Key Highlights
- Revenue run rate at Anthropic has climbed past $30 billion, a dramatic increase from $9 billion recorded at 2025’s close
- The company now serves over 1,000 enterprise clients each spending upwards of $1 million yearly, double the February count
- A strategic agreement with Google and Broadcom secures next-gen TPU chip resources beginning in 2027
- The collaboration provides Anthropic with approximately 5 gigawatts of total computational capacity
- Broadcom’s stock climbed over 3% after the partnership disclosure
The AI company Anthropic has witnessed its revenue run rate skyrocket from $9 billion at 2025’s conclusion to beyond $30 billion, according to this week’s announcement. The organization attributes this explosive growth to surging demand for its Claude AI platform throughout 2026.
The enterprise client base has expanded substantially, with more than 1,000 organizations now investing over $1 million annually in Anthropic’s services. This figure represents more than a twofold increase since February’s tally.
In response to escalating demand, Anthropic has finalized a substantial computing arrangement with both Google and Broadcom. This strategic partnership will deliver approximately 5 gigawatts of total computational power to Anthropic over the forthcoming years.
Approximately 3.5 gigawatts of this infrastructure will be powered by Google’s tensor processing units (TPUs), manufactured by Broadcom. This component of the arrangement is scheduled to commence operations in 2027.
The chip manufacturers have also established a comprehensive supply contract extending through 2031. Additionally, Anthropic will participate in developing and providing customized TPU solutions for Google under this framework.
During last month’s quarterly earnings presentation, Broadcom CEO Hock Tan referenced this partnership. He also projected that Broadcom’s AI semiconductor revenue would surpass $100 billion in the upcoming year.
The overwhelming majority of this new computational infrastructure will be located within United States borders. According to Anthropic, this expansion reinforces its November 2025 pledge to deploy $50 billion toward American computing infrastructure development.
Legal Challenges Fail to Derail Momentum
Anthropic finds itself embroiled in litigation with federal authorities. The Department of Defense classified Anthropic as a supply-chain risk after disagreements emerged regarding AI safety protocols.
The company has cautioned that this designation threatens billions in potential revenue losses. According to testimony from a company legal representative, over 100 corporate clients reached out to express concerns about maintaining their Anthropic partnerships.
Nevertheless, the organization reports sustained growth momentum. Paul Smith, chief commercial officer, noted that certain clients appreciate how Anthropic “stands by its principles” when engaging with government entities.
Implications of the Broadcom Partnership for Semiconductor Industry
Google initially developed its TPU technology to enhance search engine performance. These processors have evolved into critical infrastructure for developing and deploying artificial intelligence applications.
Broadcom functions as the manufacturing partner, transforming Google’s chip blueprints into production-ready components for fabrication. This arrangement establishes Broadcom as a competitive alternative to Nvidia, the current leader in AI computing hardware.
Following the partnership announcement, Broadcom’s stock price increased up to 3.6% during extended trading hours. Alphabet, Google’s parent corporation, gained approximately 1.6% in pre-market activity, while Nvidia experienced a modest 0.4% decline.
Anthropic’s competitor OpenAI established comparable computing agreements during the previous year with Broadcom, Nvidia, and additional partners to guarantee AI infrastructure availability.
In its regulatory filing, Broadcom noted that Anthropic’s utilization of the enhanced computing resources is contingent upon the company maintaining its business performance trajectory.





