Lenovo warns that the global memory shortage is likely permanent, as surging demand for AI infrastructure continues to absorb manufacturing capacity for DRAM and NAND flash components.

Key Points

• Lenovo executives stated at the International Supercomputing Conference that memory prices are unlikely to return to the low levels seen in early 2025.
• Massive AI infrastructure investments by hyperscalers are expected to consume new manufacturing output through at least 2030.
• Memory manufacturers like SK hynix and Micron are prioritizing high-margin HBM production for AI accelerators over commodity DDR5 and LPDDR5 memory.
• New server architectures require significantly higher memory capacity, with some configurations needing 1 TB of RAM to fully utilize available bandwidth.
• Major tech firms, including Apple, are facing supply constraints so severe they are seeking alternative sources for DRAM components.

Why it Matters

The shift suggests that the historically cyclical nature of memory pricing has been fundamentally altered by the sustained, high-volume demands of the artificial intelligence sector. For businesses and consumers, this indicates a long-term transition toward higher hardware costs as memory manufacturers prioritize enterprise-grade, high-margin products over traditional consumer components.