Bank of England Warns of AI Bubble and Risks to Financial Stability: What the Central Bank Sees in the Financial System

Britain's central bank has put AI near the top of its list of threats. In its July 2026 Financial Stability Report, the Bank of England said the technology is a growing danger to the financial system for two reasons: investors are betting heavily on its success, and it makes banks easier to attack. Risks have risen this year, though UK lenders remain resilient.

Why regulators warn of a dotcom-style risk: asset valuation and debt

The headline "Bank of England warns of AI bubble" rests on three observations:

• Stretched prices. Share prices of AI-linked tech companies have become more stretched, and the Bank's earlier reports compared US levels with the dotcom era.
• Borrowed money. Hedge funds are borrowing to buy shares, and AI builders are borrowing to fund investment in chips and buildings.
• Scale. The Bank has cited industry forecasts of more than $5 trillion of spending on AI infrastructure, with about half financed externally, mainly by borrowing.

If revenue disappoints, each asset bought at a high valuation falls, and every lender exposed to the sector takes losses. Historically, bubbles built on credit have hurt more than those built on equity alone. An economist would add that not every firm valued highly today will earn a return.

Risks to financial stability from frontier AI models: Anthropic and Mythos

The second channel is cyber. The report says frontier AI could raise the sophistication of attacks, because the newest AI models are increasingly able to exploit weaknesses in software. Regulators cite Anthropic's Mythos as an example of this capability; Anthropic limited its release under Project Glasswing so defenders could secure critical infrastructure first.

The Bank says it is unclear whether frontier AI helps attackers or defenders more. It does expect more frequent patching on any platform, and each update brings its own chance of an incident. Under Governor Andrew Bailey it has also committed to AI-specific stress test work, since many institutions that deploy AI may react the same way in a crisis. The Financial Stability Board, which reports to the G20, has raised a similar systemic concern about reliance on a few providers.

Scam risk for the public: the BDCB case and what each regulator advises

The warning is not limited to London. The Brunei Darussalam Central Bank told members of the public to beware of videos made with artificial intelligence after a fake clip misused its name. The video appeared to impersonate a senior figure and to promote a scheme the BDCB never endorsed. It is a typical scam: fraudsters manipulate footage, promise quick profits and push victims into making payments. Wide deployment of video tools makes it cheap.

The Bruneian regulator's advice applies anywhere in global finance, because risks to financial consumers look the same everywhere:

1. Verify any offer through official channels.
2. Deal only with licensed entities.
3. Treat unsolicited "investment opportunities" as suspect.

Investment risk: how to respond when firms deploy AI at scale

Regulators do not warn against AI adoption itself; they accept that AI could lift productivity. Their point is that systemic risk grows when prices, debt and one application of technology move together. Three habits help:

• check how much of a portfolio depends on a single theme;
• read what companies say about financing, not only growth;
• follow regulatory statements, which often move before markets.

Clear Report was built for this checking. Our AI-generated market reports summarise news, bank research and earnings calls by ticker, and we see financing language change before estimates do. We are a research partner, not an adviser. For context, read our Fed policy and AI capex analysis.

FAQ: global financial risk and AI

Does this affect markets outside Britain?

Yes. Global financial markets share the same AI leaders.