A prominent strategist from JPMorgan Chase, Michael Cembalest, recently raised concerns about the technical indicators of artificial intelligence (AI) stocks in the U.S. market. His analysis, detailed in the 'Eye on the Market' report on July 22, draws unsettling comparisons to the dot-com era's financial boom and subsequent bust. Cembalest points out a significant performance gap between semiconductor companies and larger AI hyperscalers, suggesting a potential instability reminiscent of past market corrections. This caution prompts a discussion on how investors might consider broadening their portfolios beyond domestic AI-centric investments.
Cembalest's research highlights a curious trend: while semiconductor stocks, which provide the foundational hardware for AI, have seen substantial gains, the AI hyperscaler companies, which are closer to the end-user demand, have experienced stagnation. This divergence, where capital expenditure beneficiaries continue to thrive even as companies reliant on final demand falter, is what Cembalest describes as a 'worrisome part of a boom cycle.' He observes that the iShares Semiconductor ETF has returned nearly 90% over the past year, significantly outperforming major AI hyperscalers such as Alphabet, Amazon, Meta, Microsoft, and Oracle. This pattern echoes the late 1990s, when communications equipment stocks soared while communications services stocks slowed, a precursor to the dot-com crash.
This historical parallel doesn't necessarily predict an imminent collapse of the AI market but serves as a warning about potential overvaluation and the risks of concentrated investments. If investors are concerned about the high valuations of AI stocks and the unpredictable returns on AI capital expenditures, diversifying away from the U.S. stock market and the AI sector might be a prudent strategy. Two international Exchange-Traded Funds (ETFs) are presented as viable options for this purpose.
The Vanguard FTSE All-World ex-US ETF (VEU) is one such option. It encompasses a vast portfolio of 3,858 global stocks from 43 countries, spanning both emerging and developed markets, all outside the U.S. With a minimal expense ratio of 0.04%, its top holdings are in Japan (15.5%), Taiwan (9%), the United Kingdom (8.1%), South Korea (7.7%), and Canada (7.5%). Over the past five years, VEU has delivered an annualized return of approximately 9.1%. However, a closer look reveals that its top four holdings—Taiwan Semiconductor Manufacturing, Samsung Electronics, SK Hynix, and ASML Holding—are all deeply involved in the semiconductor industry, collectively accounting for about 11.33% of the fund's assets. While it offers broad exposure, its significant stake in semiconductor companies means it still carries some indirect exposure to AI-related market dynamics.
Another alternative is the State Street SPDR Portfolio Developed World ex-US ETF (SPDW), which holds 2,439 stocks from 25 developed economies. It boasts an even lower expense ratio of 0.03% and has achieved an annualized return of 9.81% over the last five years. Its primary country exposures include Japan (21.6%), the United Kingdom (11.7%), Canada (10.9%), France (7.2%), and South Korea (7.2%). Notably, SPDW does not include Taiwan in its definition of a developed market, thus excluding Taiwan Semiconductor Manufacturing from its portfolio. Its top holdings, Samsung Electronics, ASML Holding, and SK Hynix, are also semiconductor and memory chip manufacturers, but their combined weight in the fund is a more modest 5.55%. This makes SPDW a potentially more diversified choice for investors aiming to reduce their AI exposure.
For investors prioritizing reduced exposure to AI-related market fluctuations, the State Street SPDR Portfolio Developed World ex-US ETF (SPDW) appears to be the more suitable option due to its lower concentration in AI-centric stocks. However, for those looking for broader international exposure, including emerging markets like Taiwan, and are comfortable with some AI-related holdings, the Vanguard FTSE All-World ex-US ETF (VEU) offers a comprehensive global portfolio. Both ETFs provide avenues for international diversification, but their specific compositions cater to slightly different investor preferences regarding AI market sensitivity.
