Chinese semiconductor industry

hvpc said:

I took a quick crack at translating the video almost verbatim. I did rearrange the order of the info and highlighted key message in bold font. Buy it's basically a transcript of the video, not mine opinion.

I did not have the luxury of Google translator and solely relied on my rudimentary Chinese listening skill. If I made mistake translating, hope others would help point them out and make corrections.


The creator/author presented the video in 5 sections

=====================
1.

• UPrecision delivered 5 sets DWS dual stages to SMEE from April ’20 to Dec ’21.
  • Note: DWS is dual stage for dry litho tool & DWSi is for immersion tool.
• The host of video therefore assumed 3-5 assembled prototype (not production) litho systems at SMEE.
• News of: 1. UPrecision & SMEE collaboration on proto I, II, & III systems along & 2. Delivery of 5 DWS to SMEE provide evidence to support assumption of at least 3 systems undergoing verification.
• UPrecision revealed SMEE’s plan to ship litho system with DWS stages around end of 2022. Customers include Yangdong, YMTC, and CXMT. Uncertainties and risks with the plan remain a possibility.

2.

• UPrecision has 327M RMB back order of DWS stages.
• SMEE must have confidence in DWS stage otherwise would not issue such big order
• Each DWS is ~20M RMB (but this couldn’t be verified).
• So, SMEE may be planning 15-20 dry Lito tools in the near future

3. What can dry systems do?

• UPrecision indicate DWS can support needs of 65nm litho systems.
• Creator of video believes it may be possible to improve beyond 65nm capability, but these are not the rumored 28nm node litho system

4. Why isn't it for immersion?

• Answer: Very simple, it (DWSi?) has not been (finished) developed yet.
• The 02 project target is an immersion litho tool, not dry tool.
• So, in another word, UPrecision has not realized its 02 project commitment yet.

5 What is the status of immersion development?

• Two Immersion hood has already been developed by Cheer and delivered to SMEE.
• But the immersion dual stage that goes hand-in-hand with the immersion hood is still in development.
• Status of immersion dual stage development:
  • UPrecision had built 20 development prototype systems & 85 test units as of December 2021.
  • It is unclear if the 20 development prototype system refers to actual dual stages or modules. Combining with news of only 2 immersion hood, creator of video hypothesize it’s 20 module. Because having 20 immersion stage to go with 2 immersionhood just isn’t logical.
• UPrecision development cost is high in 2019 & 2020 then drastically reduced in 2021. Author thinks this is sign that major milestone had been over taken so the development cost decreased drastically.
• Filing for 02 project subsidy is based on projected expenses and raw material cost. Same trend of reduced subsidy filing observed in 2021. This is another clue suggesting development of immersion dual stages has entered mid-testing(?) phase.

6.Any concerns?

• Yes. The dual stages interferometer supply source is Zygo, an American company. There is risk of supply being cut off. As there's no alternate domestic option, this is currently the biggest risk.

So, the good news: The (litho system) has finally reached volume production. The bad news: it's not THE (28nm) tool that everyone was hoping for. But, the first step has been taken so the next step will not be far away.

--- end of "transcript" ---

Click to expand...

Thank you very much for your effort! Well this is very important information because clarifies how is the current situation with SMEE, and although maybe not what people expected, it returns a much more realistic scenario and this is a good thing IMO.

So, this is what I have made up my mind regarding SMEE

1. SMEE officially shows on his site 3 machines: SSA600/20 (ArF laser), SSC600/10 (KrF), SSB600/10 (i-line lamp) , but these machines have been there since years, already developed and sampled to some customer, but for some reason (not production ready?) have been retired and are now more or less a dead-end. No active development, not in use anywhere.

2. Real first production machine will come out hopfully at the end of the year and will be a dry dual stage machine for 65nm (ArF) node.

3. Dry dual stage will be used also for KrF and i-line lamp variants, maybe due to easier standardization, so to have a single setup for all the machines. For a not so big firm like SMEE it definitely makes sense not waste time/resources on cost optimization in this phase of furious catch-up with ASML. So once you have a solution that works and could scale to 28nm and beyond, simply use that and iterate on that.

4. Immersion machine will possibly follow next year in 2023 and this will be the famous 28nm machine.

european_guy said:

Thank you very much for your effort! Well this is very important information because clarifies how is the current situation with SMEE, and although maybe not what people expected, it returns a much more realistic scenario and this is a good thing IMO.

So, this is what I have made up my mind regarding SMEE

1. SMEE officially shows on his site 3 machines: SSA600/20 (ArF laser), SSC600/10 (KrF), SSB600/10 (i-line lamp) , but these machines have been there since years, already developed and sampled to some customer, but for some reason (not production ready?) have been retired and are now more or less a dead-end. No active development, not in use anywhere.

2. Real first production machine will come out hopfully at the end of the year and will be a dry dual stage machine for 65nm (ArF) node.

3. Dry stage will be used also for KrF and i-line lamp variants, maybe due to easier standardization, so to have a single setup for all the machines. For a not so big firm like SMEE it definitely makes sense not waste time/resources on cost optimization in this phase of furious catch-up with ASML. So once you have a solution that works and could scale to 28nm and beyond, simply use that and iterate on that.

4. Immersion machine will possibly follow next year in 2023 and this will be the famous 28nm machine.

Click to expand...

how do you reconcile your assumption that those machines "are retired" "more or less a dead end" "no active development" "not in use anywhere", with the market share analysis from EqualOcean and

Meanwhile, SMEE has made a great deal of progress in the back-end advanced packaging and LED/MEMS/power devices area. In the former segment, the company holds over 80% of the domestic and nearly 40% of the global market; its share in the global LED/MEMS/power devices lithography equipment industry is around 20%.

Click to expand...

FairAndUnbiased said:

how do you reconcile your assumption that those machines "are retired" "more or less a dead end" "no active development" "not in use anywhere", with the market share analysis from EqualOcean and

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

Click to expand...

I already knew about the big order from Foxconn, this refers explicitly to SMEE packaging lithography machine. Indeed I think currently SMEE revenue is mainly from advanced packaging and LED/MEMS/power market, also equal ocean article confirms that. The article BTW is one year old, and in the last year there have been no real news on revenues from SMEE machines for front-end lithography apart from some few samples sent for customer verification.

The fact that also all the recent tenders have been won by ASML at 100%, even for not so advanced nodes, is another hint that currently SMEE does not sell litho machines for front-end. Samples sent for verification don't go with tenders but by private agreements.

Key relevant points:

Jiangsu Nanda Optoelectronic Materials Co., Ltd. (Nata) has obtained funding to set up ArF resist facilities in China. They have no resist products as yet, though do have significant business is gases used for doping semiconductors.

Kempur (based in Beijing) is the biggest photoresist manufacturer in China and has a variety of products, which includes negative photoresists and ancillaries

Shanghai Sinyang Semiconductor Materials Co Ltd (based in Shanghai) is reported to be investing US$12M dollars to set up photoresist capability in China.

Click to expand...

Biggest resist market is 1. ArF immersion 2. KrF

i-line and ArF dry are stable, EUV is growing fast (unsurprisingly).

SMIC • Planned 2021 Capex is approximately $4.3 billion, the majority which is for non-FinFET capacity expansion, and the remaining is for FinFET, the infrastructure of the new Beijing joint-venture project, and other. US$4.3B 2021 Capex vs. US$5.7B in 2020

Click to expand...

Chinese companies entering the market as photoresist suppliers to the Chinese market and the world market which is expected to impact the current market dynamics.

Chinese subsidized will likely cause downward pricing pressure within the market as they have government support, at times financial, which can allow them to provide ‘competitive’ pricing.

Chinese semiconductor materials suppliers have expanded beyond China.

For example, Kemper photoresist selling in the US (outside it’s Chinese market) because they want to have large, international customer base.

Click to expand...

Jingrui, the parent company of ReiHong, purchased an used immersion scanner from Korea for USUS$11 million.

Shanghai Xinyang has announced they will raise over 1 Billion RMB from up to 35 investors to fund “high-end photoresist research and development and industrialization projects for integrated circuit manufacturing”. The money will be used to fund ArF dry resist and KrF thick film resist.

Ningbo NATA optoelectronics has announced the commissioning of a 193nm resist development facility.

Kemper Microelectrons, based in Shunyi District, Beijing China is already making I and G line products and thus have the skills to move into more advanced photoresist types (248nm and 193nm with exposure capability located in China.)

There are now four companies in China working on ArF resist. A dominate China market position could make one these companies a rival marketshare leader (displacing for example Fujifilm and/ or Shin Etsu.)

Click to expand...

Rettam Stacf said:

But what is the advantage of having all those if China is, at least perception wise, not going to leverage them to "return the favor" or those things don't really matter to the US ?

Someone argued earlier that it is a "nuclear option" that can only be used once. But that is not true. China an salami slice too by embargo just one item at a time and for a particular country or region etc. The fact that the US keeps heaping on sanctions after sanction of semiconductor manufacturing equipment means that they do not have any concern that embargoing such items by China has any effect on them.

Click to expand...

American sanctions are actually quite deadly to domestic firms. They have built competitors in China that will likely have most of the Chinese market (largest market in the world) to themselves.

It is actually unwise for China to exert any similar pressure on other countries unless it becomes sanction proof. Or more likely, it's facing economic warfare from Western countries for taking away their manufacturing jobs.

FairAndUnbiased said:

in fact, even if US mines the rare earths, the refining capability is in China, and the refining is the value added step, but the greatest cost is in the ore mining. so by trying to bypass China, they've actually made it even more profitable for China to refine the rare earths while they take on the burden of mining it. And recycling is actually even more valuable than original mining, since just like steel, it is much easier to melt down and purify already refined rare earths than to refine it from rock at the ppm level.

Click to expand...

Right, America imo is doing the right thing in trying to build its own refining capacity for its military end users. But anything beyond that would be too costly and wall street would not approve.

As such, other countries are likely to keep relying on Chinese supply chain, because China is not using it for economic warfare and keep the process easy and profitable. So despite CIA/MI6 complaining on a daily basis, businesses are simply not listening. As we've seen recently, SK has only become more dependent on Chinese supply chain this year.

With Russian war on Ukraine, they are now dominating noble gas, another important part of supply chain.

It would be best for Chinese companies to continue to make their product/services as easy for international partners as possible.

immersion system: complete
28nm photoresist: done
light source: done
at least later this year or Q1 next year
This 28nm lithography project has been postponed for 3 years since COVID...

olalavn said:

immersion system: complete
28nm photoresist: done
light source: done
at least later this year or Q1 next year
This 28nm lithography project has been postponed for 3 years since COVID...

Click to expand...

The immersion subsystem supplier stated that 2 unit of DUVi already in production line. What do you think?

european_guy said:

Thank you very much for your effort! Well this is very important information because clarifies how is the current situation with SMEE, and although maybe not what people expected, it returns a much more realistic scenario and this is a good thing IMO.

So, this is what I have made up my mind regarding SMEE

1. SMEE officially shows on his site 3 machines: SSA600/20 (ArF laser), SSC600/10 (KrF), SSB600/10 (i-line lamp) , but these machines have been there since years, already developed and sampled to some customer, but for some reason (not production ready?) have been retired and are now more or less a dead-end. No active development, not in use anywhere.

2. Real first production machine will come out hopfully at the end of the year and will be a dry dual stage machine for 65nm (ArF) node.

3. Dry dual stage will be used also for KrF and i-line lamp variants, maybe due to easier standardization, so to have a single setup for all the machines. For a not so big firm like SMEE it definitely makes sense not waste time/resources on cost optimization in this phase of furious catch-up with ASML. So once you have a solution that works and could scale to 28nm and beyond, simply use that and iterate on that.

4. Immersion machine will possibly follow next year in 2023 and this will be the famous 28nm machine.

Click to expand...

For SMEE to succeed in the frontend sadly ASML would need to experience a little bit of pain no much just a little. SMEE can't compete directly with ASML, even Nikon and Canon can't do it, I think that a least for government founded or subsidize projects the Chinese goverment would have to put a requirement of a least the KrF and I-line models as mandatory, the ArF dry model as preferable and Immersion at discretion. If ASML immersion tools are banned for sale in china the the decision for pushing forward SMEE dry tools at the expense of ASML will be easier to make.

tokenanalyst said:

For SMEE to succeed in the frontend sadly ASML would need to experience a little bit of pain no much just a little. SMEE can't compete directly with ASML, even Nikon and Canon can't do it, I think that a least for government founded or subsidize projects the Chinese goverment would have to put a requirement of a least the KrF and I-line models as mandatory, the ArF dry model as preferable and Immersion at discretion. If ASML immersion tools are banned for sale in china the the decision for pushing forward SMEE dry tools at the expense of ASML will be easier to make.

Click to expand...

Let's analyze this from technical and geopolitical point of view.

From technical point of view banning ASML would make sense only if SMEE can guarantee enough production capacity to cover the 100% of the Chinese market, and if current customers of ASML are already independent from a maintenance / spare parts point of view. Otherwise it simply means to shoot itself in the feet.

Far better for China to allow SMEE to develop along with ASML: there is well enough space in the market for both of them. China does not need to ban ASML to make some foundries to buy SMEE when ready....China can and did it in different, more subtle ways already many times in the past.

Form geopolitical point of view such a decision IMO would be even dumber:

1. It would be a huge propaganda gift for US that can scare away companies from investing in China

2. It would be a huge blow to China credibility on the international theater, a blow that would not be possible to recover.

3. It would condemn not only SMEE, but all Chinese equipment manufacturers to being lock in the Chinese market forever because US would very easily succeed in banning them everywhere else (this is a real danger anyhow though).

4. Would make SMEE to become lazy and slow due to missing technical competition with a stronger actor (that is always useful in the long term)

5. Would force decoupling of Chinese semiconductor businesses from the world (huge loss for China in the long term from many, really many points of view).



IMO China should play the strong and confident character here. A strong character does not revert to cheap retaliations or to using the same dirty tricks when faces a bully.

I tell you one thing, I can say this because I know western mentality. One of the things that really drives US mad against China...is that China does not react back with the same dirty tricks and arbitrary nonsense that instead faces from US. This is a strong hint of being strong and confident in itself, and it is what really drives US crazy.

european_guy said:

2. It would be a huge blow to China credibility on the international theater, a blow that would not be possible to recover.

Click to expand...

You ought to take phrases like this entirely out of your thought process. China could line up every single Western CEO in the country against a wall and blow their brains out, and all the next bunch of CEOs would do is increase their investments in China. There's only one direction investment in China goes - no matter what it does - and that's up.

european_guy said:

5. Would force decoupling of Chinese semiconductor businesses from the world (huge loss for China in the long term from many, really many points of view).

Click to expand...

The words "loss" and "China" don't belong in the same sentence.