The current equity market landscape presents a unique challenge for portfolio diversification, as a significant portion of market gains is concentrated in a narrow group of AI-related stocks. Some estimates suggest that the "AI trade" now accounts for as much as 60% of the market, with even companies like TOTO, a Japanese high-tech toilet maker, acquiring an AI angle due to their tangential involvement in semiconductor ceramics. This concentration means that traditional asset class diversification, which historically relied on inverse movements between bonds and equities, is becoming less effective, as evidenced by its failure in 2022.
This trend has led investment specialists to emphasize sector-based diversification. While large-cap technology stocks dominate, investors are encouraged to look beyond the obvious AI plays to the broader value chain. For instance, the "picks and shovels" of AI, particularly hardware-focused names in semiconductors and memory chips, are seeing substantial investment, with cash flowing from US hyperscalers to Asian technology manufacturers like Samsung, SK Hynix, and TSMC. These three semiconductor companies alone are forecast to generate $300 billion in combined free cash flow this year, driven by expected hyperscaler capital expenditures approaching $800 billion in 2026.
The AI boom, however, carries risks, with warnings of an "AI bubble" from figures like OpenAI CEO Sam Altman. The market is also experiencing a shift in correlations, with equity and bond returns increasingly moving in lockstep, reducing the traditional cushioning effect of bonds. The rolling 52-week correlation between equities and bonds has risen to +0.44, compared to a 10-year trend of -0.2. This necessitates a re-evaluation of portfolio construction, with some investors using derivatives to mitigate directionality risk and focusing on fund managers who are style-agnostic.
