Artificial Intelligence: Managing Concentration Risk

AI concentration has become a cross-asset portfolio risk, so investors should retain exposure to the technology while pairing it with higher-quality and less correlated assets.

• The AI revolution is progressing through innovation, infrastructure buildout, and economy-wide adoption, while US AI infrastructure investment could exceed US$9 trillion over the next six years, equal to 3.2% of annual US GDP.
• AI adoption reached 78% of businesses across seven countries by mid-2026, up from 69% six months earlier, but only 2% of US and UK firms accounted for 50% of current AI spending.
• Current productivity gains remain modest at 0.2% annually in sales per employee among surveyed adopters, although respondents expect 0.6% annual growth over the next three years.
• Hyperscalers represent 17.8% of S&P 500 market capitalization and AI-adjacent companies another 28.1%, while AI-related businesses account for about 19% of global equity market capitalization, worth US$21.8 trillion.
• The concentration extends to credit and private markets: AI-related issuers sold about US$330 billion of bonds in the first half of 2026, representing roughly 16% of outstanding US investment-grade corporate debt, while AI accounts for about a quarter of direct lending and over 15% of private equity investments.

Takeaway: Investors should assess overlapping AI exposure across public equity, bonds, private markets and infrastructure, then consider equal-weight strategies, non-directional alternatives and a barbell combining AI assets with US Treasuries, high-quality sovereign bonds, mortgage-backed securities and other investment-grade credit.