Artificial intelligence has permeated every corner of the financial markets, from stocks and bonds to private equity and infrastructure. This widespread expansion is creating a "hidden" risk concentration in the portfolios of large institutional investors, who are now being called to shift their focus from traditional asset diversification to the diversification of risks themselves.

AI is Everywhere

The issue goes far beyond the stocks of tech giants. According to Bloomberg data, AI now influences:

  • Corporate Bonds: Through large debt issuances by so-called hyperscalers.
  • Private Equity: With investments in AI startups accounting for approximately 87% of venture capital funding.
  • Infrastructure and Energy: Due to the explosive demand for data centers and the massive electricity consumption they require.

Indicatively, about 40% of the total market capitalization of the S&P 500 now concerns companies linked to AI infrastructure, while half of this year's investment-grade bond issuances are related to the same technology.

The Institutional Investor's Dilemma

Large funds, such as the New York City Retirement Systems ($327 billion), are expressing concerns that the same investment risk is now appearing across many different asset classes. Chief Investment Officer Monte Tarbox specifically mentioned a decision to refrain from private equity products with high AI exposure, as the total portfolio's exposure might already be excessive.

The difficulty lies in calculating this exposure. The LACERA fund ($94 billion) estimates its AI exposure ranges from 8% to 19%, depending on the methodology used. Meanwhile, an Invesco survey of 90 sovereign wealth funds showed that more than half consider market concentration to be the most significant risk associated with the technology.

The Shift to the Total Portfolio Approach

To address this challenge, large funds like CalPERS are adopting the "Total Portfolio Approach." Instead of examining asset classes in isolation, they treat the portfolio as a whole to identify common sources of risk. Despite the risks, AI remains attractive, as Bloomberg Intelligence's relevant index has outperformed the broader market by 11 percentage points annually over the last two years. The challenge for managers is to maintain returns without becoming entirely dependent on a single technological trend.