Chinese semiconductor industry

Status
Not open for further replies.

tokenanalyst

Brigadier
Registered Member

Straight to the Shareholders' Meeting | Huahai Qingke: Continue to increase the R&D of high-end equipment and promote the layout of diversified product​

Jiwei.com reported that on January 31, Huahai Qingke Co., Ltd. (stock abbreviation: Huahai Qingke, stock code: 688120) held the first extraordinary general meeting of shareholders in 2023. The Proposal on Implementing New Projects with Part of the Over-raised Funds and Self-owned Funds" and the "Proposal on Using Part of the Over-raised Funds to Permanently Supplement Working Capital" were reviewed and voted on. Aijiwei attended the meeting as an institutional shareholder and voted in favor of all the proposals.

Overweight high-end equipment research and development

On January 5, Huahai Qingke held the 33rd meeting of the first board of directors and the 26th meeting of the first board of supervisors, deliberated and approved the "About using part of the raised funds to invest in project surplus funds, some over-raised funds and Proposal on Implementing New Projects with Self-Owned Funds", agreeing that the company will use the remaining raised funds of 213.0044 million yuan, over-raised funds of 286.9956 million yuan and self-owned funds of 317.543 million yuan to invest in wholly-owned subsidiary Huahai Qingke (Beijing) Technology Co., Ltd. (hereinafter referred to as "Huahai Qingke Beijing") increased capital and provided it with loans for the implementation of the "Huahai Qingke integrated circuit high-end equipment research and development and industrialization project". The project construction period is expected to be 26 months.

Huahai Qingke said that the implementation of the project is a measure for the company to fully grasp the development opportunities of the semiconductor equipment market and to expand production capacity forward-looking to meet the rapidly growing market demand. And high-end semiconductor equipment such as wet process equipment, fill the gaps in domestic related subdivision fields, realize the company's platform strategic layout, and create more space and new profit growth points for the company's future development.

Sustained rapid growth in performance

On January 20, Huahai Zero2IPO released an announcement on the expected increase in annual performance in 2022. It is estimated that the annual operating income in 2022 will be 1.53 billion to 1.695 billion yuan, an increase of 725 million to 890 million yuan compared with the same period last year. An increase of 90.09% to 110.59%; it is estimated that the annual net profit attributable to the owners of the parent company in 2022 will be 437 million to 517 million yuan, an increase of 239 million to 319 million yuan compared with the same period last year, and a year-on-year increase of 120.40% to 160.75%.

Since its establishment, Huahai Qingke has always adhered to the development route of independent innovation, and has achieved effective breakthroughs and systematic breakthroughs in key technologies such as nano-scale polishing, online detection of nano-precision film thickness, nano-particle ultra-clean cleaning, big data analysis and intelligent control. Layout, developed Universal series CMP equipment, Versatile series thinning equipment, HSDS/HCDS series liquid supply system, accumulated rich technical reserves in the professional field, product technology and performance have reached the domestic leading level, and have entered the domestic major integration The large circuit production line has achieved a good market reputation with excellent product performance, high cost performance and timely and fast after-sales service, which has improved the company's market development capabilities for new customers and new products.

Please, Log in or Register to view URLs content!
Please, Log in or Register to view URLs content!
 

latenlazy

Brigadier
In the havoc thread there is this guy 郑城城 that seems an insider.

He even tells to havoc. But is difficult to me to understand. What do you think @tonyget ?

郑城城
验证后才能上量,现在验收标准是能够稳定商业生产,要求高的多,一些子设备在准备量产了

havok -> 郑城城
我会注意发的内容
验证后才能上量,现在验收标准是能够稳定商业生产,要求高的多,一些子设备在准备量产了

Can only move to mass production after verification, right now the verification requirement is stable/reliable commercial production, a very high bar, a bunch of equipment are being readied for mass production.
 

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
永不言败:采用平面光栅的双工台能否支持10至7nm工艺节点?
havok:21年6月是初样,这是一个不停改进的过程

永不言败:Can dual-station with planar gratings support 10 to 7nm process nodes?
havok:June 2021 is the first sample, this is a process of continuous improvement
This makes sense what he is saying.

It confirms what @tokenanalyst said earlier that planar grating inferometer was completed by 2021. havok seems to know what he is talking about, but he is also quite inconsistent with things he say. I assume this means the 14nm scanner is on track, but it will probably be like 1980i. Which means they can get to 7 to 10nm with further refinement.

by the way, ASML continues to improve 1980i to have higher wpm, which may or may not be available to China after the latest sanctions.
In the havoc thread there is this guy 郑城城 that seems an insider.

He even tells to havoc. But is difficult to me to understand. What do you think @tonyget ?

郑城城
验证后才能上量,现在验收标准是能够稳定商业生产,要求高的多,一些子设备在准备量产了

havok -> 郑城城
我会注意发的内容
Right, I think he is pointing out the inconsistent narrative in some of the things havok is saying. Let's face it, he contradicts himself sometimes.

Loongson 3A5000 plays Windows online games very smoothly.... in the future this chip will grow a lot

Please, Log in or Register to view URLs content!

Phytium is ARM , Loongson is LoongArch.

LoongArch is niche so the designers supposedly added instruction sets to aid emulation to other architectures. The app store for Loongson UOS even had options to install x86 wine apps and arm android apps.
Yes, Loongson is the future as I've said recently.

Huawei would be well served to work with the only 100% domestic CPU maker. Anyone else would be concerned about getting put on entity list. They've made LA compatible with Euler and Harmony. They've also integrated LA SoC with harmony, but haven't done anything like that with also major Chinese chipmakers.

Even more than that, LA is the only CPU offering competitive performance with 12 to 28 nm process in China. Kunpeng-920 is a great chip but it's 3 years old and uses a process that SMIC probably can't mass produce right now. So, what good is that?

Huawei should work with Loongson to get it even more compatible with ARM and x86.

验证后才能上量,现在验收标准是能够稳定商业生产,要求高的多,一些子设备在准备量产了

Can only move to mass production after verification, right now the verification requirement is stable/reliable commercial production, a very high bar, a bunch of equipment are being readied for mass production.
That's kind of comforting. They are not so desperate that they are willing to let go of qualification standard. Things are setup for success imo, but they need some time.

So, they've gone 100% with 55/65 nm process by now.
2024 will be 28nm
2025 will be 14 and then 7nm. Seems like if they can achieve this, they did well
 

latenlazy

Brigadier
This makes sense what he is saying.

It confirms what @tokenanalyst said earlier that planar grating inferometer was completed by 2021. havok seems to know what he is talking about, but he is also quite inconsistent with things he say. I assume this means the 14nm scanner is on track, but it will probably be like 1980i. Which means they can get to 7 to 10nm with further refinement.
Havok sounds a lot more consistent when you distinguish between R&D verification from product validation. R&D verification clears the design. Product validation clears the bar to commit to mass production and installation. For complex equipment especially there’s sometimes a bit of extra development work between completion of R&D verification and completion of product validation. It’s a bit like overtime for R&D. You know the design works as intended and can do the things you set out with your requirements but there are things here and there that you may want to fix or refine or revise because they make a big difference for the operation of the equipment in real world use cases. R&D verification is ultimately all about internal testing on the design but it’s usually a good idea not to commit to the production line for a complex device until it’s received external testing by the people who you want to sell to and are the ones who will actually use the product.
 

olalavn

Senior Member
Registered Member
This makes sense what he is saying.

It confirms what @tokenanalyst said earlier that planar grating inferometer was completed by 2021. havok seems to know what he is talking about, but he is also quite inconsistent with things he say. I assume this means the 14nm scanner is on track, but it will probably be like 1980i. Which means they can get to 7 to 10nm with further refinement.

by the way, ASML continues to improve 1980i to have higher wpm, which may or may not be available to China after the latest sanctions.

Right, I think he is pointing out the inconsistent narrative in some of the things havok is saying. Let's face it, he contradicts himself sometimes.




Yes, Loongson is the future as I've said recently.

Huawei would be well served to work with the only 100% domestic CPU maker. Anyone else would be concerned about getting put on entity list. They've made LA compatible with Euler and Harmony. They've also integrated LA SoC with harmony, but haven't done anything like that with also major Chinese chipmakers.

Even more than that, LA is the only CPU offering competitive performance with 12 to 28 nm process in China. Kunpeng-920 is a great chip but it's 3 years old and uses a process that SMIC probably can't mass produce right now. So, what good is that?

Huawei should work with Loongson to get it even more compatible with ARM and x86.


That's kind of comforting. They are not so desperate that they are willing to let go of qualification standard. Things are setup for success imo, but they need some time.

So, they've gone 100% with 55/65 nm process by now.
2024 will be 28nm
2025 will be 14 and then 7nm. Seems like if they can achieve this, they did well
UOS will support HAP in the future.... Hongmeng's ecosystem will creep into all parts of China... it's the real IOT system... Loongson and Kunpeng, PanguM will be Huawei's future
 

paiemon

Junior Member
Registered Member
Havok sounds a lot more consistent when you distinguish between R&D verification from product validation. R&D verification clears the design. Product validation clears the bar to commit to mass production and installation. For complex equipment especially there’s sometimes a bit of extra development work between completion of R&D verification and completion of product validation. It’s a bit like overtime for R&D. You know the design works as intended and can do the things you set out with your requirements but there are things here and there that you may want to fix or refine or revise because they make a big difference for the operation of the equipment in real world use cases. R&D verification is ultimately all about internal testing on the design but it’s usually a good idea not to commit to the production line for a complex device until it’s received external testing by the people who you want to sell to and are the ones who will actually use the product.
From reading his posts, Havok sounds inconsistent because he toggles between saying that the project has met the design requirements (i.e. verification) vs the product is good to go, which is design validation. The DUVi prototype may have met the design/performance requirements like power source, resolution, throughput, etc specified in the project scope through testing and evaluation at SMEE. However, it does not mean that the product/prototype satisfies the need of the customers, hence the need for the end-users or an approximation of them such as ICRD to conduct that evaluation. SMEE certainly took user feedback in terms of establishing its design and requirements, but only realistic validation scenarios will confirm if the fabs will go with it. So he is technically correct in statement that the Project 002 has been completed from a design and development standpoint (thus far assuming no major iterations are needed due to user feedback).
 

latenlazy

Brigadier
From reading his posts, Havok sounds inconsistent because he toggles between saying that the project has met the design requirements (i.e. verification) vs the product is good to go, which is design validation. The DUVi prototype may have met the design/performance requirements like power source, resolution, throughput, etc specified in the project scope through testing and evaluation at SMEE. However, it does not mean that the product/prototype satisfies the need of the customers, hence the need for the end-users or an approximation of them such as ICRD to conduct that evaluation. SMEE certainly took user feedback in terms of establishing its design and requirements, but only realistic validation scenarios will confirm if the fabs will go with it. So he is technically correct in statement that the Project 002 has been completed from a design and development standpoint (thus far assuming no major iterations are needed due to user feedback).
Yes that’s what I’m saying. Good to go could just mean good to go to trial with customers. A product can go to trial with customers and also not be ready for mass production. From my read of him it sounds a lot like the product is currently undergoing validation, which is about where I would expect it to be given that it was supposed to have “finished” verification last year. This is part of the “integration” process I talked about years back. We have an instrument but we haven’t cleared adoption seems to be the consistent line behind what he’s saying. The validation process itself is a bit of a unique situation because the bans make it so that adoption really isn’t dependent on passing validation, and if the need arises the instrument could probably be utilized without any post validation revisions that might come up from customer trials.

I think the cleanest way to present the situation is that they have a machine that they can use if they really need to but if they can afford the time to refine it before production integration then it’s worthwhile to take that time to get a better product.
 
Status
Not open for further replies.
Top