A Morgan Stanley report warned that surging AI demand is creating a structural shortage in the global memory market, driving sharp increases in memory prices and raising costs across the technology sector.
The report said demand for high-bandwidth memory (HBM), DRAM, and enterprise solid-state drives is accelerating as hyperscalers ramp spending on AI infrastructure.
Memory prices have risen more than sixfold over the past year, marking a sharp departure from the long-term trend of declining semiconductor costs. Analysts argued that the market is facing a multi-year supply bottleneck, with new capacity taking years to build and qualify.
According to the report, AI workloads are increasingly consuming available memory, prompting manufacturers to prioritize higher-margin products for data centers and AI systems.
That shift is expected to leave less supply available for traditional markets such as smartphones, personal computers, automotive electronics, and industrial equipment.
The report estimates that smartphones and PCs could face memory supply shortfalls by 2027 if current trends continue.
Analysts said large cloud providers are increasingly securing capacity through long-term supply agreements and prepayments, tightening availability for other customers and reducing the role of traditional commodity pricing in the memory market.
The investor note also highlighted broader economic implications, noting that rising memory costs are already showing up in producer price inflation and corporate cost structures.
Although the direct impact on consumer inflation may remain limited, elevated memory prices could increase hardware costs, pressure margins, and slow technology deployment across several industries.
On stock positioning, the report favored memory manufacturers and related infrastructure suppliers, citing stronger pricing power and earnings visibility.
Companies exposed to consumer hardware and other non-AI end markets were identified as facing the greatest risk from higher memory costs and supply constraints.
source https://www.investing.com/news/stock-market-news/how-ai-started-turning-memory-into-a-structural-bottleneck-4741048
