Chinese semiconductor industry

antonius123 said:

Bro, by 2023, smartphone with 2nm chipset will be introduced, and smartphone with 3nm SOC will be available in many smartphones already.

Click to expand...

@antonius123 bro it may cost an arm to purchase(especially a Apple product)...lol with the present economic situation, my calculation 2025 is the date of TSMC 2nm introduction so a generation behind if China is able to deliver.

Addendum: I need to ask some experts here if 7nm 3d stacking chipset is possible? if yes then its equivalency is that of a 5nm chip?

antonius123 said:

Bro, by 2023, smartphone with 2nm chipset will be introduced, and smartphone with 3nm SOC will be available in many smartphones already.

Click to expand...

Then it means temporarily ceding the market for flagship smartphones while continuing to contest the midrange smartphone domain, until such a point that the domestic semiconductor industry (including not only fabs, but also the suppliers of lithography systems, and subsuppliers that provide components for those systems, and EDA, basically the entire industry) reaches technological parity.


This focus on the most bleeding edge processes for the short term is a little bit detrimental to the long term war which is one of technological and semiconductor self sufficiency.
While it is important for industry to focus on that sort of bleeding edge technology, to get there requires mastery, scaling and profits in mature processes with wide applications, to be able to channel reinvestment into advancing your own R&D and products.
Leading edge SOCs are mostly used on flagship smartphones -- a profitable market, yes, but not the only one that exists.

As I wrote in my previous post, the question is a simple one -- how long will it take for China to close the technology gap, and will it be sufficient to achieve the win conditions for the technological war in the context of national geopolitical competition?


I don't think any of us can provide an answer yet.

ansy1968 said:

@WTAN @krautmeister @FairAndUnbiased @foofy Sir it had a lot of technical jargon and the author seems confident, need your opinion if his confidence is justify?

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

922 views1 week ago

CC

Click to expand...

Why did they disclose this, as it could well be national secret?

antonius123 said:

Exactly.

There is Wassenaar Arrangement that US can use to prevent China from obtaining such components including light source etc, so it is not about competitive market force anymore. If the coverage still not include these items now, it could be, by US intention to thwart China. So funding alone won't be enough.

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

Click to expand...

The economic losses from these sanctions will have to be recouped fully, by penalizing the big bully someday.

antonius123 said:

Exactly.

There is Wassenaar Arrangement that US can use to prevent China from obtaining such components including light source etc, so it is not about competitive market force anymore. If the coverage still not include these items now, it could be, by US intention to thwart China. So funding alone won't be enough.

Click to expand...

the issue is that this only works if China has absolutely nothing. But that's not the case. China has SMEE, AMEC, Naura, etc so it can focus efforts into government institutes developing the parts that commercial industry can't deliver yet.

It is not that big a step to go from 90 nm to 7 nm since you know what you need to improve on. It's a huge step to go from nothing to even 180 nm since you need to do everything else. Luckily, 'everything else' has been solved.

Bltizo said:

Then it means temporarily ceding the market for flagship smartphones while continuing to contest the midrange smartphone domain, until such a point that the domestic semiconductor industry (including not only fabs, but also the suppliers of lithography systems, and subsuppliers that provide components for those systems, and EDA, basically the entire industry) reaches technological parity.


This focus on the most bleeding edge processes for the short term is a little bit detrimental to the long term war which is one of technological and semiconductor self sufficiency.
While it is important for industry to focus on that sort of bleeding edge technology, to get there requires mastery, scaling and profits in mature processes with wide applications, to be able to channel reinvestment into advancing your own R&D and products.
Leading edge SOCs are mostly used on flagship smartphones -- a profitable market, yes, but not the only one that exists.

As I wrote in my previous post, the question is a simple one -- how long will it take for China to close the technology gap, and will it be sufficient to achieve the win conditions for the technological war in the context of national geopolitical competition?


I don't think any of us can provide an answer yet.

Click to expand...

If DUV and then EUV machines are perfected in the next 3 years, I reckon another 5-10 years to build out 200+ fabs in China

That should take China to over 50% of global semiconductor production.

I reckon there's a 80% chance of this scenario happening.

Bltizo said:

If the US chooses to wholesale block semiconductors of specific nodes to Chinese companies, that's just something that Chinese companies will have to accept --- it means that those companies will have to temporarily exit the market of the most advanced products until domestic semiconductor foundries catch up.

Click to expand...

I don't see a reason for China to take such a tolerant attitude toward this. There are many ways to retaliate both symmetrically and asymmetrically, and that's considering only the economic sphere. There are also crazier ways of retaliation, like the PLA Rocket Force getting involved in the "analytical deconstruction" of the Taiwanese semiconductor industry.

If you think that's extreme, like you said this is a war. Your enemy should never get too comfortable with your rationality and sanity in a war.

Bltizo said:

The question is a pretty simple one -- how long will it take for China to close the technology gap, and will it be sufficient to achieve the win conditions for the technological war in the context of national geopolitical competition?
I think it is still an open question.

Click to expand...

You mean the answer isn't 2025 and yes?

More seriously, I think it depends on your definition of "close the technological gap." I believe @WTAN (feel free to correct me if I'm wrong) mentioned that China has already reached parity in the optical components of EUVL and the light source should be ready by the end of the 14th FYP. So, knock on wood, China should have a first-generation EUV capability by 2025 and move into the sub-7nm realm with indigenous technology. Others will still remain ahead, but there won't be an impassible chasm between them and Chinese companies. Moving past 2025, there are very interesting candidate technologies - like industrial particle accelerators - that could leapfrog China to the cutting edge:
Note especially the last sentence of the article.

As for leveraging technological parity, let me quote a previous post of mine on this topic:

ZeEa5KPul said:

I don't think the US has even begun to realize how badly it dun goof'd with its semiconductor sanctions. China having an indigenous EUVL capacity does a lot more than just give it technological independence, it allows it to use its massive share of the semiconductor market coercively. Armed with its EUVL fabs, the Chinese government can phone up the CEOs of Qualcomm et al. and tell them that if they want to continue selling chips in China, i.e., selling chips at all, they'd better use Chinese fabs to make those chips.

China can use a complete semiconductor tech tree to leverage being the largest semiconductor consumer to become the largest semiconductor producer almost overnight.

Click to expand...

antonius123 said:

Bro, by 2023, smartphone with 2nm chipset will be introduced, and smartphone with 3nm SOC will be available in many smartphones already, and smartphone with 5nm will be considered mid range, and 7nm will be considered between mid to lower range.

Click to expand...

Processor performance is not even in the top 5 when it comes to consumer priorities.
More than half of the people don't even desire more performance and only 25% of people are frustrated with the speed of their phone.

The top 5 uses of a smartphone are texting, social media, phone calls, watching videos and web surfing. These five collectively make more than 75% of what people spend time with while using their smartphones. None of these depend on processor performance. Huawei can still be competitive while being 4 generations (approximately 70% performance difference) behind. They need to get their application ecosystem sorted though. That is much more important.

BoraTas said:

Please, Log in or Register to view URLs content!

Processor performance is not even in the top 5 when it comes to consumer priorities.

Please, Log in or Register to view URLs content!

More than half of the people don't even desire more performance and only 25% of people are frustrated with the speed of their phone.

The top 5 uses of a smartphone are texting, social media, phone calls, watching videos and web surfing. These five collectively make more than 75% of what people spend time with while using their smartphones. None of these depend on processor performance. Huawei can still be competitive while being 4 generations (approximately 70% performance difference) behind. They need to get their application ecosystem sorted though. That is much more important.

Click to expand...

More people are probably more concerned with storage size than processing speed, unless they are using it for gaming.

I am guessing that the new drive (need) for performance from consumers will be Augmented Reality (AR) applications in mobile phones. Not today but in probably 2-3 years it will be a significant factor

Currently AR is underdeveloped which is only used for some niche applications