2022 China AI Chip Industry Research Overview

With the rapid development of the emerging AI industry, traditional chips can no longer meet the on-chip performance and computing power requirements of the AI ​​industry. Therefore, building efficient AI chips and effectively combining chip technology with AI technology has become a hot topic. EqualOcean believes that AI chips, as the foundation and core of AI and related applications, are areas with potential. In the first quarter of 2022, EqualOcean released the 2022 China AI Chip Industry Research Report, which analyzes the main types of AI chips in detail, shows the development status of China’s AI chip industry, and explores its development challenges and opportunities.

Trends in the AI ​​chip industry

The AI ​​algorithm must be implemented on the computer equipment, and the chip is the essential part of the operation of the computer equipment. The development of AI chips mainly depends on two areas: the first is the mathematical model and the algorithm established by imitating the human brain, and the second is the semiconductor integrated circuit, namely the chip. Advanced algorithms require sufficient computing power, i.e. the support of high-performance chips. The development of on-chip AI is divided into three stages: in the first stage, due to insufficient computing power of the chip, the neural network algorithm could not start; In the second stage, the computing power of the chip is improved, but it still cannot meet the requirements of the neural network algorithm; In the third stage, the GPU and AI chips of the new architecture promote the launch of AI. With the emergence of the third generation of neural networks, China has gradually bridged the barrier between neuroscience and machine learning, and AI chips are developing closer to the human brain.

In 2021, the 14th five-year plan (2021-2025) for national economic and social development and the long-term goals until 2035 pointed out that during the 14th five-year plan (2021-2025), the AI ​​of new China’s industry generation will focus on key areas such as high-end chips. It has established a good policy environment for the AI ​​chip industry from a national perspective. According to their backgrounds, all localities have also released plans to promote the AI ​​chip industry. As of September 2021, more than 20 provinces including Beijing, Tianjin, Shanghai, Jiangsu and Fujian have issued AI-related policies to further support and guide the development of the AI ​​and chip industry.

In terms of industry data, compared to 2020, the number of AI investments has decreased, but the scale of single investment shows an upward trend. The AI ​​chip industry also continues to have capital inflows, and the amount of one-time financing exceeds CNY 100 million. As of January 2022, there have been 92 financing events in AI chip-related fields in China in 2021, with around CNY 30 billion. Political support and market demand remain the main drivers for the development of AI chips. According to EqualOcean calculations, in 2025 the market scale of China’s AI core industry will reach 400 billion yuan, of which the market scale of core layer chips and related technologies will be about 174 billion yuan.

Breaking down the AI ​​chip industry

1. Technical layer

AI chips can be divided into GPUs, FPGAs, ASICs, and brain-inspired chips based on their technical architecture; According to their location in the network, they can be divided into cloud AI chips, edge and edge AI chips; According to their purpose in practice, it can be divided into training chip and information chip.

AI hardware acceleration technology has gradually matured. In the future, more innovations may come from combining technologies at the circuit and device level, such as in-memory processing and brain-inspired computing, or for unique computing modes or new designs, such as than spark matrix computing and approximate computing; Or optimize the architecture for the characteristics of the data rather than for the model. At the same time, if the algorithm does not change significantly, according to the main AI acceleration methods and the development trend of semiconductor technology, it will soon reach the limit of a digital circuit . It will be based on approximate, analogical, even material or fundamental research innovations.

456.jpg.jpg

2. Application layer

With the maturity of technology, the application scenarios of AI chips will be in the cloud and large data centers and move to the edge with computing power and be deployed in smart homes, smart manufacturing, digital finance and other fields. At the same time, with the increasingly rich types of smart products, smart terminals, smartphones, security cameras and self-driving cars will become increasingly popular.

Currently, most AI training and reasoning workloads occur in the public cloud and private cloud, and the cloud remains the center of AI. Driven by the demand for privacy, network security, and low latency, the AI ​​training and reasoning workload on gateways, devices, and sensors is emerging in the cloud. Better computer chips and a new AI learning architecture will be key to solving these problems. The internet is an industry with high demand for cloud computing power. Therefore, in addition to traditional chip companies, chip design companies and other participants, Internet companies have joined the AI ​​chip industry to invest in self-developed cloud AI chips.

According to EqualOcean data calculation, the large-scale growth of China’s self-driving industry will reach 24% in 2022, and the growth of smart camera product shipments will exceed 15%. Shipments of smart products such as cellphones, tablets and VR/AR glasses have also increased significantly, increasing the demand for smart chips. At the same time, the types of intelligent terminal products are also gradually diversifying. Consumer hardware such as smart audio, service/commercial robots, industrial/digital control equipment and communication products are increasingly abundant. Different product types also place more demands on chip performance and cost.

Opportunities and Challenges

In chip design and manufacturing, China still lacks design software, and there is still a gap between advanced processes and equipment and the world’s top level. Some products and equipment in this area are still very dependent on imports. The quantity and quality of data determine the accuracy of the AI ​​model. At present, most of the data usually belongs to different institutions or departments, such as government departments, financial industry or medical industry. It is difficult to integrate them into a whole, which has caused significant obstacles to improving AI technology.

After fully recognizing the importance of data, local governments have set up big data management offices to effectively use data insecurity, government affairs, legal affairs and other government-level areas. At the same time, formulate better data management policies to make data better serve the local economy and effectively solve the data island problem. Since the outbreak, people have paid more attention to the network and accumulated more data. In 2021, the big data of major Internet companies reached thousands of Pb levels, the data volume of major enterprises in traditional industries also reached the Pb level, and the data generated by individuals reached the TB level. Currently, China accounted for 23% of global data volume in 2018 and is expected to reach 27.8% in 2025.

China’s digitalization transformation direction has led to the gradual improvement of the underlying technology, and the international influence is also increasing year by year. At the same time, it has gradually established a dominant position in big data, chip design and application launch. Industrial development has also attracted more foreign talent to return home for entrepreneurship and employment. The industry chain structure could be rebuilt in the future, and more enterprises, universities and organizations could form a joint force to jointly promote the new development of AI and chips. At the same time, AI-related applications will also be widely used to develop healthcare, education, media, finance and customer service. Manufacturing and transportation will be important application scenarios for AI.

At the end of the line

Overall, the development of AI chips still needs the accumulation and precipitation of basic science. Therefore, the integration of industry, university and research is an effective way. Make full use of enterprises, universities, scientific research institutions and other different educational environments and teaching resources, combine the teaching of theoretical knowledge with the practice of industrial engineering and the practice of scientific research, cultivate and accumulate high-quality talents in the field of AI, and maintain the sustainable development of China’s AI and chip industry.

View the full report here.

Comments are closed.