The Bank of England's Financial Policy Committee has noted that accelerating investment linked to artificial intelligence poses a growing threat to financial stability, partly because an increasing proportion of that expenditure is being funded through borrowing and private credit markets. The FPC observed that funding linked to artificial intelligence had been expanding at pace and was forecast to maintain that upward momentum.

Rising reliance on borrowed capital to finance artificial intelligence expansion means that turbulence in the sector could ripple across a broader pool of investors and credit markets. The report noted that investor confidence may have been influenced by uncertainty surrounding the long-term viability of AI-driven revenue growth and associated capital expenditure trends.

Minutes from the most recent gathering noted that worldwide debt issuance tied to artificial intelligence had reached approximately USD 450 billion by early September, while the possibility of a pronounced downturn spreading into core markets remained a concern. According to Morgan Stanley's calculations, the figure represented more than twice the combined new debt issued throughout the entirety of 2025.

JP Morgan analysts have projected that debt-funded expenditure on artificial intelligence infrastructure could total approximately USD 4.1 trillion over the four-year period from 2026 to 2030, figures the FPC has referenced in its assessment. The panel noted that growth in AI-related funding had also become apparent in private markets, which are anticipated to take on a larger role in financing such investment. Projections from Morgan Stanley put private credit financing of data centre capital expenditure at roughly USD 700 billion across the three-year period from 2026 to 2028.

The FPC drew attention to vulnerabilities stemming from a combination of growing leverage, insufficient transparency and, in some cases, interconnected financing arrangements tied to artificial intelligence. Such dynamics may hamper accurate risk evaluation and amplify potential losses should anticipated returns on artificial intelligence spending fall short.

These apprehensions reach into stock markets as well as debt financing arenas. The FPC noted that equities linked to artificial intelligence and chipmakers retreated during July, with certain leveraged investors compelled to close out positions as market turbulence mounted. Forced selling and portfolio adjustments intensified the swings across financial markets.

The panel noted that while trading conditions had remained stable and no broader systemic pressures were evident, the possibility of a steeper downturn spreading into core markets had not been ruled out. The FPC observed that borrowing levels among hedge funds remained above historical norms, while asset valuations stayed elevated on the back of robust earnings forecasts.

Any notable shift in market expectations around AI-related earnings — whether driven by doubts about the pace of technological development or its broader uptake — has the potential to prompt a pronounced reassessment of asset valuations.

The panel additionally flagged concerns about the growing capabilities of AI systems. Models launched in the third quarter of 2026 demonstrated ongoing progress in autonomous task execution and the detection and exploitation of software vulnerabilities within controlled test settings.

The FPC noted that anticipated productivity improvements driven by artificial intelligence are increasingly shaping growth outlooks and public finance projections. Shifts in those projections have the potential to affect valuations linked to artificial intelligence, along with government bond markets. Minutes from the gathering noted that the volume of capital flowing into artificial intelligence is forging deeper ties between the technology sector, debt instruments, private lending, and wider risks to financial stability.