Opinionated - China Chipping Away to Semiconductor Dominance

[ZeEa5KPul]

That's what they say about ZTE.
I respect the Chinese courage to face up to the SOLE Military and Financial world superpower.

Indians are JUST LOUDMOUTH COWARDS, threatening to impose tariffs on the US for months now but NO BALLS to execute.
Indians are Cheerleaders for the STRONG to OPPRESS the WEAK.
The OPPRESSED, and the VICTIMS OF COLONIZERS look up to CHINA to fight IMPERIALISM.
.

Seriously. I have never in my life seen anything as disgusting and repugnant as the Indian mentality. The most jealous, cowardly, bitter haters walking the face of the Earth.

Oh, well, haters gon' hate; winners gon' win.

 

[Han Patriot]

TSMC and Samsung don't face pressure from US. Nor do Taiwan and South Korea have dreams of being a global power some day.



All of this easily available online.



It's almost dead. There doesn't seem to be any sign that US is relenting.

So if it's early available online, post it and compare. YMTC is a few magnitudes bigger than Jinhua, do you know that. Well unlike you submissive indicas, we Chinese are fighters, the more you deny us, the stronger we become. Years of Spoon-feeding had turned Indian into a technological stooge.

What's wrong in having ambition to be a superpower? Just because we want to be on parity with the US that's wrong? All our efforts are futile? We will never catch up? They will ban us, they will kill switch us? I can sense the envy in you. Ever seen any Indians etchers, moved? Whether we can do it doesn't matter, we will never give up. You laughed when we started HSR, when we sent a man to space, when we had Olympics, when we had quantum SATs, when we created huawei, you are still laughing now. Continue, we will still not give up. One day you wake up, you will realize your while high tech infrastructure being made in China.

And you wonder why your TBMs, ultrasyoer xritical power equipment, radiation airport scanners are made in China? Because we dare to dream, we don't just talk.

 

[cirr]

Semi matters of significance 2019

①. 三维闪存芯片 长江存储量产自研Xtacking架构64层三维闪存芯片，打破了美日韩厂商的垄断。
②. 车规级碳化硅Mosfet芯片 比亚迪微电子推出自研车规级Mosfet芯片，打破了德法日厂商的垄断。
③. CMOS图像传感器 北京豪威推出自研0.8um，4800万像素级的CMOS图像传感器，与索尼之间的技术差距缩短到不到一年。
④. X86架构桌面CPU 上海兆芯研发的KX6000桌面CPU正式量产，达到英特尔第七代Core i5的水平。
⑤. GPU 上海兆芯推出国产最高水平的GPU极瑞3000。
⑥. 手机芯片 华为海思推出业界最高性能的5g手机芯片麒麟990。
⑦. 电视芯片 华为海思推出业内性能最强的智能电视芯片海思Vxxx。
⑧. DRAM 合肥长鑫正式量产8GB LPDDR4型内存芯片。
⑨. 晶圆代工 中芯国际14纳米工艺晶圆代工线正式量产。
⑩. FPGA芯片 紫光同创推出28纳米工艺制程7000万门级titan2系列高性能FPGA芯片。

 

[Beidou2020]

New machine raises country's image in photolithography

The Institute of Optics and Electronics under the Chinese Academy of Sciences announced on Thursday that it has developed its own photolithography machine. [Photo/IC]

"Make it small" has been the mantra that drives big developments in microelectronics, from transistors to processors. So much so that high-tech companies' competitiveness rests on the ability to reliably create components on the scale of nanometers, or one-billionth of a meter.

At the heart of microengineering lies photolithography, one of the key techniques used to create the circuitry patterns on semiconductor chips. This allows engineers to pack and replicate complicated circuits with millions of components into a very tiny space using light.

This technique is so advanced that only a handful of companies from Europe and Japan can produce the machine capable of such a feat. But recently, China entered the game.

The Institute of Optics and Electronics under the Chinese Academy of Sciences announced on Thursday that it has developed its own photolithography machine, thus overcoming one of the major engineering obstacles limiting China's development in chips, nanocomponents and optical instruments.

The new machine, which scientists began building in 2012, can etch circuitry patterns less than 22 nanometers using ultraviolet light. Combined with other techniques, it can be used in the future to create chips of around 10 nanometers.

"The machine is of great value in manufacturing general chips and other materials that require microengineering, including some integrated circuits," said Hu Song, deputy chief designer of the project.

The machine can also be used to make small components for applications such as sensors, detectors and biochips, Hu said.

It is already being used by several institutions including Sichuan University and the University of Electronic Science and Technology of China.

However, the new machine's production capability is still small, hence it is still limited to producing key components for research, Hu said.

In the coming years, the teamwill focus its efforts on increasing the machine's productivity to industrial scale. There are still substantial gaps in the microengineering sector between China and developed countries, but China is catching up fast, Hu said.

Companies capable of photolithography will have an overwhelming edge in producing microelectronics, he said. Chip manufacturing giant Intel claimed it could produce more than 5 billion nanoscale transistors every second, according to its company's fact sheet.

The world's largest photolithography supplier is a Dutch company called ASML Holding. Some of ASML's main competitors are Canon and Nikon.

Since the late 1990s, China has been blocked from importing cutting-edge photolithography technologies and other chip manufacturing equipment from developed countries.

This situation is aggravated because ASML owns substantial patents covering imprint lithography, which includes photolithography, Hu said.

As a result, Chinese companies have to rely on relatively outdated and inefficient techniques to produce microelectronics, thus their product is often inferior to that of developed countries.

To overcome the monopoly, the Institute of Optics and Electronics discovered a new physics phenomenon in 2003 and used the property as the basis for the new machine.

China now has 47 domestic and four international patents regarding the new technology, Hu said.

"We no longer fear a foreign technical blockade because we have full intellectual ownership of the new technique."

 

[Paul2]

What a long thread we have here. Semiconductors are such and industry about which 99% stuff known in the popular culture is wrong.

I work in an electronics engineering consulting, and once pursued career in microelectronics. Microelectronics was just too tough to enter for me, and I went to OEM electronics instead 12 years ago. Still kicking.

You guys put up quite a lot of misconceptions. What pains me is that the supposedly qualified people who were entrusted the colossal 1 trillion CNY "Made in China 2025" fund also don't seem to navigate well in the industry.

I meet people from Qinghua Unigroup on almost every industry event around, and I wonder if 10+ of their senior people have time to just hang out aimlessly there, what the rest of their employees do.

I once had a small talk about fab equipment with one of their lad, allegedly also in "fab strategy." It blew away my mind that the guy who was claimed to be an industry veteran, had zero idea about process control, metrology, and workings of the industry in general. The only thing in his sight was the idea "we will get scanners matching AMLS's and our fabs will magically become competitive." When I started the talk about that, it felt it was a total revelation to him . No wonder, Zhao Weiguo did not last on the job.

I have a lot of criticism for such policy making, and handling of semiconductor industry in general in China.

1. For one, China needs own semi industry, but making it is not a job for freaking bureaucrats
2. Semiconductor industry is already enjoying completely unthinkable subsidies from the state, with the biggest being 0% tax rate with close to no preconditions.
3. The function of a business is to freaking make money — and you can't make a global leader company that doesn't make tons of money. Something so simple, yet not so clear to policymakers.
4. To make money, you need to make a viable businesses — and not thinking of achieving "technology A, B, or C" with mere hopes and speculations of its commercial values. The story how state funds spent many tenths of billions on 10+ failed fab businesses in the past is all about that.

My comments on specific cases:

1. For once, stop that stupidest endeavour of making own x86 chips. I once heard the notion "Chinese chips suck because because they can't run Microsoft Windows" and that, again, came out of a mouth of an industry official. I will rephrase that as "Microsoft Windows sucks because it can't run on Chinese chips." All that endeavour with Chinese x86 is completely meaningless.
2. At least policymakers should police illegal non-compete agreements in the industry. It's not a big secret that Allwinner - once an unstoppable steamroller entered a non-compete with Qualcomm for a completely pitiful amount of money in exchange for not rolling over them in a phone business. Rockchips is said to also did something similar with Intel, but it's less clear for what they made a non-compete for.
3. Qualcomm is a racketeering business, but all its money come from the fact that electronics with their chips is all made in China. Qualcomm can be totally annihilated by a just and sound anti-trust action from Chinese industry regulators.
4. Current fab strategy as seem by policy makers in Beijing is wrong. Even if some of mainland fabs will make a verbatim copy of TSMC's fab, it will not be a successful business. THERE IS NO COMMERCIAL RATIONALE IN COPYING TSMC. If they, after all, continue on this path, they need to fund companies with a chance of gaining commercial viability, and there is none in chasing TSMC in advanced nodes, without having a mature commodity process business. This is where the policy is 100% wrong — old proverb says "You don't eat the snake from the head."
5. Memory business... now memory prices are plunging, and the gigantic investments into memory fabs will go to nowhere. If Samsung can't make money on memory now, how will newcomers do?
6. Fab equipment — the only way to make money there is to develop and sell fab equipment to competitors abroad. There is not a single worthy client inside China (SMIC included) to kickstart a profitable tool business. However, there is rationale to continue research there, in anticipation of time when it will make sense for China to do it. And one more thing, scanners are not the best thing to begin to begin with in this industry, metrology and process control tools are as important, if not more.
7. Stop wasting money on gimmick stuff like "AI chips," FPGAs, when China struggles to make a single decent microcontroller...
8. Power chips — do it
9. Advanced packaging — do it
10. "Making China a great semiconductor superpower" — these people really know how to pick policy names... In general, the effect of existing constituent policy components, made a very small effect on industry in overall
    1. To begin with, there is not enough engineers for that. A giant amount of worthy engineers leave China, and state funds invites loosers from Taiwan to run our semiconductor industry instead. That's unsustainable. High profile professionals will never ever want to live in China without us making China a nice place to live in outside of "Laowai ghettoes" for foreign cadres
    2. Semiconductor industry is a global industry — there is no "Chinese semiconductor industry," "Taiwanese semiconductor industry," or "American semiconductor industry." You can only see China becoming competitive in semiconductors as a player on the global market. Otherwise, it will be Chinese semiconductor industry vs. Global semiconductor industry. Invite more foreign companies, make them stuck here, make China invaluable and irreplaceable for them.
    3. Stop chasing others, make others chase you — develop own areas of competency, one in which we can take a lead, not ones in which we will keep loosing.
    4. If you want to develop electronics industry, it's a good idea to not to bulldoze electronics factories... A poke to the devastating 2009-2012 factory demolitions in Shenzhen.

 

[somsak]

What a long thread we have here. Semiconductors are such and industry about which 99% stuff known in the popular culture is wrong.

I work in an electronics engineering consulting, and once pursued career in microelectronics. Microelectronics was just too tough to enter for me, and I went to OEM electronics instead 12 years ago. Still kicking.

You guys put up quite a lot of misconceptions. What pains me is that the supposedly qualified people who were entrusted the colossal 1 trillion CNY "Made in China 2025" fund also don't seem to navigate well in the industry.

I meet people from Qinghua Unigroup on almost every industry event around, and I wonder if 10+ of their senior people have time to just hang out aimlessly there, what the rest of their employees do.

I once had a small talk about fab equipment with one of their lad, allegedly also in "fab strategy." It blew away my mind that the guy who was claimed to be an industry veteran, had zero idea about process control, metrology, and workings of the industry in general. The only thing in his sight was the idea "we will get scanners matching AMLS's and our fabs will magically become competitive." When I started the talk about that, it felt it was a total revelation to him . No wonder, Zhao Weiguo did not last on the job.

I have a lot of criticism for such policy making, and handling of semiconductor industry in general in China.

1. For one, China needs own semi industry, but making it is not a job for freaking bureaucrats
2. Semiconductor industry is already enjoying completely unthinkable subsidies from the state, with the biggest being 0% tax rate with close to no preconditions.
3. The function of a business is to freaking make money — and you can't make a global leader company that doesn't make tons of money. Something so simple, yet not so clear to policymakers.
4. To make money, you need to make a viable businesses — and not thinking of achieving "technology A, B, or C" with mere hopes and speculations of its commercial values. The story how state funds spent many tenths of billions on 10+ failed fab businesses in the past is all about that.

My comments on specific cases:

1. For once, stop that stupidest endeavour of making own x86 chips. I once heard the notion "Chinese chips suck because because they can't run Microsoft Windows" and that, again, came out of a mouth of an industry official. I will rephrase that as "Microsoft Windows sucks because it can't run on Chinese chips." All that endeavour with Chinese x86 is completely meaningless.
2. At least policymakers should police illegal non-compete agreements in the industry. It's not a big secret that Allwinner - once an unstoppable steamroller entered a non-compete with Qualcomm for a completely pitiful amount of money in exchange for not rolling over them in a phone business. Rockchips is said to also did something similar with Intel, but it's less clear for what they made a non-compete for.
3. Qualcomm is a racketeering business, but all its money come from the fact that electronics with their chips is all made in China. Qualcomm can be totally annihilated by a just and sound anti-trust action from Chinese industry regulators.
4. Current fab strategy as seem by policy makers in Beijing is wrong. Even if some of mainland fabs will make a verbatim copy of TSMC's fab, it will not be a successful business. THERE IS NO COMMERCIAL RATIONALE IN COPYING TSMC. If they, after all, continue on this path, they need to fund companies with a chance of gaining commercial viability, and there is none in chasing TSMC in advanced nodes, without having a mature commodity process business. This is where the policy is 100% wrong — old proverb says "You don't eat the snake from the head."
5. Memory business... now memory prices are plunging, and the gigantic investments into memory fabs will go to nowhere. If Samsung can't make money on memory now, how will newcomers do?
6. Fab equipment — the only way to make money there is to develop and sell fab equipment to competitors abroad. There is not a single worthy client inside China (SMIC included) to kickstart a profitable tool business. However, there is rationale to continue research there, in anticipation of time when it will make sense for China to do it. And one more thing, scanners are not the best thing to begin to begin with in this industry, metrology and process control tools are as important, if not more.
7. Stop wasting money on gimmick stuff like "AI chips," FPGAs, when China struggles to make a single decent microcontroller...
8. Power chips — do it
9. Advanced packaging — do it
10. "Making China a great semiconductor superpower" — these people really know how to pick policy names... In general, the effect of existing constituent policy components, made a very small effect on industry in overall
    1. To begin with, there is not enough engineers for that. A giant amount of worthy engineers leave China, and state funds invites loosers from Taiwan to run our semiconductor industry instead. That's unsustainable. High profile professionals will never ever want to live in China without us making China a nice place to live in outside of "Laowai ghettoes" for foreign cadres
    2. Semiconductor industry is a global industry — there is no "Chinese semiconductor industry," "Taiwanese semiconductor industry," or "American semiconductor industry." You can only see China becoming competitive in semiconductors as a player on the global market. Otherwise, it will be Chinese semiconductor industry vs. Global semiconductor industry. Invite more foreign companies, make them stuck here, make China invaluable and irreplaceable for them.
    3. Stop chasing others, make others chase you — develop own areas of competency, one in which we can take a lead, not ones in which we will keep loosing.
    4. If you want to develop electronics industry, it's a good idea to not to bulldoze electronics factories... A poke to the devastating 2009-2012 factory demolitions in Shenzhen.

I think AI chip is the best bet.

 

[qwerrty]

1. At least policymakers should police illegal non-compete agreements in the industry. It's not a big secret that Allwinner - once an unstoppable steamroller entered a non-compete with Qualcomm for a completely pitiful amount of money in exchange for not rolling over them in a phone business. Rockchips is said to also did something similar with Intel, but it's less clear for what they made a non-compete for.

this. i am so happy that donald trump is forcing china to stop doing those jv bs.
when unisoc announced jv with intel a year ago, many people were like wha? why?.. i'm so glad they broke up now. it made no sense in the first place.. lol

Stop wasting money on gimmick stuff like "AI chips," FPGAs, when China struggles to make a single decent microcontroller...

totally disagree with this. there are a lot of stuff that don't need overpowered jack of all trades cpu or gpu. there's a lot of money to be made in this area. china for the first time in semiconductor industry is moving in right direction. all the big tech companies in both china and the US investing heavily in this area is proof

 

[Paul2]

unexpendable

Stop wasting money on gimmick stuff like "AI chips," FPGAs, when China struggles to make a single decent microcontroller...

totally disagree with this. there are a lot of stuff that don't need overpowered jack of all trades cpu or gpu. there's a lot of money to be made in this area. china for the first time in semiconductor industry is moving in right direction. all the big tech companies in both china and the US investing heavily in this area is proof

I hope you have an understanding what microcontrollers are. Those are tiny, self-contained computers with just SRAM, primitive CPU, some low speed I/O, and, some times, integrated flash. Those exactly are the truly unexpendable parts of everyday things, and machinery, and China is nearly wholly dependent on outside imports.

See, microcontrollers are parts of a great many light and heavy industry goods: from appliances, to industrial machinery. Pretty much every electronics that is not a smartphone or computer.

Imagine embargo on them, and the chain reaction that will follow in the few industries where China have dominance. If USA, or Taiwan will embargo an Nvidia GPU, or Intel CPU, nobody but videogame addicts will complain, but if they will close the tap on MCUs, they will cut of the oxygen to a big, and by far the most competitive portion of Chinese industry.

There is only one Chinese company that somehow achieved commercial viability in a mid-market MCU segment, but event its boss is not Chinese.

china for the first time in semiconductor industry is moving in right direction. all the big tech companies in both china and the US investing heavily in this area is proof

It's not a proof. Americans by far surpass the world in investing into dumb, useless stuff; driven by speculative gain expectations of reactionary social classes of bankers, lawyers, and MBAs.

What value you can make with "AI" chips? Ai chip makers themselves have poor idea about that, besides using them for some ridiculous Instagram filters. Billions of RMB for that? No thanks.

Same with programmable logic — very niche product. For any mass market good, ASICs flatten them down. The only good use case for them is doing something that nobody does en masse, and that there are not many such things by the definition.

It pains me hearing again and again the logic of "if Americans are investing/doing this, we must too" on a premise of them thinking that "Americans must know that better." Great heaps of money were lost by the industry and those people themselves.

See — the flexible display idiocy: Steve Jobs made a joke about flexible displays, and those two (Samsung and BOE/Royole/Huawei team) actually went and banged away many billions on RnD and manufacturing facilities to actually make it, because they actually thought of that as a smart idea!

 

[qwerrty]

Imagine embargo on them, and the chain reaction that will follow in the few industries where China have dominance. If USA, or Taiwan will embargo an Nvidia GPU, or Intel CPU, nobody but videogame addicts will complain, but if they will close the tap on MCUs, they will cut of the oxygen to a big, and by far the most competitive portion of Chinese industry.

no, they won't die. c'mon man. it's not some wonder technology with no alternatives. lol. nearly all MCUs are arm based less complex to design than general purpose chips. china has a lot of startups doing those. you dont see many on the market, because of patent protection. if banned, they just copy em. they are not super complicate like semiconductor manufacturing tool. lol

https://wallstreetcn.com/articles/3478479
http://www.navinfo.com/product/chip

 

[Paul2]

no, they won't die. c'mon man. it's not some wonder technology with no alternatives. lol. nearly all MCUs are arm based less complex to design than general purpose chips. china has a lot of startups doing those. you dont see many on the market, because of patents. if banned, they just copy em. they are not super complicate like semiconductor manufacturing tool. lol

https://wallstreetcn.com/articles/3478479
http://www.navinfo.com/product/chip

Simple they are, but not less important. You still have to have somebody making them and having a commercially viable business while doing so. With my knowledge of the industry, Chinese companies position today are even worse than 20 years ago. 20 years ago there were companies copying 8051 and making meagre profits, but these days all of them are gone because 8051 stopped making money.

Most "startups" in that area in China are there just to beg for government handouts, and investors money. I looked up your links, and I can tell that I haven't seen any of their chips making any market presence, and that with me being working in OEM electronics for 12 years.

It's still all NXP, Atmel, and STM even in cheapest product niches.

 

[qwerrty]

Simple they are, but not less important. You still have to have somebody making them and having a commercially viable business while doing so. With my knowledge of the industry, Chinese companies position today are even worse than 20 years ago. 20 years ago there were companies copying 8051 and making meagre profits, but these days all of them are gone because 8051 stopped making money.

Most "startups" in that area in China are there just to beg for government handouts, and investors money. I looked up your links, and I can tell that I haven't seen any of their chips making any market presence, and that with me being working in OEM electronics for 12 years.

that's why we need a ban on china, so those startups can start to explode. lol. a lot of people on cn forums don't like it when xi offers to buy more chips from the US in part of the trade deal. this is bad for "china core" dream

 

[grandmaster]

1. For once, stop that stupidest endeavour of making own x86 chips. I once heard the notion "Chinese chips suck because because they can't run Microsoft Windows" and that, again, came out of a mouth of an industry official. I will rephrase that as "Microsoft Windows sucks because it can't run on Chinese chips." All that endeavour with Chinese x86 is completely meaningless.

I like this point. Why do the policy makers have such narrow mind and are too rigid? the way is always opened, but why china cinch itself with such kind of thinking? Everyone knows Microsoft OS is not secure, no one know anything about Microsoft work with NSA behind the screen, why do china make itself depend on American product and let them spy china? even Apple can write its own OS, why cannot China write new OS for new chip architecture? why not invest money and human resource into developing Chinese OS and basic Chinese software ecosystem? developing Chinese OS is cheaper than trying to make x86 chip. they still not learn good lesson when google withdrew from china, china search engines boomed.

 

[j20blackdragon]

First, there are too many companies/countries making microcontroller (MCU) for a ban on China to be effective.

Second, China has domestic MCU.

GD32 is a new 32-bit high performance, low power consumption universal microcontroller family powered by the ARM Cortex®-M3 RISC core
http://www.gigadevice.com/products/microcontrollers/gd32/arm-cortex-m3/

The GD32F4 device belongs to the performance line of GD32 MCU Family. It is a new 32-bit general-purpose microcontroller based on the ARM Cortex-M4 RISC core
http://www.gigadevice.com/products/microcontrollers/gd32/arm-cortex-m4/

 

[Han Patriot]

that's why we need a ban on china, so those startups can start to explode. lol. a lot of people on cn forums don't like it when xi offers to buy more chips from the US in part of the trade deal. this is bad for "china core" dream

So if it's so stupid and useless then why is Trump so worried? Our friend Paul here seems to suggest we should microcontrollers instead of CPUs. What a genius Indian.

 

[ZeEa5KPul]

What a genius Indian.

I know, right? @Bussard Ramjet's alt should start up a company in China since he seems to have everything figured out. By his own admission there's a lot of money sloshing around in Chinese semiconductors, why's he wasting his time and ours here when he could be grabbing some of it?