Shulou(Shulou.com)11/24 Report--

The unprecedented breakthrough of the general model has stirred the spring water of artificial intelligence (AI). Oriental Securities believes that the demand for chip computing upgrading brought about by the AI model, the data center industry chain will continue to benefit, and the user experience of AI applications will continue to improve.

According to Gartner, the market for chips used to execute artificial intelligence (AI) workloads is growing at more than 20 per cent a year.

The agency expects the market for AI chips to reach US $53.4 billion in 2023, an increase of 20.9% over 2022 and 25.6% to US $67.1 billion in 2024. AI chip revenue is expected to more than double the market size in 2023 to $119.4 billion by 2027. Gartner analysts pointed out that the future mass deployment of custom AI chips will replace the current dominant chip architecture (discrete GPU) to adapt to a variety of AI-based workloads, especially those based on generative AI technology.

The demand for AI chips continues to be hot.

AI chips, also known as AI accelerators or computing cards, are chips designed to handle AI-related computing tasks. They are an important cornerstone for building computing power. While the construction of large models is in full swing, the demand for computing power of large models will further drive the research and development of AI chips, and the cross-integration with health, education, transportation and other fields will gradually deepen.

When a generative artificial intelligence tool such as ChatGPT is running, it needs to process a large amount of data, which must be retrieved from the memory chip and sent to the processing unit for calculation. In the face of the blowout demand for AI chips brought by the breakthrough development of large models, technology giants and emerging enterprises are rushing into the competition track of the big models, and a new round of competition for AI chips is emerging.

Last month, the net profit of NVDA.US Q2 of AI chips surged 422%, indicating that demand for AI chips continues to be hot. Nvidia CEO Huang Renxun said, "A new computing era has begun." Companies around the world are moving from general computing to accelerated computing and generative artificial intelligence. This quarter, major cloud service providers announced the launch of a large-scale NVIDIAH100AI infrastructure to introduce NVIDIAAI to various industries, and the race for generative artificial intelligence has begun. "

From the data point of view, taking the current mainstream Nvidia A100 GPU as an example, the single training of the large model requires at least 100 calories and kilocalorie scale. According to the current market price, the cost of building a 1000-card GPU computing cluster can reach 200 million yuan.

With the best-selling world's most cutting-edge AI chip H100, Nvidia is undoubtedly one of the most conspicuous "wave riders" in this round of AI. AI craze to drive chip orders, Nvidia plans to ship about 550,000 pieces worldwide in 2023 H100 H100A800GPU delivery Chinese mainland customers or 2024Q1.

With the launch of a new generation of AI chip GH200, Nvidia will continue to expand its product portfolio by leveraging the AI spending of major data center participants. Major data center participants are expected to spend more than $10 billion on AI in 2023.

Intel wants to ride the wave of AI chips

After ChatGPT led the explosion of the entire AI chip market, Invid reached a "hot cake" in everyone's eyes, at a time when INTC.US (Intel) was going deep into the artificial intelligence hinterland, and China was one of the "main battlefields" of its AI chips. Last month, Intel announced the official launch of the second-generation deep learning accelerator, Habana Gaudi2, for the Chinese market, which can be used to accelerate AI training and reasoning.

As one of the key products of Intel AI strategy, Habana Gaudi 2 is expected to break the inherent pattern of the current chip market.

The executive vice president of Intel said that in the past six months, the AI acceleration chip has suddenly become popular again because of the big model, and then we will have more resources to invest in using local teams to speed up the landing of these Chinese customized products.

Construction of AI efficient solution by Weimei Hologram

At present, major chip companies around the world are actively carrying out the layout of AI chips. also on the AI track, as an emerging AI semiconductor supplier, AI Vision Leader WIMI.US is betting on the powerful field of artificial intelligence hardware, committed to providing high-performance, low-power intelligent semiconductor solutions for cloud computing and artificial intelligence, developing and building core competitiveness. Seize the opportunity by overtaking at the corner.

In fact, Weimei holography has devoted itself to the development and application of AI algorithms in recent years, providing customers with comprehensive solutions and achieving technical goals by integrating technical algorithms with software or hardware or both. At present, Weimei holographic has 195 patents related to image processing and display, model input / output, 3D modeling, and 325 patents related to AI, etc., to accumulate leading technology patents, carry out model innovation in the business, and actively enhance the industry value of the general field of digital intelligence processing industry.

While AI continues to be hot, AI chips can help companies achieve large-scale revenue, and Weimei holography can increase horsepower in AI, especially in cloud computing, data centers, autopilot and other areas. At the same time, with the continuous development of AI technology and the continuous improvement of AI chip performance, it is more and more suitable for various application scenarios. It is reported that according to the development trend and technology evolution of AI industry, the large-scale production of AI hardware is also constantly advancing. In the future, Weimei holography will enable more B-end customers to apply more efficient AI solutions.

End

The upsurge represented by large models has brought about an explosive growth in the demand for AI computing power, and AI computing power has become a key factor to promote the development of artificial intelligence. From the perspective of the industry, optimistically speaking, the AI chip is in the embryonic stage of letting a hundred flowers blossom and a hundred schools of thought contend, and each subject is rapidly exploring a suitable growth path in the fierce market competition. And how to further enhance the competitiveness of AI chips? Perhaps we can encourage more enterprises and social capital to enter the field of AI chips, increase investment, speed up the progress of process innovation, build a local supply chain system, and make a breakthrough in the overall development of AI.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.