Bank of England warns AI poses growing financial stability risks

Rising debt-financed AI spending and potential spillover

The BoE has tracked how the AI infrastructure boom is shifting from equity and cash to debt. In its December 2025 report, the FPC noted that “while AI infrastructure investment has mostly been financed by the cash flows of large, profitable technology companies and equity investments to date, debt financing is increasing quickly” . The BoE estimated that around half of the expected $5 trillion in AI spending over the next five years could be funded through debt . A correction in AI-linked assets could transmit losses through broader debt markets, with early warning signs already visible in credit default swaps of companies relying on debt to finance investments . Bloomberg separately reported that rising long-term interest rates threaten to upset the AI boom, just as it began to spread to chip stocks . The BoE’s July 2026 report additionally warned of “increasing complexity and opacity in debt structures used” around AI .

Correlated trading strategies and systemic risk

The BoE’s Financial Policy Committee has highlighted that widespread use of uniform AI models could heighten correlation among financial firms, amplifying shocks during a stress event . An academic paper citing the BoE notes that the FPC warns this correlation could “intensify shocks, undermining resilience” . The BoE’s own Financial Stability in Focus report from April 2025 stated that “the potential future use of more advanced AI-based trading strategies could lead to firms taking increasingly correlated positions and acting in a similar way during a stress, thereby amplifying shocks” . This is not yet deemed a systemic risk, but the FPC determined in March 2026 that “risks are likely to increase, potentially rapidly, amid growing intent among financial firms to expand their deployment” of more advanced AI .

Lack of transparency and explainability

A general-challenge the BoE has identified is the opacity of AI models. The Financial Policy Committee’s Financial Stability in Focus mentioned “opacity in AI outputs, particularly from generative or agentic AI systems” . A separate analysis of the BoE’s work notes that the FPC warns “the lack of explainability combined with increasing autonomy means financial institutions may take on risk-taking positions they do not properly understand” . The BoE’s FPC Record from March 2026 acknowledges that agentic AI — systems that act autonomously — is a growing concern . Reuters reported that British banks’ race to deploy agentic AI is raising new risks for regulators, new risks the BoE is now monitoring .

Heightened cybersecurity vulnerability

The Reuters report on the BoE’s July 2026 warning explicitly states that AI “increases banks’ vulnerability to cyberattacks” and that this was among the additional dangers highlighted since the prior report . The BIS has also identified cybersecurity as a primary concern in central bank AI adoption, ranking it at the top of operational risk priorities .

Deeper global alignment: BIS warnings and US Treasury draft

The BoE’s concerns are not isolated. In its 2026 Annual Economic Report, the Bank for International Settlements (BIS) warned that “intense competition for market leadership may fuel overinvestment further” in AI, “increasing the risk of a sharp reversal if AI payoffs disappoint” . The BIS stated that “the five largest hyperscalers are set to spend over $1 trillion” in AI capital expenditure, raising questions about “the sustainability of the current economic expansion” . The Wall Street Journal reported the BIS sees this AI investment boom could jeopardize profitability and “plunge certain economies into recession” .

Regarding a US Treasury draft: a July 2026 report from NOTUS stated that a draft report inside the Treasury Department warned that AI firms are “more deeply entrenched in the U.S. economy than their dotcom predecessors” and that a downturn could send “shockwaves throughout the entire economic ecosystem” . The Treasury leadership publicly refuted that draft, with a spokesperson saying the department “rejects the draft” . A congressional oversight letter from October 2025 also warned of AI bubble risks, citing the dotcom comparison and noting the AI industry would need to earn $2 trillion in annual revenue by 2030 to cover data center costs .

Key gaps and limitations

Several claims that were part of the original query could not be independently verified within the available source set. There was no confirmation found for the ECB ordering banks to submit AI cybersecurity plans in specific detail, nor for an FSB push for tighter AI agent safeguards, nor for EU AI Act enforcement. The specific quote about a “lack of explainability combined with increasing autonomy” came from a third-party analysis rather than a direct BoE document . The stronger claim that AI infrastructure buildout is “largely” debt-financed was not fully supported — the BoE’s own letter notes that “to date” investment has been mostly from cash flows and equity, with debt increasing quickly .

Bottom line

The BoE’s warnings are clear across multiple reports: AI-related stretched valuations, leveraged investor bets, rising debt-financed spending, correlated trading strategies, and cybersecurity vulnerabilities are being treated as real and growing financial-stability risks. The BIS independently warns of AI overinvestment and possible recessionary consequences. The Treasury draft report adds political weight in the US, though it was formally disavowed. For anyone tracking financial stability in 2025-2026, the consensus among these central-bank sources is that AI has moved from a niche tech risk to a front-and-center macro-financial concern.