Former Fidelity Manager Warns AI Bubble Risks Dot-Com Scale Collapse

Market sentiment is shifting as Polymarket traders place the probability of an AI bubble burst in 2026 at over 17%. This instability follows sharp declines in global tech shares, including a significant drop for South Korean chipmakers SK Hynix and Samsung Electronics. IBM recently suffered its steepest daily share price decline since 1968, falling nearly 25% after disclosing that AI infrastructure spending is cannibalizing software budgets and suppressing revenue growth.

Concerns regarding structural risks are mounting beyond individual company performance. A draft U.S. Treasury Department report suggests that AI firms are more deeply integrated into the broader economy than internet companies were two decades ago. This connectivity implies that a downturn could ripple through electric utilities, cloud providers, and private credit markets. Bridgewater Associates founder Ray Dalio highlighted that the sector faces a liquidity trap, where inflated paper valuations could evaporate if shareholders attempt to exit positions simultaneously. Economists Bernstein and Cummings further noted that technology investment has climbed to nearly 5% of U.S. GDP, a threshold exceeding the levels observed during the dot-com era, leaving the market highly sensitive to any shortfall in expected AI productivity.