Recently, DeepSeek's big model products have rapidly gained popularity worldwide, effectively lowering the threshold for big model applications and attracting industry attention. Experts said that DeepSeek's success shows that the pace of innovation in China's artificial intelligence industry is accelerating based on the advantages of the overall industrial system layout, system collaborative optimization, rich scene applications, etc. Since December 2024, DeepSeek has successively released the open-source big language model DeepSeek-V3 and inference model DeepSeek-R1. Due to their low cost, high performance, and open source characteristics, they quickly topped the free app download rankings in Apple's US and China app stores. Many domestic and foreign technology giants, as well as government, education, healthcare, and other fields, have joined the DeepSeek model. According to analysis, the training cost of the DeepSeek-V3 model is less than one tenth of that of the OpenAI GPT-4 model, and the training cost of the DeepSeek-R1 model is only 3% to 5% of that of the OpenAI o1 model. In terms of reasoning, the cost of using the DeepSeek-R1 model is one thirtieth of that of the OpenAI o1 model. In addition, DeepSeek has effectively improved performance at a lower cost through engineering technology innovations such as algorithm optimization and efficiency enhancement, breaking the traditional path of "stacking computing power and data consolidation". Han Jian, Director of the Institute of Information Technology and Software Industry at CCID Think Tank, believes that the success of DeepSeek marks a "low-cost, high-performance" innovation path for China in the field of large models, providing a new development paradigm for global artificial intelligence. "China's AI technology innovation, represented by DeepSeek, has made another breakthrough, not only making outstanding performance in the global big model performance competition, but also contributing China's strength to the global big model open source. At the same time, it has significantly reduced the cost of large-scale model application, and pushed China's AI large-scale model technology into a new development stage of large-scale application. " Zhao Gang, the director of the Saizhi Industry Research Institute, said. Han Jian believes that DeepSeek's contribution in reducing the cost of big models will effectively promote the empowerment of various industries with artificial intelligence technologies represented by big models, and the "AI+" action will accelerate its deepening and implementation. In addition, DeepSeek's open and open-source development path attracts upstream and downstream enterprises in the industry chain to quickly follow up, which is expected to promote the formation of a chain innovation trend of collaborative breakthroughs in chips, cloud computing, models, and accelerate the construction of a good industrial ecosystem. Lenovo Group Chairman and CEO Yang Yuanqing stated that artificial intelligence technology with higher efficiency and lower cost has accelerated the maturity process of personal intelligence, especially end side intelligence and edge side intelligence, and has also promoted and accelerated the deployment and application of artificial intelligence in enterprises. At present, Lenovo's artificial intelligence computers, artificial intelligence smartphones, and artificial intelligence tablets have fully integrated with the cloud based DeepSeek-R1 networked full blood version of the big model, and will be the first to deploy the DeepSeek big model locally on the end side. Enhancing the capability of the entire industry chain is the key to industrial competition. Zhao Gang said that China's AI industry as a whole is in the first echelon in the world, its development level is second only to the United States, and its industrial chain is complete, its innovation ability is strong, and its application scenarios are rich. China's AI industry covers the whole industrial chain of AI chips, large model algorithms, high-quality data sets, large model applications, embodied agents, etc., forming the Beijing Tianjin Hebei, Yangtze River Delta, Guangdong Hong Kong Macao Greater Bay Area and other AI industrial clusters. Han Jian introduced that in terms of technological innovation, China has become the country with the highest number of artificial intelligence patents in the world, accounting for 61%; China has recorded 302 large-scale models through generative artificial intelligence, and a number of excellent large-scale model products such as DeepSeek have emerged. In terms of factor aggregation, the construction of intelligent computing infrastructure is accelerating, and the data factor market is steadily advancing. According to the Evaluation Report on the Development of China's Artificial Intelligence Computing Power jointly released by IDC and Inspur Information a few days ago, in 2024, the scale of China's intelligent computing power will reach 72.53 trillion times/second, with a year-on-year growth of 74.1%, which is more than three times the growth rate of general computing power in the same period; The market size is 19 billion US dollars, a year-on-year increase of 86.9%. Artificial intelligence empowers the deepening of new industrialization. For example, Inspur Cloud relies on the industry knowledge enhancement and multimodal understanding capabilities of the Knowledge Industry Big Model, combined with technologies such as retrieval enhanced generation and autonomous decision-making intelligent agents, to create safety monitoring and early warning and intelligent customer service applications for Shandong Electric and Electric Group. Through the intelligent analysis and decision-making system driven by large models, the response time for enterprises to deal with danger has been shortened, providing strong support for safety production and customer service. According to data from the China Iron and Steel Industry Association, as of January this year, 80% of enterprises have built intelligent centralized control centers, and 81% of enterprises are exploring the application of big data models. Artificial intelligence is accelerating its integration with emerging fields, continuously creating new quality productivity engines. Taking intelligent driving as an example, BYD has fully integrated the DeepSeeker R1 model, promoting the installation of advanced intelligent driving technology in all models. Domestic mainstream car companies such as Geely, Dongfeng, and SAIC have also integrated the DeepSeeker model Han Jian said. Scene applications are the main driving force for industrial development. On the one hand, the threshold for developing large models and the continuous reduction of application costs are driving product innovation, scenario exploration, and industry empowerment. DeepSeek's open-source strategy is accelerating the diversification of artificial intelligence applications; On the other hand, China has advantages such as a wide range of industrial categories, a large population base, and multiple market demands, which provide rich scenarios for the application of artificial intelligence and constantly stimulate new applications. DeepSeek has narrowed the gap between China and the United States in artificial intelligence, especially in terms of inference models. But overall, we are still in the catching up stage, and there are still significant gaps in areas such as artificial intelligence chips, new generation model core architecture, and scientific applications of large models Zhao Gang said. The Evaluation Report on the Development of China's Artificial Intelligence Computing Power in 2025, which tamps the base enabling industry, believes that the improvement of algorithm efficiency brought by DeepSeek does not curb the demand for computing power, but instead, because more users and scenarios join, it promotes the popularization and application of large models, drives the construction of data center, edge and end side computing power, and drives the growth of computing power demand. DeepSeek's significant innovation at the algorithmic level has brought profound changes to the artificial intelligence industry in China and even globally. Algorithms have become the core engine driving the development of artificial intelligence, driving the growth of computing power and driving changes in computing architecture and data centers. It is reported that the Ministry of Industry and Information Technology will organize the "unveiling the list and leading the way" to support enterprises, research institutions, etc. to accelerate the research and development of computing power, algorithms, and data, strengthen the research and development layout of general and industry models, promote the iterative upgrading of intelligent product equipment, and accelerate the construction of high-quality datasets. Han Jian believes that we need to increase innovation support in areas such as chips, models, architectures, and toolchains, and build an AI innovation ecosystem that is willing, daring, and capable of innovation; Continuously strengthen the construction of artificial intelligence industry development factor systems such as intelligent computing centers and data element markets, and consolidate the foundation of artificial intelligence industry development; Deeply promote the "Artificial Intelligence+" initiative, guide large model enterprises to strengthen cooperation with industry users, and create excellent products and typical benchmarks for industry applications. We need to support enterprises in continuously improving their inference capabilities, enhancing model efficiency, and reducing model costs through open source models, model distillation, and model engineering optimization. Support the construction of high-quality datasets in various industries, fully utilize industry expert data and synthetic data, and enhance the level of model intelligence. In the fields of scientific research, biomedicine, intelligent manufacturing, etc., we will deeply promote the application of industry models and embodied intelligent agents Zhao Gang said. It is reported that the Ministry of Industry and Information Technology has supported enterprises to build large-scale models in industries such as steel, non-ferrous metals, electricity, communication, and chemical engineering, promote the deep integration of artificial intelligence and key industries, and accelerate the transformation and upgrading of intelligence. Around the fields of consumer goods, raw materials, electronic information, etc., we have organized and carried out a "deep journey" of empowering new industrialization with artificial intelligence, promoting precise docking of more than a thousand enterprises, and strengthening the matching of supply and demand in the upstream and downstream of the industrial chain. In terms of ecological cultivation, the Ministry of Industry and Information Technology supports Shanghai, Shenzhen, Beijing and other places to build national AI innovation application pilot zones, promotes the National High tech Zone to establish an AI industry collaborative innovation network, guides the establishment of China AI industry development alliance, and gathers more than 1000 enterprises for collaborative development. At present, China has cultivated a total of 416 national level specialized and innovative "little giant" enterprises in the field of artificial intelligence. Next, on the one hand, we will promote the industrialization of intelligence, consolidate the foundation of the artificial intelligence industry, strengthen the supply level of intelligent equipment products, and develop the artificial intelligence industry with high quality; On the other hand, we will promote industrial intelligence, accelerate the application of artificial intelligence, leverage the advantages of China's industrial system, promote the deep integration of artificial intelligence and manufacturing, and empower new industrialization at a high level The relevant person in charge of the Ministry of Industry and Information Technology said. (New Society)