Chinese semiconductor industry

Status
Not open for further replies.

dfrtyhgj

Junior Member
Registered Member
China and Chinese companies shouldn't deglobalize or isolate themselves but it should definitely de-Americanize at least its electronics industry and a good part of its technology sector if it wants those sectors to survive beyond 2030. It is not even a question of nationalism but of survival. There are groups of powerful neocons within the United States determined to stop the technological advance in China, they will do whatever its takes to archive that, even if they have to force the allies to comply. They are like terminators, you can't negotiate with them. If Trump wins in 2025, I think the situation could to be significantly worse than in 2018-2020. But if Chinese companies manage to de-Americanize a good chunk of their supply chain successfully, it will be more difficult for the United States to talk to their allies to isolate China.​
Europe has also proven themselves to be unreliable. All of the West is about to fall and China knows it.
 

tokenanalyst

Brigadier
Registered Member
The standard for self sufficiency from Russia = capable of doing it with foreign tools like how they imported Toshiba machine tools

The standard for self sufficiency from China = not even a single speck of paint can be from a foreign country otherwise it is the deepest dependence in the world.
A random self-sufficient goal in Technology Nodes could be:

>45nm: close 95%-99%
>14nm: 80%-90%
<14nm: above 50% as possible
 

ChongqingHotPot92

Junior Member
Registered Member
The standard for self sufficiency from Russia = capable of doing it with foreign tools like how they imported Toshiba machine tools

The standard for self sufficiency from China = not even a single speck of paint can be from a foreign country otherwise it is the deepest dependence in the world.
Yea, but the thing is the US has no lithography machine sanctions on Russia, but has them on China. The speaker (Ma) has clearly stated that China needs to incentivise domestic industries (say through subsidies) to use 90nm chips (and design their products around such outdated chips) for now (as much as this would be a temporary setback for the smartphone and PC industry) before the SMEE's 28nm and 14nm lithography machines are ready. While such strategy would set the Chinese electronics industry 10 years behind the West and South Korea, at least China could begin the process of achieving 100% self sufficiency.

Ultimately, China is going to be 2-3 generations behind TSMC and South Korea regarding chips' nm counts for the foreseeable future, but there are ways to design final products (like cellphones, computers, AI tools, weapons, etc.) around such bottleneck.

Just curious, could chip (processors in general) be one of the reasons why Chinese and Russian solid-fuel SAMs (and some other guided weapons) tend to have bigger size but shorter range compared to their most recent American counterparts? Just asking.
 

FairAndUnbiased

Brigadier
Registered Member
can you elaborate on which parts the poster was wrong?

My thoughts are the photonic chips may be a dead-end? but others with more knowledge can comment.
He is a frequent hater and naysayer to anything Chinese and nowhere near a subject matter expert in anything even closely related to semiconductor.

For example the recommendation to train more physics, chemistry and mechanical engineering graduates as a short term solution to semiconductor talent is utterly laughable. The easy problems have been solved already, how can new graduates help especially ones not specifically trained in the area?

That's why existing engineers are being poached: not because Chinese fabs don't know anything but because they already know nearly everything. Basic understanding is not useful for advancement. Maintaining - yes, cultivate domestic talent pool. But if you want to advance quickly on extreme bottlenecks you need someone to have actually done the exact specific thing before.

For instance WTF does a new grad know about 2 nm precision wafer stages? Nothing. He might know the principles of a fast actuated stage but you don't need principles at this point (developing leading edge equipment with specific bottlenecks). You need a guy who has specifically designed and seen nm precision wafer stages being made.

Same with light sources. Many people know the principles of EUV production. Many people even know the actual implementation of an EUV source. Ok now you need one in the exact form factor fitting your current equipment to put it into operation.... And you don't have that. And even if you did you don't have some highly specific lenses like Mo/Si multilayer lenses.

So you tell me how slowly training nonspecific talent helps solve an extremely immediate problem? It doesn't.
 

FairAndUnbiased

Brigadier
Registered Member
Yea, but the thing is the US has no lithography machine sanctions on Russia, but has them on China. The speaker (Ma) has clearly stated that China needs to incentivise domestic industries (say through subsidies) to use 90nm chips (and design their products around such outdated chips) for now (as much as this would be a temporary setback for the smartphone and PC industry) before the SMEE's 28nm and 14nm lithography machines are ready. While such strategy would set the Chinese electronics industry 10 years behind the West and South Korea, at least China could begin the process of achieving 100% self sufficiency.

Ultimately, China is going to be 2-3 generations behind TSMC and South Korea regarding chips' nm counts for the foreseeable future, but there are ways to design final products (like cellphones, computers, AI tools, weapons, etc.) around such bottleneck.

Just curious, could chip (processors in general) be one of the reasons why Chinese and Russian solid-fuel SAMs (and some other guided weapons) tend to have bigger size but shorter range compared to their most recent American counterparts? Just asking.

I'll address the last point first since it is the most blatantly wrong. First you have to provide evidence that your assertion is actually true. Then you
Please, Log in or Register to view URLs content!


Then you need to realize that all mission critical aerospace chips are on 180 nm or larger processes from the 1990s, because they need to resist radiation damage so they're usable during a total strategic exchange.

Larger feature sizes means resistance to lattice displacement effects. A few thousand atoms being displaced every square mm due to radiation damage is nothing for a 180 nm device but can be devastating to a 32 nm device for instance where line widths are a few hundred atoms wide.

Larger chip features means higher sustained voltage which means relatively lower susceptibility to effects such as radiation photocurrent. A 0.5 V spike for instance is devastating to a
Please, Log in or Register to view URLs content!
because it's halfway between the on and off states. A 0.5 V spike is nothing to an old school 3.3V or 5V TTL device.

You also don't know the specific sanctions on Russia. In fact Russia is under total sanctions at the moment.
 

tokenanalyst

Brigadier
Registered Member
He is a frequent hater and naysayer to anything Chinese and nowhere near a subject matter expert in anything even closely related to semiconductor.

For example the recommendation to train more physics, chemistry and mechanical engineering graduates as a short term solution to semiconductor talent is utterly laughable. The easy problems have been solved already, how can new graduates help especially ones not specifically trained in the area?

That's why existing engineers are being poached: not because Chinese fabs don't know anything but because they already know nearly everything. Basic understanding is not useful for advancement. Maintaining - yes, cultivate domestic talent pool. But if you want to advance quickly on extreme bottlenecks you need someone to have actually done the exact specific thing before.

For instance WTF does a new grad know about 2 nm precision wafer stages? Nothing. He might know the principles of a fast actuated stage but you don't need principles at this point (developing leading edge equipment with specific bottlenecks). You need a guy who has specifically designed and seen nm precision wafer stages being made.

Same with light sources. Many people know the principles of EUV production. Many people even know the actual implementation of an EUV source. Ok now you need one in the exact form factor fitting your current equipment to put it into operation.... And you don't have that. And even if you did you don't have some highly specific lenses like Mo/Si multilayer lenses.

So you tell me how slowly training nonspecific talent helps solve an extremely immediate problem? It doesn't.
Although I agree that it would be very nice to have someone who has a lot of specific experience in the problem you are working on, it is almost always impossible to find someone like that. What you need is people with experience close to the problem you are working on. For example if you are designing a wafer stage you need someone with experience in mechatronics and vibration control, probably who also knows python and c ++, they do not necessarily have to have worked in the lithography industry, their experience in lithography will be develop in the job. If you look at the ASML job listing, they almost always ask you for generic knowledge and experience. You are supposed to work as a team solving complex problems.

Group Lead Mechatronics Optical Column DUV​

Education and experience​

Education
Technical University (Master) or PhD in Mechanical / Mechatronic Engineering, Applied Physics or Electrical Engineering

Experience
Relevant leadership experience as manager or senior architect in a multidisciplinary working environment. Knowledge in electro-magnetic design and/or applied (electro) magnetic research, interests in position measurement systems, handling systems, control engineering, actuator/motor design, dynamics.

Skills​

  • Strong personal communicator
  • Academic mindset with a pragmatic approach
  • Highly motivated team player with excellent social, coordination and communication skills
  • Ability to work in an international multicultural setting
  • Entrepreneurship and initiative in a technically complex and very dynamic environment
  • Excellent/proven stakeholder management skills
  • Not afraid to build inter-continental working relations
  • Relevant technical experience build up
  • Experience of working in/with (scaled) agile team is a pre
 

FairAndUnbiased

Brigadier
Registered Member
Although I agree that it would be very nice to have someone who has a lot of specific experience in the problem you are working on, it is almost always impossible to find someone like that. What you need is people with experience close to the problem you are working on. For example if you are designing a wafer stage you need someone with experience in mechatronics and vibration control, probably who also knows python and c ++, they do not necessarily have to have worked in the lithography industry, their experience in lithography will be develop in the job. If you look at the ASML job listing, they almost always ask you for generic knowledge and experience. You are supposed to work as a team solving complex problems.
ASML has established IP for wafer stages and component suppliers. They have documentation. Job posting is for a sustaining engineering type position who delegates to an existing team that already knows what they're doing.

None of that applies to a breakthrough situation.
 

gadgetcool5

Senior Member
Registered Member
He is a frequent hater and naysayer to anything Chinese and nowhere near a subject matter expert in anything even closely related to semiconductor.

For example the recommendation to train more physics, chemistry and mechanical engineering graduates as a short term solution to semiconductor talent is utterly laughable. The easy problems have been solved already, how can new graduates help especially ones not specifically trained in the area?

That's why existing engineers are being poached: not because Chinese fabs don't know anything but because they already know nearly everything. Basic understanding is not useful for advancement. Maintaining - yes, cultivate domestic talent pool. But if you want to advance quickly on extreme bottlenecks you need someone to have actually done the exact specific thing before.

For instance WTF does a new grad know about 2 nm precision wafer stages? Nothing. He might know the principles of a fast actuated stage but you don't need principles at this point (developing leading edge equipment with specific bottlenecks). You need a guy who has specifically designed and seen nm precision wafer stages being made.

Same with light sources. Many people know the principles of EUV production. Many people even know the actual implementation of an EUV source. Ok now you need one in the exact form factor fitting your current equipment to put it into operation.... And you don't have that. And even if you did you don't have some highly specific lenses like Mo/Si multilayer lenses.

So you tell me how slowly training nonspecific talent helps solve an extremely immediate problem? It doesn't.
My point is that there are no short term solutions.

You won't be able to poach talent because most non-Chinese and especially non-Asian cutting edge lithography engineers don't want to migrate to China to work on lithography. If you look on the ASML page for senior engineers, 90% of them are Caucasian. And even for the few engineers who are Chinese enough that they may want to move to China for a high salary (and keep in mind, average salaries in this field in China aren't necessarily higher than in the West), recruitment can be banned:
Please, Log in or Register to view URLs content!
And frankly, I wouldn't rule out the US imposing exit visa bans in the future for personnel who are extremely skilled. Heck, even the West is facing a shortage of highly skilled semiconductor line engineers.

You cannot just throw money, money and R&D at the problem to solve it. It requires a process of step-by-step experimentation where each step depends on the previous step, and throwing x10 more people at the same step does not make it go faster because there is a minimum physical time required to do each step. Therefore, there is a limit to the speed you can go and China will always be behind when chasing a moving target. Especially when the West has more skilled engineers who have 20+ years of experience in the field which is what you really need. Inferior technology means inferior market position, which means less money coming into your industry, which is vicious cycle where you become dependent on government largesse but you are never actually sustaining yourself on your own revenue. In short, your technology might be bought by the military or central government, but only due to political reasons. If the political issues ever become resolved between the US and China, all your efforts could be abandoned due to inefficiency. This is what happened to the USSR and also China in the past in aerospace.

Therefore, there is no way to overcome the first mover advantage for existing technologies. And there is no way to catch up in the short term. It is like a game of "Go" where your enemy has already surrounded an area with its pieces. You cannot try to place more pieces in that area as it is too late. The traditional semiconductor game for China is lost. You can only try to create a bigger circle or take territory somewhere else.

That is why I recommend China to look towards leapfrog technologies like photonic chips where it can have a first mover or early mover advantage.
 
Last edited:
Status
Not open for further replies.
Top