BoE’s Bailey warns crowded AI trade could end in asset-price correction
The warning landed as UK 30-year gilt yields broke above 6% and US 10-year yields hit 5.34%, tightening conditions for risk assets.
Bank of England Governor Andrew Bailey warned that the AI investment boom could trigger financial market shocks, with valuations vulnerable if expectations fail to convert into profits. His comments landed into a bond sell-off that pushed long-dated UK and US yields to multi-decade highs.
Key Takeaways
- Andrew Bailey said the AI investment boom could produce market shocks and that “You could see some correction of asset prices at some point.”
- Long-end rates pushed to multi-decade highs, with UK 30-year gilt yields above 6% and US 10-year Treasury yields at 5.34% on Thursday.
- Nvidia was cited at a $5.5tn (£4.14tn) market valuation, a level tied directly to investor expectations for AI-driven profits.
- Bailey also framed AI as an operational risk amplifier, warning that deepfakes and AI-enabled vulnerability discovery can hit confidence and are hard to trace.
Multi-Decade-High Yields Raise the Stakes for Risk Assets
The macro backdrop is doing the opposite of cushioning a crowded equity theme. On Thursday, the yield on 30-year UK government bonds rose above 6%, the highest level since 1998. US 10-year Treasury yields reached 5.34%, the highest since 2002.
Those numbers matter because a bond yield is the effective interest rate investors earn from holding a bond, and yields rise when bond prices fall. A move like this is a sell-off in duration, and it tightens financial conditions even before any central bank changes policy.
For traders, the transmission channel is simple: higher long-end yields raise the discount rate used to value future cash flows. That tends to pressure long-duration equities, which is exactly the bucket that AI-linked mega-caps and AI-adjacent growth names sit in. When the long end reprices, the market gets less forgiving about “profits later” stories.
The immediate catalyst for Thursday’s move was not pinned to a single event. An analyst cited in the same discussion described “no particular trigger,” leaving the sell-off framed as part of a broader environment of high global borrowing costs.
“Everybody Is Priced to Be a Winner”: Crowding, Valuation, and the Nvidia Signal
Bailey’s core market point was not that AI is fake, but that the trade is crowded and priced for a clean outcome. “Everybody is currently priced to be a winner,” he said, before adding that “you look back at the past, not everybody is a winner.”
He used a specific historical analogy to make the mechanism legible: early leaders can disappear even when the category is real. “Google was not the first market leader in internet search. It was Netscape. Nobody can remember Netscape today. It doesn't exist,” Bailey said.
That framing lands harder when it is paired with the valuation anchor he discussed. AI chipmaker Nvidia was described as the world’s most valuable listed company, with a market valuation of $5.5tn (£4.14tn), a level attributed to investor expectations that AI will generate large profits.
Bailey also pointed to the scale of capital being committed across the stack. Alphabet, Meta, Microsoft, and Amazon were described as spending hundreds of billions of dollars on AI, and two large AI companies, Anthropic and OpenAI, were described as preparing to sell shares on US stock markets. The article characterized those potential share sales as moves many believe could lead to “hundreds of billions of dollars more” flowing into the industry, but it did not provide confirmed timelines or fundraising totals.
The failure mode in Bailey’s setup is not “AI stops existing.” It is that the market is paying today for a future where multiple firms simultaneously earn dominant, durable margins, and the repricing happens when that future turns out to be narrower than the equity curve implies.
AI as a Financial-Stability Risk: Cyber, Deepfakes, and Traceability
Bailey’s warning was not limited to valuation math. He also described AI as a cyber-risk amplifier, saying it creates a “much more powerful way of uncovering vulnerabilities,” and called it “a very powerful, potentially very powerful, weapon” “In the wrong hands...”.
He tied that to a concrete deepfake incident. In June, deepfake images depicting Bailey and Nigel Farage in a physical fight were promoted on social media platform X. Bailey said the fakes were produced in a way that left the Bank struggling to identify their origin, adding: “We've got to be able to trace these things back. And we need a lot of help from the tech sector to do that,” and calling the quality of deepfakes “alarming.”
For markets, this is a different shock path than an earnings miss. Incident-driven confidence hits can arrive without warning, and they can force risk committees to treat AI exposure as operational risk, not just a growth bet. Bailey’s traceability point is the uncomfortable part: if the origin of a market-moving fake cannot be identified quickly, the window for misinformation to do damage is measured in hours, not quarters.
Bailey also sketched the upside case for policymakers. He said AI could speed up work that supports the Monetary Policy Committee (MPC), the Bank of England committee that votes on the UK’s benchmark interest rate, while stressing: “It's not taking a decision, but it's a tool in the hands of the policy maker and that's good,”.
Bailey Puts the AI Boom on Bubble Watch
Bailey said the Bank of England is watching the “huge amounts of money being invested in AI” “very carefully,” and warned that the investment wave could produce financial market shocks. Asked whether an AI bubble could burst, he said: “You could see some correction of asset prices at some point.”
He described the investment surge as rational in direction but dangerous in pricing. “There is a large, very large, amount of investment going into this sector now, and of course that's natural because it's a major area of growth,” Bailey said. “And of course you see that the asset prices of the companies that are developing it have gone up a lot and that reflects the fact that there are high expectations of what it can deliver.”
The central bank’s posture, as he described it, is resilience rather than prediction. “We are prepared for the fact that there will be, I think, some shocks come along to markets and we have to deal with that. We have to make sure the system is resilient,” Bailey said.
The forward path is mostly about whether the macro and positioning backdrop stays hostile. If UK 30-year gilt yields remain above 6% or continue rising, and if US 10-year yields hold near or above 5.34%, the discount-rate pressure on duration-heavy risk assets does not ease. Traders will also be looking for any follow-up from the Bank of England that formalizes AI-linked valuation concentration as a financial-stability monitoring topic, whether via speeches, testimony, or stability reports.
The other open variable is supply and narrative reinforcement. The article referenced expectations of US stock-market share sales by Anthropic and OpenAI that could pull “hundreds of billions of dollars more” into AI, but without confirmed timing or size. If those details firm up, they become a real flow story rather than a belief-based one. Finally, Bailey’s cyber and deepfake warnings set a separate tripwire: a major AI-linked incident that hits a large platform or financial institution could reprice risk faster than a slow earnings narrative.
My Read: Why Crypto Traders Should Treat This as a Cross-Asset Risk Signal, Not a Timing Call
The threshold that matters is not whether AI is “a bubble,” it is whether the market can keep paying peak multiples for peak expectations while the long end is repricing to multi-decade highs. With UK 30-year yields above 6% and US 10-year yields at 5.34%, the discount-rate regime is already doing the work of tightening, which makes any crowded equity theme more likely to spill into broader risk sentiment.
Bailey’s most useful line for traders is “Everybody is currently priced to be a winner,” because it defines the asymmetry. If the winners narrow, or if an incident-driven confidence shock hits before cash flows arrive, the correction does not need a single catalyst to propagate across equities and into crypto beta. This matters in practical terms if elevated yields persist long enough to force a repricing of AI-linked mega-cap concentration rather than just a rotation within tech.
