What's new

Opinionated - China Chipping Away to Semiconductor Dominance

Jun 20, 2018 05:42 PM BUSINESS & TECH
AI-Chip Unicorn Wins $2.5 Billion Valuation
By Coco Feng

1529487738377782.jpg
Cambricon Technologies Corp. Ltd. CEO Chen Tianshi shows a pair of Cambricon MLU100 cloud chips, also China’s first cloud artificial intelligence (AI) chip, at a news conference in Shanghai on May 3. Photo: IC

Chinese artificial-intelligence (AI) chipmaker Cambricon Technologies Corp. Ltd, which is now collaborating with Huawei Technologies Co. Ltd. on chips for its smartphones, has completed a new round of fundraising that valued it at $2.5 billion.

The Beijing-based startup raised “hundreds of millions of dollars” in the latest round, it said in a Wednesday statement (link in Chinese) published on its official WeChat account.

The leading investors are state-backed venture capital funds that include China state-owned Capital Venture Investment Fund, SDIC Venture Capital and China Reform Fund. Lenovo Capital, Incubator Group and Alibaba Group Holding Ltd. also increased their respective investments in Cambricon.

Established in 2016 with financial and technological support from the state-owned Chinese Academy of Sciences, Cambricon develops chips with AI capability and deep learning processors. It supplied technology for Huawei’s smartphone chip Kirin 970, which is embedded in several flagship models, including the Mate10 and the P20. Cambricon and Huawei are expected to continue the collaboration on the next-generation chip.

Cambricon completed a $100 million funding round in August, with a valuation at more than $1 billion, which the company claimed made it the first AI-chip unicorn — a startup company valued at over $1 billion.

China has for years struggled to make its domestic chip industry more competitive with the U.S.’ and Japan’s. China’s lagging capabilities have been highlighted since the U.S. banned American chip companies from selling components to Chinese telecom equipment-maker ZTE Corp., which caused ZTE to halt its main business operations. The two sides reached a settlement after ZTE agreed to a $1 billion fine, but the U.S. Senate on Monday passed a bill that would reimpose the sanction.

Cambricon is among a group of Chinese startups aiming to break into the business of high-tech chip design, now dominated by global giants such as Intel Corp. and Qualcomm Inc. Other Chinese AI-chip startups, such as DeePhi Tech and Horizon Robotics, are reported to have raised tens of millions of dollars each.
 
.
huawei is investing heavily in silicon photonics chip design and fab tools at their r&d centres in europe and japan, according to huawei's linkedIn.

will say goodbye to optical components suppliers from the US, japan and china too :D

Silicon Photonics will revolutionize optical communication

Traditionally transceivers are built with optical subassemblies containing multiple components made of materials that are often expensive to purchase and challenging to process. Also these components need to be individually packaged and then repackaged into an hermetically sealed optical assembly.

In Silicon Photonics this optical subassembly is replaced by an optical chip. On this chip all components of a traditional subassembly are replaced by structures in silicon. This chip can then be fabricated just like an electronic chip enabling low cost high volume manufacturing.

The precision of the manufacturing process of optical chips in Silicon Photonics allows for huge miniaturization making it possible to develop transceivers that enable higher port density in optical networking equipment and consume less power than traditional transceivers.
 
. . .
China develops micro flexible capacitors with high performance, safety
Source: Xinhua| 2018-07-02 20:50:47|Editor: Liangyu


SHENYANG, July 2 (Xinhua) -- Chinese researchers have developed new micro capacitors with high energy storage density and excellent thermal stability, according to Dalian Institute of Chemical Physics, Chinese Academy of Sciences.

The lithium ion micro-capacitors (LIMCs) are mechanically flexible without performance degradation under repeated bending, and can operate safely even at temperatures of 80 degrees Celsius.

The ever-increasing boom in smart, miniaturized electronics has led to an urgent need for on-chip energy storage systems that exhibit high performance, safety, flexibility and robust integration.

The solid-state planar LIMCs can boost high voltage and capacitance and their high cycling stability allows them to maintain almost 99 percent of their capacitance after 6,000 electric cycles.

The micro devices are expected to be commercially available in 2022.
 
.
nasdaq.com
Baidu Unveils High-Performance AI Chip, Kunlun, at Baidu Create 2018
4 minutes
The search engine provider releases China's first cloud-to-edge AI chip for training and inference
BEIJING, July 03, 2018 (GLOBE NEWSWIRE) -- Baidu Inc. (NASDAQ:BIDU) today announced Kunlun, China's first cloud-to-edge AI chip, built to accommodate high performance requirements of a wide variety of AI scenarios. The announcement includes training chip "818-300"and inference chip "818-100". Kunlun can be applied to both cloud and edge scenarios, such as data centers, public clouds and autonomous vehicles.

Kunlun is a high-performance and cost-effective solution for the high processing demands of AI. It leverages Baidu's AI ecosystem, which includes AI scenarios like search ranking and deep learning frameworks like PaddlePaddle. Baidu's years of experience in optimizing the performance of these AI services and frameworks afforded the company the expertise required to build a world class AI chip.

In 2011, Baidu started developing an FPGA-based AI accelerator for deep learning and began using GPUs in datacenters. Kunlun, which is made up of thousands of small cores, has a computational capability which is nearly 30 times faster than the original FPGA-based accelerator. Other key specifications include: 14nm Samsung engineering, 512 GB/second memory bandwidth, as well as 260TOPS while consuming 100 Watts of power.

In addition to supporting the common open source deep learning algorithms, Kunlun chip can also support a wide variety of AI applications, including voice recognition, search ranking, natural language processing, autonomous driving and large-scale recommendations.

With the rapid emergence of AI applications, dramatically increasing requirements are being imposed on computational power. Traditional chips limit how much computing power is available and thus how far AI technologies can be accelerated. Baidu developed this chip, specifically designed for large-scale AI workloads, as an answer to this demand. Baidu believes that it will allow for significant advancements in the open AI ecosystem.

Baidu plans to continue to iterate upon this chip, developing it progressively to enable the expansion of an open AI ecosystem. As part of this, Baidu will continue to create "chip power" to meet the needs of various fields including intelligent vehicles, intelligent devices, voice recognition and image recognition.
 
.
PUBLIC RELEASE: 7-JUL-2018
Mystery of phase change in sub-nanosecond-octahedra structure motif
SCIENCE CHINA PRESS

Phase change random access memory (PCRAM) has been successfully applied in the computer storage architecture, as storage class memory, to bridge the performance gap between DRAM and Flash-based solid-state drive due to its good scalability, 3D-integration ability, fast operation speed and compatible with CMOS technology. Focusing on phase change materials and PCRAM for decades, we have successfully developed 128 Mb embedded PCRAM chips, which can meet the requirements of most embedded systems.

3D Xpoint (3D PCRAM), invented by Intel and Micron, has been regarded as a new breakthrough in the last 25 years since the application of NAND in 1989, which represents state-of-the-art memory technology. This technology has some remarkable features, such as the confined device structure with 20 nm size, the metal crossbar electrodes to reduce the resistance variations in PCRAM arrays, and the ovonic threshold switching selector that can provide a high drive current and a low leakage current. A good understanding of phase change mechanism is of great help to design new phase change materials with fast operation speed, low power consumption and long-lifetime.

In a recent paper published in SCIENCE CHINA Information Sciences, researchers firstly review the development of PCRAM and different understandings on phase change mechanisms in recent years, and then propose a new view on the mechanism, which is based on the octahedral structure motifs and vacancies.

Octahedral structure motifs are generally found in both amorphous and crystalline phase change materials. They are considered to be the basic units during phase transition, which are severely defective in the amorphous phase. These configurations turn into more ordered ones after minor local rearrangements, the growth of which results in the crystallization of rocksalt (RS) phase with a large amount of vacancies in the cation sites. Further driven by thermodynamic driving force, these vacancies move and layer along certain directions; consequently, the metastable RS structure transforms into the stable hexagonal (HEX) structure. Based on the results, researchers find that reversible phase transition between amorphous phase and RS phase, without further changing into HEX phase, would greatly decrease the required power consumption. Robust octahedra and plenty of vacancies in both amorphous and RS phase, respectively avoiding large atomic rearrangement and providing necessary space, are crucial to achieve the nanosecond or even sub-nanosecond operation of PCRAM.

###​

Zhitang SONG, Sannian SONG, Min ZHU et al. From octahedral structure motif to sub-nanosecond phase transitions in phase change materials for data storage. Sci China Inf Sci, 2018, 61(8): 081302


Mystery of phase change in sub-nanosecond-octahedra structure motif | EurekAlert! Science News
 
.
Chinese Chip Tester Eyes Growing Domestic Market With USD1.2 Billion Facility
TANG SHIHUA
DATE: MON, 07/09/2018 - 19:54 / SOURCE:YICAI

3.4%E5%8D%8E%E5%A4%A9%E7%A7%91%E6%8A%80%E6%8B%9F%E6%8A%95%E8%B5%8480%E4%BA%BF%E5%85%83%E5%9C%A8%E5%8D%97%E4%BA%AC%E5%BB%BA%E8%AE%BE%E9%9B%86%E6%88%90%E7%94%B5%E8%B7%AF%E5%B0%81%E6%B5%8B%E5%9F%BA%E5%9C%B0ic.jpg

Chinese Chip Tester Eyes Growing Domestic Market With USD1.2 Billion Facility

(Yicai Global) July 9 -- Chinese semiconductor packager and tester Huatian Technology is to invest CNY8 billion (USD1.2 billion) in a new industrial base in Nanjing, capital of China’s eastern Jiangsu province, as part of efforts to capitalize on expected long-term growth in the country’s chip development sector.

Huatian Tech aims to build the facility in three stages with full operation slated for 2028, the Gansu-based firm said in a statement, adding that it will provide services for storage, artificial intelligence and micro-electro-mechanical integrated circuits.

China is making strides to reduce its dependence on imports of semiconductors, most recently in response to growing trade frictions with the US. Annual imports of such products hit USD260 billion at the end of last year, up from USD200 billion at the end of 2013, a report from Bernstein Research states.

The country set out plans in May to raise up to CNY200 billion to fund the development of the technology and Huatian Tech aims to be at the front of the line for the increased demand for testing and packaging in the long-term.

The company will register and establish a wholly-owned or controlled company in Nanjing’s Pukou Economic Development Zone for the project’s construction and operation. The zone is an integrated circuit industrial base in Jiangsu Province and home to key firms in the sector like TSMC Nanjing and Ardentec.
 
Last edited:
.
China Begins Domestic Production of AMD Server CPUs
Michael Feldman | July 9, 2018 15:58 CEST

Chinese chipmaker Hygon is now manufacturing Zen-based x86 CPUs using a licensing agreement it signed with AMD in 2016.

zen-module-400x304.png

The Chinese-designed "Dhyana" x86 processors are said to be essentially identical with AMD’s own Zen-based EPYC processor, differing only in the internally stored vendor ID and family series number. If true, the Chinese silicon should be able to take advantage of the entire software stack compatible with the EPYC CPU, including operating systems, compilers, libraries and programmer tools.

The 2016 licensing agreement earned AMD $293 million in cash and will continue to pay dividends in the form of royalty payments from unit sales. AMD says the licensing of its Zen design with Hygon does not violate its x86 license agreement with Intel and thus far, Intel has not objected.

The terms of the deal restrict Hygon to only selling the Dhyana chips inside China. Nonetheless, that will almost certainly impact sales of both AMD EPYC and Intel Xeon processors in the China, which is the world’s largest consumer of semiconductor products. Having a bigger overall share of the datacenter business, Intel will likely be affected more, especially since AMD will at least realize royalty payments from the sales of the Dhyana chips. For the Chinese server market, this is all good news, since customers there will now get access to x86 silicon at lower prices than either AMD or Intel could charge.

It’s probably only a matter of time until we start seeing Dhyana-powered HPC cluster deployments in China, some of which might entail large installations. Two years ago, a series of tweets sent out by James Lin, vice director for the Center of HPC at Shanghai Jiao Tong University, revealed that one of the three exascale tracks in China was going to be based on CPUs using AMD’s x86 design. (The other two tracks are based on the ShenWei processor and the Arm architecture, respectively.) According to Lin, these x86-based systems would be manufactured by Sugon and deployed at both the Shanghai Supercomputer Center and the National Supercomputing Center in Shenzhen. More than likely, these will be pre-exascale supercomputers.

It’s hard to imagine the Zen-based Dhyana would be the basis of an exascale machine unless they were paired with a much more powerful accelerator, such as a future version of the Chinese-built Matrix-2000. But if the Zen deal works out for everyone involved, we might see additional licensing agreements in China for AMD’s next-generation Zen 2 CPU or maybe even their Radeon GPUs.

Dhyana is part of a larger strategy in China to develop a domestic chipmaking capability, initially to satisfy internal demand, but eventually to sell its wares globally. Since x86, Arm, and OpenPower can now all be licensed, the country has a broad array of architectures from which to choose. The completely open source RISC-V architecture, which is being considered by the EU for future supercomputers, could be yet another option for China. One thing seems certain, the days of proprietary chip designs and vendor lock-in appear to be coming to an end.



China Begins Domestic Production of AMD Server CPUs | TOP500 Supercomputer Sites
 
.
News & Analysis
Chips Boom as Trade War Looms
DRAM prices drive the upswing

Rick Merritt
7/10/2018 00:01 AM EDT

SAN FRANCISCO, Calif. — The semiconductor industry is poised for as much as 15% growth this year and a shot at its first $500-billion year in 2019, driven largely by rising memory prices. The big dark spot on the horizon is a growing trade war between the U.S. and China.

That was the outlook from a handful of analysts at a kickoff for the annual Semicon West event here.

“We are at or just beyond the peak in the economic growth cycle…but the potential for political uncertainty is greater than it has been in a while and that could knock us off the growth path,” said Duncan Meldrum, chief economist at Hilltop Economics, forecasting a downturn will hit the capital equipment market next year and spread to the chip sector in 2020.

It’s the second year rising DRAM prices amid tight supply have boosted the overall chip industry. The trend is expected to continue until 2020 when more supply comes online.

“When prices went up, demand did not decline as it used to. Consumers kept buying and made the DRAM guys very happy — that was exactly what the big three needed to build their 3D NAND fabs,” said Bob Johnson, a research vice president at Gartner, referring to Samsung, SK Hynix and Micron.

Looking forward, Johnson predicts DRAM density will increase at less than half its historic 32% rate. The big questions are whether in the next couple years China will start producing DRAMs, driving prices down, or alternative memory technologies will emerge, he said.

NAND flash revenues are growing at a stately 9% compound rate, but will increasingly feel cost pressures of trying to stack more layers. “If DRAM profits roll off as China comes online, NAND investments could be pinched,” Johnson said.


The chip forecast looks like stair steps going forward, said Johnson. Click to enlarge. (Chart: Gartner)

A shift from smartphones to industrial and automotive markets as growth drivers is expected to moderate the ups and downs of the overall chip sector. “It looks like stair steps going forward,” he said.

Meanwhile, the memory “super-cycle” and China’s chip ambitions are fueling unprecedented fab growth, said Clark Tseng, director of industry research and statistics for the SEMI trade group that hosts the event.

Last year was a record in front-end fab investments. Fab spending in China is expected to surpass $10 billion this year and reach near $18 billion in 2019, exceeding spending in South Korea. “We haven’t seen four years of consecutive global fab growth since the 1990s,” Tseng said.


DRAM cost/MByte may never return to its traditional 32% a year decline. (Chart: Gartner)
Tense relations between the U.S. and China remain the biggest flashpoint for the semiconductor industry after both countries approved broad new tariffs last week. It was the latest flare up in a long-simmering debate over China’s intellectual property practices and the recent decision to ban sales of U.S. components to ZTE.

“The semiconductor industry is in the middle of the tech and trade issues. The trade issues get worse daily, and there’s no settlement on the horizon,” said Robert Maire, president of Semiconductor Advisors.

Maire called tariffs ineffective, and said the Trump administration also is considering bans on sales to China of leading-edge foundry gear. Given the intertwined nature of the two economies “this is a bit like a nuclear standoff of mutually assured destruction, and it’s unclear who would be damaged more,” he said.

The ban on ZTE is already encouraging it to consider suppliers other than Intel and Qualcomm which do 23% and 65% of their business in China, respectively. Maire called a China court’s decision to block sales of Micron chips in the country “a thinly veiled” counterpunch in the trade dispute.

Long term, “I think it’s not a question of if but when China takes over Taiwan, they get TSMC and the tables are turned as they get more leverage over Apple, Qualcomm,” and others, Maire said.

The issues have a long history. Years ago, the U.S. forbade China’s SMIC from importing the latest fab gear, keeping it from leading-edge capabilities, said Maire, who worked on the foundry’s IPO. Such moves fueled China’s concerns about self-sufficiency in semiconductors and the ZTE ban “threw gas on the fire,” he said.


Spending on front-end fabs in China is expected to outpace spending in Korea. Click to enlarge. (Chart: SEMI)


— Rick Merritt, Silicon Valley Bureau Chief, EE Times



Chips Boom as Trade War Looms | EE Times
 
.
Cutting-edge chip moves at light speed
By Zhou Wenting in Shanghai | China Daily | Updated: 2018-07-11 09:37
f_art.gif
w_art.gif
in_art.gif
more_art.gif


A test production line is expected to be completed this year following Chinese breakthroughs in the development of a new type of computer chip - one that replaces electrons with light, making it incomparably faster than current chips.

The production line will put China among a handful of countries in the world capable of manufacturing integral silicon photonics chips, which will clear the bottleneck created by the physical limits of conventional chips, according to scientists in Shanghai who made the breakthroughs.

Preliminary testing confirms that the research group has mastered the means to make the chips, and orders are coming in domestically and from abroad, the group's leader said.

"Our long-term goal is to help domestic enterprises stop relying on imports," said Yu Mingbin, leader of the research group at the Shanghai Institute of Microsystems and Information Technology under the Chinese Academy of Sciences. "And the global market also has big potential demand for our product."

The chips are seen as particularly useful in fields relating to information transmission, such as smartphones, computers, autonomous vehicles and ultra high-definition TV. Processing speeds are dramatically improved while energy consumption is reduced.

"Even if we keep updating conventional electronic chips, it's like replacing ox carts with trucks. But now we have a solution more like airplanes, which have obvious advantages in distance and speed," said Yu.

"Individual end users will also notice big changes in their electronic gadgets - faster computers and mobile phones, and less time to download movies," he said.

The Shanghai research team said fewer than 10 companies in the world are currently capable of producing such chips, and all are based in the United States. Some European countries, Singapore and Japan are also racing to master the techniques.

"But it's not too late for China to catch up," Yu said. "We're trying to build a library of devices, the most basic units in a chip. So whatever order we receive, we'll find a solution through some combination of the units."

The Shanghai Municipal Science and Technology Commission said research on the chips became a major city-level science and technology project last year. The city wants to be a world-class base for the novel technology, with the joint efforts of domestic enterprises, universities and research institutes.

Gan Pin, deputy director of the commission, said the city's ultimate goal is to establish at least 20 enterprises that would be responsible for the whole industry chain of the new chips, from design to manufacturing and testing.
 
.
Study turns up large family of organic-inorganic materials that could find low-cost, lightweight applications in electronics
by Mitch Jacoby
JULY 12, 2018 | APPEARED IN VOLUME 96, ISSUE 29

09629-leadcon-2strucs.jpg
Credit: Yu-Meng You/Southest U.
A metal-free perovskite (left) resembles barium titanate (right) in terms of structure and ferroelectric properties.

Ferroelectric materials do something unique under an applied electric field: They exhibit spontaneous polarization, meaning positive and negative charges quickly separate within the crystal. That property makes the materials ideal for numerous applications in energy storage, medical imaging, and computer memory that retains information even when a device is off.

By supplementing a design strategy with trial-and-error experimentation, researchers have now discovered a new family of metal-free ferroelectric materials. The finding may pave the way to lightweight, flexible, and low-cost memory devices, capacitors, and other types of electronics (Science 2018, DOI: 10.1126/science.aas9330).

09629-leadcon-3strucs.jpg
Credit: Yu-Meng You/Southest U.
A metal-free triiodide perovskite with a diazabicyclo group and a related bromide compound (top and bottom respectively) resemble barium titanate (middle) in terms of structure and ferroelectric properties.



Common ferroelectrics currently used in devices include barium titanate (BaTiO3 or BTO), lead titanate, and other ceramics with the perovskite structure and composition, which follows the stoichiometry ABX3. Although widely used industrially, these materials are costly to produce and often contain lead and other toxic heavy metals. For those reasons, scientists have searched for metal-free perovskites and have succeeded in making some. But those compounds tend to exhibit weak ferroelectric properties, precluding applications.

Now, a research team led by Yu-Meng You and Ren-Gen Xiong of China’s Southeast University have produced a family of 23 metal-free perovskites, some of which are on par with BTO in terms of ferroelectric properties.

Conventional perovskites consist of large and small metal cations, represented by A and B, respectively, in the general chemical formula. The size difference between the ions is one of the important attributes in these materials’ electronic behavior. Other factors, such as van der Waals forces and intermolecular interactions between ions in the crystal, also play key roles.

Armed with that information, the team developed a design strategy for making metal-free perovskite ferroelectrics with the general formula A(NH4)X3, in which A is a divalent organic cation and X is a halogen. They took the synthesis plan to the lab, where they reacted a large number of organic reagents with inorganic ammonium compounds and halogen acids.

Among the 23 metal-free perovskites produced in the study, one named MDABCO–NH4I3, which contains a diazabicyclo group, looks especially promising. That perovskite exhibits a large spontaneous polarization value, 22 microcoulombs per square cm, which is close to BTO’s value of 26. The new compound also looks to be especially stable in that it remains ferroelectric up to a phase transition temperature of 448 K, exceeding BTO’s limit of 390K.

“These results are stunning,”—a milestone reached 70 years after development of conventional ferroelectric perovskite oxides, says Wei Li, a specialist in inorganic functional materials at Nankai University. Li adds that the organic content, which can be modified to tune crystal properties, makes these materials easy to synthesize, lightweight, and inexpensive—all of which may soon lead to applications.

Chemical & Engineering News
ISSN 0009-2347
Copyright © 2018 American Chemical Society​



Perovskite ferroelectrics go metal-free | Chemical & Engineering News

Heng-Yun Ye, Yuan-Yuan Tang, Peng-Fei Li, Wei-Qiang Liao, Ji-Xing Gao, Xiu-Ni Hua, Hu Cai, Ping-Ping Shi, Yu-Meng You & Ren-Gen Xiong. Metal-free three-dimensional perovskite ferroelectrics. Science (2018). DOI: 10.1126/science.aas9330
 
.
chinese people on social media don't like this at all. they are demanding government to shoot it down, calling deephi and all the people that approved this deal are traitors.

http://www.sohu.com/a/241849706_324615?_f=index_betapagehotnews_1


medium.com
FPGA-Maker Xilinx Buys Chinese Chip Startup DeePhi Tech
Synced
3-4 minutes
Xilinx, the world’s leading designer and supplier of programmable logic devices, today announced its acquisition of DeePhi Tech — a Beijing-based chip unicorn with a focus on machine learning, specializing in deep compression, pruning, and system-level optimization for neural networks.

Xilinx is best known for inventing Field Programmable Gate Arrays (FPGAs), a type of processor particularly powerful in small-scale but intensive data access. The FPGA chip allows users to program the circuit path through its tiny logic block to handle any kind of digital function. With advantages in low-latency streaming and data-intensive tasks, FPGAs are well suited to cloud computing industries. Microsoft has been deploying FPGAs in its Azure servers for many years. AI researchers believe FPGAs have huge potential in the area of training, where Nvidia’s GPU is currently a dominant player.

(From left to right) DeePhi’s Song Han, Song Yao, Yu Wang, Yi Shan.
Xilinx has been keeping an eye on DeePhi since it was founded in March 2016. DeePhi provides end-to-end solutions using a deep-learning processing unit (DPU) platform and Deep Compression, a technique that aims to compress neural nets by an order of magnitude without losing prediction accuracy. Deep Compression was introduced by DeePhi Chief Scientist Song Han in 2015. The company also specializes in pruning and system-level optimization for neural networks. DeePhi’s founders and employees are mostly from Tsinghua University.

DeePhi has attracted approximately US$100 million in three financing rounds from investors including GSR Ventures Capital, Samsung Venture Capital, Xilinx, MediaTek, and Ant Financial. It is China’s upcoming unicorn in AI hardware, alongside Cambricon Technologies and Horizon Robotics.

On May 22nd, 2017, Xilinx announced an investment in DeePhi after the company won the Best Paper award at the FPGA 2017 with ESE: Efficient Speech Recognition Engine with Sparse LSTM. The DeePhi engine performed 43 times better than CPUs with 40 times the performance per unit of power. Performance was three times that of GPUs with 11 times less power consumption.

DeePhi’s technologies have already been deployed on Xilinx FPGAs, as one of the company’s earliest product solutions.

Financial terms of the transaction were not disclosed. After the acquisition, DeePhi’s 200 employees will continue to operate out of the same Beijing offices.

“We are excited to continue our strong partnership with Xilinx and work even more closely to deliver leading machine learning solutions to our customers in China and around the world,” said DeePhi CEO Song Yao in a statement released today.

“Xilinx is accompanying DeePhi Tech along its journey to explore the potential of machine learning and is supporting our innovation as one of our early investors. We look forward to continuing our joint efforts with Xilinx to bring our solutions to the next level in performance,” said DeePhi CTO Yi Shan.
 
. . .
Back
Top Bottom