Nvidia(US:NVDA) Zhi Tong Cai Jing·2025-11-20 04:19Core Insights - Nvidia's shift to using low-power memory chips (LPDDR) in AI servers may lead to a doubling of server memory prices by the end of 2026 [1][2] - The global electronic supply chain is currently facing a shortage of traditional memory chips due to manufacturers focusing on high-end memory chips suitable for AI [1] - The transition to LPDDR is expected to create sudden demand that the industry may struggle to meet, as each AI server requires significantly more memory chips than a smartphone [1] Industry Impact - Major memory suppliers like Samsung, SK Hynix, and Micron are experiencing shortages in traditional DRAM products as they shift production towards high-bandwidth memory (HBM) to support the AI boom [1] - Counterpoint Research indicates that the shift in Nvidia's memory sourcing makes it a customer comparable in scale to large smartphone manufacturers, creating a significant challenge for the supply chain [2] - The anticipated increase in server memory prices will raise operational costs for cloud service providers and AI developers, exacerbating budget pressures already strained by GPU upgrades and record spending on power infrastructure [2]