Chinese semiconductor industry

[FriedRiceNSpice]

Let's just assume SMIC is 2yrs behind in 2002. Then it can be expected they would fall even further behind, assuming free market competition regardless of whether or not their are sanctions or not. Why would anyone choose a chip that is two years behind when they have the option of choosing a cutting edge children for the same cost? Being two years behind is still a huge disadvantage for SMIC when it comes to market competition. As a result, SMIC would have less revenues and profits to invest back in R&D, and the tech gap will only increase. That is why so many chip fabs fell out of the game over the last two decades, fabbing chips is a winner takes all business. To attribute SMIC falling behind to the success of sanctions is a logical fallacy. Conversely, with the imposition of total sanctions, SMIC is actually in a more advantageous position, as now it has no competition within certain domestic market segments, and will have the revenues as well as governmental support drive its R&D efforts.

 

[krautmeister]

'Cannot be manufactured without the threat of the imposition of American sanctions' not the same as illegal to manufacture anywhere in the world!

Just because some gangster demands protection money so your bar doesn't burn down and the local police can't do anything about it doesn't make it illegal to open a bar!

Huawei's founder, Ren Zhengfei's daughter, Meng Wenzhou, was "illegally" arrested against Canada's own laws because of the gangster, the US government. You're saying it's illegal, but it's still happening. You're using "lawyer talk" but in the world of US hegemony, the US government does what it wants, when it wants, even against the rules they created. These American vassals might cry every once in a while to show they aren't slaves but in the end, they do as they are told.

Furthermore, unless the software was purchased on a term license it would be perfectly legal to continue to use it, rule of law and all that!

American content in lithography machines worldwide is currently preventing Huawei and other entity listed companies from buying only either older generation semiconductor nodes or no semiconductors at all.

You're just going along with US extraterritorial propaganda!

What you call propaganda, Meng Wenzhou calls jail since she was kidnapped by the Canadian government because of US extraterritorial justice.

 

[horse]

SMIC qualified its 130nm process in 2001. TSMC qualified it in 1998. A gap of 3 years.
SMIC had volume production of 130nm 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 around 300-400).

You are just saying you are an expert to what happened 20 years ago.

That is simply amazing!

 

[gelgoog]

SMIC qualified its 130nm process in 2001. TSMC qualified it in 1998. A gap of 3 years.
SMIC had volume production of 130nm 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.

Now you are getting all mixed up. 180nm is not 130nm. And no, SMIC did not have volume production at 180nm in 2002.
Now I ask you, which product was SMIC manufacturing that was sold in the market 2001 or 2002?
I have given you plenty of primary sources. You only give the second hand reports.

And SMIC not only hired engineers, they licensed technology back then. For example.

The 180nm process you are talking about was licensed from Chartered Semiconductor.
Chartered was a Singaporean company which is now part of GlobalFoundries.

Please, Log in or Register to view URLs content!

"Shanghai,China [2001-12-21]
...
Semiconductor Manufacturing International Corporation (SMIC), China's first advanced open-IC foundry, and Chartered Semiconductor Manufacturing (NASDAQ: CHRT and SGX-ST: Chartered), one of the world's top three silicon foundries, today announced they have entered into an alliance. Upon execution of the detailed agreements, SMIC will receive 0.18-micron baseline logic process technology transfer and be granted patent license rights from Chartered, as Chartered will receive an equity stake and access to capacity in SMIC. Financial terms were not disclosed.
...
SMIC recently entered into pilot production producing 8-inch wafers at 0.25-micron (and below) technology, and it is expected to reach full capacity producing 85,000 8-inch wafers per month by the end of 2004. SMIC devices in the initial production phase are expected to include SRAM, ASIC RAM, LOGIC, and other kinds of chips with applications in digital TV, VCD/DVD"

8-inch = 200mm.

Notice how that news is from 21 Dec 2001 and back then SMIC was manufacturing at 250nm with 200mm wafers according to the news article. Prior to that, in September 2001, the VIA C3 was being sold and it was at TSMC in 130nm with 300mm wafers.

Then SMIC later licensed processes from IBM.

Please, Log in or Register to view URLs content!

Let's just assume SMIC is 2yrs behind in 2002. Then it can be expected they would fall even further behind, assuming free market competition regardless of whether or not their are sanctions or not. Why would anyone choose a chip that is two years behind when they have the option of choosing a cutting edge children for the same cost? Being two years behind is still a huge disadvantage for SMIC when it comes to market competition. As a result, SMIC would have less revenues and profits to invest back in R&D, and the tech gap will only increase. That is why so many chip fabs fell out of the game over the last two decades, fabbing chips is a winner takes all business. To attribute SMIC falling behind to the success of sanctions is a logical fallacy. Conversely, with the imposition of total sanctions, SMIC is actually in a more advantageous position, as now it has no competition within certain domestic market segments, and will have the revenues as well as governmental support drive its R&D efforts.

No, they were at least 4 years behind back then. If not more.

 

[nlalyst]

You are just saying you are an expert to what happened 20 years ago.

That is simply amazing!

Haha. What would you say about historians who study ancient Chinese history?

 

[FriedRiceNSpice]

No, they were at least 4 years behind back then. If not more.

Was assuming for the sake of argument to prove the assertion that sanctions on China were successful is logically incorrect.

 

[horse]

Haha. What would you say about historians who study ancient Chinese history?

Enough with the fake discussion. SMIC is not a library.

You cannot get your head around to the fact that SMIC is a different company in a different place now today than it was a few years ago.

Why is that?

Simply because the US sanctions has made it a priority for it to succeed.

Before, it was it would be nice it would succeed, now it is one of those Chinese communist projects as important as the bomb.

Those US sanctions against Chinese semiconductors did more for the push to development than Xi Jinping talking about China 2025 ever did. Now there is real urgency, when before there was none at all.

Before it was win-win in the worldwide IC industry. Now the Americans have created a powerful enemy in IC. Who happens to be the largest market for the end product, the semiconductor chip.

What happened 20 years ago, no relevance to the business conditions of today.

 

[jfcarli]

This is true when you’re developing *new* workflows, but not when you’re reproducing old ones.

Whe you think of it, the whole systems analytical work was done. Essentially it is a matter of coding.

@latenlazy is correct . If you already know very clearly what you want to do, just get an army of coders and put then coding day and night. You will find out that your "localized" EDA will come very very quickly. And probably not a single line of coding will be needed to be copied, thus avoiding IP claims.

Knowing what you want to do, analysing the system, understanding where you are going from point a to point b, that is what is really costly and time consuming.

The architectural and the engineering part is already fully done. Just code. Coders are not THAT expensive. I am not undersestimating the coding work, but it is far, far, far from starting from scratch.

 

[nlalyst]

Now you are getting all mixed up. 180nm is not 130nm. And no, SMIC did not have volume production at 180nm in 2002.

I fixed that booboo.

Now I ask you, which product was SMIC manufacturing that was sold in the market 2001 or 2002?
I have given you plenty of primary sources. You only give the second hand reports.

Your first hand source turned out to be irrelevant, because you couldn't find a single TSMC 180nm product in 1998, the alleged year when the technology was available. All we were left with were 2nd hand sources of equal credibility. Back to square one.

Here's a "highly prized" 1st hand source from SMIC, talking about its Top Fab award in 2003:

Please, Log in or Register to view URLs content!

The first fab facility in China to receive this honor, SMIC’s rapid advancement has significantly narrowed the foundry’s IC technology gap with global players rolling out fully qualified 0.35-0.18-micron processes and 0.13-micron copper backend services in a little over 1.5 years since initial pilot production.

In 2003 TSMC filed a lawsuit against SMIC for copying 90% of their 180nm node process:
As a point of proof, TSMC referred to the

Please, Log in or Register to view URLs content!

that SMIC received from

Please, Log in or Register to view URLs content!

in 2003, highlighting the fact that just four months after installing equipment in its fab, SMIC had four processes up and running, manufacturing 18 different products. Adding to that suspicion was the claim that SMIC hired away 100+ TSMC employees that had access to the sensitive process data required to bring a fab to production. To begin the legal discovery process, TSMC analyzed SMIC .18m silicon from a Broadcom product and documented stark similarities to the identical product silicon from TSMC. With discovery came incriminating emails which are a centerpiece of the case.

The above gives us an upper range of a SMIC 180nm product available on the market: 2003. First TSMC products appeared on the market in 2000. Yet again, a 3 year gap.

 

[FriedRiceNSpice]

Whe you think of it, the whole systems analytical work was done. Essentially it is a matter of coding.

@latenlazy is correct . If you already know very clearly what you want to do, just get an army of coders and put then coding day and night. You will find out that your "localized" EDA will come very very quickly. And probably not a single line of coding will be needed to be copied, thus avoiding IP claims.

Knowing what you want to do, analysing the system, understanding where you are going from point a to point b, that is what is really costly and time consuming.

The architectural and the engineering part is already fully done. Just code. Coders are not THAT expensive. I am not undersestimating the coding work, but it is far, far, far from starting from scratch.

True, you don't actually need software engineers who have past experience with working on EDA systems. You just need a small core of domain knowledge experts that thoroughly understand the problem space. Whatever domain knowledge the senior engineers would need to know can be knowledge transfered from your core group of experts, most ICs would need only limited amounts of domain specific knowledge.

However, this process still takes time, and throwing more engineers at the problem will not speed it up further. There is a popular adage in the industry that goes along the lines of, "Nine mothers cannot have one baby in one month." Would expect it would take a few years, but still very fast relative to other technical challenges such as developing EUV machines, node processes, and light sources.