According to reports from Taiwan media, semiconductor manufacturer MediaTek held a media end-of-year meeting today and disclosed plans for 2018 to the outside world. It is expected that two P-series processors will be launched, taking into consideration performance, cost, and market demand. According to Chen Guanzhou, general manager of MediaTek, the next-generation HelioP series will focus on the development of artificial intelligence and computer vision in the functional area. As we all know, the three pillars of artificial intelligence are hardware, algorithms, and data. Among them, the hardware mainly refers to the chip used to run the AI ​​algorithm. At the end of 2017, we also welcomed an important player in the domestic AI chip manufacturing army. Below, we will take a look at how many manufacturers have announced the relevant AI chip design or production plan this year. Cambrian On August 18, 2017, Cambrian Technology announced that it had completed 100 million U.S. dollars of round A financing, from SDIC Ventures (A round led investment), Alibaba Ventures, Lenovo Ventures, National Investment, Zhongke Turing , Lantern Point Origin (Angel Wheel Collar Investment), Surabaya Investment (Angel Wheel Investor) jointly invested. After this round of financing, the Cambrian technology has become the first unicorn in the global AI chip field. Because 100 million financing became the first unicorn in the AI ​​chip field, plus the release of the Huawei Unicorn 970, the Cambrian results in the field of AI chips are well known. Kirin 970 chip adopts TSMC10nm technology, equipped with the Cambrian AI chip IP, AI computing capabilities compared to the four Cortex-A73 core has about 25 times the performance and 50 times the energy efficiency of 2 advantages, can significantly enhance the mobile phone in image recognition, The capabilities of voice interaction, smart camera, etc., bring more possibilities for AI technology in the application field and bring unprecedented Mate10 mobile artificial intelligence experience. On November 6, the Cambrian held its first press conference and launched its next-generation smart processor IP products. They are the Cambrian 1H8 for low-power scenarios and the Cambrian 1H16 for high performance and versatility, and the Cambrian 1M for terminal artificial intelligence products. Cambrian Technology CEO Chen Tianshi said: “Compared with the Cambrian 1A products, these three new products have been optimized in terms of energy efficiency, power consumption, and cost. Their scope of application covers smart driving, voice recognition, and image recognition. Multiple fields." horizon Horizon was established in 2015, and the company set out to establish a one-stop artificial intelligence solution that defines “All Things Intelligent†to make life easier, more fun, and more secure. The line is dedicated to creating artificial intelligence “brains†based on deep neural networks. Platforms - including software and chips, can achieve low-power, localized solutions to environmental awareness, human-computer interaction, decision-making and other issues. Line> Line> After two years of unremitting efforts, on December 20th, 2017, the world’s leading artificial intelligence startup Horizon successfully announced China’s first global leading embedded artificial intelligence “China Core†for smart driving Journey 1.0 processor. With the Sunrise 1.0 processor for smart cameras, it also released artificial intelligence solutions for smart driving, smart city and smart business applications. Yu Kai, founder and CEO of Horizon, said: “The AI ​​chips developed by Horizon are mainly used to solve specific problems in real-life scenarios. Only in this way can AI maximize its potential.†Shen Jian Technology In March 2016, Shenzhen Jian Technology was officially established. In April, he received angel round financing from Jinsha River Venture Capital and Gaochun Capital. In May, Shen Song’s chief scientist Han Song’s paper on “deep compression†technology won the best paper award at the ICLR Conference. In early 2017, Shenzhen Sci-Tech won the tens of millions of USD A round of financing led by Xilinx and MediaTek. △ Shen Jian Technology founder team, from left to right as the chief scientist Han Song, co-founder, CEO Yao Yao, co-founder Wang Yu, partners, CTO orders On October 24th, 2017, Shenzhen Shenjian Technology held a new product launch event and released a total of six new products, which were used in the security field. At the same time, it announced the completion of the A+ round of financing information. At present, Shen Jian Technology's products are mainly used in two major areas of security monitoring and data center. On the one hand, SZ Technology's solutions can provide services for security and data centers, and these two industries are currently relatively mature in all application fields, with faster revenues and larger market totals. On the other hand, Shen Jian Technology has channel advantages in these industries. According to Yao Song, Shen Jian Technology has already reached cooperation with many top brand manufacturers in the intelligent security field. As CEO Yao Song said: "Based on a more efficient artificial intelligence chip platform, we can put artificial intelligence algorithms into more scenarios and societies. We want to see that everything is connected through artificial intelligence chips. Future society." Yun Tian Li Fei Yun Tian Li Fei is an AI startup company that focuses on intelligent video and image recognition. It independently researches and develops AI chip intellifusion. At present, Yun Tian Li Fei's business focuses on the security field of Ping An City. By assisting the customer in collecting data, the tagged database is extracted from the data, which in turn feeds back and trains the deep learning algorithm, constantly optimizing and iterating the accuracy of the algorithm. In the process of designing the algorithm and the chip, they combined again, which can improve the computing power of the chip and reduce the complexity of the algorithm. In short, from algorithms, chips, massive data, back to the optimization algorithm to form an organic closed loop, and iteratively upgrade. "The core of Yun Tian Li Fei is to have a chip-level solution, which can reduce the hardware platform demand for the server. This overall solution can improve the performance of the chip by two orders of magnitude, and the cost is lower, and it is easier to scale. , Industrialization, and the larger the scale, you can collect more data." Chen Ning said. On March 15th, 2017, Yuntian Lifei announced the acquisition of tens of millions of dollars in Series A financing. The investors include Shanshui Yirong Media Investment Co., Ltd., Songsong Capital, Shenzhen Investment Co., Ltd., and investment in Donghai and Hongxiu Yingxin. mechanism. Yun Tian Li Fei mainly focuses on the field of visual intelligence. Its main product is through China's first commercial artificial intelligence chip, to achieve visual identification and big data analysis, that is, the public generally understand the "sky eye system." ThinkForce According to public information, ThinkForce (Shanghai Yuzhi Electronic Technology Co., Ltd.) was established in 2017 and was founded by senior experts from chip design, algorithm software and system development. In December 2017, AI chip startup team ThinkForce announced the completion of A-round financing of 450 million yuan from Yitu Technology, Yunfeng Fund, Sequoia Capital, and Gaochun Capital. To design a smart chip that integrates first-class AI algorithms and advanced manufacturing processes, and to build an artificial intelligence hardware platform to provide a one-stop industrial application solution. According to reports, the AI ​​chip that ThinkForce plans to introduce is based on the industry's advanced semiconductor process technology and uses the independently researched and developed micro-kernel ManyCore architecture to complete the implementation of AI cloud virtualization scheduling at the chip level. This architecture improves the elastic computing and scheduling of AI cloud. An order of magnitude, CPU-like virtualization brings multiple cost savings to the elastic scheduling of cloud computing. At the same time, this technology, combined with self-developed firmware and TFDL software SDK, can accelerate the calculation of various neural network models and achieve more than 5 times power and cost savings compared to Nvidia mainstream computing cards. It is reported that ThinkForce has already cooperated with many companies including IBM and Cadence. The field of artificial intelligence is surging, heroic competition, innovation and development, and the golden era of AI chips has begun.