Bridgewater Associates founder Ray Dalio stated that the current artificial intelligence (AI) investment boom carries the same bubble dynamics seen in past technology cycles, drawing a parallel to the late 1920s.

In an interview on Bloomberg Television’s “Insight with Haslinda Amin,” aired Tuesday, Dalio explained that bubbles “always come together” with major innovations. He cited his study of 500 years of market history to identify this repeating pattern, noting that markets are not fully pricing assets correctly due to the unknown future value of AI.

Wealth versus money gap

Dalio identified the divergence between wealth and money as the primary danger. “Wealth is in a sense easy to create but you can’t spend wealth,” he said. He explained that heavy borrowing to purchase assets, or forces such as a wealth tax requiring asset sales for cash, can trigger a bubble burst. According to Dalio, this dynamic is currently active.

He cautioned that AI investment today is concentrated in a limited number of companies and has become “very expensive,” with uncertain future cash flows. While he noted that AI has been “fantastic” for productivity, the valuation risks remain significant.

Early signs in Asian markets

When asked about South Korea and Taiwan, which have attracted heavy AI capital expenditure, Dalio pointed to early indicators of instability. He observed that these markets have invested beyond their own productivity needs and are now deploying capital globally. This has created what he described as early signs of “bubbles beginning” and “bubbles bursting.”

Dalio emphasized that while the risk is “not systemically threatening,” the concentration and cost of current investments warrant caution.

Context of rising capex and debt

Dalio’s comments follow warnings he issued last month on X, where he stated “we are now in a bubble,” comparing AI to historical innovations like railroads and the Industrial Revolution. His analysis aligns with broader market data showing hyperscalers’ spending projected to approach $1 trillion in 2027.

Recent reporting indicates this spending is increasingly financed through debt and credit markets rather than equity alone. Analysts have warned that the pace of spending growth poses a macro risk if demand falls short.

What the numbers show

The juxtaposition of Dalio’s historical analogy with current financing trends highlights a structural shift in how innovation is funded. In previous cycles like the railroad era, equity often bore the initial risk. However, the source data indicates a move toward debt-financed expansion for AI infrastructure. This increases sensitivity to interest rates and liquidity conditions, potentially amplifying the “wealth versus money” gap Dalio describes, as assets must be liquidated at scale to service debt obligations.

Bridgewater Associates founder Ray Dalio warned Wednesday that the current artificial intelligence boom follows historical patterns of past technology bubbles. He stated that "we are now in a bubble," arguing that while the underlying innovation is transformative, valuations have become detached from reality.

Dalio drew parallels to the late 1920s, when electricity, refrigeration, and automobiles emerged simultaneously. He noted that such moments naturally attract investors but warned that failing to distinguish the "miracle" from the investment leads to crashes similar to the 1929-1933 bust.

Market Context and Capex Surge

The warning coincides with aggressive infrastructure spending by major technology firms. Amazon.com Inc., Microsoft Corp., Alphabet Inc., and Meta Platforms Inc. are increasing capital expenditures for data centers. U.S. hyperscalers are on pace to spend roughly $916 billion over the next 12 months.

Apollo Global Management’s Torsten Slok noted this trajectory puts data-center capex at 3.1% of U.S. GDP by 2027. This figure is nearly three times the peak share telecom investment reached during the dot-com buildout.

Nvidia’s Role in the Buildout

Nvidia Corp. continues to fuel this expansion by acting as a "buyer of last resort" and partnering with Wall Street to finance customer data centers. Nvidia shares closed 0.91% lower at $223.67 on Wednesday.

Dalio emphasized that investors must evaluate who truly benefits from AI and whether current stock prices reflect excessive optimism rather than real returns.