行业点评报告：国产Scale-up/Scale-out硬件商业化提速，聚焦AI运力产业投资机遇

Investment Rating - The industry investment rating is "Positive" (maintained) [1] Core Insights - The development of AI hardware capabilities relies on three key components: computing power, storage capacity, and operational capacity. The focus is shifting towards enhancing operational capacity, particularly in domestic solutions [6][9] - The traditional computing architecture is insufficient for the demands of large-scale AI training, leading to a trend towards supernodes that enhance single-node computing capabilities and drive demand for Scale up hardware [7] - The market for operational capacity hardware is expected to grow significantly, with the global market for Scale up switching chips projected to reach nearly $18 billion by 2030, with a CAGR of approximately 28% from 2022 to 2030 [7] - Domestic manufacturers are making strides in operational hardware, with low current domestic self-sufficiency rates indicating a significant opportunity for domestic substitution [9] Summary by Sections Industry Overview - The report highlights the importance of collaboration between computing power, storage, and operational capacity for the development of AI [12] Market Trends - The demand for Scale up and Scale out hardware is increasing, driven by the need for efficient, low-energy, and large-scale AI training [7][8] - Different communication protocols are emerging for Scale up and Scale out, with major companies developing proprietary protocols alongside public ones [8] Investment Opportunities - Key beneficiaries in PCIe hardware include companies like Wantong Development (Shuduo Technology) and Lanke Technology [10] - Ethernet hardware beneficiaries include companies like Shengke Communication, ZTE, and Yutai Microelectronics [10]