It could be that they are sub-system manufacturer and SMEE the integrator. But whatever sub-system they are building combines optics and motion sub-systems.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Chinese semiconductor thread II
- Thread starter vincent
- Start date
The pellicles is a huge deal like you mentioned, most people are completely oblivious to it.There is a lot of information that you never see in the news or in think tank websites. Let suppose you found a government website for the finalization of a project for ultrapure fluid particle control sub-system that says it goes in a immersion lithography machine used in a 14nm process node. That means IMO that they have systems in production for a 28nm process node and they want to move into a more advanced process node and pretty much like ASML that means upgrading the machines. And that also tells that China domestic lithography machines are probably modular so they only have to upgrade the sub-system instead of the entire machine.
Is the same with EUV. The fact that they already moving in pellicles tells me that they have EUV masks that they want to protect from contaminants in some sort of production system that goes beyond the simple pattern lab testing.
Last time I checked it was just patents, has there been anything else happening in that regard?
In 2020 and 2021 the consensus by stooges, including some guys here was that Huawei was finished, 28nm was a hard cap, as more ASML scanners got banned. YMTC was condemned to an 128L hard cap in NAND, all SMIC and YMTC tools should had been ossified by now and so on. Is safe to said that what the news says is not even the tip of the iceberg and what we post here is just the tip iceberg. The level secrecy in China semiconductor industry is almost reaching military levels and is blindsiding a lot of stooges in the US.
How Huawei keeps advancing in chip production? How their patterned M1 metal pitch at the same level of Intel? How they are maintaining their banned "imported" tools? How they are able to advance in 3DNAND? 3DRAM? Or as a guy put it when Huawei announced its Tau´s law. How the hell they were able to make high precision bonding tools? How they advanced so quickly <his words>?
How Huawei keeps advancing in chip production? How their patterned M1 metal pitch at the same level of Intel? How they are maintaining their banned "imported" tools? How they are able to advance in 3DNAND? 3DRAM? Or as a guy put it when Huawei announced its Tau´s law. How the hell they were able to make high precision bonding tools? How they advanced so quickly <his words>?
its seems to be a collaboration between China research institution and China domestic mask fabrication companies to get these delicate films as soon as possible.The pellicles is a huge deal like you mentioned, most people are completely oblivious to it.
Last time I checked it was just patents, has there been anything else happening in that regard?
take a look at just the AI demand on server side and consumer side. the models are doubling in size every year for the next few years. So the most recent DDR5 pricing was 19k RMB for 64 GB -> that's $85 per 2 GB. Let's say we get to a more reasonable level of $15 per 2 GB. In a few years, CXMT will be up to 7m wafers per year. That is 15834m GB assuming 80% yield & 48mm2 die size for 2GB. If we go by $15 per 2 GB -> $120B in revenue per year.
What is your bearish case for server DRAM? That it will get to $8 per 2GB long term?
CXMT is going to have so much cash to expand its DRAM capacity and the entire Chinese AI supply chain is dependent on CXMT, not Huawei. Huawei uses the same AI supply chain that everyone else uses. Piotech hybrid bonding, SMIC logic die, the huge PCB supply chain, shengyi CCL, JCET for high bandwidth memory.
The most innovative AI stack design right now all require customized DRAM from CXMT.
I actually posted three times saying that Korea cut production!!
if i remember correctly, i think some people sued those chipmakers in the US court regarding supply?I actually posted three times saying that Korea cut production!!
In 2020 and 2021 the consensus by stooges, including some guys here was that Huawei was finished, 28nm was a hard cap, as more ASML scanners got banned. YMTC was condemned to an 128L hard cap in NAND, all SMIC and YMTC tools should had been ossified by now and so on. Is safe to said that what the news says is not even the tip of the iceberg and what we post here is just the tip iceberg. The level secrecy in China semiconductor industry is almost reaching military levels and is blindsiding a lot of stooges in the US.
How Huawei keeps advancing in chip production? How their patterned M1 metal pitch at the same level of Intel? How they are maintaining their banned "imported" tools? How they are able to advance in 3DNAND? 3DRAM? Or as a guy put it when Huawei announced its Tau´s law. How the hell they were able to make high precision bonding tools? How they advanced so quickly <his words>?
I know you follow at deep level of Lithography (Patents and stuff) and my knowledge is nowhere close to yours as far as the Lithography Supply Chain goes.SiCarrier and that Yuliangsheng are different companies. Yuliangsheng is in hiring spree so it seems to be legit lithography company.
I think is the contrary, the news are vastly underestimating the lithography situation in China. The production of immersion lithography has been going since 2021 in a 28nm process node, accelerated in 2024 with in a 14nm and think they are closer to 7nm node with domestic immersion lithography. The "mass" production of immersion systems has been happening since a few years ago. They had been iterating the system to get it to work in more advanced nodes.
EUV is bigger deal because I can put my hand on the fire that they already have pilot production lines with domestic EUV systems. There is not other way to explained.
One common misconception here is that any Tools good enough for 28nm level can be scaled down to sub 7 nm with small tweaking which is not correct. Etch, Deposition and other crucial steps going from 28nm to sub 7nm requires much more precision, accuracy (and the iteration does not happen overnight). This is why you'll notice even though Chinese companies tools had domesticated 28nm supply chain, sales of AMAT, LAM, TEL were exploding to China in 2024-2025 (for sub 7nm chips). By 2026 these Chinese companies tools became precise enough for sub 7nm Chips (albeit its still at 65-75% Sufficiency)
Now Lets talk about Lithography, from what I was reading in last 5-6 years, SMEE 28nm got delayed repeatedly which was supposed to debut in 2021-2022. But now in 2026 they are good enough for 7nm Chip. Now I don't kow what was happening during last 5 years. Did they have a Prototype DUVi which they were trying to get to work for 28nm or were they trying to get it to work all the way to 7nm.
Late 2023 -2024, Huawei tried to Poach SMEE employees because SMEE DUVi kept getting delayed repeatedly. So In last 2 -3 years the new UEAscend (YuliangSheng) was formed which took over SMEE DUVi to get to work in 7nm chip. So Last year I think it was getting verified in SMIC Fab through late 2025. Now in 2026 this tool is good enough for 7nm Chip (I am assuming). The Information reported 20 shipment in 2027 and ASML Ships approximately 90 DUVi each year to China.
Imagine by 2028 (with more feedback loop and iteration) UEAscend will be able to ship 50 to 60+ DUVi in China and by 2030, China will be 80-90% Self Sufficient in DUVi which is extraordinary to say the least.
Also by 2030 we will get our second or third generation of Huawei phone having chips patterened with a Chinese EUVL. The future is golden
Last edited:
SMEE appears to have operated as a relatively small company with limited access to technological expertise, capital, and strategic policy support until the sanctions imposed from 2018 onward created a stronger incentive to accelerate its development. Prior to that, the approach seems to have been more along the lines of establishing a company and allowing it to manage its own progress with limited external involvement. A plaything for a CEO with zero discipline. His sandbox where he could mess around because nobody was watching before 2018.I don't think "SMEE" exist as was before. The company was divested in two companies. One is AMIES <low end lithography machines> and the other company will only do High End FrontEnd Lithography machines.
Last edited:
To be fair SMEE was constrained by the supply Chain back then.I know you follow at deep level of Lithography (Patents and stuff) and my knowledge is nowhere close to yours as far as the Lithography Supply Chain goes.
One common misconception here is that any Tools good enough for 28nm level can be scaled down to sub 7 nm with small tweaking which is not correct. Etch, Deposition and other crucial steps going from 28nm to sub 7nm requires much more precision, accuracy (and the iteration does not happen overnight). This is why you'll notice even though Chinese companies tools had domesticated 28nm supply chain, sales of AMAT, LAM, TEL were exploding to China in 2024-2025 (for sub 7nm chips). By 2026 these Chinese companies tools became precise enough for sub 7nm Chips (albeit its still at 65-75% Sufficiency)
Now Lets talk about Lithography, from what I was reading in last 5-6 years, SMEE 28nm got delayed repeatedly which was supposed to debut in 2021-2022. But now in 2026 they are good enough for 7nm Chip. Now I don't kow what was happening during last 5 years. Did they have a Prototype DUVi which they were trying to get to work for 28nm or were they trying to get it to work all the way to 7nm.
Late 2023 -2024, Huawei tried to Poach SMEE employees because SMEE DUVi kept getting delayed repeatedly. So In last 2 -3 years the new UEAscend (YuliangSheng) was formed which took over SMEE DUVi to get to work in 7nm chip. So Last year I think it was getting verified in SMIC Fab through late 2025. Now in 2026 this tool is good enough for 7nm Chip (I am assuming). The Information reported 20 shipment in 2027 and ASML Ships approximately 90 DUVi each year to China.
Imagine by 2028 (with more feedback loop and iteration) UEAscend will be able to ship 50 to 60+ DUVi in China and by 2030, China will be 80-90% Self Sufficient in DUVi which is extraordinary to say the least.
Also by 2030 we will get our second or third generation of Huawei phone having chips patterened with a Chinese EUVL. The future is golden
-There was only one lithography optics company back then, CNEPO, now there are multiple. Gouwang optical, Maolai Optics and others
-Optic manufacturing was constrained by imports, looks like that is not the case anymore, domestic supply chain of optics manufacturing machines has been filled with domestic companies thanks to the result of China R&D efforts.
-U-Precision was the only company making semiconductor level interferometers and lithography wafer stages, now it seems there are multiple companies at least in the interferometer part.
-There were was no overlay metrology company back then, now it seems that there are multiple companies in that area.
-Domestic computational lithography capabilities and DTCO was more primitive back then, that is not the case anymore.
-Mask development, coating/developing and domestic photoresist has advanced pretty far.
-There is more competition and you don't have to stick to a single supplier for everything.
So the situation is pretty different now to what was back then.
Shengjisheng Semiconductor, specializing in thin film deposition equipment, has completed its IPO preparatory filing.
According to information released by the Ningbo Securities Regulatory Bureau on July 28, Shengjisheng Semiconductor Technology Co., Ltd. completed its initial public offering (IPO) filing on July 27, 2026, officially launching its A-share listing process. The sponsor for this IPO is Guotai Haitong Securities Co., Ltd., and the supporting intermediaries are Shanghai Jincheng Law Firm and Rongcheng Certified Public Accountants (Special General Partnership). The IPO filing has been formally accepted by the regulatory authorities.
Publicly available business and industry information shows that Shengjisheng Semiconductor was registered and established in Ningbo in March 2018. It was initiated by SMIC in conjunction with industrial capital and is a national-level specialized and innovative small giant enterprise and high-tech enterprise. With a registered capital of 1.07 billion yuan, its equity structure is relatively dispersed and there is no actual controller. The company's core business focuses on the self-developed and mass-produced specialized equipment for integrated circuit front-end manufacturing, primarily 12-inch PECVD and CVD thin film deposition equipment, as well as RTP rapid thermal processing equipment. It also provides supporting semiconductor core component manufacturing and equipment maintenance services. Many of its complete machine products have passed the process verification of leading wafer foundries and achieved batch repeat order delivery. It has entered the supply chain of advanced process production lines such as logic chips, memory chips, and silicon photonics chips, and is one of the core targets for domestic substitution of thin film deposition equipment in China.