Anthropic is expanding its AI infrastructure through partnerships with Google and Broadcom
The American startup Anthropic has entered into a strategic agreement with Google and Broadcom to gain access to new computing resources with a capacity of several gigawatts.
This move is aimed at expanding the infrastructure needed to ensure the stable operation of the Claude AI model amid rapidly growing global demand.
The company has officially confirmed “exponential” growth in the number of users and corporate clients.
In particular, the number of large clients whose spending on Claude services exceeds $1 million per year has grown from 500 to over 1,000 in just the last two months.
This growth has directly impacted the startup’s financial performance.
As of today, Anthropic’s revenue, annualized, has already surpassed the $30 billion mark. For comparison, at the end of 2025, this figure stood at approximately $9 billion.
The agreement calls for the use of the latest generation of AI accelerators, which will allow Claude to scale more efficiently. The majority of the new capacity will be commissioned in 2027.
Most of the computing nodes will be located directly within the United States. This aligns with Anthropic’s plans, as the company previously announced $50 billion in investments in U.S. infrastructure.
In addition to hardware, Anthropic plans to more actively implement Google’s cloud solutions. This involves the systematic integration of BigQuery, Cloud Run, and AlloyDB services into internal development processes.
Such collaboration is critical for maintaining the pace of training new models. Company leadership emphasizes that the partnership with Google and Broadcom will provide the necessary technological foundation to compete in the artificial intelligence market.
“The use of the latest generation of AI accelerators will allow the company to significantly expand the infrastructure needed to support its Claude model amid rapidly increasing demand,” the Anthropic press office notes.
As a reminder, North Korean IT specialists have been working on DeFi projects for seven years.
We also reported that Anthropic unexpectedly raised the issue of intellectual property protection following the accidental publication of Claude’s code.
