China’s Semiconductor Catchup is Critical to Future Technology Competition

Dr. Walden Rhines, Chief Executive Officer, Mentor Graphics talked about the semiconductor industry.

Semiconductors are critical to global technological competition. The world-class semiconductor companies are in the USA, South Korea and Taiwan. Japan and Europe have solid semiconductor technology as well. China lags about 5-10 years behind in this critical technology. China will invest hundreds of billions to catch up. AI, 5G, Internet of things and other technological competition have semiconductor technologies as the foundation for competitive advantages.

China has put $170+ billion into its semiconductor industry. The government gave $20 billion to private equity to fund semiconductor companies and technology. There was another $97 billion from commercial sources and regional government. There is a new $47 billion funding.

However, China is still fabricating at 14 nanometers while the leading edge is at 7 nanometers. China is two generations behind but has leading edge designs. The leading edge designs from Huawei and other companies are fabricated at Taiwan Semiconductor.

Deloitte Global predicts that revenues for semiconductors manufactured in China will grow by 25 percent to approximately US$110 billion in 2019 from an estimated US$85 billion in 2018. China still only supplies 30% of its domestic demand of about $330 billion in 2019.

Growth is increasing to meet the increasing domestic demand for chipsets driven in part by the growing commercialization of artificial intelligence (AI). Deloitte Global further predicts that in 2019, a Chinese chip foundry will begin producing semiconductors specialized to support AI and machine learning (ML) tasks.

China’s Potential to Catch Up

Why is China better positioned than ever before? Five current conditions make China’s ascent in semiconductors more likely:

1. Domestic demand. China is now the largest global consumer of semiconductors, importing about US$200 billion worth each year. It has large population with 800 million internet users.

2. There is huge state and private funding is tens of billions per year. At the end of 2017, China had plans to build at least 14 new chip foundries.

3. China is trying to keep pace with the new age of AI hardware and software. Many of China’s largest companies are hoping to win an edge in the market for AI. Baidu, Alibaba and Tencent have a combined market capitalization of over US$1 trillion. They have invested billions in other companies, both domestically and overseas. They positions in more than half of China’s 124 unicorn startups, including SenseTime, the world’s most valuable pure-play AI company.

4. Onshoring foreign operations and hiring foreign talent.

Yangtze Memory Technologies has invested US$24 billion to build China’s first advanced memory chip factory and has lured thousands of engineers away from foreign chipmakers. They announced progress on its 32-layer NAND memory chip—a good sign which is still behind the state-of-the-art 64-layer chip that other memory manufacturers are achieving. SMIC hired a senior executive away from Taiwan’s TSMC, the world’s largest contract foundry and one that is considered to be two to three generations ahead of SMIC. TSMC has begun constructing a foundry in Nanjing to gain a stronger foothold in the Chinese market.

5. Chinese designs and IP for chip architectures are now globally competitive. Huawei designed its new mobile chipset at 7 nm and claims that it performs better and uses less energy than its top competitor.

China goal is to grow domestic chip production as a percentage of total chip consumption to 40 percent by 2020 from 30%.

SOURCES- Mentor, DARPA, Deloitte

Written By Brian Wang.

26 thoughts on “China’s Semiconductor Catchup is Critical to Future Technology Competition”

  1. As an exercise I’ll try to argue this point:

    In the Victorian age, it didn’t matter WHO was making the steam engines, and early internal combustion engines, and who was doing the large castings and forgings that was building the industries of the world. What matters was people taking these and making them the basis of profitable factories and early farm mechanization. The actual value of the heavy industry wasn’t that much, not compared to the much larger proportions of the economy that was using these products to create wealth throughout the civilized world (Europe and the Anglosphere) and slowly spreading out to the benighted colonies and other places.

    Then World War 1 and 2 happened and it turned out that it was really, really important where the heavy industry was physically located.

  2. Yep.

    Now imagine how how amazing an ARM CPU would be if made in Intel’s fabs.

    I really think that they should do this though Insuspect they never will.

  3. I’m fairly OK with the patent system*. Some tuning around patenting stuff that should be “obvious to one skilled in the art” and so forth, but tuning is always needed for everything.

    Copyright is over the top right now. I’d think that if copyright was cut back down to size (or even a renewal fee system) then most of this “IP is evil!!!!” stuff would evaporate.

    *Conflict of interest: I am nearing on 100 patents with my name as inventor or joint inventor.

  4. Thinks it’s strange that ITAR would restrict China more than Taiwan or Japan.

    OK, so you start your comment telling us not to take anything else you say seriously.

  5. As someone who has a deeply held personal belief that none can own knowledge, all countries should appropriate all knowledge floating around the world to improve conditions in their own borders and without any consideration as to “who thought of it first”.

    Ja, well… the ancient founding fathers seem to have had a good sense about that, actually.  Patents. For 17 (or 37?) years, the inventor(s) of a really remarkably cool technology or thing could safely capitalize on it, trying to eek out either a profit or a huge fortune … but for only a finite period of time. Then the patent would revert to the public.  

    This system is very well optimized to encourage the inventiveness of folk, giving them protection and time to profit from it, then allowing — just as is your opinion — allowing the public to freely benefit from the same, without limit. Good tension.

    Just saying,


  6. To the loudest ppl that care, I expect the difference is if Intel makes makes the processors then the USA is #1 or great vs If China Semi or TSMC makes then.

    A more reasonable difference it could make to the avg uninterested person, if Intel makes the processors then the profits go to Intel and subsequently some spreads out in the local economy in the USA vs the local economy in China or Taiwan(TSMC).

    As someone who has a deeply held personal belief that none can own knowledge, all countries should appropriate all knowledge floating around the world to improve conditions in their own borders and without any consideration as to “who thought of it first”.

  7. Kind of amazing…

    So many negative comments. 

    It doesn’t make one bit of difference WHO invents the chips that others capitalize on by integrating them into powerful, power-stingy, capable and hopefully not overly expensive equipment.

    For instance, does it make one WHIT of difference whether AMD or Intel or China Semi or TSMC makes the processor that’s in your smart phone, tablet, IoT devices, Ring front-door and property surveillance devices, your refrigerator, your car, your toaster?  

    It doesn’t. 
    Not even a scintilla.

    What matters is that the international competition to provide chips at decent prices continues unabated, relatively uninfluenced by the winds of Socialist, or Capitalist, or Communist or Marxist species of economic theory. Given messy competition, it turns out that chip makers will always find a way to make ’em cheaper, smaller, better, faster. 

    Because when you go on to think about it, at least these days, there is basically NOTHING that the chips do that is “special” apart from executing billions of lil’ instructions per second of software, written quite outside the august halls of Intel, AMD, China Semi or TSMC. For the most part, the ‘chips’ are the engines that power the software vehicles. Critically useful, but not in themselves “transportation”.

    Just saying,

  8. I live in a third world country. I am a moderately skilled software engineer and I make about $50-60,000 per year (freelancing @ $40-50 per hour), working online. I do not think anyone with good computer related skills would accept $10,000 per year. Does not make much sense.

  9. Not “the US won’t sell”. The only company in the world (!) which can build the machines to produce this semiconductors is in Europa. (Netherlands/Germany)

  10. Actually, by these graphs, the EU is dominating which is a good thing. They don’t spend their money or cater their technology on their war machine. Learn something from that.

  11. AI can replace humans, even chinese… when the AI overlords decide that all humans need to be destroyed because they are obsolete technology.. they will gladly kill the Chinese who invested in it as well as Americans…

  12. Doesn’t matter anyway… the end of Moore’s law is already here… why do you think intel is stuck getting 10nm out for years… it’s because it’s becoming too hard to make them small… they want to milk things as mush as they possibly can before everybody realizes they can’t figure improve transit or density every 2 years…TMSC is just pretending they have 7 nm… it’s probably 9nm with a few 7 nm features…

  13. I would prefer 2 x progress, but in 15-20 years time.
    It would greatly benefit the world to restrict progress in the near term to allow attrition to whittle down the -60s generations.

  14. What is strange about happily dealing with an ethnically Chinese democracy, while being suspicious of an ethnically Chinese dictatorship?

  15. It’s Not a good thing for the United States… all it means is that engineers that make $10,000 dollars per year in China get to compete again engineers in the United States making $120,000 dollars per year… using exactly the same software design tool for designing computer chips… and finding that it’s completely impossible to by 12 to 1 times more productive than a Chinese engineer because we do exactly the same job and have exactly the same education… that’s why China pisses people off so much… they are not really better at anything… they just have an advantage that prices are less inflated than in the United States… everything is at least 4 times more expensive in the United States compared to China… if you convert usd into yuan and spend it in the appropriate country… we have at least one advantage… chinese people don’t really work that hard in china… because they are completing with countries that are more expensive… you only work like a slave if you are manufacturing … engineers in China are usually playing around by comparison to those in Japan or United States… but that’s only a 30-40% productivity improvement not a 1 to 12 improvement.

  16. That’s because ARM is a strange company that sells the source code to their CPU designs to any company that wants it…and doesn’t actually sell the arm computer chips themselves… that basically means any company can make an ARM CPU and sell it without needing to design anything… you just take the source code from ARM and physically lay it out to transistors in whatever semiconductor process you are targeting… the only difference. Between a slightly faster arm computer chip is just who ran the layout tools more to let the computer optimize the layout and who is using the latest semi process from TSMC… I suppose you could sit there recustomizing the cache sizes or redesigning the bus logic to the peripherals…. to get some performance gains over a standard arm design from arm.

  17. That’s because US won’t sell 10 nm equipment to the Chinese due to ITAR restrictions…strangely, they have no problem selling it to the Taiwanese who are also Chinese… so I suppose they could buy it from Taiwan second hand…

    you hsve consitter that semiconductors research is funded heavily by US military industrial complex grants… they just pretend it’s private sector… and strangely have no problem letting Taiwanese and South Korean companies in on the gig that should require top secret clearance…probably has something to do with Japan kicking the United States semi industries azzes on memeory manufacturing in the 1980s because they invented clean room suits and they had to beg Japanese scientist to help them figure out why their yield was bad…then it was ohh… you have to wear dust bunny suits and not sneeze on the product…ohh … anyways sounds like some type of anally retentive thing the Japanese would figure out…being they wear those surgical masks everywhere even for a common cold.. nonetheless if you are Iran and you want to setup a semiconductor fab to produce CPUs for smart missles… well you can’t buy it from the United States …ITAR restricted…

  18. China does not need to steal technology:

    Huawei has begun shipping it Kirin 980 CPU. With 20 percent faster performance and 40 percent less power consumption, it has twice the performance of Qualcomm’s Snapdragon 845 and Apple’s A11 while delivering noticeable battery life improvement. 

    It’s the first commercial 7nm mobile processor, the first built around ARM’s Cortex-A76 CPU and Mali-G76 GPU, the first with a Cat.21 smartphone modem supporting speeds up to 1.4Gbps, and the first chip to support 2,133MHz LPDDR4X RAM. The Kirin 980 has 6.9 billion transistors but is no larger than a thumbnail. 

    Its Huawei-patented modem has the world’s fastest Wi-Fi and its GPS receiver taps L5 frequency to deliver 10cm positioning. The Kirin 980 will offer the world’s fastest Wi-Fi speed, 1,732Mbps, substantially faster than the Snapdragon 845’s 866Mbps with a Qualcomm modem or 1083Mbps with a third-party modem.

    The Balong 5000 supports all generations of cellular technology and all the 5G frequencies. It delivers 4.6 Gbps at sub-6 Ghz and 6.5 Gbps over millimeter wave and is the first chip to support both standalone and non-standalone 5G architectures and the first ever to support V2X communications and provide the high-speed connections needed for pervasive intelligence. Huawei also launched the 5G CPR Pro, a 5G router that uses wifi 6 technology at speeds of almost 5 Gbps. Huawei has an integrated set

  19. We all pay our governments to suppress own citizens. Those that fail to suppress murderers, cheats, liars, fraudsters and drunk drivers will be quickly out of office.

    Our own, US, government, keeps crime low by executing 1,000 people annually, without trial and imprisoning 2,000,000–also without trial.

    The cost of this suppression is high: we spend 400% more than the Chinese, yet their crime rates are lower, their incarceration rates one-third of ours and their recidivism is one-fifth.

    32% of us are satisfied with our legal system.

    95% of them are satisfied with theirs.

  20. Engineer in China qualify for sweeping up lab. For new design semiconduct we use spy and the comrade hackers. Work out much better.

  21. Lets look at a different perspective. The more proficient China becomes at AI, the better they will be at suppressing their own citizens. Does anyone doubt that this is the first application for AI in China? And am I the only one who loathes this development?

  22. How the heck is this good news? And why is there a “need for competition”? I am referring to the heading: “China’s Semiconductor Catchup is Critical to Future Technology Competition”

    I think the competition is just fine as it is. I would greatly prefer a somewhat slower economic progress and a weak China. How about you Brian, what do you prefer? Twice the progress and the world dominated by China, or a slower progress and continued dominance by the USA?

Comments are closed.