Bank of England warns AI bubble could trigger sharp correction as debt risks mount

Bank of England Sounds Alarm on Overheated AI Sector

The Bank of England has issued a stark warning about growing debt risks in the artificial intelligence sector, cautioning that a debt-fuelled AI spending boom could unravel with severe consequences for financial stability. In its twice-yearly Financial Stability Report released Tuesday, the central bank warned that UK share values are approaching the "most stretched" levels seen since the 2008 financial crisis, while equity valuations in the United States echo the dangerous territory that preceded the dot-com bubble burst 1

1

2

2

.

Stretched Stock Market Valuations Raise Concerns

The Bank of England specifically highlighted that valuations are "particularly stretched" for companies focused on artificial intelligence infrastructure, raising concerns about an AI bubble that could burst with cascading effects across global markets. The central bank estimates that AI has driven two-thirds of this year's gains on the S&P 500 index, and investment in the technology accounted for half of US economic growth in the first half of 2025 2

2

. This concentration of market gains in a single sector has raised red flags about whether current valuations reflect sustainable fundamentals or speculative excess.

Trillions of Dollars of Debt Fuel Infrastructure Expansion

While current investment in AI technology is mostly driven by cash held by "hyperscalers"—the massive tech companies building out data centers and computing infrastructure—the Bank of England warns that the financing landscape is reaching an inflection point. Around half of the expected $5 trillion of AI spending over the next five years will be financed externally, largely through debt 2

2

. This shift toward debt financing introduces new vulnerabilities into the system, particularly if the anticipated returns from AI investments fail to materialize.

Early Warning Signs Emerge in Credit Markets

The central bank pointed to troubling early indicators that investor confidence may be wavering. The five-year credit default swap spreads of Oracle—an AI company with lower free cash flow margins than some larger hyperscalers that has issued substantial debt this year to finance AI infrastructure spending—have widened dramatically from less than 40 basis points to around 120 basis points since end-July 2

2

. This contrasts sharply with the steady credit default swap spreads of US investment-grade corporates more broadly, suggesting specific concerns about debt-laden AI companies.

Market Correction Could Spill Over Into Wider Debt Markets

The Bank of England warned that a sharp correction in AI stocks would not remain contained within the tech sector but would spill over into wider debt markets with far-reaching consequences. A sharp fall in stock valuations could hit UK household wealth, feeding through to consumer spending and triggering losses on lending to firms investing heavily in AI infrastructure 2

2

. The central bank cautioned that such a scenario would ramp up borrowing costs for companies more widely, potentially affecting household wealth and borrowing costs across the economy 1

1

.

Parallels to the Dot-Com Bubble Draw Scrutiny

The Bank of England's warning adds to growing concerns drawing parallels to the dot-com bubble that burst in stock markets in the early 2000s. As firms invest heavily in building out data centers and other infrastructure needed for AI technology, questions mount about whether valuations have reached irrational levels divorced from underlying business fundamentals. The central bank stated: "If material credit losses on AI lending were to occur (directly or indirectly), this could have spillovers to broader credit conditions including in the UK" 2

2

. Market participants will be watching closely for further signs of stress in credit markets and whether the rapid expansion of AI infrastructure investment can deliver returns sufficient to justify current valuations and service the mounting debt loads.