Chinese semiconductor thread II

Wrought

Junior Member
Registered Member
Government listed goal is by 2030 it should be production using its EUV.

The window cannot stretched too long...
For example, 20 years would be unacceptable

You're missing the point, EUV is not some kind of magic wand. Samsung has EUV. Intel has EUV. Yet neither of them are genuine competition for TSMC. It is only one piece of the puzzle, necessary but not sufficient.
 

latenlazy

Brigadier
No, even assuming all goes smoothly and the EUV prototype is delivered next year, the necessary testing, validation, integration, mass production, ramp up, and so on will take many years to complete. It is perfectly correct to say that genuine competition with TSMC at the leading edge is a long ways off.

That's not to say domestic development is not happening quickly, but there is still plenty of room for improvement.
The tool is such an essential enabler relative to the alternative that even a prototype will likely get meaningfully productive use. The first EUV instruments used by TSMC were prototype instruments that received field upgrades. You don’t need EUV for all the scanning layers either.


You're missing the point, EUV is not some kind of magic wand. Samsung has EUV. Intel has EUV. Yet neither of them are genuine competition for TSMC. It is only one piece of the puzzle, necessary but not sufficient.
Yes…but the other parts of the fab process are already being employed by persisting with DUV at much smaller nodes.
 

Wrought

Junior Member
Registered Member
The tool is such an essential enabler relative to the alternative that even a prototype will likely get meaningfully productive use. The first EUV instruments used by TSMC were prototype instruments that received field upgrades. You don’t need EUV for all the scanning layers either.



Yes…but the other parts of the fab process are already being employed by persisting with DUV at much smaller nodes.

Of course it will be put into production, and of course it will represent a big step forwards. That's not the same as scaling up to TSMC level overnight. EUV will give domestic companies the tools they need to properly compete; it doesn't mean they automatically jump to first place.

.
I see Liang mong song and his process team is as good as TSMC. Using old DUV to do 5nm and that's a step beyond TSMC did.

Sure, hiring ex-TSMC employees is a great way to transfer institutional knowledge and certainly helps to catch up. Using DUV for 5nm is a suboptimal stopgap though, hardly the first choice. Once they have EUV, they will abandon it.
 

Oldschool

Junior Member
Registered Member
I see alot of troubles for TSMC. Fewer demand from customers for its high end chips as apple and Nvidia losing share in China . The tools are prohibitively expensive.
Hard to recoup it's investment.
Also US government wasted alot of TSMC money as it demands TSMC to build 6 fabs in Arizona. Who's going to buy all those chips??
 

Oldschool

Junior Member
Registered Member
Of course it will be put into production, and of course it will represent a big step forwards. That's not the same as scaling up to TSMC level overnight. EUV will give domestic companies the tools they need to properly compete; it doesn't mean they automatically jump to first place.



Sure, hiring ex-TSMC employees is a great way to transfer institutional knowledge and certainly helps to catch up. Using DUV for 5nm is a suboptimal stopgap though, hardly the first choice. Once they have EUV, they will abandon it.
Well, that suboptimal stop gap currently is making apple losing shares.
And not to use Nvidia. TSMC will feel it.
 

Wrought

Junior Member
Registered Member
Well, that suboptimal stop gap currently is making apple losing shares.
And not to use Nvidia. TSMC will feel it.

No, the fact that Huawei is using a suboptimal stopgap to produce handsets is making Apple lose market share. There is a big difference between the components and the consumer product; the former enables the latter, but they are not the same thing.

TSMC will certainly feel its loss of monopoly, but again, there is a difference between not having a monopoly and running neck-and-neck.
 

Oldschool

Junior Member
Registered Member
No, the fact that Huawei is using a suboptimal stopgap to produce handsets is making Apple lose market share. There is a big difference between the components and the consumer product; the former enables the latter, but they are not the same thing.

TSMC will certainly feel its loss of monopoly, but again, there is a difference between not having a monopoly and running neck-and-neck.
The world divides into 2 camps. Western tech and China's non Western tech and neither trust each other.
I don't think China is competing with TSMC.
TSMC serves only Western market and its gradually being phased out inside China despite having superior tech.

Likewise China low end chips not accepted in Western world.

Heading into separate paths. Big fuss about nothing
 

Hyper

Junior Member
Registered Member
I see alot of troubles for TSMC. Fewer demand from customers for its high end chips as apple and Nvidia losing share in China . The tools are prohibitively expensive.
Hard to recoup it's investment.
Also US government wasted alot of TSMC money as it demands TSMC to build 6 fabs in Arizona. Who's going to buy all those chips??
Nvidia GPU keep getting smuggled. They are peerless atleast right now.
 

Wrought

Junior Member
Registered Member
The world divides into 2 camps. Western tech and China's non Western tech and neither trust each other.
I don't think China is competing with TSMC.
TSMC serves only Western market and its gradually being phased out inside China despite having superior tech.

Likewise China low end chips not accepted in Western world.

Heading into separate paths. Big fuss about nothing

Neither US nor China has the power to control international trade flows to such a degree. Just look at how leaky US sanctions are right now, and we are talking about much smaller countries with much less trade than China.

No, the way to compete with TSMC on the global market is to undercut them at the low-end, gradually take market share, deprive them of R&D revenue, and finally surpass them technologically. But that is a long and difficult process which will take many years. It's possible that domestic semiconductor companies will never surpass TSMC at fabrication, the same way domestic automobile companies never surpassed Toyota at ICE. Instead new technology changed the rules of the game for everyone, and Toyota went from leader to loser.
 
Top