Dalio Warns: AI Bubble Is Approaching the Extreme Levels Seen in 1929 and 2000

Dalio says the AI mania has pushed the market to the edge of a bubble burst, approaching the extreme levels seen in the 1929 stock market crash and the 2000 dot-com bubble.

Billionaire investor and founder of Bridgewater Associates, the world's largest hedge fund, Ray Dalio issued one of his harshest warnings to date on the current market environment in a latest interview podcast.

He believes that the artificial intelligence (AI) mania has pushed the market to the edge of a bubble burst, evoking memories of the 1929 (stock market crash) and 2000 (internet) bubbles.

The host opened the show by mentioning previous guest Jeremy Grantham, a legendary British investor and co-founder of asset management firm GMO. Grantham stated on the show earlier that the market is facing "the largest investment bubble in American history."

Dalio responded directly: "He is right."

His warning comes as SpaceX has completed the largest IPO in history, while Anthropic and OpenAI are seeing their valuations skyrocket toward the trillion-dollar level. This is exactly the obvious warning sign of a bubble he points out — the surge in speculative offerings.

"Few things are easier to create than stocks," he described how a company can raise 50 million U.S. dollars, reach a valuation of 1 billion U.S. dollars, and create a paper billionaire, without ever actually completing a transaction worth 1 billion U.S. dollars. He believes that the surge in stock offerings is one of the two major factors that prick the bubble, and the other is rising interest rates.

SpaceX went public in June, but its share price has traded below the IPO offer price ever since. S&P predicts its free cash flow will remain negative before 2029, while Moody's notes that Elon Musk's concentrated voting rights pose a corporate governance risk. Anthropic has secretly submitted its IPO application, which is expected to go public as early as October with a target market capitalization close to 1 trillion U.S. dollars; OpenAI has also submitted a separate IPO application targeting a market capitalization of more than 1 trillion U.S. dollars, but affected by "changing market dynamics", its listing timeline has been postponed from the end of 2026 to 2027.

Dalio uses a simple thought experiment to explain the mechanism at play: you spend 100 U.S. dollars to buy shares in an AI company, then use this paper wealth as collateral to borrow money. When market sentiment shifts and everyone is desperate for cash, the share price could plummet to 25 U.S. dollars, while the loan still needs to be repaid.

His core point is to distinguish between wealth and money: "Wealth is not the same as money. You will see many people become wealthy, but you cannot spend that wealth. You have to sell your wealth to get money, because you can only spend money."

When asked if he sees signs of a bubble, Dalio answered unequivocally: "Yes, yes, yes. These are all typical signs that we are in a bubble." He emphasized that a bubble is a "matter of degree" rather than an absolute yes-or-no scenario, noting that many inexperienced investors are piling into leveraged investments, including leveraged ETFs that track the stock market.

"It is more like a game of chance," he said.

The "Big Cycle" Behind the Scenes

Peter Berezin, an analyst at BCA Research, has been arguing for months that the AI trade is "largely a profit bubble rather than a valuation bubble" — a type of bubble that historically often appears in boom-and-bust industries such as the banking sector before 2008.

This pattern repeated itself in the week from July 26 to 31: Microsoft and Amazon rose 18% and 10% respectively on strong capital expenditure outlooks, while Alphabet and Meta fell 4% and nearly 10% respectively despite both reporting strong earnings — indicating that investors no longer award valuations to companies simply for existing, and the bubble phase is drawing to a close.

Dalio emphasized that the two major factors that typically prick a bubble are rising interest rates that increase debt servicing costs, and the surge in stock offerings triggered by companies rushing to capitalize on investor enthusiasm.

He even cited a real case: a friend who runs an AI company told him that he is raising hundreds of millions of dollars precisely because he expects a market downturn — he plans to use the funds to acquire troubled competitors after the market rebounds.

His more important warning is not aimed at the market itself, but at the situation after the bubble bursts. In Dalio's view, the burst of the AI bubble is not just a financial event, but a fuse that ignites political and geopolitical conflicts, which historically often accompany the end of an 80-year cycle — meaning that the risk that really needs to be guarded against may not be the loss of the investment portfolio, but the ensuing turmoil.

The so-called "80-year big cycle" refers to debt dynamics, widening wealth gaps, domestic political conflicts, and shifts in geopolitical power. Dalio warns that when the bubble bursts, "people will turn on each other", citing the example of six prime ministers in the UK in seven years, pointing out that governments are short of funds while voters are blaming each other over how to raise money.

This article is from the WeChat Official Account "Sci-Tech Innovation Board Daily", written by Huang Junzhi, and published with authorization from 36Kr.