The myth of a broad and diversified AI market is finally crumbling, giving way to a rigid duopoly. A fresh report from The Information confirms a troubling diagnosis: OpenAI and Anthropic now command a staggering 89% of the revenue generated by the 34 largest AI startups. While the group’s total revenue surged 112% to $80 billion in just six months, capital has stopped circulating within the industry. Instead, it is pooling into the accounts of these two giants, leaving players like Perplexity, ElevenLabs, and Cognition as mere background actors.
Despite these second-tier companies surpassing the $500 million annual revenue mark, the financial chasm between them and the market leaders makes their position precarious. However, the financial stability of the frontrunners is just as phantom-like as the prospects of their smaller rivals. Sam Altman and Dario Amodei are burning over $30 billion a year on model training, essentially turning their startups into massive incinerators for venture capital.
This situation is further aggravated by a 'hidden tax' paid to Big Tech. OpenAI is obligated to hand over 20% of its revenue to Microsoft until 2030, while Anthropic shares its earnings with Google and Amazon. Ultimately, we are not witnessing the healthy growth of an industry, but rather an elegant scheme for pumping investor money back into the pockets of cloud providers. In this cycle, top-tier AI labs serve as little more than intermediaries for server rental payments.
For businesses, it is time to critically re-evaluate your AI stack. Relying on second-tier providers that lack their own infrastructure is becoming a toxic risk. In an environment of such extreme capital concentration, these companies will inevitably face R&D budget shortfalls. This will compromise the reliability and quality of their products far sooner than they can meet your expectations.