Tianfeng Securities: The AI Storage Revolution Has Arrived, "Storage as Computation" Opens a New Era of Storage

The research report from Tianfeng Securities points out that in order to break through the bottleneck of computing power and the constraints of the "storage wall," "storage-based computing" has emerged as a disruptive technological paradigm. This technology achieves a strategic expansion of the storage layer from memory to SSD by migrating vector data from expensive DRAM and HBM video memory during the AI inference process to large-capacity, cost-effective SSD media, rather than simply replacing it. Its core value lies in significantly reducing the first token latency, improving inference throughput, and greatly optimizing end-to-end inference costs, providing a feasible path for the large-scale implementation of AI. The AI storage revolution has arrived, and "storage-based computing" has spawned core opportunities, significantly saving computing power consumption and accelerating AI inference, driving SSD demand growth above traditional curves