TechanaLye assesses China as 3 years behind TSMC

[KevinK]

SemiWiki could consider paying for more professional replies/analyses instead of people just leaving broken, biased comments.

You’re clueless on the real numbers. Pretty clear you never saw all the ecosystem development that needed to happen to make EUV work. First off, Intel put in about $4B and Samsung ante’d up almost the same as TSMC in that 2012 round - were you just naive or purposefully deceptive in leaving those out of your equation ? What about the cost of developing fracturing, ILT, new pellicles, and new resists for EUV ? And all the money spent by fabs experimenting ? And all the money spent to buy dozens of first generation tools. They don’t pay for themselves at the fabs ahead of time - that is also part of the investment China would have to make, because even China’s most profitable fab could only fund a few EUV system out of their operating income.

Sorry to be combative, but you don’t seem to be paying attention to all the costs.

 

[jim_22222]

You’re clueless on the real numbers. Pretty clear you never saw all the ecosystem development that needed to happen to make EUV work. First off, Intel put in about $4B and Samsung ante’d up almost the same as TSMC in that 2012 round - were you just naive or purposefully deceptive in leaving those out of your equation ? What about the cost of developing fracturing, ILT, new pellicles, and new resists for EUV ? And all the money spent by fabs experimenting ? And all the money spent to buy dozens of first generation tools. They don’t pay for themselves at the fabs ahead of time - that is also part of the investment China would have to make, because even China’s most profitable fab could only fund a few EUV system out of their operating income.

Sorry to be combative, but you don’t seem to be paying attention to all the costs.

Can you summarize and provide a number? If ASML presents this information to their board without a number, do you think they will approve it? Why don't they just go directly for accelerators and X-rays?
for example you can say 1 trillion, 100 trillion etc.
not something like big, big money.
that's kindergarden way of telling things.

btw, the tsmc 's 1.4 b is 5%, who care's how much intel put there. just a bigger percentage. the total remains same.

 

[KevinK]

Can you summarize and provide a number? If ASML presents this information to their board without a number, do you think they will approve it? Why don't they just go directly for accelerators and X-rays?
for example you can say 1 trillion, 100 trillion etc.

I'm going to rely on someone almost as smart as me, but who is excellent at the details, complete with attribution, Perplexity. But first I'm going to reiterate that the move to EUV required investment across the entire industry, and not just the tiny upfront "good money" Intel, TSMC and Samsung fronted ASML. When I say $100B, I'm not fishing it out of the air. I'm basing that on the total cost of the 200mm wafer to 300mm wafer transition. But you're a bit ridiculous when you suggest that we only use ASML's direct R&D costs as the measure.

In summary, while a precise total is not available, the semiconductor industry's transition from 200mm to 300mm wafers likely involved cumulative investments well over $100 billion across fabs, equipment, and R&D. The transition spanned many years and continues today as 300mm capacity expands.

While a precise total cost for the industry's EUV transition is not available, it's clear that the cumulative investments across equipment makers, chip manufacturers, and research and development have likely amounted to tens of billions of dollars over the past decade. The transition is ongoing, with continued investments expected as the technology evolves and becomes more widely adopted.

https://www.perplexity.ai/search/what-was-the-total-semiconduct-KAPgwSYrRdaRQPkrxU6mHA

 

[jim_22222]

I'm going to rely on someone almost as smart as me, but who is excellent at the details, complete with attribution, Perplexity. But first I'm going to reiterate that the move to EUV required investment across the entire industry, and not just the tiny upfront "good money" Intel, TSMC and Samsung fronted ASML. When I say $100B, I'm not fishing it out of the air. I'm basing that on the total cost of the 200mm wafer to 300mm wafer transition. But you're a bit ridiculous when you suggest that we only use ASML's direct R&D costs as the measure.

In summary, while a precise total is not available, the semiconductor industry's transition from 200mm to 300mm wafers likely involved cumulative investments well over $100 billion across fabs, equipment, and R&D. The transition spanned many years and continues today as 300mm capacity expands.

While a precise total cost for the industry's EUV transition is not available, it's clear that the cumulative investments across equipment makers, chip manufacturers, and research and development have likely amounted to tens of billions of dollars over the past decade. The transition is ongoing, with continued investments expected as the technology evolves and becomes more widely adopted.

https://www.perplexity.ai/search/what-was-the-total-semiconduct-KAPgwSYrRdaRQPkrxU6mHA

Someone can argue that China's R&D activities have overtaken those of the US since 2022 based on the publication index. Its enormous size and investment cannot be overlooked. If they can improve the efficiencies of research, many things are possible.

China overtakes United States on contribution to research in Nature Index

Data on affiliations suggest that authors from China made the largest contribution to high-quality natural-science research in 2022.

www.nature.com

China is mobilizing science to spur development — and self-reliance

The National People’s Congress and the Chinese People’s Political Consultative Conference have boosted science and technology in the national agenda.

www.nature.com

 

[tooLongInEDA]

World's largest contract chipmaker Taiwan Semiconductor Manufacturing Co. has agreed to invest 1.11 billion euros (US$1.4 billion) in chip equipment maker ASML in return for a 5 percent stake, and to bankroll research and development of future technology.

TSMC to invest $1.4B in ASML chip kit, stake

Taiwan Semiconductor Manufacturing Company agrees to buy 5 percent stake for US$1.03 billion, and invest US$342 million for R&D in ASML's next-generation lithography technologies.

www.zdnet.com

You cannot exaggerate the situation and promote your view. A more accurate assessment is TSMC's investment. $14 billion would be a close estimate for ASML's investment. Other parts of the ecosystem are cheaper. Let's say $20 billion in total, roughly the amount for two fabs.

SemiWiki could consider paying for more professional replies/analyses instead of people just leaving broken, biased comments.

The whole point is that we can and do leave slightly broken and biased comments ! That way we get to share our expertise and increase our insights and understanding of complex and evolving situations where none of us have the complete picture.

And that - just my view - is the key strength of the US/Western model. The fact that we can do so politely without letting our egos and politics get in the way. It works pretty well, provided you have free speech, mutual respect and an open mind and desire to learn.

None of us are immune from broken and biased comments. Not even you.

 

[tooLongInEDA]

if they pay you, let s say 100$,
you might sit down for two hours and run some serious study of the subject.
what is happening now is no one want to spend more than three minutes and want to prove he is correct

I'm not sure you really understand the purpose and goals of this discussion forum and how it operates.

Or appreciate the fact that there are many people here who charge way north of $100 for 2 hours who are freely contributing expert insights and expertise. And don't do so for ego-driven or ideological reasons.

You might also want to ask yourself if you don't share some of the faults you're attributing to others here.

 

[count]

If it's $10b or $100b, either way, if China fails to develop EUV it won't be because of lack of money. The Chinese economy is the largest in the world measured by purchasing power parity. In manufacturing they are easily the best in the world.

 

[Daniel Nenni]

I'm not sure you really understand the purpose and goals of this discussion forum and how it operates.

Or appreciate the fact that there are many people here who charge way north of $100 for 2 hours who are freely contributing expert insights and expertise. And don't do so for ego-driven or ideological reasons.

You might also want to ask yourself if you don't share some of the faults you're attributing to others here.

I bill out at $1,000 per hour but I am free on SemiWiki, as one example.

 

[Barnsley]

If it's $10b or $100b, either way, if China fails to develop EUV it won't be because of lack of money. The Chinese economy is the largest in the world measured by purchasing power parity. In manufacturing they are easily the best in the world.

No doubt the Chinese folks are pumping Billions into Semicon , and I believe most are very grateful for this.
In reality I am more than happy for.Govts of any ilk to be pumping money into Semicon.
As long as everyone in in the industry gets a little bit and its not just focused on the top end!

 

[KevinK]

@jim_22222, this article is behind a paywall, but the South China Morning Post gives out 5 free articles. This one is a great, expert explainer of the tensions between China’s previous spectacular growth and more rocky future direction set by Xi. The short version:

* China is not a command economy. Beijing sets top level directions and dumps some money out per a set of rule, and millions of bureaucrats and hundreds of local governments go to work executing on that vision. That’s not always an efficient approach.

“Although China is effective at quickly generating massive outputs, this approach results in significant waste. The EV industry is a prime example: China has more than 450 car factories, yet one-third of them operate at less than 20 per cent capacity. Ultimately, most of these producers are likely to go bust, leading the industry to consolidate around a few giants like BYD.”

* That same mechanism applies to scientific research.

Someone can argue that China's R&D activities have overtaken those of the US since 2022 based on the publication index. Its enormous size and investment cannot be overlooked.

“In a recent article, my co-authors and I show that after the central government set ambitious targets for new patents – a standard indicator of innovation – local officials inflated the numbers by encouraging junk patents. Consequently, the share of genuinely novel innovations has declined. We refer to this phenomenon as a “low-productivity innovation drive”.

* The housing bust is reducing the ability of the ”old model” to work, due to reduced domestic demand.

“This is already evident in the real-estate meltdown, which has wiped out jobs and household wealth, leading consumers to cut back on spending. As a result, producers have been forced to export unsold goods like EVs, exacerbating trade tensions with the US and other countries accusing China of dumping its overcapacity into their markets.”

* Local government debt, is also hamstringing investment in new areas.

“Moreover, shifting to a hi-tech economy typically requires robust gross domestic product (GDP) growth and healthy public finances to enable the government to invest in industrial policies, retrain workers and establish social safety nets for those left behind. Without such support, the transition risks deepening social and economic divides. China, however, is accelerating its shift to cutting-edge technologies in the midst of an economic slump and a local-government debt crisis.“

The general picture is that China is a huge and vibrant economy that can do any one thing the government sets out to do. But there’s not just one thing anymore. China now has to navigate a set of complex objectives, without some of the benefits it had in the past (lowest cost advantage, unlimited capital, burgeoning domestic market, wide-open export market) plus a new set of liabilities (huge local government debt, higher unemployment, and embargoes on some key technologies). They could choose to build a duplicate EUV and semiconductor tool ecosystem, but it’s going to cost them much more than $10B and arrive far later than they need it for today’s products. I’m going to be watching the launch of the Huawei Mate 70 on Sept 10, to read the tea leaves.

Here‘s the article:


Yuen Yuen Ang

Opinion | Is China rising or slowing down? The answer is it’s complicated​

China is far from collapse but its gamble to anchor growth in hi-tech innovation won’t pay off unless it shores up its traditional economy​

Opinion | Why ‘peak China’ talk obscures dire reality about its economy

China is far from collapse but its gamble to anchor growth in hi-tech innovation won’t pay off unless it shores up its traditional economy.

www.scmp.com

 

[benb]

I’m going to be watching the launch of the Huawei Mate 70 on Sept 10, to read the tea leaves.

Yes, wait and see the evidence, ignore the FUD or chest thumping.

 

[jim_22222]

@jim_22222, this article is behind a paywall, but the South China Morning Post gives out 5 free articles. This one is a great, expert explainer of the tensions between China’s previous spectacular growth and more rocky future direction set by Xi. The short version:

* China is not a command economy. Beijing sets top level directions and dumps some money out per a set of rule, and millions of bureaucrats and hundreds of local governments go to work executing on that vision. That’s not always an efficient approach.

“Although China is effective at quickly generating massive outputs, this approach results in significant waste. The EV industry is a prime example: China has more than 450 car factories, yet one-third of them operate at less than 20 per cent capacity. Ultimately, most of these producers are likely to go bust, leading the industry to consolidate around a few giants like BYD.”

* That same mechanism applies to scientific research.



“In a recent article, my co-authors and I show that after the central government set ambitious targets for new patents – a standard indicator of innovation – local officials inflated the numbers by encouraging junk patents. Consequently, the share of genuinely novel innovations has declined. We refer to this phenomenon as a “low-productivity innovation drive”.

* The housing bust is reducing the ability of the ”old model” to work, due to reduced domestic demand.

“This is already evident in the real-estate meltdown, which has wiped out jobs and household wealth, leading consumers to cut back on spending. As a result, producers have been forced to export unsold goods like EVs, exacerbating trade tensions with the US and other countries accusing China of dumping its overcapacity into their markets.”

* Local government debt, is also hamstringing investment in new areas.

“Moreover, shifting to a hi-tech economy typically requires robust gross domestic product (GDP) growth and healthy public finances to enable the government to invest in industrial policies, retrain workers and establish social safety nets for those left behind. Without such support, the transition risks deepening social and economic divides. China, however, is accelerating its shift to cutting-edge technologies in the midst of an economic slump and a local-government debt crisis.“

The general picture is that China is a huge and vibrant economy that can do any one thing the government sets out to do. But there’s not just one thing anymore. China now has to navigate a set of complex objectives, without some of the benefits it had in the past (lowest cost advantage, unlimited capital, burgeoning domestic market, wide-open export market) plus a new set of liabilities (huge local government debt, higher unemployment, and embargoes on some key technologies). They could choose to build a duplicate EUV and semiconductor tool ecosystem, but it’s going to cost them much more than $10B and arrive far later than they need it for today’s products. I’m going to be watching the launch of the Huawei Mate 70 on Sept 10, to read the tea leaves.

Here‘s the article:


Yuen Yuen Ang

Opinion | Is China rising or slowing down? The answer is it’s complicated​

China is far from collapse but its gamble to anchor growth in hi-tech innovation won’t pay off unless it shores up its traditional economy​

Opinion | Why ‘peak China’ talk obscures dire reality about its economy

China is far from collapse but its gamble to anchor growth in hi-tech innovation won’t pay off unless it shores up its traditional economy.

www.scmp.com

I think west can understand it like a tech company running by a CFO. in this case it is xi and his PM. there are a lot of inefficient practices other wise they're already leading the litho technology with the size of the investment and manpower.
they don't welcome religion in their state run institutions which creates a gap to western institution.
but they are also learning and improving.
you might keep on seeing this kind of dynamic scenes, some company could fail. but it's not hw but Intel. etc

 

[Barnsley]

https://www.huaweicentral.com/huawei-kirin-7nm-chip-performance-draws-level-with-tsmc-5nm-report/

 

[KevinK]

https://www.huaweicentral.com/huawei-kirin-7nm-chip-performance-draws-level-with-tsmc-5nm-report/

Bigger news yesterday is that Huawei did NOT launch the Mate 70, reportedly due to Kirin chip production shortages and yield issues. Instead, they pivoted to the tri-fold Mate XT, that costs almost 3000$. That confirms the SMIC problems.

 

[Fred Chen]

Bigger news yesterday is that Huawei did NOT launch the Mate 70, reportedly due to Kirin chip production shortages and yield issues. Instead, they pivoted to the tri-fold Mate XT, that costs almost 3000$. That confirms the SMIC problems.

It is said to be the Kirin 9010 (same as in Pura 70) https://www.huaweicentral.com/huawei-mate-xt-tri-fold-reportedly-runs-on-kirin-9010-5g-chip/ so no new chip.

 

[Barnsley]

Bigger news yesterday is that Huawei did NOT launch the Mate 70, reportedly due to Kirin chip production shortages and yield issues. Instead, they pivoted to the tri-fold Mate XT, that costs almost 3000$. That confirms the SMIC problems.

The price is quite the news here in Singapore.

Will they even make money on that price?

 

[KevinK]

It is said to be the Kirin 9010 (same as in Pura 70) https://www.huaweicentral.com/huawei-mate-xt-tri-fold-reportedly-runs-on-kirin-9010-5g-chip/ so no new chip.

Makes sense, though I'm hearing the supply of the Kirin 9010 is also impaired. But when you have limited chip supply, it's optimal to put the chips into the product with the highest prices / margins. I would imagine that at nearly $3K, the XT is the best way to utilize limited supply.

As for the Mate 70, I think its processor is a new chip on the n+3 process (denser 7nm ?) That one is definitely having yield and production issues though I haven't heard any more since the Information article came out.

https://www.theinformation.com/articles/apples-china-rival-huawei-faces-production-snag-with-new-phone

 

[KevinK]

I think west can understand it like a tech company running by a CFO. in this case it is xi and his PM. there are a lot of inefficient practices other wise they're already leading the litho technology with the size of the investment and manpower.

You’re not paying attention to the things a CFO would.

China playing it safe with capital, researchers say, as SOEs get hefty payout
Researchers with Rhodium Group say capital in China is being diverted from more dynamic firms into legacy industries and state sectors.

“Beijing needs banks to lend to sectors and firms that can more efficiently fuel innovation and economic growth than the industries that were prioritised in the past,” read a report by the Washington-based firm.

“But decades of misallocation – with credit flowing mostly to legacy industries and state-owned enterprises (SOEs) – make changing patterns of lending difficult, particularly in a context of slower economic growth.”

Wednesday, 11 September 2024, 9:00M

China playing it safe with capital, researchers say, as SOEs get hefty payout — South China Morning Post

Too much new capital in China is being allocated to companies regarded as “safe” rather than those which could drive growth through innovation – a stated goal of Beijing’s industrial policy – research outfit Rhodium Group said on Wednesday. “Beijing needs…

apple.news

 

[jim_22222]

You’re not paying attention to the things a CFO would.

China playing it safe with capital, researchers say, as SOEs get hefty payout
Researchers with Rhodium Group say capital in China is being diverted from more dynamic firms into legacy industries and state sectors.

“Beijing needs banks to lend to sectors and firms that can more efficiently fuel innovation and economic growth than the industries that were prioritised in the past,” read a report by the Washington-based firm.

“But decades of misallocation – with credit flowing mostly to legacy industries and state-owned enterprises (SOEs) – make changing patterns of lending difficult, particularly in a context of slower economic growth.”

Wednesday, 11 September 2024, 9:00M

China playing it safe with capital, researchers say, as SOEs get hefty payout — South China Morning Post

Too much new capital in China is being allocated to companies regarded as “safe” rather than those which could drive growth through innovation – a stated goal of Beijing’s industrial policy – research outfit Rhodium Group said on Wednesday. “Beijing needs…

apple.news

call it a bad CFO if you want to
investment in dynamic innovation could end up like Chinese soccer team. if you know what does that mean

 

[samwilde]

SMIC doubled the size of its FinFET production facilities. We are supposed to believe there are chip supply issues. I doubt it.

There are claims of delays with Huawei's new operating system (HarmonyOS NEXT) and application support. Their new OS is going to remove Android backwards compatibility. Which means a lot of applications that people would expect to run on their phone will need to be ported over. Making an early release with lackluster app support could lead to market failure so the release has been delayed. The Mate series is their technology leader. It is expected to be the launch platform for HarmonyOS NEXT. I find this much more plausible than any chip shortages.