Fund managers seek diversification as AI concentration risk builds across U.S. markets
Portfolio managers are stepping up efforts to reduce heavy exposure to artificial intelligence as the theme becomes more deeply embedded across U.S. equities, credit and the broader economy. Fund groups say clients are increasingly considering hedge funds, emerging markets and real assets to guard against a future reversal in sentiment toward AI-linked companies.
Highlights
- Citigroup research finds over two-thirds of Russell 1000 companies are AI-linked, and JPMorgan notes Big Tech AI beneficiaries now comprise 16 percent of the U.S. high-grade bond market.
- Fund managers such as Amundi and BlackRock report rising demand for private assets, hedge funds, and lower-correlation strategies due to heightened AI concentration risk in multi-asset portfolios.
- Franklin Templeton observes increased investor interest in emerging markets, local-currency bonds, and shorter-maturity fixed income, while reducing bullishness on AI-heavy U.S. equities.
How investors are repositioning portfolios
As reported by Financial Times, fund managers say the growing weight of AI-linked companies in both stocks and bonds is making diversification harder and increasing demand for strategies that are less tied to the sector's performance.Citigroup research indicates that more than two-thirds of companies in the Russell 1000 index are linked to AI, either directly or because the technology is becoming part of their business models. JPMorgan data also shows that Big Tech hyperscalers and other beneficiaries of the AI spending boom account for about 16 per cent of the U.S. high-grade bond market, making them the largest grouping in the investment-grade market.
Vincent Mortier, chief investment officer at Amundi, says clients are revisiting how they diversify because the AI theme is powerful but could change direction if earnings expectations are sharply revised. Ryan Marshall, BlackRock's global head of multi-asset strategies and solutions, says investors are looking for independent, uncorrelated sources of return as AI-related concentration spreads through multi-asset portfolios and even into suppliers tied to power, infrastructure and supply chains.
Marshall says this is pushing interest toward private assets, hedge funds and other lower-correlation areas of the market. He highlights global macro and equity long-short managers in particular, while Mortier says hedge funds are becoming far more attractive to clients, although selecting the right strategy and manager remains difficult because the sector is highly diverse.
Emerging markets and defensive assets gain appeal
Beyond hedge funds, managers say clients are widening their search for diversification through emerging markets and tangible assets. Mortier says investors are increasingly considering local-currency stocks and bonds across regions including Latin America, India and China, while also looking again at real assets such as mining and renewables.Daniel Gamba, co-president and chief commercial officer at Franklin Templeton, says investment teams are assessing AI as a return driver in the same way they evaluate factors such as value, momentum and growth, with the aim of building portfolios that have less exposure to AI-related volatility. He says demand is rising for systematic hedge fund strategies and for broader factor diversification linked to AI risk.
Gamba says Franklin Templeton remains positive on U.S. equities, though less bullish than at the start of the year and less focused on the small group of companies investing heavily in AI infrastructure. He adds that the firm also favors Japan and several emerging markets, while in fixed income it is concentrating on shorter-maturity bonds and avoiding large long-end rate exposure unless investors are specifically seeking income.
Even so, managers are not calling an immediate turn in the market. Mortier says the current backdrop resembles a period when investors keep participating while conditions remain supportive, but he warns that timing any shift is extremely difficult, making diversification more necessary than ever.
Our earlier report on surging AI infrastructure spending outlined how projected outlays for data centers could reach tens of trillions of dollars by 2050, raising questions about whether revenues and productivity gains will arrive fast enough to justify today’s valuations and debt-funded expansion. It also highlighted the funding pressure behind the buildout, noting that even a modest slowdown in demand or delays in payback timelines could amplify risks given the sector’s leverage.