11 November 2026
A portfolio combining global equities, investment-grade bonds and private credit may appear well diversified. Yet beneath the surface, these investments could share a significant exposure to the same economic driver: the rapid expansion of artificial intelligence. In a recent article published on CFA Institute’s Enterprising Investor blog, Hassan Al Ataby examines how the AI investment boom is creating hidden concentration risks across asset classes and geographical markets.
The central concern is not whether artificial intelligence will deliver on its technological promise, but whether investors fully understand how much of their portfolios’ performance now depends on the continued growth of AI-related investment.
Concentration Extends Beyond Technology Stocks
The growing dominance of a handful of technology companies in equity markets is well documented. As of 10 October 2026, the ten largest companies in the S&P 500 represented approximately 39% of the index, more than double their 18% share in 2015. Many of these companies, including Nvidia, Microsoft, Alphabet, Amazon, Meta and Broadcom, are closely connected to AI development and infrastructure spending. However, moving beyond US equities does not necessarily eliminate this concentration. In Taiwan, TSMC accounted for approximately 57% of the MSCI Taiwan Index following the August 2026 index review. In South Korea, Samsung Electronics and SK Hynix together represented around 67% of the MSCI Korea Index as of 30 October. These figures illustrate how geographical diversification can conceal exposure to the same underlying technology cycle. Investors shifting allocations from US technology stocks towards Asian equities may simply be replacing one part of the AI ecosystem with another.
The AI Investment Boom Is Reshaping Bond Markets
The concentration risk is increasingly extending into fixed-income portfolios. According to estimates cited in the article, AI hyperscalers are expected to account for approximately 9% of US dollar investment-grade bond issuance in 2026, compared with around 2% between 2022 and 2024. Morgan Stanley has also forecast that global AI-related debt issuance could approach $570 billion this year. This development has important implications for passive bond investors. Unlike equity indices, which generally assign greater weight to companies with higher market capitalisations, traditional bond indices tend to allocate more weight to issuers with larger amounts of outstanding debt. Consequently, as technology companies increase borrowing to finance AI infrastructure, their weight in bond portfolios may grow automatically. Private credit introduces an additional layer of exposure. Data centres are increasingly financed through structures backed by property assets and leases involving the same technology companies that dominate equity markets. What appear to be separate investments may therefore depend on the financial strength and spending commitments of a relatively small group of businesses.
What Happens if AI Investment Slows?
The scale of the AI infrastructure expansion helps explain why these overlapping exposures matter. Moody’s estimates cited in the article suggest that capital expenditure by the six largest US hyperscalers could reach approximately $785 billion in 2026, rising to around $1 trillion in 2027. These investments are based on expectations of substantial future demand for AI computing capacity and services. If that demand grows more slowly than anticipated, the consequences could extend well beyond technology share prices. A reassessment of AI growth expectations could simultaneously affect mega-cap equities, semiconductor manufacturers, investment-grade corporate bonds and private credit linked to data-centre infrastructure. The analysis does not predict an imminent market correction. Many of the companies involved are highly profitable and financially strong. Nevertheless, even high-quality businesses can become sources of portfolio risk when valuations and investment decisions depend on exceptionally optimistic expectations. The challenge is that correlations between different asset classes may increase sharply when they are exposed to the same underlying economic shock.
Five Ways to Identify and Manage Hidden Exposure
Al Ataby identifies five practical approaches that investment professionals can adopt to better understand and manage AI-related concentration.
First, look-through analysis can help reveal the underlying exposures shared by equity, fixed-income and private-market investments, rather than assessing each allocation independently.
Second, portfolio managers should stress-test the impact of a slowdown in AI capital expenditure across the entire portfolio, considering how such a scenario might affect different asset classes simultaneously.
Third, reviewing index construction methodologies can help reduce dependence on a limited number of companies. Equal-weighted, capped or factor-based strategies may offer alternatives to traditional market-capitalisation-weighted indices, although they can underperform when market leaders continue to outperform.
Fourth, investors should examine credit concentration, identifying the ultimate borrowers, tenants and counterparties behind bond and private-credit exposures.
Finally, advisers should communicate these risks clearly to clients, helping them understand that holding numerous securities or investing across several asset classes does not necessarily guarantee genuine diversification.
Diversification Requires a Deeper Understanding of Risk
The rapid development of artificial intelligence is transforming both the corporate investment landscape and the composition of financial markets. For portfolio managers, the challenge is to distinguish between diversification in the number and type of investments held and diversification in the economic factors driving their performance. A portfolio containing thousands of securities can still be vulnerable to a single investment theme if its underlying holdings depend on the same growth expectations. The analysis therefore reinforces an important principle of portfolio construction: effective diversification requires understanding the sources of risk, not simply spreading capital across different assets. As AI continues to reshape markets, identifying and managing these hidden connections will become increasingly important for investment professionals seeking to build resilient portfolios.
Read the original article by Hassan Al Ataby, published on CFA Institute’s Enterprising Investor blog on 9 November 2026: https://rpc.cfainstitute.org/blogs/enterprising-investor/2026/hidden-ai-concentration-risk-in-diversified-portfolios