IMF warns of AI asset correction, identifies $650 billion in debt
Key Facts
The International Monetary Fund warned in its October 2026 Global Financial Stability Report of a potential correction in artificial intelligence asset prices if investment and profitability in the sector weakened or became less sustainable. The fund identified $650 billion in debt in the circular financing arrangements it reviewed as part of its analysis of risks surrounding the sector's financing. Those arrangements also included about $300 billion in equity investments, according to an analysis published by the fund. These investments and debts were tied to future revenue commitments worth $1.4 trillion. The fund said rising leverage, concentrated exposures and circular financing arrangements could amplify losses from a correction in AI asset prices. The warning concerns the sustainability of the investment and profitability underpinning those assets, and financing that could magnify losses during a correction.
The fund described overall financial conditions as still accommodative despite rising risks to global financial stability. Athanasios Vamvakidis said during the report's press briefing: "This resilience, however, does not necessarily mean safety." The report identified resurgent inflation among the factors that could drive a sharp rise in sovereign bond yields. It added concerns over government finances to the potential drivers of such an increase in yields. The fund expressed particular concern about eroding buffers in some emerging economies with low credit ratings as part of its assessment of financial vulnerabilities. It called for fiscal policies to address imbalances and put public finances back on a sustainable path in response to risks linked to government finances.
The bond risks come as hedge funds expand their presence in markets, with total assets reaching about $13 trillion in early 2026, according to the report's chapter on the sector. Their share of U.S. Treasury holdings rose to 9% in 2025 from about 4% in 2022. The fund said increased market volatility or tighter funding conditions could trigger margin calls at hedge funds. To meet those calls, funds could be forced to reduce exposures and unwind positions, amplifying market stress. Their losses could spread to banks that provide them with financing and trading services, according to the fund's analysis. Absorbing those losses could reduce banks' capacity to lend during periods of financial stress.
In the insurance and pension fund sectors, the fund warned that increased holdings of illiquid assets heighten liquidity mismatch risks. These holdings also increase the likelihood of forced asset sales during periods of stress, according to the report's executive summary. Under the adverse scenario in the banking stress test, banks accounting for 22% of global banking assets are classified as weak. That classification in the test is based on those banks' common equity Tier 1 capital ratios falling below the 7% threshold. Alongside liquidity and capital risks, the report added sophisticated cyber threats to the sources of risk facing the financial system. The fund said these threats pose growing risks as finance becomes increasingly digital and interconnected.
In AI financing, the fund's analysis said shrinking free cash flows were pushing major cloud computing operators to rely increasingly on debt. The analysis expects their capital expenditure to grow by about 20% in 2027 to nearly a trillion dollars. The report also examines a scenario of declining AI investment and its implications for global economic activity. In that scenario, the level of global output could fall by as much as 0.4% by 2028. The decline in the level of investment reaches 1.3% under the same scenario, according to the chapter on financial stability conditions. The level of consumption could fall by as much as 0.6% by 2028 through the effects of declining wealth.