IMF Global Financial Stability Report Flags AI, Yield, and Cyber Risks

The International Monetary Fund released the 2026年10月 Global Financial Stability Report on 2026年10月13日, and officials from its Monetary and Capital Markets Department presented the findings at a press briefing during the 2026 Annual Meetings in Bangkok, Thailand. Athanasios Vamvakidis, deputy director of the department, said the report discusses three key risks that could trigger market turmoil and tighter financial conditions: an AI-related market correction, the evolving risk of higher yields, and cyber threats. He said each risk on its own might be manageable but could trigger unresolved and in some cases worsening vulnerabilities, with potential systemic implications.

The full report, titled Resilience, Risk, and Transformation, and its first chapter, Beneath the Resilience: Rising Risks, Unresolved Vulnerabilities, were released at 10:15 AM ICT on 2026年10月13日 (11:15 PM ET on 2026年10月12日), according to the IMF’s publication schedule. Two analytical chapters were released ahead of the meetings: Chapter 2, Hedge Funds and Financial Stability, on 2026年10月6日, and Chapter 3, Tokenization of Financial Assets: Opportunities and Risks, on 2026年10月8日. The report provides the fund’s assessment of the global financial system and markets and addresses emerging market financing in a global context. Vamvakidis was joined at the briefing by Jason Wu, assistant director, and Caio Ferreira, deputy division chief, both of the Monetary and Capital Markets Department. Camila Perez, a senior communications officer, moderated.

Banks, Hedge Funds, and AI Funding

Vamvakidis said that under a severe stress scenario in the report (risk-off shocks from an AI market correction, energy price shocks, and bond market shocks transmitted through non-bank financial institutions), the banking sector as a whole remains resilient. A tail of banks, however, would find themselves undercapitalized under stress, and those banks are the most exposed to non-bank financial institutions.

He said risks have migrated to NBFIs since regulation was tightened in the banking sector after the global financial crisis. Pensions and insurance show increasing exposure to illiquid assets that could be a problem under market stress, he said. In private credit, the fund sees increasing leverage among borrowers, which could restrict credit through liquidity pressures in a stress scenario. Hedge funds have been expanding rapidly and could amplify shocks through leverage and crowded positioning.

Ferreira said the concern about hedge funds comes mainly from increasing leverage and an increasing overlap of portfolios, which could compound outflows during stress and impact markets more globally. The report recommends greater use of central counterparties, he said, because CCPs reduce counterparty risk and can support market liquidity during times of stress.

On AI investment, Wu said the report examined 14 public transactions and found about $300 billion in equity investments and about $650 billion in debt investments, including in off-balance-sheet vehicles, in exchange for $1.4 trillion of future revenue commitments. He said what makes the structure unusual compared with previous deals is the future revenue commitment component, hence its circularity. He described a “triple whammy” concern: if AI monetization is questioned, the equity investments will deplete in value, the debt may be difficult to service, and the future revenue commitments may not materialize, with certain companies and nodes particularly exposed.

Vamvakidis said valuations are not stretched because price-to-earnings ratios have fallen as earnings keep beating forecasts, which he described as the key difference from the dot-com bubble. He noted very strong concentration in AI, calling the sector the single most important engine of resilience at this point, and said the exposure of the rest of the world to U.S. assets, particularly AI, is the highest it has ever been. That exposure supports the global economy through wealth effects today, he said, but a correction would have broader implications.

Asked whether an AI correction could broaden into a systemic credit crunch, Vamvakidis said the market has differentiated between sectors this year, so the first-round effect of an AI correction would not necessarily hit other stocks. But given the real economy’s reliance on AI wealth effects and investment, he said, a correction would tighten financial conditions and could feed back into broader asset prices, and macroeconomic policy room is much more limited than in response to previous shocks.

Rates, Cyber Threats, and Policy Priorities

On the rates sell-off, Vamvakidis said some moves have been sharp, largely because of positioning adjustments and triggered stop losses, but the market has functioned well and there is liquidity. Rate levels and volatility are back to the decades before the global financial crisis, he said, describing it as a return to the previous normal. New pressures this year came from energy shocks following the Middle East conflict, already sticky inflation, and a substantial repricing of central banks, with upside growth surprises also shifting the yield curve upward.

Countries with fiscal imbalances have been more vulnerable to the moves, and he said there is a positive correlation between changes in yields and fiscal imbalances. The market is concerned about absorbing huge issuance from sovereigns and increasingly from the private sector at a time when central banks are shrinking their balance sheets, but rates have not overshot fundamentals and, he said, “market discipline is back.”

Vamvakidis said emerging-market spreads are historically very tight despite the rate sell-off, which he called unusual, and warned of depleted buffers in some low-rated emerging economies and increasing reliance on less transparent alternative funding sources in some frontier economies. Wu added that debt transparency reassures investors and helps keep markets open for emerging economies that need to borrow in hard currencies.

Ferreira identified three global forces making cyber threats a financial stability concern: increasing digitalization that expands the attack surface, growing concentration and reliance on specific technologies such as particular software or cloud service providers, and the availability of tools to cybercriminals, including AI. He cited estimates that about 40 percent more people working on cybersecurity would be needed globally, and said emerging markets are broadly less prepared because they have devoted fewer resources to prevention. He recommended that emerging markets improve incident reporting, domestic and cross-border crisis coordination, and third-party oversight, and that advanced economies identify critical service providers, map their financial and operational links with financial institutions, and test the resilience of those nodes. Wu noted that the 2024年10月 Global Financial Stability Report contained a dedicated chapter on cyber risk, including third-party criticalities.

Vamvakidis said the policy priority is to strengthen resilience before new shocks hit. Fiscal policy should address fiscal imbalances and bring debt dynamics back to a sustainable path, monetary authorities should focus on inflation mandates, communicate clearly, and preserve their independence, and supervisors should reinforce monitoring of banks and particularly non-bank financial institutions, he said.

In country-specific responses, Wu said World Economic Outlook colleagues have projected 4.4 percent growth for Morocco next year, and that Morocco’s banking system remains sound, liquid, profitable, and fairly well capitalized, while the country is exposed to external forces including high energy prices. On Nigeria, he said improvements on the inflation front suggest a less tight monetary stance could be appropriate, with the central bank remaining data-driven and explaining its decisions clearly. He said priorities after Nigeria’s banking recapitalization include adequate capital and liquidity requirements, timely recognition of credit losses and bad loans, and attention to the sovereign-bank nexus in Africa, and that the market for Eurobonds remains very much open.

On Thailand, Wu said AI hardware exports have been an offsetting factor to higher energy prices and have created current account buffers, and that in an AI correction scenario Thailand could see tighter risk sentiment, currency depreciation, and potential capital outflow pressures, so the exchange rate should be allowed to act as the shock absorber. On China, he said AI has become more global and competition between models is healthy, outlining two scenarios: a gradual one in which investors become more realistic about AI expectations as the frontier is pushed, and one in which late and sudden information about weaker monetization or excessive investment triggers a sell-off.

In concluding remarks, Vamvakidis said that if strong risk appetite persists, the global economy will be able to handle the pressure in rates markets, and that the risk is that policies themselves trigger a reduction in risk appetite. He said policies should strengthen resilience and increase buffers to deal with an unforeseen risk-off market correction.