IMF Chief Says the World Economy Is Standing on One Leg: AI

AI funding illustration: a hand placing a red chip on the tallest stack

IMF Managing Director Kristalina Georgieva told CNBC on Monday, on the sidelines of the ASEAN summit in the Philippines, that the global economy is currently "standing on one leg, and that leg is AI." In her assessment, if that leg wobbles, it could trigger financial stability problems.

The headline number she offered was that circular financing among AI companies exceeds $1 trillion. Circular financing refers to investors putting money into AI companies, which then flows back to the investors themselves through purchase orders and similar arrangements. The IMF estimates that about two-thirds of it is debt-financed. In Georgieva's words, the scale is "not yet huge, but it is growing very fast."

Nvidia sits at the center of the web

Nvidia is one of the central players in these deals: it invests in downstream model developers and new cloud providers while also being their biggest chip supplier. Over the past month, Nvidia's market value has fallen by more than $1 trillion. Its former CEO Jensen Huang resigned after being drawn into a controversy over his private life, AI stocks worldwide were sold off, and the market began to look more closely at the role Nvidia plays in financing AI infrastructure.

In a recent report, UBS analyst Arcuri estimated that if Nvidia pulled back on external financing, the funding supply across the entire AI chain could easily shrink by more than $50 billion. Rising interest rates are at the same time pushing up the cost of borrowing for companies.

Two more events this week will test investor sentiment: Nvidia is about to report earnings, and Anthropic's IPO pricing is also close at hand.

Putting the IMF's year of statements together

The IMF's stance on AI has been moving steadily toward caution. In October this year, it raised its 2026 global growth forecast to 3.1%, citing AI investment as one reason growth had held up. Energy strain from the war involving Iran has dragged on other engines of growth, leaving major economies even more dependent on AI investment.

Also in October, the IMF laid out a worst-case scenario in its report: if AI fails to meet market expectations in the near term, US equities could fall 20% in 2026, and by 2028 global output could be 1.1% below the baseline forecast. At the time that was a hypothetical. A month later, the pullback in Nvidia's market value has brought markets close to the edge of that scenario.

Outside data is stacking up in the same direction. The IMF estimates that about 40% of the global workforce will be affected by AI, and Capital Economics calculates that as much as 80% of recent global export growth is concentrated in AI-related goods. We previously reported on the Federal Reserve's financial stability survey, in which the share of respondents citing AI rose from about three in ten a year earlier to 87%, and debt issuance tied to advanced computing reached $680 billion in the first ten months of this year. The three institutions use different yardsticks, yet they point to the same place: AI investment is increasingly propped up by borrowed money.

She did not say "bubble"

Georgieva's wording this time was notably restrained. She described recent weeks of market volatility as an "orderly correction" and said it is unlikely to turn into a systemic risk that shocks the global economy. She also said investor exuberance has run ahead of what the technology can actually do, and that "some correction" is necessary. Her advice came down to one line: it must be watched closely.

Several key details she did not provide. The IMF has not published a methodology explaining which transactions and which companies the $1 trillion covers, and there is no breakdown of how much of the two-thirds in debt matures within the next year or two. For readers in China, the protagonists of circular financing are US chipmakers and model developers; whether Chinese AI companies have similar investment-for-orders arrangements, and on what scale, has not been publicly tallied by any institution, so there is not yet comparable data to map this risk directly onto the domestic market.

Sources: CNBC interview coverage, ITHome, IMF World Economic Outlook and Global Financial Stability Report, CocoLoop, UBS research report. The scale of circular financing and the debt share follow the IMF's estimates; market value changes follow CNBC's reporting.