Opinionated - China Chipping Away to Semiconductor Dominance

[TaiShang]

Foxconn Partners Jinan Authorities to Set Up USD546 Million Chip Fund
CHEN JUAN
DATE: MON, 10/08/2018 - 12:24

Foxconn Partners Jinan Authorities to Set Up USD546 Million Chip Fund​

(Yicai Global) Oct. 8 -- Foxconn Technology Group, the world’s biggest contract electronics maker, has agreed to set up a CNY3.75 billion (USD546 million) investment fund with the municipal government of Jinan in order to bolster its development in the semiconductor sector.

Most of the money will go toward Foxconn’s main chip project after the Taiwanese firm brings in one major semiconductor maker and five integrated circuit designers to Jinan, local news outlet Qilu Evening News reported on Sept. 28. It did not disclose how much each party will contribute to the fund.

Foxconn has been gradually extending into the chip sector. In August it penned a deal with the municipal government in Zhuhai, Guangdong province, to cooperate in semiconductor design services, and its affiliate Sharp has already begun making analog integrated circuits at its plant in Fukuyama, Japan.

The Jinan government hopes the deal will attract other semiconductor makers to the city and invest there, forming an industrial agglomeration and facilitating local development of the sector, the report added.

@Viet , as the cheap production leaves Mainland, quality ones come in. This is the natural cycle (or ideal one) for a developing country.

 

[qwerrty]

based on the number specs, this ascend 910 is 2x more powerful than baidu's kunlun chip and probably the ali-npu too

---
medium.com
Huawei Leaps into AI; Announces Powerful Chips and ML Framework
Synced
6-8 minutes
Known for its advanced telecommunication technology and cost-effective smartphones, Chinese tech giant Huawei is now aggressively expanding its artificial intelligence footprint. The company made a series of AI-related announcements today at the Huawei Connect 2018 Conference in Shanghai, introducing two AI chips and a machine learning framework. Huawei’s AI push is expected to intensify its battle with domestic rivals Alibaba, Tencent and Baidu in the AI market.

Blockbuster news had been expected after The Information revealed Huawei’s Project Da Vinci mandate “to bring AI to everything from telecom base stations to cloud data centers to smartphones to surveillance cameras.” A person familiar with the matter told Synced that the Da Vinci Project led by Huawei Deputy Chairman and President of Huawei chip affiliate HiSilicon Eric Zhijun Xu is a high priority project getting a huge amount of attention.

In today’s keynote address Xu stressed the importance of an AI development strategy: “Like electricity and railways during the industrial revolution, artificial intelligence is the new general purpose technology of the 21st century. Huawei’s AI strategy is to invest in basic research and talent development, build a full-stack, all-scenario AI portfolio, and foster an open global ecosystem.”

AI chips: Ascend 910 & Ascend 310
The Internet giants’ race to create custom AI chips has so far produced Google’s powerful TPUs and Microsoft’s FPGAs. Earlier this year, Alibaba and Baidu announced their respective development plans for AI chips Ali-NPU and Kunlun.

Huawei jumped on the bandwagon today with the unveiling of Ascend 910 and Ascend 310, two 7nm-based AI chip IPs that run on the cloud for training and inferencing. Both chips are built on Huawei’s homegrown Da Vinci architecture, which features scalable memory, compute, and on-chip interconnection.

Billed as the single chip with the greatest computing density, Ascend 910 delivers performance of up to 256 teraFLOPS under FP16 and 512 teraOPS under IN8 with a maximum power consumption of 350W. In comparison, Nvidia’s most powerful GPU Tesla V100 delivers up to 125 teraFLOPS with a max power consumption of 300W, while Google’s TPU 2.0 with four ASICs can reach 180 teraFLOPS.

Huawei also announced a large-scale distributed training system, Ascend Cluster, which combines 1024 Ascend 910 chips to reach 256 petaFLOPS for deep learning. Both the chips and the cluster will be available in Q2 2019. Chinese media is reporting that Huawei is touting its new cloud computing chips to Microsoft Azure China, although Huawei has officially denied this.

Ascend 310 is an efficient 12nm SoC (System on a Chip) designed for low-power computing, with a power consumption of 8W and a performance of 8 teraFLOPS under FP16 and 16 teraFLOPS under IN8.

Huawei Executive Ping Guo said the company is pumping more than US$1 billion annually into R&D for data centers.

ML framework MindSpore
Huawei also rolled out a set of open-source AI development tools on its cloud service platform which will help developers and engineers simplify AI workflow from training machine learning models to deployment on local devices. The tools will be available on Huawei’s AI service platform Cloud Enterprise Intelligence and its AI engine for smart devices HiAI.

Huawei’s new ML framework MindSpore provides device-edge-cloud training and inferencing based on a unified distributed architecture for machine learning, deep learning, and reinforcement learning, etc. It supports models trained on other frameworks such as TensorFlow and PyTorch, and provides flexible APIs decoupled from the core system.

Also announced today at Huawei Connect was Compute Architecture for Neural Networks (CANN), an operator library for chipsets. A standout component of CANN is its highly automated operators development toolkit Tensor Engine, which enables a DSL interface, auto optimization, auto generation, and auto tuning. CANN also includes TVM, an automated end-to-end optimizing compiler for deep learning. Huawei boasts that CANN can triple development efficiency.

ModelArts meanwhile is Huawei’s new machine learning platform-as-a-service, providing full-pipeline services, hierarchical APIs, and pre-integrated solutions.

A battle for the Chinese AI market
Huawei clearly wants to get a piece of China’s burgeoning AI market. A recent Tsinghua University report projects the Chinese AI market, which was worth US$3.55 billion in 2017 (up 67 percent from 2016), will grow by another 75 percent in 2018. Chinese society is quickly adapting to and embracing AI technologies, from facial recognition authentication for bank accounts, to smart speakers, home appliances and autonomous vehicles.

Chinese tech giants are ratcheting up their games, hoping to pull ahead of the competition in AI model and application development and deployment. Alibaba recently released new AI chips and updated its cloud-based enterprise and government AI-powered solutions. Tencent announced a new robotic research center and opened a new AI platform, AI.QQ.COM, aiming to build an ecosystem that unites the company’s diverse AI technical capabilities.

Huawei’s announcements emerge from the company’s strong focus on AI democratization, akin to efforts by US tech giants like Google and Microsoft. The company’s home-developed AI chips and open-source framework will help developers industry-wide create richer and more powerful AI applications. As China’s second-largest cloud vendor, Huawei is committed to attracting developers to its cloud platform by creating easy-to-use tools. The HiAI platform so far has a developer community of some 400,000.

Huawei’s comprehensive AI strategy marks a turning point for the company and the beginning of its AI transformation. With all leading global tech companies now betting heavily on AI, why not Huawei?

Journalist: Tony Peng | Editor: Michael Sarazen

 

[JSCh]

NEWS AND VIEWS | 10 OCTOBER 2018
LED technology breaks performance barrier | Nature.com
Light-emitting diodes made from perovskite semiconductors have reached a milestone in the efficiency with which they emit light — potentially ushering in a new platform for lighting and display technology.

Paul Meredith & Ardalan Armin

Light-emitting diodes (LEDs) have revolutionized lighting and displays, not least because they use energy more efficiently than any previous light-emitting technology. Micro-LEDs made from inorganic, ‘compound’ semiconductors are emerging that deliver unprecedented resolution for displays, whereas organic semiconductor LEDs (OLEDs) provide unparalleled colour quality and near-180° viewing angles, and could potentially be used to develop flexible, lightweight displays. In this issue of Nature, two papers1,2 report what could be the birth of a new family of LEDs based on semiconductors called perovskites. Remarkably, the efficiencies with which the perovskite LEDs (PLEDs) produce light from electrons already rival those of the best-performing OLEDs3, and have been achieved in less than four years since the report4 of the first PLED — suggesting that there is plenty of room for even further improvement in their performance.

Perovskites have shot to scientific stardom in the past few years, mostly because they show great promise for solar cells5, but their potential for use in other applications, such as light sensors6 and LEDs4, is rapidly emerging. Crucially, perovskites can be processed from solution (for example, using low-cost, low-tech printing methods), and work well in the designs for optoelectronic devices that are easiest to make. This might allow perovskite-based devices that have large areas (several square centimetres) to be made extremely cheaply, and with low embodied energy (the total energy involved in the entire life cycle of a device).

Cao et al.1 and Lin et al.2 have independently developed PLEDs that break an important technological barrier: the external quantum efficiency (EQE) of the devices, which quantifies the number of photons produced per electron consumed, is greater than 20%. There are several similarities between the devices reported by the two groups. Perhaps most notably, the active (emissive) perovskite layer is about 200 nanometres thick in both cases, and is sandwiched between two relatively simple electrodes. This design is called a planar structure, and is the most basic manifestation of diodes made from thin films of materials (Fig. 1). The electrodes are appropriately modified to ensure that electrons and holes (quasiparticles formed by the absence of electrons in atomic lattices) are efficiently pumped into the perovskite. As in all LEDs, when electrons meet holes, they can release energy in the form of photons through a process known as radiative recombination.

Figure 1 | Improved light-emitting diodes (LEDs) based on perovskite semiconductors. a, LEDs have previously been made from perovskites by sandwiching a thin layer of the semiconductor between a gold electrode and a transparent electrode. However, only about 20% of the light generated in the perovskite escapes from the device. b, Cao et al.1 report perovskite LEDs (PLEDs) in which the semiconductor layer consists of separated submicrometre-sized crystals, partitioned from the gold electrode by a thin layer of an organic material. This design increases the amount of light that escapes. c, Lin et al.2 report PLEDs based on a different perovskite, in which the semiconductor crystals are partly enclosed by an organic compound and the gold electrode is replaced by an aluminium one. This device optimizes the efficiency with which charges (not shown) that are pumped into the perovskite are converted into photons.​

Another similarity between the devices is that the perovskite layers were prepared using solutions, from which the semiconductors crystallized to form the emissive components of the LEDs. Cao et al. used a perovskite known as formamidinium lead iodide (FAPI), mixed with an amino-acid additive (aminovaleric acid) to control the size and orientation of the resultant perovskite crystals. FAPI has been quite widely explored as a semiconductor for solar cells, but Lin et al. report a new composite material in which crystals of the perovskite CsPbBr3 (Cs, caesium; Pb, lead; Br, bromine) are partly enclosed by a shell of an organic compound (methyl ammonium bromide; MABr).

Achieving high EQEs in any LED requires the elimination of non-radiative losses — electron–hole-recombination pathways that do not produce photons. Both Cao and colleagues’ and Lin and colleagues’ PLEDs deliver on this equally well. But the two groups also used other, subtly different methods to improve the EQE.

Cao et al. targeted the outcoupling problem, which is well known to those working with thin-film LEDs (such as PLEDs and OLEDs). The outcoupling problem is that the optical physics of planar diodes causes 70–80% of the light generated by the semiconductor to be trapped in the device. Various strategies have attempted to address this issue in OLEDs, such as using diffraction gratings7 and buckling the device8.

But Cao and colleagues took a simpler approach: they optimized their perovskite-processing conditions so that the emissive layer spontaneously forms as distinct submicrometre-scale crystal platelets (Fig. 1). The authors’ computational modelling shows that this submicrometre structuring increases the fraction of light that makes it out of the emissive layer to 30%, compared with 22% for an equivalent ‘flat-layer’ perovskite device (a device in which the perovskite layer does not have submicrometre structuring). In combination with the reduction in non-radiative losses, this results in an EQE of 20.7%.

By contrast, Lin et al. used a flat emissive layer, but tried to optimize the balance of electrons and holes injected into the perovskite, to make the most efficient use of every charge. This seems to be facilitated by the MABr shells that enclose the perovskite crystals. The resulting PLEDs have an EQE of 20.3%.

But caution is advised before ordering your PLED ultrahigh-definition television. OLEDs, and indeed all optoelectronic devices based on organic semiconductors, suffered for many years from stability issues. The first polymer OLEDs9could emit light for only seconds, and subsequent advances were needed to ensure that smartphone screens and OLED televisions last for tens of thousands of hours. The lifetime of LEDs can be measured by the T50 metric, which is the time for the performance of the device to drop by half. The T50 values of Cao and colleagues’ and Lin and colleagues’ PLEDs are currently modest: 20 hours and 100 hours, respectively.

Furthermore, displays require a minimum of three colours (and preferably more) to create high-quality colour images. Developing a range of colours for OLEDs was a big challenge. Cao and co-workers’ PLED emits in the near-infrared region of the electromagnetic spectrum, and Lin and co-workers’ PLED emits green light — which is definitely a good start. Multiple colours of PLEDs could be generated by altering the composition of the devices, but the same developmental journey as was needed for OLEDs lies ahead.

The two papers also highlight problems that occur every time new optoelectronic materials emerge as a technological platform: inconsistent characterization and a lack of standards. Because Cao and colleagues’ PLED emits light from outside the visible spectrum, they report the metrics of their devices radiometrically — they use a measure that simply takes into account the total emitted power. By contrast, Lin and colleagues describe the emission of their green PLED using photometric measures, which are weighted by the response of the human eye. The two groups also report the peak EQEs at different brightnesses, and therefore at different driving currents. This makes direct comparison somewhat problematic.

Caveats aside, the two papers are a milestone in PLED development. For now, LEDs based on compound semiconductors remain the dominant technology: they outclass the competition in many respects, including cost, efficiency, colour and brightness. They will be hard to beat. But that should not stop the pioneers of perovskite (or, indeed, organic) LEDs from trying.

Nature 562, 197-198 (2018)

doi: 10.1038/d41586-018-06923-y



LED technology breaks performance barrier | Nature.com

1. Yu Cao, Nana Wang, He Tian, Jingshu Guo, Yingqiang Wei, Hong Chen, Yanfeng Miao, Wei Zou, Kang Pan, Yarong He, Hui Cao, You Ke, Mengmeng Xu, Ying Wang, Ming Yang, Kai Du, Zewu Fu, Decheng Kong, Daoxin Dai, Yizheng Jin, Gongqiang Li, Hai Li, Qiming Peng, Jianpu Wang & Wei Huang. Perovskite light-emitting diodes based on spontaneously formed submicrometre-scale structures. Nature (2018). DOI: 10.1038/s41586-018-0576-2
2. Kebin Lin, Jun Xing, Li Na Quan, F. Pelayo García de Arquer, Xiwen Gong, Jianxun Lu, Liqiang Xie, Weijie Zhao, Di Zhang, Chuanzhong Yan, Wenqiang Li, Xinyi Liu, Yan Lu, Jeffrey Kirman, Edward H. Sargent, Qihua Xiong & Zhanhua Wei. Perovskite light-emitting diodes with external quantum efficiency exceeding 20 per cent. Nature (2018). DOI: 10.1038/s41586-018-0575-3

 

[JSCh]

China's Guangdong, Hong Kong, Macau Join Hands to Make Chips
FENG YUQING
DATE: FRI, 10/12/2018 - 20:41 / SOURCE:YICAI

China's Guangdong, Hong Kong, Macau Join Hands to Make Chips​

(Yicai Global) Oct. 12 -- China's Greater Bay Area that includes regions of Guangdong, Hong Kong and Macau has set up a semiconductor alliance to promote the development of the country's chip industry.

The Semiconductor Industrial Alliance was kicked off in the southern city of Guangzhou yesterday. Parties from these regions will jointly build service platforms including chip testing, electronic design automation, intellectual property, talent training and incubation to improve their competitiveness in the sector.

Guangdong has always made up one-third of the country's electronics industry, said Zou Sheng, former deputy secretary of the party committee at the Economic & Trade Commission of Guangdong. Yet, the southern province imported USD80 billion worth of chips last year, he said, adding that building the whole industrial ecosystem around semiconductors requires more work.

The biggest challenge is talent and investment, said Yu Chengbin, the founding chairman of the Macau Association of Microelectronics. Shortages in professionals in the field will trigger a heated competition in hiring and keeping the best talent. The development of the chip industry requires commitment, as well as long-term investment, he added.

The alliance's participants have different strengths, such as Guangzhou in application and manufacturing, Shenzhen and Zhuhai in product design, and Macao in research. Hong Kong is competitive in the field of securing intellectual property rights.

Guangzhou's South China University of Technology has expertise in digital circuit design while Macao leads in analog chips, so collaboration combines different strengths and demands, Yu said.

 

[JSCh]

Shanghai's Biggest Chip Project Starts Production
TANG SHIHUA
DATE: FRI, 10/19/2018 - 13:59 / SOURCE:YICAI

Shanghai's Biggest Chip Project Starts Production​

(Yicai Global）Oct. 19 -- Huali Microelectronics, Shanghai's largest investment project for making integrated circuits, has finished its almost two-year construction period and started operations yesterday.

Total investment of the project which focuses on the production of logical nodes has reached CNY38.7 billion (USD5.5 billion), state-backed Shanghai Observer reported. The construction of the 12-inch advanced wafer silicon assembly line started in December 2016.

The value of integrated circuit industry will top CNY200 billion by 2020, according to Shanghai's municipal plan. The sector was valued at CNY120 billion last year, making up about 20 percent of the country's total.

Huali's assembly line covers nodes in a scale from 28 to 14 nanometers with a monthly capacity of 40,000 wafers. The parent's manufacturing capability will cover these central chip process technologies ranging from 14 nm to 0.5 nm.

The levels of production and engineering will be among the world's top five, the report added.

 

[TaiShang]

China's Greater Bay Area that includes regions of Guangdong, Hong Kong and Macau has set up a semiconductor alliance to promote the development of the country's chip industry.

Amazing

Only one last part of Greater China is missing now

 

[JSCh]

China Policy Needs a Reset | EE Times
Analyst calls for more U.S. investment and engagement
By Rick Merritt, 10.19.18

Handel Jones just got back from China, and he doesn’t like how the techno-politics are playing out.

China’s middle class is getting riled up by American tariffs and moves against ZTE. They are increasingly passionate about a government move to invest in its technological independence from the U.S.

The veteran semiconductor analyst said that President Trump needs to take a page from President Reagan’s playbook and significantly increase federal investment in U.S. technology. “The DARPA [ERI] program is good, but we need something that’s 10 or even 100 times that size,” said Jones.

Reagan invested in big tech programs like the B-1 bomber, Star Wars missile defenses, and the global positioning system. They reinforced a position of technological superiority of the U.S. over the former Soviet Union.

Today, China is outspending the U.S. It is ramping programs worth hundreds of billions of dollars in areas such as AI and semiconductors and other projects under its Made in China 2025 initiative.

The spending in chips alone is breathtaking. China’s largest foundry, SMIC, is expected to get $10 billion to ramp up 14-, 10- and 7-nm nodes in new fabs that could kick out 70,000 wafers/month by late 2021.

Like many China initiatives, it’s not clear if the bold bet on SMIC will succeed. The foundry apparently has not yet been able to produce a FinFET process, although it struck a deal in 2015 to develop a 14-nm node with help from imec and Qualcomm.

Now, SMIC reportedly will rely on Taiwan’s UMC for 14-nm technology and its follow-ons. Market success is another hurdle given that TSMC is already seeing a steep decline in its 10-nm sales, noted Jones.

SMIC is one of China’s largest semiconductor gambits, but it has several others. Huada Semiconductor, a maker of chips for smart cards, could get $7 to $14 billion as part of a plan to bring up 28-nm processes with new local partners and management, said Jones.

Shanghai Jita Semiconductor, a subsidiary of Huada and China Electronics Corp., announced plans to build 200-mm and 300-mm fabs for analog and power semiconductors in Shanghai in a deal worth a total of $5.18 billion, according to a September report from the U.S. SEMI trade group.

Meanwhile, Huali is expanding in Shanghai, Grace is building a new fab in Wuxi, and Yangtze Memory announced in August bold plans to take significant market share in NAND flash.

The China investments come at a bad time for the American semiconductor industry.

Intel, the largest semiconductor firm in the U.S., is struggling to produce 10-nm chips. GlobalFoundries recently decided to halt plans for 7-nm chips, and one of its more promising investments is in FD-SOI technology, which it is ramping up in Germany and China.

President Reagan used big U.S. investments to get an edge over the Soviet Union, which was said to have nearly bankrupted itself trying to keep up. The situation is very different with a well-heeled and still growing China.

Rather than alienate the China government and its people, the U.S. should adopt a more collaborative approach, suggested Jones. He advocated an approach of working with China while also ratcheting up federal technology investment.

— Rick Merritt, Silicon Valley Bureau Chief, EE Times

 

[JSCh]

4.7-bln-USD China AMOLED display facility goes into operation
Source: Xinhua| 2018-05-17 23:53:27|Editor: Mu Xuequan


SHIJIAZHUANG, May 17 (Xinhua) -- A 6th-generation flexible AMOLED display screen production line, which involved an investment of nearly 30 billion yuan (4.7 billion U.S. dollars), went into operation Thursday in north China's Hebei Province.

The production line of Visionox Co., in an industrial park in Langfang City, south of Beijing, can meet the demand of nearly 100 million smartphones for high-end screens.

Visionox, which has more than 3,500 OLED patents, grew out of the OLED project team of Tsinghua University. It was the first Chinese firm to participate in the formulating of the global flexible display screen standard, according to Visionox president Zhang Deqiang.

The production line is expected to start mass production in the second half of the year with a monthly production of 30,000 glass substrates (1,500mm by 1,850mm).

Chinese firms previously relied heavily on imports for high-end display products, such as curved and foldable screens.

Visionox, Hefei Government Agree to Build Sixth-Gen AMOLED Screen Factory
TANG SHIHUA
DATE: MON, 10/22/2018 - 15:22 / SOURCE:YICAI

Visionox, Hefei Government Agree to Build Sixth-Gen AMOLED Screen Factory​

(Yicai Global) Oct. 22 -- Screenmaker Vision Technology plans to team up with the Hefei municipal government to construct a factory capable of producing 30,000 substrates a month for sixth-generation active matrix organic light emitting diode displays as it looks to keep pace with its South Korean counterparts.

The pair plan to sign an agreement that will see Visionox and the government’s investment arm set up a joint venture to build and operate the plant, which should achieve mass production within 25 months after construction begins, the Chaoyang-based firm said in a statement on Oct. 19. The new firm will have CNY22 billion (USD3.2 billion) in registered capital, with the Hefei government claiming an 82 percent stake. Total investment in the project is estimated at CNY44 billion.

Visionox’s screens can be used to make flexible screens for mobile devices, a segment expected to see impressive growth in the near future. It already has a 5.5-gen production line in Kunshan, Jiangsu province, and plans to open a sixth-gen factory in Gu’an, Hebei, by the end of this year.

The compound annual growth rate of the flexible terminal market will be higher than 22 percent over the next five years, Li Yaqin, general manager at tech data consultancy Sigamintell, told Yicai Global. There will be historic development opportunities, and Visionox’s new plant will help it meet demand and become more competitive, he added.

Samsung holds a more than 90 percent global share of the flexible AMOLED display market for smartphones, Li said. Chinese online retailer JD.Com is also breaking into the sector, and has two production lines for sixth-gen flexible displays and another on the way, he continued, saying Visionox’s new factory will expand the range of suppliers and help the industrial chain grow and mature.

Visionox was born from the OLED research team at Tsinghua University and has been engaged in the sector for 22 years. It ranked third globally in production of AMOLED panels for smartphones in the first half of this year, behind South Korea’s Samsung and LG Display, after producing 3.7 million units, Sigmaintell data shows.

Under the agreement, Visionox will build a research and development hub in Hefei to serve as a benchmark for peers nationwide, while the local government will offer support via favourable policies in land, energy, talent and other sectors, the statement added.

 

[qwerrty]

SmartSens Corporate Video

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

digitimes.com
China-based Kuntech stepping into flexible AMOLED market
Martin Yao, Taipei; Steve Shen, DIGITIMES
2-3 minutes

China-based Kuntech stepping into flexible AMOLED market

Monday 22 October 2018

China-based Kuntech Semiconductor Technology has joined the race for ramping up production of flexible AMOLED panels to meet potential demand from the smartphone and other IT sectors, according to industry sources.

Kuntech announced recently that it will invest CNY40 billion (US$5.77 billion) to establish a flexible semiconductor service and manufacturing base in Shaanxi, with plans to kick off the operations of the production base in the fourth quarter of 2020 and to officially mass-produce flexible AMOLED panels in the third quarter of 2021.

Under the project, Kuntech will set up a 6G flexible AMOLED panel line with a capacity of 30,000 1,500mm by 1,800 mm substrates a month. The production and service base will also accommodate a technology research and certification center and a well-established supply chain, company president Austin Jwo said at a press conference.

Being a flat panel maker dedicated to the development of self-sufficient flexible display technology, Jwo continued that Kuntech aims to offer total solutions consisting of manufacturing and related services.

Demand for flexible AMOLED panels will come from the smartphone sector initially as handset vendors including Samsung Electronics, Huawei, Xiaomi and Oppo are expected to roll out foldable smartphone models in early 2019, said the sources.

Recent market rumors also indicated that Samsung is expected to unveil its first flexible smartphone, reportedly named Galaxy F, in the upcoming CES 2019.

The application of flexible AMOLED panels will then extend to most of IT products such as notebooks and tablets, resulting in an explosive growth in demand for the flexible panels, added the sources.

Shipments of flexible AMOLED panels are expected to reach 335.7 million units by 2020, topping those of rigid AMOLED panels at 315.9 million units, and flexible AMOLED panels are predicted to make up 52% of total AMOLED panel shipments in 2020, up from 38.9% in 2018, according to IHS Markit.


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

digitimes.com
Consortium formed in China to rev up development of RISC-V processors
Shinee Wu, Taipei; Willis Ke, DIGITIMES
3 minutes

More than 50 IC designers, academic units and research bodies in China have recently established the China RISC-V Industry Consortium, seeking to develop the country's independent intellectual property rights for core IC technologies and ushering in a new development stage for China's RISC-V industry.

RISC-V is an open instruction set architecture (ISA) based on the established reduced instruction set computing (RISC) principles, and can be freely used for any purpose, allowing anyone to design, manufacture, and sell RISC-V chips and software.

Members of the new consortium include China IC designers, such as VeriSilicon, Espressif Systems, Ingenic Semiconductor, UNISOC (Spreadtrum & RDA), Zhaoxin Semiconductor, Horizon Robotics, and C*Core Technology, yet excluding Hisilicon and C-Sky Microsystems.

Consortium chairman Dai Weimin, currently chairman of VeriSilicon, said at the inauguration ceremony for the new organization that the largest opportunity RISC-V will bring to China is to place the country and the rest of the world on the same starting line.

Based on a recent market research report issued by GSMA, the global IoT market scale including connections, applications, platforms and services will amount to US$1.1 trillion by 2025, with commercial applications to command some 50% of the market. GSMA also estimates that there will be 1.8 billion IoT connections done by 2025, and China will become a huge market for IoT technologies and applications, which will naturally drive the birth of RISC-V ISAs.

China IC players are strong in application-type SoCs but weak in IP and fabrication ends, vice consortium chairman Hu Zhenpo said, adding that RISC-V can help address the processor IP weakness and build independent ecosystems to materialize more technology innovations and differentiations for the China semiconductor industry.

China IC designers are actively developing processor chips based on RISC-V architectures. Andes Technology, for instance, has launched four CPU cores - N25F, A25, NX25F and AX25 - featuring the architectures. At the moment, the company is teaming up with SiFive to develop more RISC-V ISAs, pioneering the development of such RISC-V chips in China.

On September 17, Huami also released Huangshan MHS001 chip, touted as the world's first RISC-V processor for wearable devices with a computing performance 38% higher than ARM Cortex-M4.

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

dealstreetasia.com
Shanghai rollable display maker Royole hits $5b valuation after new round
Mars Woo
3 minutes


Shanghai-based rollable display firm Royole Corporation has reached a valuation of $5 billion after raising an undisclosed amount from Yingke Innovation Asset Management for its ongoing Series E round.

According to a China Money Network report, Royole Corporation did not disclose the amount of the new capital injection but said it is in the process of securing more investments for the ongoing round.

Founded in 2012 by a team of scientists headed by Tsinghua and Stanford graduate Bill Liu, the company creates and manufactures next-generation human-machine interface technologies and products, such as flexible displays, flexible sensors, and smart devices.

On its website, Royole sells t-shirts and caps embedded with flexible wearables, where instead of the traditional prints, the wearer can display the moving graphics he or she wants using a smart phone.

Last year, Royole raised a total of $800 million in Series D to support its R&D efforts and accelerate production and sales.

The round included $240 million equity financing from Chinese investment firms including Hanfor Capital Management Co., Ltd, Warmsun Holding Group, Shanghai Pudong Development Bank Co., Ltd., Zhonghai Shengrong and Tanshi Capital.

It also involved $560 million debt financing from China CITIC Bank, Agricultural Bank of China, Industrial and Commercial Bank of China Ltd, Bank of China and Ping An Bank Co., Ltd.

In 2015, Royole raised $172 million in a Series C round co-led by IDG Capital, Shenzhen Capital Group, and Green Pine Capital.

According to Crunchbase data, Royole’s total funding amount has reached $1.1 billion from six funding rounds. The Shenzhen-based company’s valuation has gradually climbed from $3 billion in 2016 to $5 billion currently.

“Within six years of starting business, Royole has grown to a team of over 2000 highly skilled employees from around the world, and has earned a global valuation of $5 billion, adding to our reputation as one of the world’s fastest growing tech startups,” the company said on its website.

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

digitimes.com
Tsinghua Unigroup new fab construction kicks off
Shinee Wu, Taipei; Jessie Shen, DIGITIMES
2-3 minutes

Tsinghua Unigroup new fab construction kicks off

Thursday 18 October 2018

China's state-owned Tsinghua Unigroup has already begun to implement its new fab projects in Nanjing and Chengdu as part of its memory business expansion. The new manufacturing sites are designed for monthly output of 300,000 12-inch wafers each, according to sources familiar with the matter.

Tsinghua Unigroup already has an existing manufacturing site in Wuhan, currently operated by subsidiary Yangtze Memory Technologies (YMTC). Tsinghua Unigroup in 2017 disclosed plans to establish IC manufacturing sites in Wuhan, Chengdu and Nanjing, with total investment reaching US$70 billion.

Tsinghua Unigroup broke ground for its US$30 billion plant in Nanjing on September 30, the sources said. The plant will be engaged mainly in the manufacture of 3D NAND flash and DRAM chips, and will be built in two phases.

The first phase of the Nanjing plant will bring in monthly production capacity of 100,000 wafers, with about US$10.5 billion set to be invested in the facility construction, the sources noted.

On October 12, Tsinghua Unigroup broke ground for another new memory plant, the sources indicated. Located in Chengdu, the plant will be built with US$24 billion in total investment and will have 12-inch 3D NAND flash production lines installed.

In addition, Unigroup Guoxin Microelectronics under Tsinghua Unigroup plans to transfer its entire 100% stake in Xian UniIC Semiconductors, which specializes in the design and development of DRAM chips, to UNIC Memory Technology under Tsinghua Unigroup. It marks Tsinghua Unigroup's move to integrate resources to enhance its memory business competitiveness, according to market observers.

Founded in August 2017, UNIC Memory Technology will be able to enhance its DRAM offering after merging with UniIC Semiconductors, the observers indicated. Meanwhile, Guoxin Microelectronics will be allowed to stay focused on its security chip business, the observers said.


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

digitimes.com
Serious talent shortage posing challenges for China IC development
Cindy Yu, Taipei; Willis Ke, DIGITIMES
2-3 minutes

Despite China's semiconductor industry aggressively carrying out diverse development projects, increasingly serious talent shortage is casting clouds over the development, making it a pressing issue for IC players in the country to introduce talent from abroad, especially amid the escalating US-China trade conflicts, according to industry sources.

The sources said many of the first-generation executives at China's Semiconductor Manufacturing International Corp (SMIC), including founder Richard Chang and CEOs David Wang and Simon Yang, left the company to set up IDM or steer the development of DRAM or flash memory chips at other China semiconductor businesses.

As semiconductor investment projects are no longer concentrated in Beijing, Shanghai and Shenzhen, more and more mid- and high-ranking professionals are badly needed to handle management, operation and technology development at IC fabs at other Chinese cities.

Industry experts estimate that China's semiconductor industry will see a talent shortfall of up to 400,000 people by 2020, more than the existing total employment force in the industry.

Wang Zhihua, deputy dean of the Institute of Microelectronics at Tsinghua University, estimated that if China's semiconductor wants to realize the production value of CNY1 trillion (US$144.27 billion) by 2020, a total of 700,000 employees are needed to support the production if based on the per capita output value of CNY1.4 million a year, compared to the current workforce of only 300,000.

In Jiangsu province alone, the semiconductor talent shortfall is expected to exceed 100,000 by 2020, not to mention the shortfalls seen in new semiconductor bases including Wuhan, Chengdu and other cities in central and mid-western China.

Besides cultivation of local talent, the best way to address talent shortfalls is to enforce massive introduction from abroad. Huawei, for instance, has introduced at least 6,000 engineers from overseas, and SiEn (Qingdao) Integrated Circuits has sourced one third of its engineers from Taiwan.


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

China recruiting engineering talent in Taiwan IC design industry
Wednesday 17 October 2018

Senior engineers and executives at Taiwan-based IC design houses are being targeted by China-based firms looking to enhance their homegrown technology capability, according to industry...

https://digitimes.com/news/a20181016PD208.html

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

image-sensors-world
RoboSense Announces China’s Largest-ever Round of Financing for a LiDAR Company
3-4 minutes

PRNewswire, Thomas-PR: RoboSense announces the completion of China’s largest-ever single round of financing for a LiDAR company – a combined investment of over $45 million (RMB 300 million). The investors include Cainiao Smart Logistics Network Ltd. (“Cainiao”), the logistics arm of the Alibaba Group; SAIC Motor Group (Shanghai Automotive Industry Corporation), the largest publicly-traded auto manufacturer in China’s A-Share; and BAIC Group (Beijing Automotive Industry Holding Co.) electric vehicle company.

Robosense claims to have an over 50% market share of all LiDAR sold in Asia becoming the market leader in the region. The new funding will be used to increase RoboSense’s market share and the R&D of autonomous vehicle technologies, including its solid-state LiDAR, AI sensing algorithms, and other advanced technologies, as well as accelerating product development, long-term manufacturing and market penetration.

“The rapid development of autonomous driving has ignited a huge demand for LiDAR,” said Mark Qiu, co-founder of RoboSense. “RoboSense is embracing this market demand through partnerships with multiple industry leaders. It is our great pleasure to be endorsed and funded by industry giants from many different fields. This round of funding is not only for capital assistance, but also for strategic resources. We are looking forward to continuously working with our partners to lead the large-scale commercialization era of the autonomous driving industry.”

IHS Markit predicts that by 2035, global sales of self-driving cars will reach 21 million vehicles, up from nearly 600,000 vehicles in 2025. IHS believes that by 2035, nearly 76 million vehicles with some level of autonomy will be sold globally.

Autonomous logistics vehicles are expected to become one of the first markets for autonomous vehicle technology. Based on data from Deloitte’s China Smart Logistics Development Report, the intelligent logistics market will reach $145 billion (RMB 1 trillion) by 2025.

In the past two years, RoboSense has had explosive growth:

• In April 2017, the company started the mass production of its 16-beam automotive LiDAR.
• In September 2017, the company mass-produced its 32-beam LiDAR, released a LiDAR-based autonomous driving environmental sensing AI system, and provided a software and hardware combined LiDAR environment sensing solution.
• In October 2017, RoboSense launched its MEMS solid-state LiDAR, publicly exhibited for the first time at CES 2018 in January 2018.
• In April 2018, RoboSense partnered with the Cainiao Network to launch the world’s first MEMS LiDAR autonomous logistics vehicle – the G Plus.

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

 

[JSCh]

Chip industry's growth will reduce foreign reliance
By Fan Feifei | China Daily | Updated: 2018-11-07 11:06

Chip industry's growth will reduce foreign reliance. [Photo/IC]

Echoing President Xi Jinping's call to boost the healthy development of the country's new generation of artificial intelligence (AI), Chinese technology companies attending the ongoing World Internet Conference in Wuzhen, East China's Zhejiang province, are intensifying efforts to develop their own artificial intelligence (AI) chips, as the country aims to become self-reliant in key technologies and lessen its dependence on foreign microprocessors.

In September, Chinese tech giant Alibaba Group Holding Ltd announced the establishment of a dedicated chip subsidiary aimed at creating customized AI chips and embedded processors to further support the tech giant's cloud and internet of things businesses, as well as drive the development of industry-specific applications.

The company plans to launch its first AI chip AliNPU, which could potentially support technologies used in autonomous driving, smart cities and smart logistics in the second half of next year, said Zhang Jianfeng, Alibaba chief technology officer, at the company's annual cloud computing conference in Hangzhou, Zhejiang province.

In April, Alibaba announced that it had fully acquired Hangzhou C-Sky Microsystems Co, which develops embedded central processing units and chip architecture, underscoring its commitment to the chip industry. "We are confident that our advantages in algorithm, data intelligence, computing power and domain knowledge on the back of Alibaba's diverse ecosystem will put us in a unique position to lead real technological breakthroughs in disruptive areas, such as quantum and chip technology," Zhang said.

China's largest telecom equipment maker and smartphone vendor Huawei Technologies Co unveiled two new chips for AI applications called Ascend 910 and Ascend 310.

Ascend 310 is designed for the low-power computing needs of smart devices, while Ascend 910, which will be available in the second quarter of 2019, is aimed at data centers. Huawei claims its chip can process more data at a faster rate than its competitors and help link networks in a matter of minutes.

The two chips will pit Huawei against major players such as Intel Corp, Qualcomm Inc and Nvidia Corp. "Huawei's AI strategy is to invest in basic research and talent development, build a full-stack, all-scenario AI portfolio and foster an open global ecosystem," said Xu Zhijun, rotating chairman of the company, at the Huawei Connect 2018 Conference in Shanghai in October. Xu added that these chips will not be sold as standalone products, because the company will develop products such as AI servers, AI accelerators and autonomous driving solutions based on these processors and sell them along with the chips to clients.

Huawei wants to get a piece of China's burgeoning AI market. A recent Tsinghua University report projects the Chinese AI market, which in 2017 grew 67 percent year-on-year to $3.55 billion, will increase by another 75 percent in 2018.

Besides, Chinese internet search giant Baidu Inc unveiled AI chip Kunlun, which can be built to meet the high performance requirements of a wide variety of AI fields and products including data centers, public clouds and autonomous vehicles.

These companies expect their respective chips to help their AI applications run better while lowering costs. The chips could also reduce these companies' dependency on processor makers whose products excel at performing the functions modern AI applications require.

Luo Weidong, chairman of Shenzhen Saiya Capital Management Co said: "The ZTE issue is like a wakeup call for the country's technology industry. The country will attach greater importance to the independent research and development of chips."

And Shu Qiquan, general manager of Shanghai Qianbo Fund, said domestic chip manufacturers still lag behind leading foreign competitors in chip design, material, and equipment. "The chip industry is in urgent need of talent and capital, and the authorities should provide financing channels and capital support for chip companies. The chip design, memory, graphic processing unit, packaging and testing, semiconductor equipment, and materials have great investment value."

 

[j20blackdragon]

Nightmare for Intel.

Sugon x86 exascale supercomputer.

Hygon x86 processors.

Hygon DCU accelerators.

 

[Bussard Ramjet]

Nightmare for Intel.

Sugon x86 exascale supercomputer.

Hygon x86 processors.

Hygon DCU accelerators.

Does China have a company like Flir? That makes infrared sensors for both military and civilian applications?

 

[qwerrty]

biometricupdate.com
Horizon Robotics raises up to $1 billion to address facial biometric and autonomous vehicle markets
Nov 27, 2018 | Chris Burt
2-3 minutes

Horizon Robotics has raised up to $1 billion in a Series B funding round which values the company at between $3 billion and $4 billion, the Financial Times reports. The funding round for the Intel-backed company includes investment from an international chip company, Senior Business Development Manager Stone Li said.

Previous investors in Horizon include Hill House Capital, Sequoia Capital, Yuri Milner and Sinovation Ventures.

The company is benefiting from China’s push to reduce its reliance on imported semiconductors, which the Financial Times reports it spends more on than oil. The state-owned SMIC is still producing chips multiple generations behind those of Taiwan’s TSMC or Intel.

Horizon says that one of its chips runs facial recognition algorithms for on-camera matching from databases storing up to 50,000 faces at the edge. In another project, the company is developing self-driving cars in the Eastern Chinese city of Wuxi in partnership with Audi.

“You can debate how many cars will be self-driving by when,” Mark Li, semiconductor analyst at Bernstein Research, told the Financial Post. “In my view it will be a while, but along the way you will have more and more self-driving functions in the cars — and in addition some chips could be used elsewhere, like facial recognition surveillance.”

The investment climate for Chinese AI may be cooling, but some companies in the industry continue to raise billions, according to the report.

UBS estimates that the market for shifting AI from the cloud to devices will generate $15 billion by 2021.

Horizon launched a new line of cameras with embedded facial recognition technology earlier this year.

Does China have a company like Flir? That makes infrared sensors for both military and civilian applications?

for civilian:

guide infrared
north guangwei (gwic)
magnity
dali-tech

.

 

[cirr]

China's Indigenous Ultraviolet Ultra High Resolution Lithography (capable of 22nm-10nm processes)

我国成功研制出世界首台分辨力最高紫外超分辨光刻装备

邹维荣、吕珍慧 装备科技 今天

Pls click on the link below if the images fail to show.

https://mp.weixin.qq.com/s/n0Izw3cIwspvXXGEa9HtGA

我国成功研制出世界首台分辨力最高紫外超分辨光刻装备

可加工22纳米芯片

▲超分辨光刻装备核心部件纳米定位干涉仪以及精密间隙测量系统。

军报记者成都11月29日电（吕珍慧、记者邹维荣）国家重大科研装备研制项目“超分辨光刻装备研制”29日通过验收，这是我国成功研制出的世界首台分辨力最高紫外超分辨光刻装备。该光刻机由中国科学院光电技术研究所研制，光刻分辨力达到22纳米，结合多重曝光技术后，可用于制造10纳米级别的芯片。

▲超分辨光刻设备核心部件超分辨光刻镜头。

中科院理化技术研究所许祖彦院士等验收组专家一致表示，该光刻机在365纳米光源波长下，单次曝光最高线宽分辨力达到22纳米。项目在原理上突破分辨力衍射极限，建立了一条高分辨、大面积的纳米光刻装备研发新路线，绕过了国外相关知识产权壁垒。

▲超分辨光刻设备加工的4英寸光刻样品。

▲采用超分辨光刻设备加工的超导纳米线单光子探测器。

光刻机是制造芯片的核心装备，我国在这一领域长期落后。它采用类似照片冲印的技术，把一张巨大的电路设计图缩印到小小的芯片上，光刻精度越高，芯片体积可以越小，性能也可以越高。但由于光波的衍射效应，光刻精度终将面临极限。

▲中科院光电所科研人员展示利用超分辨光刻设备加工的超导纳米线单光子探测器。

为突破极限、取得更高的精度，国际上目前采用缩短光波、增加成像系统数值孔径等技术路径来改进光刻机，但也遇到装备成本高、效率低等阻碍。

项目副总师胡松介绍，中科院光电所此次通过验收的表面等离子体超分辨光刻装备，打破了传统路线格局，形成了一条全新的纳米光学光刻技术路线，具有完全自主知识产权，为超材料/超表面、第三代光学器件、广义芯片等变革性领域的跨越式发展提供了制造工具。

▲项目副总设计师胡松研究员介绍超分辨光刻装备研制项目攻关情况。

▲中科院光电所科研人员操作超分辨光刻设备。

据了解，该光刻机制造的相关器件已在中国航天科技集团公司第八研究院、电子科技大学太赫兹科学技术研究中心、四川大学华西医院、中科院微系统所信息功能材料国家重点实验室等多家科研院所和高校的重大研究任务中取得应用。

▲中科院光电所科研人员操作超分辨光刻设备。

@Bussard Ramjet India?

 

[qwerrty]

is it ready for commercial use?

the latest euv lithography tool from asml has 13nm resolution capable of volume production chips at 7nm and below. china is still behind. still, this is big big news for china. even supapowah amelika buys from asml.