BlockBeats News, June 18th - MIT cognitive science Ph.D., AI researcher Gary Marcus warned that OpenAI's potential IPO and cash burn issues could pose a risk to some tech stocks that are highly reliant on AI computing power. This view provides a bearish narrative for the current hot AI infrastructure trades.

Marcus believes that the valuations of companies like Nvidia, Oracle, and CoreWeave are partly based on the expectation that OpenAI will continue large-scale chip procurement and data center capabilities. If the OpenAI IPO process faces difficulties, valuations come under pressure, or spending is cut due to price wars, these suppliers may face downward revenue revisions.

OpenAI has submitted a confidential S-1 filing this month, laying the groundwork for a potential IPO. Meanwhile, the market is also watching its high computing cost, competition with Anthropic, and whether enterprise clients will reduce usage due to high AI implementation costs.

Marcus's concerns are not only at the stock level. He suggested that if there is a high degree of linkage between AI data center financing, cloud computing contracts, and OpenAI demand, a contraction in core client spending could lead lenders to reassess the credit quality of AI-related assets.