Chinese semiconductor industry

coolgod said:

TMSC has fabs outside Taiwan, I don't see why SMIC can't build fabs or buy fabs outside mainland China in the next 10 years.

Click to expand...

Because you need to think of the location. TSMC has a fab in China and the US. Do you think SMIC is going to be allowed to open a fab in either of those locations?

North America = Zero Chance
South America = Lack of Workers
Western Europe = Zero Chance
Eastern Europe = Lack of Workers
Russia = Possible Option but just export chips from China
Africa = Lack of Workers
Middle East = Lack of Water
ASEAN = Too close to China
East Asia (Not China) = Zero Chance.

FriedButter said:

Because you need to think of the location. TSMC has a fab in China and the US. Do you think SMIC is going to be allowed to open a fab in either of those locations?

North America = Zero Chance
South America = Lack of Workers
Western Europe = Zero Chance
Eastern Europe = Lack of Workers
Russia = Possible Option but just export chips from China
Africa = Lack of Workers
Middle East = Lack of Water
ASEAN = Too close to China
East Asia (Not China) = Zero Chance.

Click to expand...

I think if the question is "are there any prospects of selling LPP EUV machines outside of China," the answer should be "yes, but in the medium to long term" and of course not just to SMIC but to potential other international fabs that other nations are likely to aim and set up themselves.

Future potential customers for Chinese LPP EUV may well be companies and organizations that have yet to be founded, and if we are talking about the idea of SMIC setting up fabs outside of China, then that would depend on the geopolitical relationship with other host nations in the medium to long term future.


This isn't to say a large customer base in the future is guaranteed (far from it), but I do think at present that we should not dismiss the idea that LPP EUV machines will not sought to be exported in future simply due to the sheer world of possibilities that it offers.

FriedButter said:

Because you need to think of the location. TSMC has a fab in China and the US. Do you think SMIC is going to be allowed to open a fab in either of those locations?

North America = Zero Chance
South America = Lack of Workers
Western Europe = Zero Chance
Eastern Europe = Lack of Workers
Russia = Possible Option but just export chips from China
Africa = Lack of Workers
Middle East = Lack of Water
ASEAN = Too close to China
East Asia (Not China) = Zero Chance.

Click to expand...

Russia is a good location to build fabs. Russia already has operational fabs from domestic Mikron. Russian fabs service the entire Eastern European and Central Asian market.

Malaysia is also a semiconductor middle power with Lam and Amat facilities and multiple analog fabs such as Infineon and Xfab as well as domestic SilTerra.

Don't need to build everywhere, there's low hanging fruit already. You don't need super high GDP per capita to have a semiconductor industry.

FriedButter said:

Because you need to think of the location. TSMC has a fab in China and the US. Do you think SMIC is going to be allowed to open a fab in either of those locations?

North America = Zero Chance
South America = Lack of Workers
Western Europe = Zero Chance
Eastern Europe = Lack of Workers
Russia = Possible Option but just export chips from China
Africa = Lack of Workers
Middle East = Lack of Water
ASEAN = Too close to China
East Asia (Not China) = Zero Chance.

Click to expand...

What do you think will happen to TSMC's oversea Fabs when Taiwan reunities or if TSMC bankrupts?
I can also see China exporting LPP EUVs to Samsung or TSMC in the next 10 years, never say never.

Remember, Samsung and TSMC don't have to buy China's LPP EUV machines, but the fact China offered it on the international market means ASML have to lower their selling price. Anything China can to do pay back ASML by killing its bottom line is a win in my book.

FairAndUnbiased said:

Russia is a good location to build fabs. Russia already has operational fabs from domestic Mikron. Russian fabs service the entire Eastern European and Central Asian market.

Malaysia is also a semiconductor middle power with Lam and Amat facilities and multiple analog fabs such as Infineon and Xfab as well as domestic SilTerra.

Don't need to build everywhere, there's low hanging fruit already. You don't need super high GDP per capita to have a semiconductor industry.

Click to expand...

The problem with building anything in Russia, is how incompetent local government and mid-level federal officials can be. Otherwise yeah, Russia has the human capital, economic power, and good relations with China to become a major customer.

coolgod said:

What do you think will happen to TSMC's oversea Fabs when Taiwan reunities or if TSMC bankrupts?

Click to expand...

The only other fab is Arizona, which the US will simply take over.

coolgod said:

I can also see China exporting LPP EUVs to samsung or TSMC in the next 10 years, never say never.

Click to expand...

Aren’t we talking about SMIC building oversea fabs in 10 year timeframe? Exporting LPP EUVs is a different matter, which has a more optimistic market in the same timeframe.

FairAndUnbiased said:

Russia is a good location to build fabs. Russia already has operational fabs from domestic Mikron. Russian fabs service the entire Eastern European and Central Asian market.

Click to expand...

While it is a good location. The problem comes down to feasibility. Everything is already in China. Is it really worth spending billions to build a fab and to set up a new supply chain when you can simply export it?

coolgod said:

I can also see China exporting LPP EUVs to Samsung or TSMC in the next 10 years, never say never.

Click to expand...

Samsung and TSMC contributed $1.5B to ASML's R&D efforts for EUV in exchange for a 3% and 5% minority stake in ASML.

Intel contributed $4B to ASML's R&D effort for EUV in exchange for 15% stake in ASML. So basically you are left with SK Hynix and Micron as potential customers of Chinese EUV. Micron under US gov't pressure will unlikely adopt, and SK Hynix could conceivably adopt Chinese EUV for their China-only fabs.

In short, Chinese EUV will never "kill ASML EUV", not when Samsung+TSMC+Intel are the main EUV adopters and own up to 23% minority ownership stake in ASML and are ASML's largest funders of EUV R&D to begin with.

HighGround said:

The problem with building anything in Russia, is how incompetent local government and mid-level federal officials can be. Otherwise yeah, Russia has the human capital, economic power, and good relations with China to become a major customer.

Click to expand...

The most modern Russian fabs are in Moscow region. Zelenograd. The Moscow Mayor is quite competent.

Still, I think joint ventures with Russian fabs are way more likely. If China is to expand their semi industry worldwide through, Russia, Brazil, India or Vietnam would be likely candidates I think. But the whole economics of semis mean it makes more sense to centralize operations in China. IMHO one of the worst mistakes the Soviets did in semis was actual lack of centralization of effort. It meant they never reached the economies of scale to compete with the US semi industry.

Phead128 said:

Samsung and TSMC contributed $1.5B to ASML's R&D efforts for EUV in exchange for a 3% and 5% minority stake in ASML.

Intel contributed $4B to ASML's R&D effort for EUV in exchange for 15% stake in ASML. So basically you are left with SK Hynix and Micron as potential customers of Chinese EUV. Micron under US gov't pressure will unlikely adopt, and SK Hynix could conceivably adopt Chinese EUV for their China-only fabs.

In short, Chinese EUV will never kill ASML EUV, not when Samsung+TSMC+Intel are the main EUV adopters and own up to 23% minority ownership stake in ASML and are ASML's largest funders of EUV R&D to begin with.

Click to expand...

But Huawei, SMIC and SMEE combination can give Apple, TSMC and ASML a run for its money. And the ultimate winner are the peripheries, SK HYNIX ,UMC, Huahong and even Samsung ( they need something special to compete with TSMC) can pick and choose which one offer the affordable and better equipment.

FriedButter said:

Where would SMIC go and still be competitive? There isn’t any viable options that doesn’t span decades into the future.

Click to expand...

SMIC already has customers all over the world, including American companies.

Phead128 said:

Samsung and TSMC contributed $1.5B to ASML's R&D efforts for EUV in exchange for a 3% and 5% minority stake in ASML.

Intel contributed $4B to ASML's R&D effort for EUV in exchange for 15% stake in ASML. So basically you are left with SK Hynix and Micron as potential customers of Chinese EUV. Micron under US gov't pressure will unlikely adopt, and SK Hynix could conceivably adopt Chinese EUV for their China-only fabs.

In short, Chinese EUV will never "kill ASML EUV", not when Samsung+TSMC+Intel are the main EUV adopters and own up to 23% minority ownership stake in ASML and are ASML's largest funders of EUV R&D to begin with.

Click to expand...

Samsung and TSMC have 100% stake in themselves. If using ASML's product makes them uncompetitive then they'll be compelled to switch. It all depends on the quality of the product and the ease of switching.