Steve Eisman, the investor who made his name betting against subprime mortgages, is warning that the AI boom's revenue is too concentrated in two companies. He says if cheaper alternatives gain traction, it could destabilize Big Tech's growth.
The OpenAI-Anthropic dependency
Eisman's warning centers on a simple but uncomfortable fact: the AI industry's revenue is heavily dependent on OpenAI and Anthropic. These two companies are the main drivers of AI spending, and their frontier models command premium prices. That concentration, Eisman argues, makes the entire sector vulnerable.
If either company stumbles, or if their models lose their edge, the ripple effects would be felt across Big Tech. The companies that have poured billions into AI infrastructure and research are counting on OpenAI and Anthropic to keep generating demand. That's a fragile foundation, in Eisman's view.
The cheaper alternative threat
The real risk, Eisman warns, comes from cheaper alternatives. If smaller or newer players can deliver models that are good enough at a fraction of the cost, the pricing power of OpenAI and Anthropic could erode quickly. That would hit revenue growth not just for those two companies, but for the entire AI ecosystem.
Big Tech firms have built their growth strategies around AI, and they've been willing to pay top dollar for the best models. But if a cheaper model can do the job, why would they keep paying a premium? That's the question Eisman is raising, and it's one that investors are starting to take seriously.
The Big Tech growth risk
Eisman's warning lands at a time when Big Tech's growth is increasingly tied to AI. The sector has been a major driver of stock market gains, and any slowdown in AI revenue could have outsized effects. If cheaper alternatives disrupt the current revenue model, the growth story that has powered the market could lose its momentum.
That doesn't mean the AI boom is over, but it does mean the foundation is narrower than it looks. Eisman's point is that the market is pricing in continued dominance from OpenAI and Anthropic, and that assumption may not hold.
The question of whether cheaper alternatives can actually match the performance of the frontier models is still open. But Eisman's warning is a reminder that the AI boom's revenue is not as diversified as it might seem.
