政策解读《“人工智能+制造”专项行动实施意见》实施对算力产业链未来发展趋势的影响分析
Lian He Zi Xin·2026-01-20 11:06

Policy Impact - The implementation of the "Artificial Intelligence + Manufacturing" policy will provide clear guidance and development opportunities for the computing power industry chain[4] - In the short term, the policy is expected to stimulate market confidence and direct funds towards computing infrastructure, increasing demand for computing hardware and intelligent computing center construction[5] - In the long term, the policy aims to address structural imbalances, core technology bottlenecks, and insufficient application scenarios in the computing power industry[15] Current Market Analysis - As of September 2025, the total scale of computing power centers in China has grown from 5.2 million racks to 12.5 million racks, with an annual growth rate of 30%[6] - There is a structural imbalance in the computing power market, with general computing power being relatively surplus while intelligent computing power is in short supply[6] - The geographical mismatch in computing power resources is evident, with high demand in the eastern regions and underutilization in the western regions[7] Core Technology Challenges - Domestic manufacturers hold a low market share in high-end AI training chips, with major markets dominated by foreign companies like NVIDIA and AMD[8] - The software ecosystem is monopolized by international standards like CUDA, creating barriers for domestic chip manufacturers to gain market acceptance[9] Application and Development Opportunities - The policy encourages the construction of high-level intelligent computing facilities and the deployment of computing resources in industrial scenarios, which will benefit related enterprises[5] - The expected market expansion will create a trillion-level incremental market as computing demand shifts from the internet to broader industrial applications[15] - The policy will promote the development of domestic high-end computing hardware and software ecosystems, enhancing the competitiveness of the computing power industry[13][14]