Anthropic’s business is scaling up faster than expected this year. The company says its annualised revenue run rate has crossed $30 billion in 2026, a steep jump from about $9 billion at the end of 2025.

That growth is being fuelled largely by enterprise demand. More than 1,000 customers are now spending over $1 million annually with the company, roughly double the number it reported just a few months ago.

With this momentum, Anthropic has moved ahead of OpenAI in terms of annual revenue run rate, which is estimated to be around $25 billion.

The update comes soon after Anthropic raised $30 billion in a Series G funding round earlier this year, taking its valuation to around $380 billion. Reports suggest the company is also exploring the possibility of an IPO as early as 2026, although nothing has been finalised yet.

At the same time, Anthropic is investing heavily in infrastructure to keep up with demand. It has signed a new agreement with Google and Broadcom to secure large-scale, next-generation TPU capacity. This additional compute power is expected to come online starting in 2027 and will support its Claude models as usage continues to grow.

According to CFO Krishna Rao, the partnership reflects a steady approach to expanding capacity while keeping pace with rapid customer growth and ongoing AI development.

Most of the new infrastructure will be based in the United States, building on Anthropic’s earlier plan to invest $50 billion in computing capacity announced last year.

The agreement also deepens its ties with Google Cloud and Broadcom. Anthropic relies on a mix of hardware platforms—including chips from Amazon Web Services, Google TPUs, and NVIDIA GPUs—to run its systems efficiently.

The company added that Amazon remains its primary cloud and training partner through Project Rainier. Its Claude models are currently available across major platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure.

Also Read: Anthropic Accidentally Exposes Claude Code Source Code