Global investors reassess diversification amid AI concentration

Chris Hogbin, chief executive of Lazard Asset Management, said the world is moving into a “different geopolitical regime,” one in which risk premiums are embedded at the start of a deal rather than discounted at the end. That shift matters economically because it raises the cost of capital, complicates cross-border dealmaking and makes earnings more vulnerable to policy shocks, sanctions and supply-chain disruptions.

For investors, the bigger issue is that classic diversification is working less effectively than before. Richard Oldfield, chief executive of Schroders, said geopolitics is helping drive up oil prices, feeding inflation and forcing governments to spend more, which in turn lifts sovereign debt and weakens the usual offset between bonds and equities. In other words, the old assumption that equity weakness can be cushioned by fixed income is less dependable when the same macro shock hits both.

The comments also reflect how much of the market’s resilience has rested on AI rather than broad-based earnings growth. Oldfield said the first seven months of 2026 were dominated by concentration in AI stocks, leaving the impression of uniform market strength even as gains were narrowly led. That concentration has been visible in the price action of the biggest technology names: Microsoft has risen 18.9% from its Nov. 2 close to its Nov. 4 high before easing back, while Nvidia has slumped about 10% in the past week from $206.69 to $186.64. Alphabet has fallen nearly 17% in that span to $284.78. The volatility underscores how reliant index performance has become on a small group of megacap AI names.

That dynamic is reinforced by the broader AI trade itself. While corporate spending on artificial intelligence is beginning to generate measurable revenue gains, those benefits are still uneven and often trail the productivity improvements companies claim. SEC filings from the sector also point to the scale of the bet: Nvidia said its commitments totaled $36 billion as of late July, while Alphabet warned that heavy investment in new businesses and technologies could harm results if returns lag expectations. The risk for markets is that capital is flowing toward infrastructure and winners in the AI stack faster than profits are broadening out.

Adalytica’s Global Stability Sentiment snapshot, which tracks market awareness of geopolitical risk, remains neutral at 57 but has fallen 43 points over seven days, suggesting investors are becoming more alert to the possibility that complacency is giving way to repricing. At the same time, Adalytica’s S&P 500 trade signals show extreme awareness of the broader market backdrop even as sentiment sits at neutral, a combination that typically points to crowded positioning and a fragile advance.

The bullish case is that AI earnings and investment still support equity markets into 2027, as Robeco chief executive Karin van Baardwijk argued, especially if the U.S. economy remains intact. The bearish case is that rising leverage, higher rates and geopolitical spillovers eventually hit credit spreads and expose the weakest balance sheets, particularly among companies funding AI buildouts with borrowed money.

For investors, the takeaway is less about abandoning AI or risk assets than about assuming that correlations, policy risk and valuation support will behave differently from the last cycle. Selectivity, liquidity and balance-sheet strength are becoming more important than broad index exposure, and that could leave active managers with more room to outperform — if they can price the new regime correctly.