- A report from BofA Securities indicates that the recent sell-off in Japanese AI-related stocks is a healthy sector rotation rather than the start of a widespread market decline. It advises investors to diversify their portfolios amid rising macroeconomic risks and high valuations.
- Market sentiment is being suppressed by concerns over the sustainability of AI data center spending, rising memory and computing power costs, and competition from Chinese open-source models. Additionally, falling oil prices and rising U.S. real interest rates have accelerated the outflow of funds from the AI sector.
- Seasonal factors and the reversal of highly concentrated positions have exacerbated the weak performance of momentum stocks from late June to early August. Although the sector is expected to gradually recover after the first-quarter earnings reports, investors should avoid high P/E ratio stocks in the short term and allocate horizontally to non-AI sectors such as banking and defense.
Capital Flows and Market Concentration Reversal
BofA Securities analysis shows that the recent weakness in the AI sector is primarily due to the excessive concentration of high-beta, high-valuation stocks after a prolonged rise. As momentum stocks entered a seasonal weak period from late June to early August, the weakening of earnings revision momentum triggered profit-taking. This shift in capital flows has prompted investors to reassess market concentration risks, but the current level of concentration has not yet reached the extreme turning points seen during the internet bubble or the post-World War II financial crisis, suggesting this is merely healthy capital redistribution rather than a systemic collapse.
Computing Power Costs and Open-Source Model Competition Suppress Sentiment
On a micro level, there are challenges related to multiple costs and technological paths. Market doubts have arisen over whether AI-driven data center capital expenditures can continue to deliver high returns, while the ongoing rise in memory and AI computing power costs is squeezing the profit margins of related companies. Additionally, investors are turning to more affordable Chinese open-source AI models. This technological substitution effect and marginal changes in costs have weakened the premium capabilities of high-valuation hardware and software service providers, putting pressure on semiconductor heavyweights like Tokyo Electron.
Pricing Logic of Macro Rates and Oil Price Linkage
Changes in macroeconomic variables are reshaping the valuation anchors of risk assets. Although the recent decline in oil prices and the rise in U.S. real interest rates have jointly driven this round of sector rotation, the core tail risk facing the market remains the possibility of an unexpectedly sharp rise in oil prices, which could force the Federal Reserve (Fed) to maintain a hawkish stance. If nominal and real interest rates are pushed further up, AI-related assets with high valuations will face a deeper correction. However, political factors ahead of the U.S. midterm elections may exert some restraint on oil price increases.
Enhancing Asset Allocation Breadth and Defensive Substitution
Against the backdrop of benchmark indices like the Nikkei 225 (JP225) masking the true scale of rotation, investment strategies should shift from full concentration in a single track to broader allocation. BofA Securities suggests that investors retain core high-quality AI exposure while avoiding targets with excessively stretched P/E ratios. Funds should be horizontally shifted to non-AI sectors benefiting from falling oil prices and improved market breadth, with a focus on increasing allocations to stable growth industries such as banking, defense, construction, real estate, and food, to build a diversified portfolio with shock-resistant capabilities.
