The artificial intelligence sector carries a concentration risk that "Big Short" investor Steve Eisman says the market has not fully reckoned with. His warning is specific: the AI boom has grown increasingly dependent on the fortunes of just two companies, OpenAI and Anthropic. He calls that dependency the industry's Achilles' heel.
The constraint Eisman is naming sits at the model layer. Commercial AI value chains run from chip suppliers and data center operators through application developers and enterprise software vendors that have restructured their roadmaps around large language model access. That entire structure leans on a small group of foundation model providers. His argument is that the group has effectively narrowed to two, and two is a dangerously short list for a sector that has drawn the capital commitment it has.
Investment theses built on AI adoption carry a hidden assumption: that OpenAI and Anthropic remain viable and willing to serve the broader ecosystem around them. A setback at either firm travels upward through every stack layer that depends on their models. The more the sector concentrates adoption around two providers, the more that exposure compounds.
Eisman made his name identifying structural fragility in markets that looked stable until they did not. The current concern follows that same line of reasoning. The risk he is describing is not that the underlying technology fails. It is that a major commercial build-out has been structured around the continued health of two specific companies, and the sector is pricing that as certainty. Whether the market has adequately priced the Achilles' heel is, by Eisman's read, an open question.