Morgan Stanley Warns of Storage Stock Valuation Trap Amid AI-Driven Cycle

Huo Xing Finance reports that, on October 13, Morgan Stanley’s latest report indicated that Samsung and SK Hynitz currently have forward P/E ratios of only about 3 to 4 times—appearing cheap but potentially concealing a valuation trap at the peak of the cycle. Historically, memory stocks have often peaked before earnings peaks, by approximately four months in 2008 and 2012, and about eight months in 2018. Although the AI-driven memory supercycle continues, supported by tight supply-demand dynamics, rising prices, and improving profitability, the substantial stock price gains and concentrated upward revisions to earnings expectations have already fully priced in optimism, leaving limited room for further upside. Morgan Stanley warns that the stronger the fundamental backdrop of the memory industry, the more likely the risk of a cycle peak is underestimated; high levels of activity do not necessarily translate into sustained outperformance. The report cautions that high profits will inevitably attract new capacity, with supply expected to gradually ramp up starting in the second half of 2027, potentially triggering a new cycle inflection point. Since their rally began in September 2025, Asian memory stocks have risen approximately ninefold, with consensus EPS forecasts rising on average by 432%. The low P/E ratios of Samsung and SK Hynitz may be a classic trap characteristic of cyclical stocks near their earnings peak. To justify the upside implied by current stock prices, DRAM contract prices would need to rise another roughly 70% quarter-over-quarter—a significant challenge. Additionally, evolving data center architectures may shift profit distribution across the supply chain: in NVIDIA’s NVL72 configuration, storage accounts for about 29% per rack; when scaled to eight interconnected racks, this share declines to 19%, while networking equipment’s share rises from 10% to 20%, meaning memory manufacturers may not benefit proportionally from increased AI infrastructure investment.