Steve Eisman, the investor who made his name betting against the subprime mortgage market and who now hosts The Real Eisman Playbook podcast, used a CNBC appearance to argue that the loudest warnings about artificial intelligence are coming from the people with the most to gain from acting scared. His claim: the extinction-risk rhetoric from frontier AI leaders is a strategic instrument.
Eisman said “there’s no evidence at all that AI has achieved artificial general intelligence at all. And most of the evidence seems to point that it never will. Or if it does, it’s many, many years in the future.” He then went further, alleging motive. Eisman said “the idea that this whole terminator thing is nonsense… these companies are very nervous. They realize that there are no moats around their business whatsoever, and they’re trying to manufacture a crisis that will create regulation and that they think they can then manipulate to create the moats to create the duopoly that they want.”, according to The Real Eisman Playbook
Chain Dependency Argument
Eisman then linked the moat question to a concentration risk running through every AI-exposed stock. NVIDIA (NASDAQ:NVDA | NVDA Price Prediction), Microsoft (NASDAQ:MSFT), Alphabet (NASDAQ:GOOGL), Amazon (NASDAQ:AMZN), Meta Platforms, and GPU-cloud pure-play CoreWeave all sit downstream of a very small number of frontier labs. Eisman said “the entire AI chain, from nvidia to the hyperscalers to anthropic and open AI all depends on the future health of anthropic and openai… So if something were to happen to one of those two companies, then the chain would really fall apart. And between the two, openai is the weaker company, according to The Real Eisman Playbook. I think that’s one reason why they postponed their IPO.”
OpenAI IPO Delay: Two Readings of the Same Event
The delay is real. Fortune reported on September 12, 2026 that Sam Altman said OpenAI would not go public this year because an IPO would come at an “ill-advised moment.” The Guardian reported on September 12, 2026 that the OpenAI IPO will not happen in 2026 amid AI safety fears, according to Altman. The Economic Times reported on September 13, 2026 that OpenAI delayed its IPO beyond 2026 as Altman backed a call to slow AI development.
OpenAI’s stated reason is AI safety. Eisman’s reading is that safety talk is the instrument, and that the postponement reflects competitive weakness. Same event, two irreconcilable readings of motive. Readers can weigh both.
Open Weights and the Hugging Face Deal
Eisman’s premise about open-weight models eroding incumbent moats received an unusual market validation. The Information reported on August 26, 2026 that NVIDIA agreed to buy open-source AI platform Hugging Face for $12.9 billion. Yahoo Finance reported on September 3, 2026 that NVIDIA confirmed the acquisition. Wired characterized the deal on September 3, 2026 as a bet on open-source AI. The Register argued on September 3, 2026 that Hugging Face is too important to fall into NVIDIA’s hands, evidence the deal itself is contested. The dominant chipmaker paying that price for the open-weight repository supports Eisman’s premise about open weights mattering. It does not prove his claim about motive.
Eisman used the phrase “token mixing is over” to describe what he sees happening as open-weight models take share. That is his terminology and his analysis.
Security Incidents and What Eisman Did
OpenAI announced in July 2026 that it was partnering with Hugging Face to address a security incident during model evaluation. Anthropic published a post on July 30, 2026 investigating three real-world incidents in its cybersecurity evaluations. The company whose platform was breached has since been acquired by NVIDIA. Eisman said he has taken some of his AI position off the table.
For scale on what is riding on the answer, NVIDIA’s own Q2 FY2027 8-K shows Data Center revenue of $89.02 billion, up 117% year over year, with a full-year approximately 70% revenue growth outlook for fiscal 2028. Microsoft’s commercial remaining performance obligations rose 84% to $678 billion. CoreWeave reported a revenue backlog of approximately $104 billion. That is the chain Eisman argues rests on two labs.