Skip to content
Advertisement
Financial markets show signs of strain as the AI trade starts to crack

Financial markets show signs of strain as the AI trade starts to crack

Bond selloffs, sliding currencies and growing doubts over AI returns are testing investors who leaned on one optimistic bet

For a long stretch, artificial intelligence was the one story that kept investors cheerful through wars, swelling government debt and trade fights. That story is now developing hairline fractures.

Reuters reported from London on Nov 3 that global financial markets are straining under all three pressures at once. The AI trade, long the market’s emotional support animal, is no longer immune.

The warning signs are scattered across the map. Japanese and French bonds have sold off, and emerging-market currencies have plunged.

Until recently, a multitrillion-dollar spending spree on AI had papered over much of that stress. Helped along by Washington’s light-touch approach to business, that spending pushed stock markets to record highs and lifted expectations for the global economy.

The numbers behind the wobble

October delivered the clearest evidence yet that patience is thinning. More than $2 trillion was erased from the Nasdaq 100 during the month as concerns about AI spending mounted.

Intel took the hardest single hit. Its shares fell 28.6% on October 30 after the company reported heavy AI-related cash burn.

The pressure extends to the biggest spenders in the sector. Three of the four major hyperscalers, Amazon, Meta and Alphabet, posted negative free cash flow in their latest reported quarter.

Advertisement

Free cash flow is the money left over after a company pays for its operations and its investments in things like data centers. When it turns negative, a business is spending more than it brings in, and the gap has to be filled somehow.

Amazon’s shortfall alone came to nearly $35 billion.

When AI meets the bond market

The International Monetary Fund flagged the risk in a report published in October. It warned that an AI-driven selloff could ripple outward, with particular danger for economically vulnerable countries that have few fiscal tools to cushion a downturn.

One detail from the IMF stands out. Since 2024, the hyperscalers’ share of global investment-grade bond issuance has climbed from under 1% to 10%.

The IMF also pointed to $300 billion in equity deals and $650 billion in debt tied to weaker AI firms.

Meanwhile, government borrowing costs are flashing their own warnings. French bond yields have reached levels last seen during the 2012 euro-zone debt crisis, and the 10-year US Treasury yield has hovered near 5%.

Why weak government balance sheets matter

Reuters noted that governments carrying high debt loads are in a weak position to fight a new crisis. The IMF’s concern about countries with limited fiscal options points in the same direction.

What this means for investors and markets

The first issue is concentration. When a single theme drives a large share of stock market gains, a stumble in that theme drags the whole index with it, as the Nasdaq 100’s October losses showed.

The second issue is how stress travels. With hyperscalers now a meaningful slice of investment-grade bond issuance, trouble in AI no longer stays in the equity market.

A disappointment in AI returns could push up borrowing costs for tech giants. It could also tighten conditions for the weaker AI firms carrying the $650 billion in debt the IMF identified.

The third issue is timing. Rising sovereign yields, a 10-year Treasury near 5% and sliding emerging-market currencies leave less room for error.

Intel’s 28.6% drop shows how harshly investors now punish cash burn without a clear return.

There are several signals worth tracking from here. Watch hyperscaler free cash flow and capital spending guidance in upcoming earnings, the pace of tech bond issuance, the direction of US Treasury and French bond yields, and the health of emerging-market currencies.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

More in AI

See all