News on China's scientific and technological development.
- Thread starter Quickie
- Start date
Hi localizer,
Its good to know from the source itself, I'm happy to hear more from you regarding the state of Chinese IC development. This past few weeks is very disheartening , with constant negative news of sanctioning chinese company. Hopefully Trump action will enlighten our brethren to come home, to work for a chinese company or establish their own.
That is not good, right now I'm working an IT job where I'm "working" really just 15 minutes to half an hour a day, rest of the time goofing off and surfing the web, making under $100k
hell of a long way to fall to being slave labored at American prison rates of $.40 cents an hour...
cross post from JustAnotherPerson (PAKISTAN DEFENSE FORUM) about China present situation
That only means how much they are going to lose and is going to be A LOT. supply and demand, the Chinese know how to make semiconductors and equipment, they trust you and you break the trust, now the demand will change faster than most people think. They are closing the technical gap fast, The irony is all thanks to Neocons in the White House, for years the Chinese government failed to change the behavior in the industry but the Americans give them the push that they needed. Before the money that they invested was going nowhere now is sticking more and more.
This is a supply and demand problem, your problem is that you are seeing this with your ideological lenses that for some reason you never take off. This is about country trying to fix a supply problem cause by overzealous politicians, so whatever they gain is going to be your loss.
The fact is China has companies in every single stage in the semiconductor industry the only thing they need is costumers to grow.
The fact is that if you do your research and you start look under the waters the Chinese are doing significant advances in semiconductor equipment and closing the technical gap faster, according to people inside China they are in war mode, that an industry that once was a death end job now are seeing the faster grow in salaries than any other industry.
The fact is the Chinese semiconductors companies that use to exist in the margin of the China electronic industry are now seeing an significant increase in demand and opportunities
The fact is the China do a significant amount of R&D in semiconductors and they have a lot of knowledge in that area.
The fact IP is becoming less of issue this days, the irony with open ISAs like RISC-V the nobody can control is making easier to anyone to develop semiconductors without worrying about IP issues.
Take off your ideological lenses, the Neocons are fighting a war without knowing what they are getting into, they dont understand the industry and they are grossly underestimating the capabilities of their adversary.
The winning strategy for United State was keep doing what they were doing in the past. selling, period. nothing more and nothing less. As long as a supply problem doesn't exist would have taken longer for the Chinese to mature the semiconductor industry in China.
Contributor
Founder & Director of the Quantitative Finance Program and Hanlon Financial Systems Center at the Stevens Institute of Technology (New Jersey) and Advisory Board Member at Hanlon Investment Management"
There is contrast between people who write articles in economic magazines who are not in the industry and people who write articles in semiconductors journals, the first only see the industry from outside, the shell of it. The second are getting nervous because they are seeing the activity below the water and they dont like it. They know once China have the capabilities it will probably over-invest hurting the whole industry in the process.
Remenber the Chinese have nothing to lose so whatever the win 10%, 20% or 100% is going to be a loss for the U.S.
That only means how much they are going to lose and is going to be A LOT. supply and demand, the Chinese know how to make semiconductors and equipment, they trust you and you break the trust, now the demand will change faster than most people think. They are closing the technical gap fast, The irony is all thanks to Neocons in the White House, for years the Chinese government failed to change the behavior in the industry but the Americans give them the push that they needed. Before the money that they invested was going nowhere now is sticking more and more.
This is a supply and demand problem, your problem is that you are seeing this with your ideological lenses that for some reason you never take off. This is about country trying to fix a supply problem cause by overzealous politicians, so whatever they gain is going to be your loss.
The fact is China has companies in every single stage in the semiconductor industry the only thing they need is costumers to grow.
The fact is that if you do your research and you start look under the waters the Chinese are doing significant advances in semiconductor equipment and closing the technical gap faster, according to people inside China they are in war mode, that an industry that once was a death end job now are seeing the faster grow in salaries than any other industry.
The fact is the Chinese semiconductors companies that use to exist in the margin of the China electronic industry are now seeing an significant increase in demand and opportunities
The fact is the China do a significant amount of R&D in semiconductors and they have a lot of knowledge in that area.
The fact IP is becoming less of issue this days, the irony with open ISAs like RISC-V the nobody can control is making easier to anyone to develop semiconductors without worrying about IP issues.
Take off your ideological lenses, the Neocons are fighting a war without knowing what they are getting into, they dont understand the industry and they are grossly underestimating the capabilities of their adversary.
The winning strategy for United State was keep doing what they were doing in the past. selling, period. nothing more and nothing less. As long as a supply problem doesn't exist would have taken longer for the Chinese to mature the semiconductor industry in China.
Contributor
Founder & Director of the Quantitative Finance Program and Hanlon Financial Systems Center at the Stevens Institute of Technology (New Jersey) and Advisory Board Member at Hanlon Investment Management"
There is contrast between people who write articles in economic magazines who are not in the industry and people who write articles in semiconductors journals, the first only see the industry from outside, the shell of it. The second are getting nervous because they are seeing the activity below the water and they dont like it. They know once China have the capabilities it will probably over-invest hurting the whole industry in the process.
Remenber the Chinese have nothing to lose so whatever the win 10%, 20% or 100% is going to be a loss for the U.S.
An example of what Huawei is doing to solve its Chip supply problem.
from cnTechPost
Huawei may start using dual processor solutions from Mate40
2020-08-09 12:50:06 GMT+8 | cnTechPost
0
Huawei will use dual processor options starting with the Mate40 at the earliest, with both MediaTek and Samsung as possible suppliers, cited Qualcomm as well as MediaTek insiders as saying.
Huawei's stockpile on the 5G chip Kirin 9000 can support the Mate40's higher end version, but Huawei's flagship series shipments are usually above the 10 million level, a chipmaker insider said.
The stockpile on the Kirin high-tier series is between 8 million and 9 million, and on the standard version Huawei is likely to use processors from other companies, the person said.
Huawei began to increase the purchase of MediaTek chips in May, and even offered for $10 more for each chip, the person said.
"Three weeks ago, Huawei updated the latest flagship phone chip roadmap for Chinese customers. Many specifications at a glance is dominated by MediaTek," the insider said.
MediaTek staff said they could not disclose any details because of a "confidentiality agreement". But reports said Huawei recently ordered 120 million chips from MediaTek, and six of the mobile phones released this year have used MediaTek chips.
The main contenders for the 5G chip platform are now Huawei's Hisilicon, Qualcomm, MediaTek and South Korea's Samsung.
Although there were reports that Samsung and Huawei are exploring a deal, Samsung may soon be able to manufacture advanced chips for Huawei's 5G devices, but the process will see difficulties, said.
from cnTechPost
Huawei may start using dual processor solutions from Mate40
2020-08-09 12:50:06 GMT+8 | cnTechPost
0
Huawei will use dual processor options starting with the Mate40 at the earliest, with both MediaTek and Samsung as possible suppliers, cited Qualcomm as well as MediaTek insiders as saying.
Huawei's stockpile on the 5G chip Kirin 9000 can support the Mate40's higher end version, but Huawei's flagship series shipments are usually above the 10 million level, a chipmaker insider said.
The stockpile on the Kirin high-tier series is between 8 million and 9 million, and on the standard version Huawei is likely to use processors from other companies, the person said.
Huawei began to increase the purchase of MediaTek chips in May, and even offered for $10 more for each chip, the person said.
"Three weeks ago, Huawei updated the latest flagship phone chip roadmap for Chinese customers. Many specifications at a glance is dominated by MediaTek," the insider said.
MediaTek staff said they could not disclose any details because of a "confidentiality agreement". But reports said Huawei recently ordered 120 million chips from MediaTek, and six of the mobile phones released this year have used MediaTek chips.
The main contenders for the 5G chip platform are now Huawei's Hisilicon, Qualcomm, MediaTek and South Korea's Samsung.
Although there were reports that Samsung and Huawei are exploring a deal, Samsung may soon be able to manufacture advanced chips for Huawei's 5G devices, but the process will see difficulties, said.
Quickie
Colonel
China already has within its own border the counterparts of Google, FB, Youtube, and other popular social apps.
China can offer an open-source version of them to countries outside China where they can make changes, upgrade, and maintain the source code as an international effort. Any advertising money made would stay within the country.
That would be China's nuclear option as that would spell the end of the U.S. stranglehold on the mobile software ecosystem.
China can offer an open-source version of them to countries outside China where they can make changes, upgrade, and maintain the source code as an international effort. Any advertising money made would stay within the country.
That would be China's nuclear option as that would spell the end of the U.S. stranglehold on the mobile software ecosystem.
An interesting proposal from FairAndUnbiased (pakistan defense forum) on where HiSilicon should proceed.
I believe HiSilicon still has many niches to explore within the semiconductor market. Not being able to fabricate leading edge logic at TSMC does not preclude its ability from competing in other markets. However, it also has to make financial sense.
My proposal: HiSilicon should become an IDM with an initial fab focus on higher value portions of the analog/mixed signal/RF (modem, microcontroller, FPGA, etc) market, while keeping fabless for its highest value digital IP.
Why go this route?
1. HiSilicon has extensive experience in designing for leading edge nodes, which is very expensive and difficult. It is much cheaper and easier to design for the analog/mixed signal/RF market. However, being easy with no demand doesn't mean much. But there is demand - Huawei need modems, microcontrollers and FPGAs just as much as they need leading edge logic devices. Currently, they can outsource ALL of these (including almost leading edge logic) to SMIC. But then they will never attain true technical independence. Everyone has to start somewhere.
2. You can still make money in the analog/mixed signal/RF market, for little investment. That is because they do not need leading edge tools - you can buy old tools from 10-20 years ago for 1/5th the cost and still make things work. For example, Analog Devices still makes $5 billion in sales every year (half of HiSilicon) in 1980's fabs using 150 mm and 200 mm wafers running on .350-.180 um tech.
3. This is an area that has lots of competitors, but is easy to get started due to cost. Other than Texas Instruments, every one is much smaller than HiSilicon. Some have been in this field for decades, but the starter advantage doesn't matter much when they're essentially either standing still or not following industry best practices. HiSilicon has tons of experience in both designing ARM based chips and using them in applications - designing a high end ARM based microcontroller or a RF front end should not be too much of an issue for them.
Summary: HiSilicon should become an IDM, starting with high end analog/mixed signal applications like microcontroller and RF, then move back into its old niche once their fabs prove profitable. It can source its equipment from depreciated Japanese assets, Samsung/TSMC surplus and buy new from Chinese manufacturers when appropriate. To keep money on its Kirin IP, it can license it to other fabless companies.
Once HiSilicon proves it can be profitable as a manufacturer, it can slowly (over 5-10 years) start buying more leading edge assets and/or paying for fab time elsewhere. It will be hard to compete with TSMC, but competing with TSMC is difficult even for Intel and Samsung. Instead it will be to ensure that Huawei and HiSilicon can supply their own parts and design for manufacture directly on their own equipment.
I believe HiSilicon still has many niches to explore within the semiconductor market. Not being able to fabricate leading edge logic at TSMC does not preclude its ability from competing in other markets. However, it also has to make financial sense.
My proposal: HiSilicon should become an IDM with an initial fab focus on higher value portions of the analog/mixed signal/RF (modem, microcontroller, FPGA, etc) market, while keeping fabless for its highest value digital IP.
Why go this route?
1. HiSilicon has extensive experience in designing for leading edge nodes, which is very expensive and difficult. It is much cheaper and easier to design for the analog/mixed signal/RF market. However, being easy with no demand doesn't mean much. But there is demand - Huawei need modems, microcontrollers and FPGAs just as much as they need leading edge logic devices. Currently, they can outsource ALL of these (including almost leading edge logic) to SMIC. But then they will never attain true technical independence. Everyone has to start somewhere.
2. You can still make money in the analog/mixed signal/RF market, for little investment. That is because they do not need leading edge tools - you can buy old tools from 10-20 years ago for 1/5th the cost and still make things work. For example, Analog Devices still makes $5 billion in sales every year (half of HiSilicon) in 1980's fabs using 150 mm and 200 mm wafers running on .350-.180 um tech.
3. This is an area that has lots of competitors, but is easy to get started due to cost. Other than Texas Instruments, every one is much smaller than HiSilicon. Some have been in this field for decades, but the starter advantage doesn't matter much when they're essentially either standing still or not following industry best practices. HiSilicon has tons of experience in both designing ARM based chips and using them in applications - designing a high end ARM based microcontroller or a RF front end should not be too much of an issue for them.
Summary: HiSilicon should become an IDM, starting with high end analog/mixed signal applications like microcontroller and RF, then move back into its old niche once their fabs prove profitable. It can source its equipment from depreciated Japanese assets, Samsung/TSMC surplus and buy new from Chinese manufacturers when appropriate. To keep money on its Kirin IP, it can license it to other fabless companies.
Once HiSilicon proves it can be profitable as a manufacturer, it can slowly (over 5-10 years) start buying more leading edge assets and/or paying for fab time elsewhere. It will be hard to compete with TSMC, but competing with TSMC is difficult even for Intel and Samsung. Instead it will be to ensure that Huawei and HiSilicon can supply their own parts and design for manufacture directly on their own equipment.
Chinese software advantage (at least in China) would be language.
Japan still likes to use Yahoo due to better search results in Japanese. Koreans use Naver. Japan and Korea also likes to use LINE for chatting. Russians use Yandex and VK.
Baidu should give better Chinese search results.
Chinese should have access to Google and Youtube and stuff simply for competition. I find Baidu mostly useless for high end research.
Baidu should survive due to better Chinese search results.
Japan still likes to use Yahoo due to better search results in Japanese. Koreans use Naver. Japan and Korea also likes to use LINE for chatting. Russians use Yandex and VK.
Baidu should give better Chinese search results.
Chinese should have access to Google and Youtube and stuff simply for competition. I find Baidu mostly useless for high end research.
Baidu should survive due to better Chinese search results.
They are still building tablets, using Kirin 710 made out of SMIC 14nm process. Its still possible to make midrange phones with SMIC process. I expect the market for flagship phones to decline due to the global economic conditions anyway.
Huawei has enough stock of Kirin 9000 to make their Mate 40. They have built up stock for about two years. By the time this has depleted, SMIC hopes to have a 7nm process running then. At the very least SMIC is hoping their 7nm would be available by the end of this year.
If Huawei runs out of Kirin 9000 they would have to buy from Mediatek, and possibly Samsung Exynos. Qualcomm is lobbying for an exception to sell chips to Huawei.
US tech companies hope to get Trump voted out --- I can expect them to be donating to pro Biden PACs --- and then they will renegotiate and restructure their purchases to China.
HiSilicon and Huawei can end up losing talent that will flow to other Chinese tech companies --- Tencent and Alibaba are looking for big hires --- and even some to the Chinese defense companies, especially RF engineers who can end up working on things like radars, military communications, and electronic warfare.