News
IMF Warns Of Disorderly Bond Market Sell-Off As Debt, Inflation And AI Borrowing Rise
The International Monetary Fund (IMF) has warned that government bond yields could rise sharply and unpredictably as mounting public debt, resurgent inflation and growing borrowing by artificial intelligence (AI) companies increase pressure on sovereign bond markets.
In its latest Global Financial Stability Report, released on Tuesday during the IMF-World Bank annual meetings in Bangkok, Thailand, the fund warned of “a sharp rise in sovereign bond yields,” noting that elevated debt levels and refinancing risks had made bond markets more vulnerable to fiscal and inflation shocks.
Government bond yields in several advanced economies have climbed to multidecade highs, driven by rising energy costs linked to the Middle East conflict and expectations of tighter monetary policies as central banks seek to contain renewed inflationary pressures.
The IMF warned that a disorderly increase in yields could have severe consequences because of the central role sovereign bonds play in the global financial system. Higher yields could reduce asset valuations by increasing discount rates, including the valuations of AI-related assets, while weakening the value of government bonds used as collateral.
The warning follows the fund’s assessment last week that global public debt had reached levels close to the highest recorded since World War II and was on course to exceed 100 per cent of global gross domestic product.
The report also identified rising borrowing by AI companies as an additional source of pressure on bond markets, as technology firms increasingly compete with governments for investors’ capital.
According to the IMF, hyperscalers’ share of global investment-grade bond issuance has risen to 10 per cent from less than one per cent since 2024, reflecting the enormous financing requirements of the global AI expansion.
Technology giants, including Google parent Alphabet and Meta Platforms, alongside other hyperscalers and AI developers, have raised hundreds of billions of dollars in debt this year to finance their AI infrastructure and expansion plans.
Marc Seidner, chief investment officer of non-traditional strategies at Pacific Investment Management Company, said AI-related debt issuance had become significant enough to compete with government bonds for investor funds, although it was not yet the dominant driver of rising yields.
The IMF also raised concerns about the growing exposure of hedge funds to US Treasury securities, warning that the market could face a disorderly unwinding of investment positions.
The fund said leveraged basis trades had partly unwound without significant pressure on short-term lending markets, but warned that the leverage remaining in the system continued to pose risks.
Citing Federal Reserve data, the IMF said hedge funds’ share of US Treasuries increased to about nine per cent in 2025 from approximately four per cent in 2022, largely through highly leveraged arbitrage strategies.
Caio Ferreira, the IMF’s deputy division chief for financial supervision and regulation, said the concern stemmed from hedge funds’ increasing leverage and the growing overlap in their investment positions.
He warned that multiple funds holding similar trades could amplify market disruptions if they attempted to exit their positions simultaneously.
However, the IMF said market liquidity had remained resilient in the United States, Europe and the United Kingdom. Japan was an exception, as liquidity in its government bond market had continued to deteriorate amid increased yield volatility following the end of its yield curve control policy.
To reduce refinancing pressures, the fund urged debt managers to extend the maturities of government borrowing. It warned that replacing long-term debt with short-term financing on a large scale could increase rollover risks.
Beyond the pressure from AI companies’ borrowing, the IMF warned about interconnected financing arrangements between AI firms, describing the structure as a form of “circular” financing.
According to the report, 14 publicly disclosed deals since 2023 channelled about $300 billion in equity investments, led by hyperscalers, into financially weaker AI companies, alongside $650 billion in debt.
In return, investors secured approximately $1.4 trillion in future revenue commitments through agreements covering services and chip orders.
Jason Wu, assistant director of the IMF’s Monetary and Capital Markets Department, said the future revenue commitments distinguished the arrangements from previous investment deals and created a circular financing structure.
He warned that if the commercialisation of AI failed to meet expectations, the interconnected investments could expose companies and investors to simultaneous losses.
“If there are questions about the monetization in AI, the equity investments will deplete in value, the debt may be difficult to service, and the revenue commitments in the future may not materialize,” Wu said.
He described the potential consequences as a “triple whammy,” in which equity investments lose value, debt repayments become difficult and anticipated future revenues fail to materialise.
The IMF said the financing arrangements and revenue commitments were concentrated among a limited number of AI companies, increasing the potential for problems at individual firms to spread across the market.
Wu warned that difficulties involving particularly exposed companies could intensify the risks created by the interconnected financing arrangements, especially if AI revenues fell short of expectations.
Related
