Bank of England warns AI investment could trigger market shocks

- Andrew Bailey warned that AI investment could trigger market shocks, citing Nvidia at $5.5tn and heavy spending by major tech firms.
- UK 30-year bond yields stand above 6 per cent and US ten-year yields at 5.34 per cent as borrowing costs reach multi-decade highs.
- The Bank has struggled to trace deepfake images of Bailey and Nigel Farage and called for more help from the tech sector.
- Bailey said AI would support, not replace, analytical work supporting the Monetary Policy Committee.
The Bank of England’s governor Andrew Bailey has warned that AI investment could trigger financial market shocks, adding that the central bank expects such shocks and must ensure the financial system remains resilient. In the same interview he acknowledged the technology could strengthen economic growth, but placed the emphasis on preparation rather than prediction.
The mechanics of the concern are specific. Bailey said much of the money invested in or lent to AI firms is creating very high market valuations, and that a correction in asset prices is possible. He cited Nvidia as the world’s most valuable listed company at $5.5tn, while Alphabet, Meta, Microsoft and Amazon are spending hundreds of billions of dollars on AI. Anthropic and OpenAI are preparing to sell shares on the US stock market, a step expected to draw further large sums into the sector. Bailey compared the current enthusiasm with the early internet, noting that Netscape once led the search market and no longer exists. The Bank is monitoring investment “very carefully”, a posture that matters for prudential supervision: supervisors who see concentrated exposure to a single technology will examine how much of it sits on bank balance sheets, in pension funds, or in infrastructure debt. No new capital requirement has been announced, so the enforcement path runs through existing tools such as stress testing rather than fresh rules.
The warning coincides with borrowing costs at multi-decade highs in several countries. Thirty-year UK bond yields stand above 6 per cent, with US ten-year yields at 5.34 per cent, which adds pressure to public finances during a period of inflation and heavy AI spending.
Bailey also identified risks beyond asset prices. AI can make cyber attacks more powerful by uncovering software vulnerabilities, and deepfakes can mislead the public. He said the Bank had struggled to trace deepfake images of himself and Nigel Farage, and called for more help from the tech sector. In education and academic work, where the authenticity of source material is a first-order concern, that difficulty is already operational: verification and provenance of AI-generated content can no longer be treated as edge cases by institutions that rely on documentary evidence. He noted AI could speed up analytical work supporting the Monetary Policy Committee, while emphasising it would support, not replace, policy decisions.
What remains to be evidenced is whether the anticipated repricing materialises, and whether the deepfake-tracing problem yields to better provenance tools. The AI firms’ expected share sales will show how much new capital enters the sector, while the Bank’s stress tests will show how much of that exposure the financial system can absorb.