Aschenbrenner’s bearish AI chip bets add fuel to growing bubble warnings
Aschenbrenner's timing is pointed. CNBC reported that AI chip valuations, by at least one concentration measure, now rival the French Mississippi Bubble of the 1720s and surpass the Nasdaq during the dot-com era, with shares of Micron, AMD, SK Hynix, Marvell, and Intel all exhibiting parabolic price action since late March. Bank of America's Michael Hartnett captured the mood bluntly: "Exponential price action, market concentration, collapsing vol, stocks bossing bond yields higher — why melt-up everyone's new base case."
Short sellers more broadly have grown emboldened. At the Sohn Investment Conference on May 14, Joyce Meng, founder of Fact Capital, warned: "A rising tide lifts all boats, and a twisting tide takes down a lot of names in the same neighborhood." Soren Aandahl, CIO of Blue Orca Capital, added historical context: "Railroads changed the world. The internet changed the world. But many of the early purveyors of these technologies went completely bust."
Rising Treasury yields are providing a structural headwind, compounding the sentiment damage from bearish institutional positioning. Bloomberg reported, citing assessments from 32 investment managers across the US, Asia, and Europe, that while 80% remain bullish on equities, elevated bond yields pose a growing and widely acknowledged risk to AI-growth valuations — which depend heavily on long-duration earnings assumptions.
Goldman Sachs economist Dominic Wilson cautioned that markets are entering "a more complicated phase," warning the rally may eventually "build a valuation overhang" if expectations climb too aggressively, even as AI infrastructure demand continues to benefit chip suppliers.
The AI trade will be getting a major test on Wednesday, when powerhouse NVIDIA reports earnings. The results will be a pivotal near-term test for AI-sector valuation. Markets will scrutinize data-center revenue growth and forward guidance for evidence that demand can justify current multiples.
