AI chips provide basic computing power for artificial intelligence, divided into GPUs, FPGAs and ASICs, with the rapid development of AI technology, AI application scenarios will expand to intelligent manufacturing, home, education, medical and financial industries; The continuous upgrading of smart product categories also puts forward higher requirements for chip computing power, AI chip computing power and demand will increase greatly, according to Frost & Sullivan estimates, the global AI chip market size will grow at a compound growth rate of 26.3% from 2022 to 2026.
AI chips are also known as AI accelerators or compute cards, which are modules dedicated to processing a large number of computing tasks in AI applications (other non-computing tasks are still handled by the CPU). At present, AI chips are mainly divided into GPU, FPGA, ASIC. The GPU, graphics processor, and GPU are currently the leading chips for AI applications, because it has strong computing power and high universal applicability, and the development cost is low and widely used.
As an AI model with a large number of complex computing needs, ChatGPT has become one of the key hardware that can help it complete computing tasks. With the development of deep learning technology, GPUs have become an important hardware component in AI algorithm training. At the same time, with the development of cloud computing technology, the requirements for GPU computing power in the cloud are constantly increasing.
In recent years, the state has paid great attention to the development of the artificial intelligence chip industry and issued a series of industrial support policies, establishing an excellent policy environment for the artificial intelligence chip industry and promoting the development of the industry. Under the guidance and support of policies, the Chinese smart chip market continues to develop rapidly. At present, in the Chinese intelligent industry chain, the proportion of application-layer enterprises has reached 83%; The proportion of technical enterprises is 5%; The proportion of application-tier enterprises is 12%.
AI chips are the basic industries in the artificial intelligence industry chain and are very important for the development of artificial intelligence-related industries, so ensuring the stability of the AI chip market and improving the supply capacity is very important for social and economic development. In the future, with the increase in demand for intelligent transformation and upgrading of mainland industries, the demand for AI chips will also usher in upgrades. At present, the gap between the mainland AI chip technology and the world's advanced level is still large, and the space for future technological progress and market scale expansion is large.
Through a lot of time to review and sort, 5 "AI chip + GPU" potential leading companies were screened, which are worth collecting and studying!
Hengshuo shares
The company's storage-computing integrated AI chip is still in the research and development stage, has not achieved mass production and sales, and the fundraising project CiNOR storage-computing integrated AI reasoning chip focuses on the following four terminal market application scenarios: face recognition, voice keyword recognition, electrocardiogram detection and power equipment fault voiceprint detection.
Obit
In order to implement the strategic deployment of the company's artificial intelligence (AI) business, improve the company's AI chip design capabilities and AI algorithm data processing capabilities, it is planned to invest no more than 50 million yuan to establish a wholly-owned subsidiary Guangdong Eastern European Bit Artificial Intelligence Research Institute Co., Ltd. in Zhuhai High-tech Zone. The company and Beijing Aerospace Automatic Control Research Institute reached a cooperation agreement in improving the production capacity of supporting products, commercial space launch service consulting, and AI chip research and development.
Yuntian Lifei
In the field of AI chips, the company's visual artificial intelligence neural network processor chip is a neural network processor chip for terminal-side and edge-side visual artificial intelligence processing.
The dawn of science and technology
Sun Ninghui, director of the Institute of Computing of the Chinese Academy of Sciences, said that this year it is planned to embed it in the form of a "Cambrian" IP box on the client and mobile phone. Angel investment funds are available to develop a cloud-based chip. The cloud-based chip is expected to launch next year. The collaboration between Dawning and Cambrian will make deep learning technology available to more industries and facilitate mass production and supply.
The last one: On the basis of having the characteristics of ultra-low sleep power consumption, low bandwidth, long battery life and wide coverage that are different from traditional modules, the new IM2210-NB also has two advantages, one is small size, and the other is Chinese core. This module is provided by Zhanrui under Unigroup, and New H3C provides technology research and development and product research capabilities. The release of the new module allows the communication module to truly achieve national production, and protect information security from the lowest chip.
June has the most potential for the following reasons:
Reason 1: At present, the company's data center is getting bigger and bigger, and its performance has also ushered in a significant improvement, and it has expanded its edge computing data center business in Shanghai.
Reason 2: The company integrates market hot spots, low + low price + small cap stocks, concentrated chips, and the main force continues to increase positions.
Reason three: The valuation of the stock is very low, but also at the bottom sideways for 4 years, the current main force has increased its position significantly, the stock price is also a breakthrough platform, chips are concentrated, I am also optimistic that this vote is expected to start the main rising wave and become the next Topway information!