Longsys Frames Edge AI as a Custom Storage and Memory Design Problem

Longsys Chief Scientist Jian Chen used an FMS 2026 keynote to describe edge AI as a storage and memory architecture challenge. As AI workloads move closer to devices, the company says products must be tuned around performance, power, form factor, security, and the specific CPUs, SoCs, and models customers use. The company presented a storage foundry model built around full-stack engineering across firmware, packaging, hardware design, testing, manufacturing, and software adaptation. Its edge AI concepts include caching, AI-oriented SSD, and memory-module approaches intended to address high DRAM costs, limited capacity, and bandwidth constraints. For RamTrend, this is not a broad pricing event, but it is a useful demand signal. Edge AI may increase the need for customized NAND flash, SSD, UFS, and DRAM configurations rather than only standardized client storage. If edge AI deployments scale, suppliers with packaging and system-integration capabilities could capture higher-value memory and storage work even outside cloud data centers.