Banks and institutions warning of an AI bubble: Deutsche Bank, Bloomberg and others

Which banks and official bodies have issued an AI bubble warning?

Two warnings pass the standard of this series: a named author, a date, and a source that answered when checked. The Bank of England’s Financial Policy Committee published one on 8 October 2025. Deutsche Bank’s head of FX research, George Saravelos, wrote the other in a September 2025 research note that the press reported.

Other banks and agencies have commented on AI valuations. Their remarks are left out here because they were not confirmed on a source for this page. The list is short on purpose. A shorter list of checked warnings serves a reader better than a long list of unchecked ones.

What did the Bank of England say about AI bubble valuations?

The Financial Policy Committee wrote that on a number of measures equity market valuations appear stretched, particularly for technology companies focused on AI (AJ Bell, 8 October 2025). It said the risk of a sharp market correction had increased. Bloomberg Law reported the same update and linked the risk to stretched AI valuations and to concerns about the Federal Reserve’s independence.

The Committee named triggers. Disappointing progress on AI capability or adoption, or increased competition, could push valuations down. It also pointed to material bottlenecks in power, data, or commodity supply chains. These are observable items, which lets a reader watch them and does not require a prediction of timing. Because the Committee publishes minutes of each meeting, you can compare the next update with this one and see whether the list of concerns grew or shrank.

What did Deutsche Bank’s George Saravelos say about AI spending?

Saravelos wrote that AI machines appear to be saving the US economy right now. He added that without tech-related spending the US would be close to, or in, recession this year, and that growth was not coming from AI itself but from building the factories to generate AI capacity (Futurism, 24 September 2025). The note is a research view from one bank, and other desks may hold different views, so it counts as one voice among several.

His warning is about dependence. He wrote that capital investment needs to remain parabolic to sustain the effect, and he called that highly unlikely. The argument concerns the economy more than prices: if the build-out slows, growth slows. It fits the bubble debate because spending and valuations rise together.

What role does Bloomberg play in reporting an AI bubble warning?

Bloomberg is a reporter of warnings more than an author of this one. Bloomberg Law carried the Bank of England update on 8 October 2025, with the phrase sharp market correction. This page uses it to confirm the date and the wording of the Committee’s warning, and it relies on that report only for those two facts.

Readers who search for Bloomberg and the AI bubble will find many pieces. Judge each by the same test: is the claim the outlet’s own analysis or a quote from someone else, is it dated, and can you reach the original? A report that links back to its source is stronger than one that paraphrases it.

How does an institutional AI bubble warning differ from an investor’s?

A fund holding puts speaks for a position. A central bank speaks for financial stability, and its aim is to describe risk that could spread to credit, savings, and jobs. The two voices answer different questions. The investor asks where a price is going, and the institution asks what happens to the system if it moves.

Institutions also write carefully. The Committee’s wording is conditional: valuations appear stretched, and the risk has increased. Neither phrase names a date. A reader who expects a crash call from a central bank misreads the genre, which signals vulnerability and keeps the timing open. Read institutional warnings as a map of weak points, and use them to decide what to monitor each quarter.

What do the Bank of England and Deutsche Bank warnings share?

Both warnings rest on concentration. The Bank of England links stretched valuations to the exposure of technology companies, and Deutsche Bank links economic growth to AI capital spending. In each case a large share of results depends on one theme, and a disappointment in that theme would travel widely. Concentration is a useful word because it points to something you can measure in your own portfolio or supplier base.

The EQOS Asia AI study describes a related fragility in business terms. It scores the enterprise AI market at 0.337, in the fragile zone, and records valuations up ×4 in 18 months (EQOS Asia AI study, April 2026). The score and the warnings use different methods and point the same direction.

How can you weigh a bank warning about the AI bubble?

Ask three questions. What does the institution measure: prices, spending, or credit? What triggers does it name, and can you watch them? What is its role: a regulator, a research desk, or a news outlet? Record the answers beside the date. The notes then show which warnings were about prices, which about the economy, and which were about neither.

This page is general information as of October 2026, not investment advice. If an AI exposure sits inside a decision of yours in Singapore, ASEAN, or Hong Kong, the EQOS Asia pricing page lists the diagnostic formats that score the exposure with a method you can recompute.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile