Source: China Fund Network

Abstract: Today's DeepSeek concept stocks deduce the tide of daily limit, and its market focus has been focused on the related concept stocks that have been connected to the DeepSeek large model for successful adaptation, and it is reported that Shanghai Tiantian Zhixin Semiconductor Co., Ltd., in which Shenzhou High-speed Rail (000008) participates, and Gitee AI officially launched the DeepSeek R1 model service.

During the holidays, the DeepSeek storm swept the world! According to media reports, it ranked first in the app store download charts in 140 markets around the world. In this context, on the first trading day after the Spring Festival of A-shares, the concept of DeepSeek has become the hottest hot outlet in the A-share market! Judging from Wednesday's market, Daily Interactive (300766), Zhejiang Oriental (600120), Huajin Capital (000532), Meig Intelligent (002881), Parallel Technology (839493), Hangzhou Iron and Steel Co., Ltd. (600126), Ankai Micro (688620), Anheng Information (688023), Qingyun Technology (688316), 360 (601360), Tianyu Digital Technology (002354), Montnets Technology (002123), Nanwei Software (603636), Tors (300229), Wondershare Technology (300624), Sunshine Lighting (600261), Philips (300287), Zhixin Precision (301512), Hangjin Technology (000818), Jingyeda (003005), Super Communication (603322), UCD (688158), Hengwei Technology (603496), Topway Information (002261), Yunsai Zhilian (600602), Meiliyun (000815), Capital Online (300846), Sichuan University Zhisheng (002253), Weining Health (300253), Dataport (603881), Kaipu Cloud (688228), Insigma (600797), Oriental Materials (603110), Kunlun Wanwei (300418), Danghong Technology (688039), PIA (688306), Kingsoft Office (688111), Beixinyuan (300352), The stock prices of nearly 50 DeepSeek concept stocks, including Yunchuang Data (835305), Qianxin (688561), Yuntian Lifei (688343), Runhe Software (300339), Kailer (301070), and iSoftStone (301236), are closed at the limit or up by more than 10%! Obviously, the DeepSeek concept stock, which has attracted much attention recently, has become the hottest hot direction in the current A-share market! Judging from Wednesday's list of gainers, the main line of DeepSeek concept stocks is concentrated in the relevant listed companies that have been connected to the DeepSeek large model for successful adaptation, including Daily Interactive (300766), Qingyun Technology (688316), Hangzhou Iron and Steel Co., Ltd. (600126), 360 (601360), Anheng Information (688023), Ankai Micro (688620), MeiG Intelligent (002881), Montnets Technology (002123), Wondershare Technology (300624), Tors (300229), Kingsoft Office (688111), Parallel Technology (839493), etc. Therefore, the logic of the market is worth paying attention to. For example, the stock price of Parallel Technology (839493), which has risen by nearly 60% in the past three trading days, may be due to the company's cooperation with Shanghai Tiantian Zhixin Semiconductor Co., Ltd. to fully carry out the adaptation and launch of DeepSeek's multiple models, including R1. Then, as a listed company that participates in Shanghai Tiantian Zhixin Semiconductor Co., Ltd., Shenzhou High-speed Railway (000008) may also be worth paying attention to.

DeepSeek concept stocks have become the hottest hot spot in the A-share market after the Spring Festival: the main line of leading the rally is concentrated in the relevant listed companies that have been connected to the DeepSeek large model for successful adaptation. During the Spring Festival, the general large model represented by DeepSeek ushered in explosive attention in China, and the usage surged by 300% compared with the same period last year. This AI boom has not only swept the technology circle and capital market, but also penetrated into the daily lives of ordinary people, bringing a new technological experience to traditional festivals. Moreover, the DeepSeek storm has swept the world, and according to media reports, it ranks first in the app store download charts in 140 markets around the world. Liang Wenfeng, the founder of DeepSeek, returned to his hometown for the New Year and was welcomed like a hero, with banners and arches in his hometown to welcome. Securities and investment banks held meetings overnight, more than 60 roadshows in 8 days, and even Chinese New Year's Eve and Spring Festival were still on the roadshow to issue research reports.

DeepSeek, founded on July 17, 2023, as Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd., focuses on the development of cutting-edge large language models (LLMs) and related technologies. On January 20, DeepSeek unveiled its inference model, DeepSeek-R1, which quickly attracted the attention of the global industry. The model cleverly uses reinforcement learning technology in the post-training stage, and although there is only a small amount of labeled data, it still significantly improves the inference ability. In math, code, natural language reasoning and other tasks, DeepSeek-R1's performance has been comparable to OpenAI's model.

On the same day, NVIDIA was the first to announce that its NVIDIA NIM service now supports the DeepSeek-R1 model. This was followed by Amazon and Microsoft confirming the inclusion of the model in Amazon Web Services and Azure AI Foundry. The implementation of these platforms will undoubtedly open the door to a wider range of applications for enterprises using DeepSeek-R1, and provide users with access to the most cutting-edge AI technology.

DeepSeek's breakthrough has sparked a strong reaction from the global AI industry. On January 31, the three major U.S. technology giants, Nvidia, Amazon and Microsoft, gathered on the stage of artificial intelligence technology competition and jointly announced access to the newly launched Chinese large language model DeepSeek-R1. This move is seen as an important milestone for the global AI industry, as well as China's rise in the field of artificial intelligence and the improvement of its technological strength. HUAWEI CLOUD also announced on February 1 that it has partnered with the Silicon-based mobile team to launch the DeepSeekR1/V3 inference service based on HUAWEI CLOUD's Ascend cloud service. According to the statistics of the domestic AI product list, the daily activity of the DeepSeek application (excluding website data) has surpassed the daily activity of ChatGPT in the same period after its launch within 5 days, becoming the fastest growing AI application in the world. At present, the four major cloud giants in China have officially supported DeepSeek, and overseas cloud giants such as AWS and Microsoft Intelligent Cloud have officially announced their support.

DeepSeek's "low-cost + open-source" model has had a significant impact on the technology stock market. There is no doubt that DeepSeek has become a global phenomenon.

Judging from Wednesday's list of gainers, because there are almost no A-share listed companies that have an obvious equity relationship with DeepSeek, the main line of DeepSeek concept stocks is concentrated in the relevant listed companies that have been connected to the DeepSeek model for successful adaptation, including Daily Interactive (300766), Qingyun Technology (688316), Hangzhou Iron and Steel Co., Ltd. (600126), 360 (601360), Anheng Information (688023), Ankai Micro (688620), MeiG Intelligent (002881), Montnets Technology (002123), Wondershare Technology (300624), Tors (300229), Kingsoft Office (688111), Parallel Technology (839493), etc.

As shown in the chart above, the next step of the A-share market's market hype around the concept of DeepSeek will undoubtedly continue to unfold along this main line! Therefore, some opportunities with very hard logic to make up for the rise are also worth paying attention to. For example, the stock price of Parallel Technology (839493), which has risen by nearly 60% in the past three trading days, may be due to the company's cooperation with Shanghai Tiantian Zhixin Semiconductor Co., Ltd. to fully carry out the adaptation and launch of DeepSeek's multiple models, including R1. Then, as a listed company that participates in Shanghai Tiantian Zhixin Semiconductor Co., Ltd., it is worth paying attention to.

Shanghai Tiantian Zhixin Semiconductor Co., Ltd. and Gitee AI officially launched the DeepSeek R1 model service, and the company cooperated with Parallel Technology (839493) to fully carry out the adaptation and launch of DeepSeek's multiple models, including R1: The stock price of Parallel Technology (839493) has risen by nearly 60% in the past three trading days

According to the latest report from Shanghai Securities News, on February 4, Shanghai Tiantian Zhixin Semiconductor Co., Ltd. (hereinafter referred to as "Tiantian Zhixin") and Gitee AI jointly announced that with the efficient cooperation of both parties, it took only one day to successfully complete the adaptation with DeepSeek R1, and has officially launched a number of large model services, including DeepSeek R1-Distill-Qwen-1.5B and DeepSeek R1-Distill-Qwen-7B, DeepSeek R1-Distill-Qwen-14B, etc.

Not only that, Tiantian Zhixin is actively working with many partners such as Parallel Technology (839493) and Computing Power Interconnection to fully carry out the adaptation and launch of DeepSeek's multiple models, including R1, to ensure that users can enjoy the latest models in a more timely and convenient manner. The parallel technology platform that will be launched in the future is not only for ordinary users and developers, but also an AI service platform mainly for enterprise users, allowing business users to enjoy the latest and most convenient AI services in the cloud.

Shanghai Tiantian Zhixin Semiconductor Co., Ltd. is a provider of general-purpose GPU high-end chips and super computing power systems. The company is committed to developing independent, controllable, high-performance general-purpose GPU products, building an industrial ecology, and providing high-end computing solutions for the whole industry.

Tiantian Zhixin's general-purpose GPU products have the characteristics of independent controllability, high performance, versatility, flexibility, etc., support mainstream AI ecosystems and various deep learning frameworks at home and abroad, and have supported the implementation of hundreds of scenarios of artificial intelligence large models in dozens of industries. The application scenario public service platform project of Tiantian Zhixin National Innovation Application Pilot Zone was supported by the Ministry of Industry and Information Technology.

The R1-based adaptation of Tiantian Zhixin and Gitee AI took only one day, which once again demonstrated the advantages of Tiantian Zhixin's mature software ecology and general-purpose GPU architecture, and highlighted the characteristics of Tiantian Zhixin's general-purpose GPU products that are easy to use, easy to use, out-of-the-box, and zero-cost migration. This is an important breakthrough in GPU technology innovation and ecological construction in China, which will promote the long-term development of China's AI independent computing power.

Previously, as soon as DeepSeek V3 671B was released, Tiantian Zhixin responded quickly and completed the adaptation with the model. After rigorous testing and verification, the accuracy of the adapted model is highly consistent with the accuracy of the paper, showing excellent stability and reliability, which can be quickly put into application scenarios and provide stable and efficient service support for our customers.

The

adaptation of domestic GPUs and DeepSeek can realize the deep integration of deep learning frameworks and domestic independent hardware, give full play to the advantages of domestic GPU computing power, improve the computing efficiency of DeepSeek model training and inference, and promote the independent and controllable development of the domestic artificial intelligence industry chain, reduce dependence on foreign hardware platforms, reduce technical risks and costs, and help domestic GPUs expand the market, attract more upstream and downstream enterprises to cooperate, improve the industrial chain ecology, and promote the acceleration of application implementation.

In the face of technological innovation, market opportunities and industrial chain synergies brought by DeepSeek R1, Tiantian Zhixin actively responded and made every effort to promote technological breakthroughs and business expansion in the field of GPU chips in China. In the future, Tiantian Zhixin will work with industry partners such as DeepSeek, Gitee AI, Parallel Technology (839493), and Computing Power Interconnection to jointly accelerate the process of autonomy in China's AI and GPU fields and enhance China's position in the global technology competition.

It is worth noting that the direct reason for the stock price of Parallel Technology (839493), which has risen by nearly 60% in the past three trading days, is that the company has cooperated with Shanghai Tiantian Zhixin Semiconductor Co., Ltd. to fully carry out the adaptation and launch of DeepSeek's multiple models, including R1. Then, as a listed company with a stake in Shanghai Tiantian Zhixin Semiconductor Co., Ltd., Shenzhou High-speed Rail (000008) is worth paying attention to. According to public information, Shenzhou High-speed Railway (000008) indirectly participates in Tiantian Zhixin through Beijing Zhongguancun M&A Fund of Funds Investment Center (Limited Partnership).

From the equity penetration chart, it can be understood that the largest shareholder of Shanghai Tiantian Zhixin Semiconductor Co., Ltd. is Fujian Dacheng Phase I Investment Partnership (Limited Partnership), with a shareholding ratio of 9.42%. Beijing Zhongguancun M&A Fund of Funds Investment Center (Limited Partnership), in which Shenzhou High-speed Rail (000008) is a shareholder, holds a 15.37% stake in Fujian Dacheng Phase I Investment Partnership (Limited Partnership).

In short, the next step of the A-share market is to revolve around the market hype of the DeepSeek concept, or will continue to revolve around the main line of relevant listed companies that have been successfully adapted to the DeepSeek large model. Then listed companies with related concepts are worth paying attention to, such as parallel technology (839493), Shenzhou high-speed rail (000008), etc.

Sina Statement: This news is reprinted from Sina's cooperative media, and the publication of this article on Sina.com is for the purpose of conveying more information, which does not mean that it agrees with its views or confirms its description. The content of this article is for informational purposes only and does not constitute investment advice. Investors act accordingly at their own risk.