China News Service, Beijing, November 7 (Reporter Liu Liang) The 2026 Data Storage Industry Conference hosted by the China Electronics Industry Standardization Technology Association was recently held in Zhengzhou. Participants discussed the changes and evolution direction of the role of storage in the era of artificial intelligence (AI).
Zhang Hongtu, chairman of the China Electronics Industry Standardization Technology Association, pointed out that AI is moving from training to reasoning, from single interaction to continuous operation of intelligent agents. Massive data must not only be stored, but also flow, transformed into knowledge, and retained in memory. The industry is facing both historic opportunities and the challenge of leaping from “scale expansion” to “system foundation building”.
"Storage is entering a super cycle driven by AI." Jin Lei, director of the Electronic Systems Division of the Electronic Information Department of the Ministry of Industry and Information Technology, said that the driving force of this cycle comes from fundamental changes in the demand structure, and the value of storage is being redefined. The Electronics Department of the Ministry of Industry and Information Technology attaches great importance to the development of the storage industry and regards it as the core direction for cultivating new productivity in the field of electronic information. It has been systematically included in the overall planning of the "15th Five-Year Plan for the Development of Electronic Information Manufacturing Industry" and will further strengthen overall coordination and focus on promoting relevant work in aspects such as planning and layout, technical research, and ecological cultivation.
Currently, the "memory wall" problem is attracting more and more attention. The so-called "memory wall" means that the memory access speed seriously lags behind the processor calculation speed. As AI computing power grows, the “memory wall” has become an important bottleneck in AI infrastructure.
Jin Hai, professor at Huazhong University of Science and Technology and Changjiang Scholar Distinguished Professor, sorted out the technological evolution logic and engineering implementation difficulties of intelligent storage systems from an academic perspective, and proposed that breaking through the "memory wall" problem requires collaborative innovation in new materials, devices, system structures and other fields.
Yuan Yuan, vice president of Huawei and president of the data storage product line, said that to promote the development of artificial intelligence in my country, it is necessary to "strengthen computing through storage" to achieve efficient data supply, retention, and reuse and sharing, and to improve the utilization rate of computing power clusters. At the same time, through "preserving data to save data", the rich industry data is transformed into the unique advantages of my country's AI.
Shi Jing, assistant to the president of Sugon Beijing Company and general manager of the distributed storage product department, proposed based on industrial practice that storage innovation should focus on the actual output of AI, promote the reuse of data and calculation results with an "AI data factory", improve token output efficiency and reduce AI application costs through storage and calculation collaboration.
As an annual series of research by the Special Committee, the "Data Storage Industry Development Research Report (2026)" jointly compiled by the Data Storage Special Committee and member units was officially unveiled at the conference. The report analyzes industry positioning, current situation and challenges, systematically sorts out the top ten trends and innovative practices of the year, and proposes future-oriented development initiatives. At the same time, the conference released the "Data Storage Industry and Technology Panorama (2026 Edition)", which completely sorted out the upstream and downstream ecology and systematically presented the technology systems at each layer of the storage industry chain.
