Morgan Stanley Sees Memory-Chip Pullback as AI-Fueled Buying Opportunity

AI data-center expansion is absorbing a growing share of DRAM and NAND capacity, making infrastructure spending the main force behind the current memory cycle. Morgan Stanley says that dynamic distinguishes the upturn from traditional cycles tied more broadly to PCs, smartphones and consumer electronics. While those markets are sending mixed signals, memory supply remains insufficient for AI requirements, raising the prospect that shortages and elevated pricing will persist through 2028 and continue supporting chipmakers’ earnings.

In a July 20, 2026 note, Morgan Stanley analyst Joseph Moore said discussions with data-center procurement managers the previous week showed no easing in the supply squeeze. The bank expects like-for-like memory prices to rise at least 25% in the third quarter from the second, above its own and third-party forecasts, and warned shortages could intensify in 2027 and 2028. It views the recent selloff, driven partly by weaker consumer-market signals, as an attractive entry point.