What's new

Opinionated - China Chipping Away to Semiconductor Dominance

China's Leading Memory Chip Players to Start Trial Production This Year
Wang Zhen
Date: Fri, 04/20/2018 - 14:38 / source:Yicai

2.4%E4%B8%AD%E5%9B%BD%E4%B8%89%E5%A4%A7%E5%AD%98%E5%82%A8%E8%8A%AF%E7%89%87%E4%BC%81%E4%B8%9A%E9%A2%84%E8%AE%A1%E4%BB%8A%E5%B9%B4%E4%B8%8B%E5%8D%8A%E5%B9%B4%E9%9B%86%E4%B8%AD%E8%AF%95%E4%BA%A7.jpg

China's Leading Memory Chip Players to Start Trial Production This Year

(Yicai Global) April 20 -- China's three leading companies in the field of memory chips will start the intensive trial-production in the second half of this year, ahead of an expected wide-scale roll out next year, says a new research report.

Yangtze Memory Technology Co., Hefei Changxin Integrated Circuit Manufacture Co., and Fujian Jinhua Integrated Circuit Co. are expected to spearhead the development of the country’s memory chip sector, market intelligence provider TrendForce said in a new report.

China’s chip industry has garnered much media attention in recent days following escalating trade frictions between China and the US, culminating in Chinese telecoms equipment maker ZTE Corp. receiving a seven-year ban from purchasing components and services from US suppliers.

Wuhan-based Yangtze Memory has been focusing on NAND Flash memory technologies and completed the first stage of its workshop construction last September. Trial-production is expected to start in the fourth quarter of this year with an initial monthly output of up to 10,000 chips.

Hefei Changxin and Fujian Jinhua have exerted efforts to develop Dynamic Random Access Memory (DRAM) chips, and the pair is expected to start trial-production and mass-production in the third quarter of 2018 and the first half of 2019, respectively.

China's memory chip industry will find it difficult to challenge the existing order of the global market in the short-term due to the relatively small scale of mass-production capacity in the initial stage. Chinese companies will face more challenges such as patent disputes, compared with the international memory chip giants, said TrendForce.

In the long term, as China's memory chip industry gradually matures, the two DRAM manufacturers are expected to ramp up to full capacity in 2020 or 2021, with total monthly output expected to reach 250,000 chips. Yangtze Memory plans to build three NAND Flash workshops in the long-term, with a total output of up to 300,000 chips per month.

If you are spending $250 billion to import this goddamn commodity, you should have already started supporting and subsidizing domestic industries. The US government has extended a hand of help to force China to finally act, but, they may not be as magnanimous in the future. China needs a complete re-check of its critical dependencies and find legal loopholes to overcome WTO and cover those dependencies through domestic effort.
 
Alibaba buys Chip Manufacturer C-SKY Microsystems
Source: Xinhua| 2018-04-20 13:49:06|Editor: Chengcheng



HANGZHOU, April 20 (Xinhua) -- Alibaba Group Holding has acquired Hangzhou C-SKY Microsystems Co., Ltd., an integrated circuit design house, in a bid to increase its own chip-making capability, Alibaba said Friday.

Founded in 2001, Hangzhou-based C-SKY Microsystems develops embedded CPU and chip architecture. The company said it is the only embedded CPU volume provider in China with its own instruction set architecture.

"The acquisition is an important step for Alibaba's chip development," said Zhang Jianfeng, chief technology officer of Alibaba.

Alibaba said the purchase of C-SKY Microsystems will help unify the two companies' R&D capability amid China's campaign to gain self-reliance in key technology.

Alibaba has previously invested in five chip manufacturers, including U.S. AI chip designer Kneron and Barefoot Networks.
 
Horizon Robotics Acquiring Another Major Semiconductor Company, Says CEO Yu Kai
April 23rd, 2018, 8:00 AM EST

Horizon Robotics, China's leading AI chip startup, will complete raising several hundred millions of U.S. dollars in May, attracting another global semiconductor company after Intel Corp, and some Chinese and foreign car-makers, to become its new strategic investors. Yu Kai, founder and CEO of the company, told Bloomberg that he is aiming to equip some 30 million autonomous vehicles in China with the company's chips and systems by 2025. (Source: Bloomberg)
https://www.bloomberg.com/news/vide...ing-another-major-semiconductor-company-video

-----------------

menafn.com
Chinese tech giants speed up chip development
MENAFN
2 minutes
(MENAFN - Asia Times) Amid the United States chip embargo on Chinese phone-maker ZTE, the nation's technology giants have turned their focus to the rapid development of chip-related businesses, National Business Daily reported.

Alibaba said it is planning to buy Hangzhou C-SKY Microsystems, which is said to be the only self-developed embedded CPU IP core company as a whole.

Zhang Jianfeng, CTO of Alibaba, thinks IP Core is the mainstay of basic chip capabilities, while entering the IP Core area will lay a foundation for the Chinese chip sector to be self-dependent in the future.

It is worth noting that Alibaba has invested in a number of chip companies such as the Cambricon, Barefoot Networks, DeePhi Tech, Kneron and ASR.

Meanwhile, domestic tech giants including Tencent, iFlytek, and Xiaomi have also entered the cloud chip field in the form of investment or self-research, such as Huawei's semiconductor subsidiary HiSilicon and Baidu's XPU.

Evergrande, a real estate giant, also announced its involvement in the integrated circuit industry.

-
 
Horizon Robotics Acquiring Another Major Semiconductor Company, Says CEO Yu Kai
April 23rd, 2018, 8:00 AM EST

Horizon Robotics, China's leading AI chip startup, will complete raising several hundred millions of U.S. dollars in May, attracting another global semiconductor company after Intel Corp, and some Chinese and foreign car-makers, to become its new strategic investors. Yu Kai, founder and CEO of the company, told Bloomberg that he is aiming to equip some 30 million autonomous vehicles in China with the company's chips and systems by 2025. (Source: Bloomberg)
https://www.bloomberg.com/news/vide...ing-another-major-semiconductor-company-video

-----------------

menafn.com
Chinese tech giants speed up chip development
MENAFN
2 minutes
(MENAFN - Asia Times) Amid the United States chip embargo on Chinese phone-maker ZTE, the nation's technology giants have turned their focus to the rapid development of chip-related businesses, National Business Daily reported.

Alibaba said it is planning to buy Hangzhou C-SKY Microsystems, which is said to be the only self-developed embedded CPU IP core company as a whole.

Zhang Jianfeng, CTO of Alibaba, thinks IP Core is the mainstay of basic chip capabilities, while entering the IP Core area will lay a foundation for the Chinese chip sector to be self-dependent in the future.

It is worth noting that Alibaba has invested in a number of chip companies such as the Cambricon, Barefoot Networks, DeePhi Tech, Kneron and ASR.

Meanwhile, domestic tech giants including Tencent, iFlytek, and Xiaomi have also entered the cloud chip field in the form of investment or self-research, such as Huawei's semiconductor subsidiary HiSilicon and Baidu's XPU.

Evergrande, a real estate giant, also announced its involvement in the integrated circuit industry.

-

Lots of major companies investing, especially investing in semiconductors for AI and IoT applications. That's one area China can achieve meaningful presence and sovereignty.
 
Where is Lenovo and ZTE?

Why aren't they interested with processor chip, while their products really dependent on it
 
Chinese chip sector ‘may overtake Western peers’
By Li Xuanmin Source:Global Times Published: 2018/4/22 19:48:39

New uses, sectors can boost nation’s efforts, market position

07189bc9-de88-4079-860c-4c0e9134ca50.jpeg
An employee works at an encapsulation production line of Tongfu Microelectronics Co in Nantong, East China's Jiangsu Province in February 2018. Photo: VCG

Chinese technology companies are investing heavily in all aspects of the chip-making sector, giving them a chance to surpass their Western peers in "new emerging tracks" such as artificial intelligence (AI) and the Internet of Things (IoT), industry insiders and company executives said over the weekend.

There has been widespread discussion in recent days of the need for China to vigorously develop its own semiconductor sector, after the US government banned the sale of US chips and other components to Chinese telecom company ZTE Corp.

Domestic technology giant Alibaba Group announced over the weekend that it had acquired Hangzhou-based embedded chip maker C-Sky Microsystems, an important move in its chip-making efforts, according to a statement Alibaba sent to the Global Times on Sunday. Terms of the deal were not disclosed.

Alibaba has previously invested in network chips, smartphone chips and IoT chips, involving five companies: Cambricon, Barefoot Networks, Deephi Tech, Kneron and ASR Microelectronic, the statement noted. The group also set up a professional team to conduct research and development of AI chips.

"The headwinds we now face can only be tackled by scaling up investment in research and development. But there is something we can change, and that is the future," an industry insider told the Global Times over the weekend, speaking on condition of anonymity.

For China, making a breakthrough in the chip industry will be difficult now, he added, citing foreign dominance of semiconductor research and production as well as trade rules developed over the past decades.

Zhang Jianfeng, Alibaba's chief technology officer, noted that in some new industries such as AI and the IoT, domestic chip-makers are at the same level as their US rivals.

"Technologies in chips made around the globe have not yet matured to support the new industries, and that's where the chance for Chinese companies lies," he said.

Within several years, demand for new types of chips will surge, Zhang said. He forecast that 20 billion to 30 billion smart devices will require new chips, a market that will be "hundreds of times" larger than the current one.

"Producers with enough data and application scenarios will have an edge in chips' core competiveness," he said.

Chinese companies' gaps with their Western peers are shrinking thanks to the injection of capital into the integrated circuit sector and government support policies, Chen Feng, vice president of Chinese fabless semiconductor maker Rockchip, told the Global Times on Sunday.

"An increasing number of self-supplied chips are now being installed in cellphones and high-end electronic devices made in China," Chen said.

Diao Shijing, director of information technology at the Ministry of Industry and Information Technology, said on Saturday that China's chip industry has made "significant progress and is now catching up with the technology of the world's first-echelon companies, particularly in design capacity," the China Media Group reported.

The market is also expanding rapidly, with domestic chips being adopted from daily life and the industrial sector to AI and smart cars, Diao noted.

From 2013 to 2017, the industry achieved an average annual growth rate of 21 percent, five times that of the global market, according to Diao.
 
Market Should Be Decisive in Developing China’s Semiconductor Chip Industry
Wu Danqing

Date: Mon, 04/23/2018 - 15:42 / source:Yicai
3.3%E7%A0%94%E5%8F%91%E4%B8%AD%E5%9B%BD%E2%80%9C%E8%8A%AF%E2%80%9D%E7%9C%9F%E6%AD%A3%E6%9C%89%E7%94%A8%E7%9A%84%E6%98%AF%E5%B8%82%E5%9C%BA%E5%8C%96%E5%8A%9B%E9%87%8F.jpg


Market Should Be Decisive in Developing China’s Semiconductor Chip Industry
[Yicai Global is committed to providing an open forum to air a diverse range of views. The opinions expressed herein are the author’s alone. Yicai Global has redacted this article to conform to our style]

(Yicai Global) April 23 -- China has developed atomic and hydrogen bombs and man-made satellites, but why does it still not have its own high-end integrated circuit products?

Chinese companies have become global leaders in many advanced technologies, but why does the country still rely heavily on imports of high-performance engines?

The US commerce department banned ZTE Corp. from buying American components April 16, and the policy triggered heated debate over semiconductor innovation in China. Chinese manufacturers are with maximizing profits and failed to pay due heed to technological research and development. Tens of billions of yuan have been invested in bike-sharing, food delivery and group buying startups, but very few investors seem interested in high-tech businesses. The dilemma facing ZTE today results from the company’s long neglect of basic research and technical innovation.

I have worked in the IC industry for 16 years, and I do not agree with them on this matter. The chip industry is where first movers have significant advantages over new entrants, so Chinese companies cannot rely purely on investment -- financing from funds, venture capital firms or private investors -- if they are ever to catch up with the US semiconductor giants. China needs to work out its own ‘Manhattan Project’ to leapfrog its Western rivals.

Late-mover disadvantage of Chinese chipmakers

Chinese companies have made considerable headway in the telecommunication field in recent years. ZTE and Huawei Technologies Co. have developed very mature products and have been actively involved in setting international industry standards. However, both still depend on dominant US companies for core chipsets, a circumstance dictated by the current global division of labor and the unique features of the semiconductor market.

US manufacturers cannot beat Chinese competitors in the system integration business, so they have gradually abandoned the equipment market and instead diverted more resources to upstream chip development. This explains why most Chinese firms focus on equipment manufacturing and system integration, while their American counterparts specialize in core component development, giving the latter significant first-mover advantage in the chipmaking scene.

For the sake of clarity, let us consider the ‘battle’ between the world’s two major chipmakers, Intel Corp. and Advanced Micro Device Inc. in the 1990s. Intel developed the 486 chipset first and sold the product at a relatively high price. By the time AMD released its 486 chips about six months later, Intel had recouped its initial investment and made a fortune on the product, and it decided to mark the product down to the cost price, undercutting its main rival. It then reinvested the profits in the development of new-generation products. AMD, as the late mover, invested heavily in 486 chip research and development, but its profits were nowhere near that of Intel’s. The number of transistors in a dense integrated circuit doubles and the production price halves about every two years. American semiconductor companies applied this law to consolidate their competitive advantages and profit from it, and thus far it has been a virtuous cycle.

Like AMD, Chinese chipmakers are at a disadvantage relative to first movers. Every time they developed a new product with substantial investment of money and other resources, the US companies lowered their prices and kick off new product development. As a result, it was very hard, if not impossible, for new entrants to make money at all, considering the prohibitive investment involved. No company can sustain such losses forever.

Can Chinese chip companies achieve leapfrog development? Chipmaking is a capital- and tech-intensive business with three salient characteristics. First, technical accumulation is of paramount importance, meaning that it is almost impossible to skip a specific product and develop a newer generation of chips directly. European and American companies started semiconductor R&D in the 1960s and 70s, so they are far ahead in this respect. Second, leapfrogging in R&D necessarily involves higher spending on human resources. Third, it is impossible for an information technology firm to reduce R&D costs unless it manages to reach a certain production scale. Given the relatively limited number of users, Chinese chipsets have many defects and vulnerabilities that are difficult to find (and therefore solve), making them less competitive in terms of both product performance and pricing.

In short, most Chinese chipmakers have been operating at a loss, as dictated by the first-mover advantage enjoyed by Western companies. In my opinion, they cannot achieve leapfrogging development relying on government subsidies alone.

Chinese chipmakers should not count on government subsidies

The Chinese government has been subsidizing the semiconductor industry for many years. A dedicated R&D fund was established in the 1990s to support home-grown brands such as Shanghai Huahong Integrated Circuit Co., Semiconductor Manufacturing International Corp. and Hongli Clean Energy Technologies Corp.. Later, the government invested an aggregate of more than CNY10 billion (USD1.59 billion) in a special fund to promote core electronic device, high-end universal chip and basic software development, but the actual results were less than expected.

With a change of strategies, China started to purchase foreign companies to acquire existing semiconductor technologies, but the plan was soon detected by foreign governments. The US blocked a series of takeover deals proposed by Chinese companies. Their attempts to acquire core communication, storage, processor, sensor, radio frequency or high-end simulation technologies were unsuccessful, despite the generous offerings. Both approaches have proved non-viable.

As to why very few private investors are interested in high-tech businesses, I think all investors are after profit, and nothing can be said against this. Private investors are keen on internet startups such like food delivery and cab-hailing services, because they can make quick money from these businesses. As an emerging industry Internet + has limitless potential for development, which is something that appeals to all venture capitalists. It would be unrealistic to expect a reasonable investor to invest in such mature markets as chips. IC product development entails prohibitive financial input, but the commercial gains can be insignificant or even negative, so we should not count on VC companies to fund chip R&D activities.

Government subsidies and investment alone cannot close the gap between Chinese and Western semiconductor products. That said, we may have seen the first glimpse of light at the end of the tunnel with the recent rapid development of several Chinese chipmakers. HiSilicon Technologies Co., Spreadtrum Communications (Shanghai) Co., RDA Microelectronics, Inc. and Goodix now rank among top-tier global semiconductor companies. HiSilicon’s Kirin series, above all, are as good as similar high-end Western products in performance as well as price. It is simply a miracle! The price paid by Huawei is that it invested billions of yuan in product development and sustained heavy losses for nine years. The company subsidized the loss-making chip business using earnings from its communications divisions. Another important reason is that Huawei is not a public company; otherwise it could not possibly have weathered the massive losses.

China is increasingly aware of the strategic importance of the chipmaking industry and local governments are showing renewed enthusiasm in semiconductor products. It is said that hundreds of integrated circuit production lines have been built, or are being built, in the country. Having said that, I am personally skeptical as to whether these efforts will translate into a breakthrough in any core technologies.

The chipmaking industry chain is quite long, encompassing raw materials and equipment manufacturing (upstream), processing (middle-stream) and chip design (downstream). Most production lines are still used for contract manufacturing. The business model is based on extensive inputs (or factors of production) -- a legacy from the industrial age -- and is therefore irrelevant to our discussions here. We already have very strong contract manufacturers like SMIC and Huahong, so I believe most new production lines represent repetitive investment in low-end production, which can easily lead to serious waste of low-end production capacity.

The semiconductor industry is capital- and tech-intensive. Over the decades, Europe and the US have created an almost perfect market system, so the Chinese government should devise its own ‘Manhattan Project’ through effective planning, focusing our limited financial resources on strengthening the weakest links.

China should encourage contributions by private investors

I believe that China’s semiconductor industry has the necessary conditions to overcome the challenges. First, we have an enormous domestic market. China is already a major player in the equipment manufacturing space. Second, we have more labor resources relative to Western economies. During the early years of economic reforms, our country relied on migrant workers to boost the export processing business, but we have passed the turning point now. However, the migrant worker ‘dividend’ has been replaced by the engineer dividend. China produces millions of science and engineering graduates every year, and many foreign multinationals have trained many qualified Chinese technical professionals. Third, we have enough funding. Chinese companies have all the necessary financial resources, but the real question is how to effectively leverage these resources to deliver the best results.

In the light of the unique characteristics of the chipmaking industry and the Chinese status quo, I would like to propose the following strategy for catalyzing the domestic semiconductor industry.

Firstly, the government should stick to the market economy model and engender enthusiasm about the chipmaking industry among private investors, aiming to integrate policy support and private investment into a synergistic force. Ideally, research institute subsidies should be replaced by user-oriented subsidization -- that is, subsidies can be offered for downstream equipment manufacturers for every domestically-produced chipset they purchase, for two reasons: the prohibitive R&D expenditure associated with semiconductor products dictate that production costs cannot be lowered unless the products are installed and used in a significant amount of telecommunication systems; and (ii) a new chipset is not fully developed until it has been thoroughly adapted to various equipment. Additionally, such an approach is advisable because the subsidies are not allocated to any specific companies, and equipment manufacturers choose chip suppliers themselves. Chipmakers must compete against each other to be eligible for subsidies, so government funds are used to finance R&D of the most competitive products, without tampering with normal market competition.

Secondly, priority should go to startups to lower chip development and production costs. What China really needs now is not low-end contract manufacturing, but rather producers of core high-end chipsets. Therefore, instead of investing in dozens of production companies, the government should focus its resources on a small number of innovative startups. It can set up a fund for startups specializing in high-end semiconductor products, subsidizing their business development through tax incentives. Private investors must be mobilized to contribute to the formation of a virtuous cycle, because sustainable profits are the key to long-term development of the domestic chipmaking industry.

Lastly, human resources acquisition is very important. Some of the best Chinese engineers are still working in Europe and the US, so well-targeted policies should be introduced to attract talented overseas Chinese professionals. This is critical especially because human capital is what matters the most for the development of semiconductors, software and all other high-tech businesses.

It has been said that the supply ban the US government slapped on ZTE is catastrophic for the Chinese telecommunication industry, but some people argue that it opens new opportunities for Chinese chipmakers. In my opinion, we should not be too pessimistic about the future of our semiconductor industry. What happened to ZTE poses both challenges and opportunities. On the one hand, it taught us a bitter lesson about the consequences of excessive reliance on foreign technology. On the other hand, it toughened our determination to grow away from the real-estate-based growth model toward technology and innovation driven development. The transition can be a very painful process, but it is the only way out. We must believe in the market economy system as the main driving force for the development of a strong domestic semiconductor industry.

(The author is a senior R&D director from a leading international semiconductor company)
https://www.yicaiglobal.com/news/ma...eveloping-china’s-semiconductor-chip-industry
 
Chinese companies have become global leaders in many advanced technologies, but why does the country still rely heavily on imports of high-performance engines?

I guess the reason is partly a blind faith on free market and interdependency. But, even if market is supposedly rational and profit seeking, there is always larger strategic concerns and calculations by the state.

Thus, when the US regime says jump, those might multinationals have to ask "how high?"

That's a good practical lesson to free market fundamentalists and laizzes faire believers.
 

lvGyuD3.jpg


Horizon Robotics Exerts Tight Grip Over Artificial Intelligence Stack


James Morra | Apr 23, 2018

As the race intensifies to run machine learning tasks in embedded devices instead of the cloud, several companies are trying to set themselves up with custom chips to ease the shift. Horizon Robotics is not only tackling chips but also software and the cloud, with an eye toward beating rivals in applications like security cameras and autonomous cars.

“The chip is the local brain that directly senses the surrounding environment, while the algorithm is the miner of the data,” said Kai Yu, founder and chief executive of Horizon Robotics, and the former head of Baidu’s artificial intelligence unit, called the Institute of Deep Learning, in an interview with Electronic Design.

“We want to empower end devices with A.I. capacity and make them smart without relying on the cloud alone,” Yu said, adding that the “chip and algorithms are used to perceive and filter big data, perform real-time processing and transmit valuable data to the cloud for further mining and modeling. Each component works together.”

The central component in Horizon Robotics’ SoCs is the brain processing unit, a custom block of circuitry that specializes in algorithms trained on vast libraries of images, hundreds of hours of video, or other data. The silicon slab can also be slipped into chips like field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs), Yu said.

The Beijing, China-based company is aiming to plant its chips into tens of millions of smart cameras over the next two years. Developing and deploying the technology so quickly led it to raise $100 million in venture capital last year from investors including Sequoia Capital, state-owned China Jianyin Investment, Harvest Fund Management and Intel Capital.

One of Horizon Robotics' chips, Sunrise, runs facial recognition or other inference algorithms with up to a trillion operations per second. The company recently released a security camera using it at the International Security Technology Show in Las Vegas. The camera can identify and follow 200 objects within each frame of video, recognizing the face of customers in a clothing store, for example, or plucking a criminal suspect from a crowded sidewalk.

With Sunrise, based on 40-nanometer technology, the camera runs at 30 frames per second while consuming 1.5 watts. With support for 50,000 different faces and 99.7 percent accuracy, the system can avoid the latency and bandwidth issues introduced by steaming data to servers in the cloud, where training and inference typically occur.

While Horizon Robotics has the upper hand over American rivals in China, the challenge is in keeping its hardware and software on the same page. While software engineers can change lines of code relatively fast, chip designers need several months and millions of dollars to prototype chips and get them back from the foundry. “It’s been a priority since the get-go,” Yu said.

He added that four-fifths of the company’s more than 300 employees have research and development backgrounds. Other founders of Horizon Robotics include Chang Huang, a founder of Baidu’s A.I. business unit, and Ming Yang, a founder of Facebook’s A.I. research team. Feng Zhou, a former principle chip architect for Huawei’s HiSilicon business, leads chip development.

In many ways, the company’s pincer attack on machine learning mirrors China’s national strategy. Not only has the country pledged $150 billion to close the technology gap between American and Chinese chip suppliers – and reduce its roughly $275 billion in annual chip imports – but it is also pushing to become the world leader in artificial intelligence by 2030.

The machine learning movement could level the playing field for China’s chips, particularly as the focus shifts to custom over commodity products. Horizon Robotics uses 40nm technology that entered production almost a decade ago, but it is jumping into embedded inference with companies like Qualcomm, which has started sampling 10nm chips that can be installed in networks of security cameras.

China’s ambitions include the deployment of 30 million autonomous cars within the next decade. Horizon Robotics is also trying to tap into the momentum behind that market. The company has partnered with Robert Bosch and Ford’s Chinese partner Chongqing Changan Automobile, among others, to put its automotive camera processor, called Journey, through its paces.

Horizon Robotics designed Journey to spot pedestrians, lane markings and other vehicles on the road and help driverless cars avert accidents. The chip may ultimately compete with systems from Santa Clara, California-based Nvidia and Intel’s Mobileye business. Bloomberg reported that the company is looking to have autonomous test cars on Chinese roads by 2019.

“China right now is a huge market for a lot of innovation happening in machine learning, from city management and security to autonomous driving,” said Yu, a member of the country’s strategic artificial intelligence advisory board. “China is a great playground for us to develop and mature these products for the rest of the world.”


Code:
http://www.electronicdesign.com/embedded-revolution/horizon-robotics-exerts-tight-grip-over-artificial-intelligence-stack


--------------------


shine.cn
Investors eye chip investment after US imposes sanctions on ZTE
Zhu Shenshen
2 minutes

The chip sector has become the hottest investment target in China after the US imposed sanctions against ZTE, Shanghai Daily learned during an investment forum held in the city today.

The integration between chip and artificial intelligence and smart manufacturing also have become hot spots, investors told the three-day Chinaventure Investment Conference Annual Summit in Shanghai, which ends tomorrow.

"The ZTE case is a catalyst to push up chip investment. It may be the best time and opportunity for Chinese chip firms," Wu Yenan, a partner at Rising Investments, told the forum.

Chinese funds can still find unique opportunities in automotive electronics and related software, said Hao Dan, chairman of HTHS Capital, a Shenzhen-based fund managing 20 billion yuan (US$3.2 billion) of assets.

The US action against China's leading telecom equipment maker ZTE has raised wide-range concerns. The Chinese government and industry officials have called for urgent action to develop China's own chip technologies.

On Friday, Alibaba said it would acquire Hangzhou C-SKY Microsystems Co, an integrated circuit design house, to increase its own chip-making capability.

Rising Investments, a fund focusing on chip investment, said it will expand its investment portfolio to new segments like RF (radio frequency) chip this year.

-
 
First Chinese Terahertz Imaging Chip Debuts, Can Replace Imports
LIAO SHUMIN
DATE: TUE, 04/24/2018 - 13:48 / SOURCE:YICAI

(Yicai Global) April 24 -- The first made-in-China terahertz body-scanner imaging chip has been officially released, state Xinhua News Agency reported yesterday.

The wave signal radiated by the human body is so weak that the rice grain-sized terahertz chip must have the characteristics of ultra-high sensitivity, ultra-low noise, and ultra-wideband to detect the faint fleeting signal the body’s radiation emits and image it. These electromagnetic pulses, lasting one millionth of a millionth of a second, are called terahertz waves.

The chip, unveiled at the First Digital China Summit held in Fuzhou in China’s southeastern Fujian province from April 22 ton 24, detects these emanations and analyzes them with its internal algorithm to image the human body, which helps security staff quickly detect hazardous concealed articles and poses no radiation risk to those undergoing security checks. It thus effectively resolves a security problem, explained Wang Qiang, deputy director of the 13th Research Institute of China Electronic Technology Group Corp., its developer.

Other countries previously controlled the core technology for imaging chips for security inspection devices. China has now achieved independent research and development capability and controllability in key links of the terahertz chip, such as material growth, manufacturing techniques, simulation modeling and circuit design.

The terahertz detector modules equipped with these chips can replace imported ones and have advantages in price and performance over similar foreign products. Their advent significantly boosts the level of China's terahertz body scanners, Wang said.
 
Last edited:
good news for china's ai chip companies..

i can see DJI will stop using intel's movidius chips in their future drones :D




---
asia.nikkei.com
China's Hikvision downplays US risks amid ZTE woes - Nikkei Asian Review
CHENG TING-FANG and LAULY LI, Nikkei staff writers


A full ban on American-made components similar to that facing ZTE would have only a limited impact on Hikvision, Huang said.

"In the surveillance industry, the components we need are much less than those needed for smartphones or telecom equipment. We think we should be OK if we cannot buy anything from the U.S.," she said. "We do see more and more domestic lens providers and sensor makers that could supply quality components so that we do not need to buy from foreign suppliers."

Hikvision buys chips from U.S. producer Nvidia to facilitate artificial intelligence features for video surveillance camera systems, Huang said, but she sees alternatives in solutions provided by smaller local chip startups. Her company also announced plans late last year to develop its own AI chips but it will still rely on external suppliers until the chip is available.

The Chinese company, which was founded in 2001 and listed in 2010, is regarded as a major entity in advancing the country's AI industry, which Beijing hopes will reach $150 billion by 2030. Many analysts said Hikvision's rapid rise in recent years also fits with China's push to surveil and control its 1.4 billion citizens.

Hikvision generated net profit of 9.41 billion yuan in 2017, up more than 26%, on a 31% increase in revenue. Sales for January-March 2018 jumped nearly 33% on the year to 9.36 billion yuan, while net profit advanced more than 22% to 1.81 billion yuan.

Analysts also see just a mild impact on Hikvision if it is barred from using U.S. components. Jay Huang, an analyst at Bernstein Research, said his agency expects only up to 4% of the company's revenue could be influenced by such a ban.

"The ZTE event may trigger Hikvision to accelerate its reprioritization and diversification of suppliers," the analyst said. Huawei chip arm Hisilicon Technologies and Chinese homegrown AI chip startups such as Cambricon Technologies and Horizon Robotics could all provide related solutions that Hikvision now buys from Intel's Movidius and Nvidia, he said.

full..
Code:
https://asia.nikkei.com/Business/Companies/China-s-Hikvision-downplays-US-risks-amid-ZTE-woes
 
we have real estate company makin ai chips. why not online news company do it too? lol.. who's next? didi? weibo? dji? noodle soup restaurant? :D

--
technode.com
The company behind Jinri Toutiao is creating its own AI chips · TechNode
Masha Borak
2-3 minutes

The recent US ban on selling components to China’s state-owned communication technology ZTE has become a boon for China’s chip makers. Bytedance—the company behind AI-powered news aggregation platform Jinri Toutiao—has set its eyes on developing its own AI chips.

“Bytedance has the largest number of users in the world whose videos need to be analyzed and processed and uploaded, and we are purchasing a large number of chips. At present, we are actively seeking breakthroughs in the chip-related field,” Vice-President of Bytedance Yang Zhenyuan told 36Kr (in Chinese). He did not, however, provide any specifics on the products being developed.

The news comes only a few days after Alibaba announced the creation of its neural network chip, the Ali-NPU, which will be used in AI applications for businesses through the Ali Cloud. Just one day after the announcement, Alibaba also revealed it will fully acquire local chipmaker C-SKY Microsystems.

Yang also told reporters that although the “ZTE event” reflects the weakness of China’s high-tech industries in key chips and core components, the country still has the opportunity to seek new breakthroughs in the emerging direction of AI chips. However, it will take a long time to reach international levels in the high-end chip field; the R&D cycle will be very long, Yang added.

In the short term, China has the ability to achieve a balanced situation through some areas of leadership (such as artificial intelligence, software ecology and other fields), which can lay the foundation for long-term independent research and development, Yang added. While the US technology giants Facebook and Google are independently developing AI chips, domestic internet giants are also laying out a map, he noted.

Bytedance is not just eyeing AI development, it is also preparing an aggressive expansion overseas. CEO of Bytedance Zhang Yiming revealed at the sixth anniversary of Toutiao that the main keyword for the company in 2018 will be globalization. Their goal is to have more than half of their users from overseas in the next three years.

----------------------------------------------------------------------

Qualcomm-backed startup announces AI processor family
March 07, 2018 // By Peter Clarke
2-3 minutes

Kneron said it has adopted a filter decomposition technology that allows the division of large-scale convolutional computing blocks into a number of smaller ones to compute in parallel. Together with the reconfigurable convolution accelerating technology, the computing results from the small blocks will be integrated to achieve better overall computing performance. Model compression allows unoptimized models to be shrunk a few dozen times. The multi-level caching technique reduces the use of CPU resources and further improves the overall operational efficiency.

Kneron NPU IP Series allows ResNet, YOLO and other deep learning networks to run on edge devices including hardware IP, compiler, and model compression. It supports various types of CNNs such as Resnet-18, Resnet-34, Vgg16, GoogleNet, and Lenet, as well as mainstream deep learning frameworks, including Caffe, Keras, and TensorFlow.

Albert Liu, Kneron´s founder and CEO said: "Since the release of its first NPU IP in 2016, Kneron has been making continuous efforts to optimize its NPU design and specifications for various industrial applications. We are pleased to introduce the new NPU IP Series and to announce that the KDP 500 will be adopted by our customer and enter to the mask tape-out process in the upcoming second quarter."

Kneron was founded in 2015 and completed a Series A financing round worth more than $10 million in November 2017. Alibaba Entrepreneurs Fund and CDIB Capital Group are the lead investors, and Himax Technologies, Qualcomm, Thundersoft, Sequoia Capital and Cy Zone are co-investors.

Code:
http://www.eenewsanalog.com/news/qualcomm-backed-startup-announces-ai-processor-family/page/0/1
 
Last edited:
中国电科38所发布“魂芯二号A”芯片:实际运算性能业界同类最强

2018-04-25 来源:安徽日报

关键字:芯片国产芯片魂芯中国电科38所魂芯二号A

据安徽日报报道,4月23日,中国电科38所在福州举行的首届数字中国建设峰会上发布了实际运算性能业界同类产品最强的数字信号处理器——“魂芯二号A”。该芯片由38所完全自主设计,在一秒钟内能完成千亿次浮点操作运算,单核性能超过当前国际市场上同类芯片性能4倍。

20180425095257826.jpg


资料图

高性能芯片被誉为 “工业粮草”,代表了一个国家信息技术水平。一直以来,我国在高性能数字信号处理器(DSP)方面始终依赖进口。 12年前,38所就开始进入数字信号处理器芯片领域。2012年,该所推出我国自主研发的首款实用型高性能浮点通用DSP芯片——“魂芯一号”,性能高于同期市场同类DSP芯片4~6倍,并成功应用在我国空警-500预警机雷达等多个国防科技装备上,成为我国首款广泛应用于国防科技装备的高端自主数字信号处理器。

“魂芯二号A”采用全自主体系架构,研发历时6年,突破了控制器设计等多个技术难题,获得国家技术发明专利、软件著作权等科技成果30余项;拥有当前业界性能最强的DSP核,实现了对国内外同类产品性能指标的超越。相对于“魂芯一号”,“魂芯二号A”性能提升了6倍,通过单核变多核、扩展运算部件、升级指令系统等手段,使器件性能千亿次浮点运算同时,具有相对良好的应用环境和调试手段;单核实现1024浮点FFT (快速傅里叶变换)运算仅需1.6微秒,运算效能比德州仪器公司TMS320C6678高3倍,实际性能为其1.7倍,器件数据吞吐率达每秒240Gb。

作为通用DSP处理器,“魂芯二号A”将广泛运用于雷达、电子对抗、通信、图像处理、医疗电子、工业机器人等高密集计算领域。目前,正在多种重大装备以及图像处理领域中推广使用。

中国电科首席科学家、“魂芯二号A”总设计师洪一介绍,“魂芯二号A”的推出,使得软件无线电从理想走向现实,芯片功能逐渐取决于软件算法的更新成为可能,为我国建立自主体系高端DSP产品谱系奠定坚实基础。

@Bussard Ramjet :coffee::enjoy:
 
中国电科38所发布“魂芯二号A”芯片:实际运算性能业界同类最强

2018-04-25 来源:安徽日报

关键字:芯片国产芯片魂芯中国电科38所魂芯二号A

据安徽日报报道,4月23日,中国电科38所在福州举行的首届数字中国建设峰会上发布了实际运算性能业界同类产品最强的数字信号处理器——“魂芯二号A”。该芯片由38所完全自主设计,在一秒钟内能完成千亿次浮点操作运算,单核性能超过当前国际市场上同类芯片性能4倍。

20180425095257826.jpg


资料图

高性能芯片被誉为 “工业粮草”,代表了一个国家信息技术水平。一直以来,我国在高性能数字信号处理器(DSP)方面始终依赖进口。 12年前,38所就开始进入数字信号处理器芯片领域。2012年,该所推出我国自主研发的首款实用型高性能浮点通用DSP芯片——“魂芯一号”,性能高于同期市场同类DSP芯片4~6倍,并成功应用在我国空警-500预警机雷达等多个国防科技装备上,成为我国首款广泛应用于国防科技装备的高端自主数字信号处理器。

“魂芯二号A”采用全自主体系架构,研发历时6年,突破了控制器设计等多个技术难题,获得国家技术发明专利、软件著作权等科技成果30余项;拥有当前业界性能最强的DSP核,实现了对国内外同类产品性能指标的超越。相对于“魂芯一号”,“魂芯二号A”性能提升了6倍,通过单核变多核、扩展运算部件、升级指令系统等手段,使器件性能千亿次浮点运算同时,具有相对良好的应用环境和调试手段;单核实现1024浮点FFT (快速傅里叶变换)运算仅需1.6微秒,运算效能比德州仪器公司TMS320C6678高3倍,实际性能为其1.7倍,器件数据吞吐率达每秒240Gb。

作为通用DSP处理器,“魂芯二号A”将广泛运用于雷达、电子对抗、通信、图像处理、医疗电子、工业机器人等高密集计算领域。目前,正在多种重大装备以及图像处理领域中推广使用。

中国电科首席科学家、“魂芯二号A”总设计师洪一介绍,“魂芯二号A”的推出,使得软件无线电从理想走向现实,芯片功能逐渐取决于软件算法的更新成为可能,为我国建立自主体系高端DSP产品谱系奠定坚实基础。

@Bussard Ramjet :coffee:
nice to to see they invest in dsp chip too and powerful one too. now huawei and zte dont have to buy from cadence or ceva :enjoy:
 

Country Latest Posts

Back
Top Bottom