shipped to any Russian Entity List since it wouldn't fall under the remit of the EAR. Can the US change it? Yeah, they can but do describe US sanctions activity correctly. The Venezuela stuff though is eh., it's definitely questionable.
 

gelgoog

Brigadier
Registered Member
The fact they put AeroComposit and Aviadvigatel on the Entity List at the same time is quite telling since those companies manufacture the most advanced features in the MC-21.

AeroComposit was a company created to make the wings for the MC-21. It is co-located where the airplane production happens. While the Su-57 uses composite materials, they aren't being produced by AeroComposit. The manufacturing plants are quite far away. The MC-21 is manufactured at Irkutsk and the Su-57 is manufactured at Komsomolsk-on-Amur.

Aviadvigatel mostly manufactures engines for transport aircraft and civilian power plants. Their only strictly military product is the MiG-31 engine but since that aircraft isn't in production they would at best be refurbishing engines. The engine for the MiG-31's successor, the MiG-41, is going to use the Izd 30 core so it will likely be produced by Saturn not Aviadvigatel. Sure the engines are also used in military transport aircraft, but they are transport aircraft all the same. The claims the companies are in the Entity List because they supply the Military are, IMHO, BS.

The sanctions were clearly specifically aimed at MC-21 production.

Same way they claim Huawei is connected to the Chinese Military. Or that SMIC produces ships for the Chinese Military. Or even recently when they tried to put Xiaomi under sanctions until they were shot down.
 
Last edited:

nlalyst

Junior Member
Registered Member
So you are telling me you would rather believe a Wiki's claims of when TSMC started producing chips rather than TSMC themselves? Great.
It is little surprising that TSMC wouldn't have ramped 180nm to mass production on the start. No one does.

If you read the article I posted before, it claims SMIC only had 250nm in mass production in 2002. Not only that, they were using 200mm wafers while TSMC already had 300mm wafer production.

Same deal with SMIC 10nm right now. Only 14nm is in mass production. When they start candidate production of 7nm, 10nm might be in mass production, rise repeat.
So this is now a game of my source is better than yours? I am sorry to disappoint you, but all my sources check out.

Here is another
Please, Log in or Register to view URLs content!
quoting SMIC that they are in volume production of 0.18 micron in 2002, which appears consistent with the earlier claim of process qualification in 2001:
SMIC's internal R&D team has developed the process technology from technologies SMIC has licensed. With a small cell-sized embedded SRAM it is intended to be the start of a 0.18-micron family that will include EEPROM and mixed-signal, RF and SoC capabilities. SMIC has already implemented this 0.18 um logic technology into volume production, and has achieved competitive levels of defect density and process cycle time, the company said.

We have established that TSMC only started production of 0.18 micron products in 1999, so the gap was just 3 years, whereas with 14nm (SMIC's most advanced node in production) it was 6 years. Therefore, regression and tightening of sanctions worked as intended.

Final evidence, from a
Please, Log in or Register to view URLs content!
:
1620500743400.png
 
D

Deleted member 15949

Guest
So this is now a game of my source is better than yours? I am sorry to disappoint you, but all my sources check out.

Here is another
Please, Log in or Register to view URLs content!
quoting SMIC that they are in volume production of 0.18 micron in 2002, which appears consistent with the earlier claim of process qualification in 2001:


We have established that TSMC only started production of 0.18 micron products in 1999, so the gap was just 3 years, whereas with 14nm (SMIC's most advanced node in production) it was 6 years. Therefore, regression and tightening of sanctions worked as intended.

Final evidence, from a
Please, Log in or Register to view URLs content!
:
View attachment 71828
You also need to adjust for the fact that later nodes took *everyone* longer to catch-up because they are more technically difficult and more capital intensive
 

gelgoog

Brigadier
Registered Member
Now he's comparing TSMC mass production date with SMIC initial candidate production date...
If you compare like with like TSMC had initial candidate 180nm capability in 1998.

In 2002 TSMC had 130nm with 300mm wafers.
The VIA C3 processor was manufactured in 130nm at TSMC in 2001.

Also, SMIC was able to do 180nm and other early processes by licensing process technology from companies like IBM.
Good luck licensing 14nm from anyone.

SMIC made their own 14nm process. Yet you think this means they are less able than when they licensed a process from IBM at least 4 years after the market had it available.
 
Last edited:

horse

Major
Registered Member
Now he's comparing TSMC mass production date with SMIC initial candidate production date...
In 2002 TSMC had 130nm with 300mm wafers.
RAND research is a think tank and they aren't God.
I am not engineer, and I really do not know much about semiconductor manufacturing other than it is very complicated.

But I do know one thing, that this is an entire industry that all depends on the know-how.

This know-how has to be gained by the companies and individuals themselves while working with the machinery and materials. There is no other way.

To read that SMIC had all that near cutting edge know-how 20 years ago, now that is one fake story that we did not expect to read.

SMIC only recently was able to move up the nodes quickly because Liang Mengsong was hired and he brought his knowledge and connections to the company.

The equipment, the materials, the know-how, all three are required to make chips, more so to make the most advanced chips.

China was a very different place 20 years ago. China was a very different place even 10 years ago. The speed of development is too much for outsiders and non-China specialist to handle. Their heads cannot come to grips with it, because they never seen it before, the China speed.
 

nlalyst

Junior Member
Registered Member
SMIC qualified its 180nm process in 2001. TSMC qualified it in 1998. A gap of 3 years.
SMIC had volume production of 180nm process in 2002. TSMC had volume production in 1999. A gap of 3 years.

SMIC had volume production at the 14nm node in 2020. Intel had it in 2014. Despite all efforts and incentives, the gap has increased. If the EUV ban stays in place, that gap is set to increase further because China appears to be at least 10 years behind in EUV technology.

How was SMIC able to close the gap so rapidly 20 years ago? In large part by stealing TSMC manufacturing technology, probably through the 100+ TSMC engineers it lured over (Taiwanese media claimed the number was up to 500).

EDIT: fixed the wrong node
 
Last edited:
Status
Not open for further replies.
Top