Who says there is an AI bubble? The voices and their reasons

Who says there is an AI bubble, and who counts as a voice?

A voice earns a place on this list when it speaks in public, names a mechanism, and can be checked. Five voices meet that rule as of October 2026: an investor known for the 2008 mortgage trade, a macro fund founder, the long-time chief of Berkshire Hathaway, the chief executive of OpenAI, and the staff of the Bank of England. Their positions range from a bet with money to a remark about prices.

Each voice gets one section below with dated statements and a link to its own page. Every quote comes from a named publication that answered when checked. Opinions without a source stay out of this page, so each line can be traced back to its origin and re-read by you.

What does Michael Burry say about the AI bubble?

Michael Burry runs the fund Scion. A regulatory filing for 30 September 2025 listed put options with a notional value of about $187 million on Nvidia and $912 million on Palantir (CNBC, November 2025). The filing gave no strike prices or expiry dates, so the size of the real bet stays unclear. The full reading of his case sits on the page about Michael Burry and the AI bubble.

Burry also argued on X that hyperscalers understate depreciation by assuming long chip lives, and he estimated about $176 billion of understatement from 2026 through 2028. His case rests on accounting, not on a price target. Three deeper pages split it into the Big Short comparison, the puts, and the depreciation argument.

Where does Ray Dalio place the AI bubble in his stage view?

Ray Dalio founded Bridgewater. In a November 2025 CNBC interview he put the AI bubble at about 80% of the euphoria seen before the 1929 crash or the 2000 dot-com peak, and in a January 2026 post on X he called AI a boom in the early stages of a bubble (Fortune, 6 January 2026).

His framing is a stage, not a verdict. He also flagged Federal Reserve policy as a key unknown for 2026, because easy money can feed a bubble for longer. Ray Dalio’s prediction explained covers the 80% figure, the stage logic, and what his view does and does not say about timing.

What does Warren Buffett’s silence tell us about the AI bubble?

Warren Buffett has no quoted AI bubble call on this page, and no quote is invented here. What is public is Berkshire’s behaviour: cash and short-term investments reached a record $381.7 billion at the end of September 2025, and the company bought back no shares in the third quarter (Yahoo Finance, November 2025). Buffett was due to step down as CEO at the end of 2025.

Reading behaviour is weaker evidence than reading a statement. A large cash pile can mean caution about prices, a lack of deals at a fair price, or both. The page reading Warren Buffett’s silence sorts those readings and shows which ones the data can support.

What does Sam Altman admit about the AI bubble?

Sam Altman, chief executive of OpenAI, told reporters over dinner in August 2025: "Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes." He added that some start-up valuations are "insane" (Fortune, 19 August 2025). Those words come from the person who runs the best-known AI company, which gives them unusual weight.

His stake matters too: OpenAI raises money from the same investors. The page on Sam Altman and the AI bubble separates what he conceded about prices from what he still defends, which is the long-term value of the technology.

Which banks and institutions flag the AI bubble as a risk?

The Bank of England’s Financial Policy Committee wrote on 8 October 2025 that equity valuations "appear stretched" on several measures, particularly for technology companies focused on AI, and that the risk of a sharp market correction had increased (AJ Bell, October 2025). It named disappointing AI progress and power bottlenecks as possible triggers.

Institutional warnings read differently from investor bets: they describe system risk, not a trade. The page on banks and institutions warning of an AI bubble covers Deutsche Bank and Bloomberg alongside the Bank of England, and states which of their claims were checked.

How can you weigh each AI bubble claim by its source and stake?

Ask three questions of every claim. Where does it come from: a filing, an interview, or a repost? When was it said: a view from August 2025 may have aged by now. What does the speaker gain: a fund with puts gains if prices fall, and an AI chief gains if funding stays open. Stake does not make a claim false, and it tells you which way to double-check.

A claim that passes all 3 questions earns a place in your notes next to a number. A claim that fails one stays an opinion until you find the source. The AI bubble checklist turns this habit into ten tests, and the EQOS Asia pricing page lists how a decision diagnostic applies the same discipline to your own exposure.

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