‘Big Short’ trader Steve Eisman: It's very clear Anthropic's revenue growth has slowed dramatically

Summary

Steve Eisman argued that today’s AI systems are far from human-level intelligence. He described them as complex word-prediction tools and said fears of near-term catastrophic AI are overstated. While acknowledging that AI leaders may have long been concerned about risks, he suggested they may have been influenced by science fiction and mistaken about how close such dangers are.

Eisman suggested that warnings about AI could also serve companies’ commercial interests. Regulation might limit open models, including those from Chinese competitors, and help established firms build protective barriers. He argued that companies such as Anthropic and OpenAI currently lack strong “moats,” leaving the broader AI business vulnerable to a race to the bottom. Although he sees risks in the market’s dependence on these companies, he said he was not yet betting on a major collapse.

On Anthropic, Eisman viewed its prospectus as containing concerning signs, despite its positive reception among many observers. He cited figures from sources he considered reliable but could not independently confirm: sequential revenue growth allegedly fell from 147% in the second quarter to 45% in the third, while revenue relative to computing-infrastructure costs declined from 2.08 to 1.7. He attributed the slowdown and weaker efficiency to the end of “token maxing.” He also mentioned an unconfirmed report that Anthropic cut prices by 50% in October.

Presenters and Contributors

• Steve Eisman, guest and host of The Real Eisman Playbook
• CNBC Squawk Box interviewer(s), not identified in the subtitles