Broadcom Cements AI Infrastructure Leadership with Anthropic-Google Alliance
Broadcom has significantly tightened its grip on artificial intelligence infrastructure through a strategic new agreement with AI developer Anthropic and technology giant Google. This landmark deal positions the semiconductor manufacturer as a critical enabler of the ongoing AI boom, as companies race to secure unprecedented computing capacity for their models.
Massive Compute Commitment Through 2027
Under the newly announced arrangement, Anthropic will gain access to multiple gigawatts of Google tensor processing unit capacity starting in 2027, with Broadcom supplying the essential infrastructure to make this possible. Anthropic has confirmed that approximately 3.5 gigawatts will be involved, representing the company's largest compute commitment to date.
Broadcom, already a major partner to Google on custom TPUs, has now extended this relationship through a long-term agreement covering future TPU generations and networking components for Google's next-generation AI racks. This deeper integration places Broadcom firmly within the physical layer of the AI market, where demand is currently experiencing its most rapid acceleration.
Anthropic's Explosive Growth Drives Infrastructure Needs
The timing of this agreement coincides with Anthropic's remarkable business expansion. The company's revenue run-rate has now surpassed $30 billion, a substantial increase from approximately $9 billion at the end of 2025. Meanwhile, the number of business customers spending over $1 million annually has doubled in under two months to exceed 1,000.
Krishna Roa, Anthropic's chief financial officer, emphasized that this represents the company's "most significant compute commitment to date," necessary to maintain growth momentum and support future Claude model development. Most of the additional compute capacity secured through this agreement is expected to be located within the United States, extending Anthropic's previous commitment to invest $50 billion in domestic AI infrastructure.
Industry Shift Toward Long-Term Infrastructure Deals
This partnership reflects a broader industry trend where technology firms are increasingly entering into extended arrangements to secure access to cutting-edge computing infrastructure. By collaborating with hyperscale providers like Google, Broadcom has strategically positioned itself within this evolving supply chain.
The company's involvement in TPU production and deployment directly connects it to the expansion plans of leading AI developers like Anthropic. Broadcom had previously indicated that demand from Anthropic could eventually exceed 3 gigawatts of compute capacity, highlighting the extraordinary scale of AI workload growth.
Such partnerships represent significant revenue opportunities for chipmakers, as spending on AI infrastructure continues its relentless upward trajectory without showing any signs of deceleration.
Anthropic's Multi-Cloud Strategy Continues
Despite this expanded agreement with Google and Broadcom, Anthropic maintains its multi-cloud deployment approach across various hardware platforms. The company continues to utilize Google TPUs, Nvidia GPUs, and Amazon Web Services infrastructure simultaneously.
This strategic approach allows Anthropic to match specific workloads to different hardware types while preserving operational flexibility as market demands evolve. Amazon remains the company's primary cloud and training partner, including through substantial projects like its AI supercomputing cluster.
The new agreement with Google and Broadcom provides additional capacity alongside these existing partnerships as Anthropic scales its infrastructure to accommodate growing enterprise usage across multiple platforms.
