[Morgan Stanley: Storage Stock Cycle Peaks May Appear to Have the Lowest Valuations but Are Actually in the Most Dangerous Position] BlockBeats News, October 13 – Morgan Stanley's latest report points out that Samsung and SK Hynix currently have forward price-to-earnings ratios of only about 3 to 4 times. While they may seem cheap, they could potentially conceal valuation traps at the cycle peak. Historically, storage stocks often peak ahead of earnings, with peaks occurring approximately 4 months early in 2008 and 2012, and about 8 months early in 2018. Although the AI-driven storage supercycle is still ongoing, with tight supply-demand dynamics, rising prices, and improving profitability providing support, the sharp rise in stock prices and concentrated upward revisions in earnings expectations suggest that optimistic forecasts have been fully priced in, leaving limited room for further gains. Morgan Stanley believes that the stronger the fundamentals of the storage industry, the more likely the risks of a cycle peak are underestimated. High prosperity does not necessarily translate into sustained excess returns. The report warns that high profits will eventually attract new capacity, with supply expected to gradually come online starting in the second half of 2027, potentially triggering a new cycle inflection point. Currently, Asian storage stocks have risen approximately 9 times since their rally began in September 2025, with consensus earnings per share expectations up an average of 432%. The low price-to-earnings ratios of Samsung and SK Hynix may be a typical trap for cyclical stocks near their earnings peaks. To support the implied upside in current stock prices, DRAM contract prices would need to rise approximately 70% quarter-on-quarter, which is highly challenging. Additionally, the evolution of data center architectures may alter the profit distribution across the supply chain: in NVIDIA's NVL72 single rack, storage accounts for about 29%, but when expanded to 8 interconnected racks, this share drops to 19%, while the share of network equipment rises from 10% to 20%. Storage manufacturers may not benefit proportionally from the growth in AI infrastructure investments. [Original Link]