Memory Prices Set to Double as AI Demand Overwhelms Supply Chains
DRAM and NAND flash shortages worse than expected as hyperscalers strain fabs
The memory shortage has proven more severe than most analysts predicted, with AI ambitions pushing chip fabrication facilities to their limits. High-bandwidth memory (HBM) used in AI accelerators has been in particularly tight supply, but the crunch is now spreading across all memory categories.
Cloud service providers building out AI infrastructure have been placing massive orders for memory, competing with traditional server and consumer electronics demand. The situation has been exacerbated by limited fab capacity expansion, as building new semiconductor facilities takes years and billions in investment.
Consumers and enterprise buyers are already seeing price increases for SSDs, RAM, and other memory products. The shortage is expected to persist through at least mid-2026 as new capacity comes online gradually.
Analysis
Why This Matters
Memory costs flow through to every computing device from smartphones to data centers. Doubling DRAM prices will impact cloud computing costs, PC prices, and ultimately consumer technology prices across the board.
Background
The memory industry has historically been cyclical, with periods of oversupply followed by shortages. However, AI demand represents a structural shift that may keep prices elevated longer than previous cycles.
Key Perspectives
Memory manufacturers like Samsung, SK Hynix, and Micron are benefiting from higher prices, but face pressure to expand capacity. Tech companies reliant on memory are warning of margin pressure.
What to Watch
New fab announcements and AI chip demand trends will determine how long the shortage persists. Government semiconductor subsidies may accelerate capacity expansion.