The A800 Tensor Core GPU from NVIDIA is a card for deep learning acceleration. This product has been in production since the third quarter of 2022 and has recently received a lot of attention. According to 36Kr, the price of NVIDIA's A800 chip has risen sharply. It is understood that the market price of NVIDIA A800 80GB PCIe standard version GPU was 90,000 yuan 15 days ago, and now it has risen to about 110,000 yuan. Despite soaring prices, NVIDIA's high-end chips are still "sought-after".
AI chips are the core of AI computing power, and as the amount of data continues to grow, algorithm models become more and more complex, processing objects are diversified, and the requirements for computing performance are getting higher and higher. In this case, AI chips are gradually becoming the most ideal choice for handling various computing tasks in AI applications. Compared with CPUs, GPUs are better at parallel computing, which also makes GPUs widely concerned and highly sought after in the AI hardware market. With the widespread application of AI technologies such as ChatGPT, the demand for AI computing hardware is also rising.
Scientific research institutions in the mainland are also actively carrying out self-research work on large models, and the speed of self-research is accelerating. Therefore, due to the continuous rapid growth of the number of algorithms, the demand for computing power is still strong. It can be said that if the product is the "traffic entrance" for AI-empowered enterprises to open the second round of performance growth curves, then computing power is the key to the technology giants competing for the "entry ticket". Global technology giants such as Microsoft and Google have entered an important stage of the AI battle, using large models to enhance the competitiveness of their products and start a new round of growth cycle. The high-demand model of large models will inevitably put forward higher requirements for the underlying computing power.
Impact on domestic related manufacturers
Due to foreign control and restrictions on related chips, resulting in the increasing cost of domestic AI-related enterprises, more and more domestic AI companies have begun to choose other computing power chips to replace NVIDIA or AMD's computing power chips, such as many companies have begun to use Huawei's Ascend chips. Although the total performance is not as good as NVIDIA's computing power chip, it is already a good substitute for comprehensive cost, acquisition difficulty, and computing power.
Therefore, whether it is the previous chip regulation or this price increase, it is actually an opportunity for domestic computing power chip companies, and domestic related enterprises will accelerate the research and development of computing power chips and will also get more share in the market.
And the author also brought you several computing power chips and related companies after in-depth review and investigation, which are worth collecting and studying.
Beijing Junzheng
The company has strong technical capabilities in computing technology, related chip products in the computing power continues to improve, at the same time, the company's chip products in recent years AI computing power continues to improve, A algorithm technology continues to enrich, well meet the market demand for these two types of chips in AI performance.
Hengshuo shares
The company is developing the second CINOR storage and computing integrated AI acceleration chip, the control engine of the chip requires an MCU for control and scheduling, the company CINOR can be verified and corrected on its own MCU platform, greatly reducing the development cycle of CINOR AI chip, can provide customers with "MCU + memory + AI" product solutions.
Rockchip
Leading AloT chip supplier in China: The company has rich experience and technical reserves in large-scale SOC chip design, digital-analog hybrid chip design, image signal processing, high-definition video coding and decoding, artificial intelligence and system software development, and has formed multi-level, multi-platform and multi-scenario professional solutions; The company's main product intelligent application processor chip, is a kind of SOC chip, which is a system-level ultra-large-scale digital IC.
Cambrian
The company focuses on the research and development and technological innovation of artificial intelligence chips, and is committed to building core processor chips in the field of artificial intelligence, which can provide basic computing platforms for various artificial intelligence technologies such as vision, speech, natural language processing, and traditional machine learning technology; The Cambrian 1A processor developed by the company is the world's first commercial terminal intelligent processor IP product, and the Siyuan 100 (MLU100) chip is China's first high-peak cloud intelligent chip.
The last one
The average cost in the near future is 76.78 yuan. The company's operating conditions are acceptable, and most institutions believe that the long-term investment value of the stock is high, which can be paid more attention.
Leading domestic high-end processor enterprises; The company's processors have advanced technology, excellent system architecture, rich software and hardware ecology and other advantages, rich software and hardware ecology, and can be widely used in big data processing, artificial intelligence, business computing and other applications.
Disclaimer: The content of the article is for informational purposes only and does not constitute investment advice
Risk warning: The stock market is risky, and you need to be cautious when entering the market