Former Fidelity fund manager George Noble has warned that a collapse of the AI bubble could cause 17 times more damage than the collapse of dot-com companies, which wiped about $5 trillion off the Nasdaq.
According to Polymarket, the probability of an AI bubble bursting in 2026 has climbed above 17% after recently falling from 30% to 14%. Contracts using different resolution criteria put the probability between 16% and 24%, with traders taking into account falling tech stocks, revenue concerns and weakness in global markets.
Noble linked his predictions to the large sums being pumped into AI infrastructure, arguing that the financial fallout could extend far beyond tech companies if expected returns don’t occur.
“The consequences could actually be much greater,” Noble said of increased capital spending on AI.
AI Bubble Odds Rebounded Above 17%
Further pressure on semiconductor and technology stocks added to these concerns. The Wall Street Journal reported that U.S. stock futures fell Thursday as AI anxiety spread to Asian markets, where SK Hynix and Samsung Electronics fell nearly 9%.
The two South Korean chipmakers plan to spend billions of dollars on semiconductor factories and AI capabilities. Their declines came as investors questioned whether the revenue generated by AI services would justify the sector’s rising infrastructure bill, the report said.
IBM added to the unease after its shares suffered their biggest daily fall since 1968, falling nearly 25% earlier this week. Market data cited in the report showed IBM closed down 2.7% at $211.20 on Wednesday, taking its multi-session decline beyond 26%.
In its warning, IBM said spending on AI infrastructure was cutting into companies’ budgets in favor of software, contributing to weaker-than-expected revenue growth. The selloff wiped tens of billions of dollars from IBM’s market value and weighed on other software and information technology stocks, the report said.
A draft report from the U.S. Treasury Department also examined how an AI slowdown could spread through the economy. Drawing on research from the University of Texas at Austin cited by NOTUS, the report finds that AI companies have become more closely intertwined with the U.S. economy than Internet companies were during the Internet period.
Under the report’s pessimistic scenario, disappointing productivity or earnings could hurt private credit, chipmakers, cloud providers, electric utilities and corporate finance data centers. The Treasury did not predict an imminent crash, but it listed power shortages, financing limits, supply chain disruptions and geopolitical tensions among the risks facing the sector.
Liquidity Demands Could Reveal Inflated AI Valuations
Ray Dalio has previously argued that liquidity, rather than technology weakness, could break the AI boom. In a television interview reported by Bloomberg, the founder of Bridgewater Associates explained that investors often confuse rising asset values with money they can easily spend.
Dalio used private companies to illustrate the risk: a company can be valued at a billion dollars after raising much less real capital, but shareholders can’t use that paper wealth without selling. According to him, tensions would arise if many investors tried to transform these valuations into liquidity at the same time.
Bernstein and Cummings pointed to another growing pressure under the boom. In a recent Substack article, economists wrote that the AI bubble “continues to inflate,” while technology investments have reached nearly 5% of U.S. GDP, above levels seen in the dot-com era.
Their analysis also found that large tech companies were committing enough capital to AI projects to reduce their cash reserves. Combined with Noble’s warning and Dalio’s liquidity concerns, these numbers leave investors wondering whether AI profits can make up for the money already committed to the sector.
