According to the latest survey by tech media The Information, AI market revenue is rapidly concentrating in top companies. In this feast of computing power and capital, the two giants in large models, OpenAI and Anthropic, have shown particularly strong performance, collectively capturing about 89% of the annual revenue in the entire sector.

Data shows that the total annual revenue of the 34 listed AI startups is approaching the $8 billion mark, achieving an explosive growth of 112% in just six months. Although the overall market is growing rapidly, most of the profits and revenue benefits are firmly held by these two super giants.

Revenue Surge with Diverse Strategies

Among the two giants, Anthropic's growth momentum has been particularly notable, with its annualized revenue expected to surge to $5 billion by the end of June this year, far exceeding the $1 billion level at the start of the year. However, since part of its revenue is recognized on a gross basis through cloud computing partners such as Amazon and Google, the actual net income will be somewhat diluted.

In comparison, OpenAI's ability to generate revenue remains at the industry's top level, with its monthly revenue reaching $2 billion by the end of March this year, translating to an annualized revenue of $24 billion. According to their business agreements, OpenAI must share 20% of its revenue with Microsoft until 2030, and it is expected to pay Microsoft around $6 billion this year.

Dependence on the Foundation is Hard to Break in the Short Term

Although application-layer stars like Perplexity, ElevenLabs, and Cursor have also performed well, with annual sales all crossing the $5 billion threshold, the report points out that there are still hidden concerns behind this prosperity, as they need to pay billions of dollars in model access fees to OpenAI and Anthropic each year.

This "growth and dependence coexist" ecosystem gives the underlying model suppliers significant influence and revenue-generating power. As the head-to-head effect continues to intensify, the survival space for smaller model manufacturers and pure application-layer companies is being squeezed, and the global AI industry's arms race is evolving into a prolonged battle between two dominant forces.