In a significant move within the AI chip landscape, Anthropic PBC has expanded its partnership with technology giants Broadcom Inc. and Google LLC to accelerate its artificial intelligence capabilities. This renewed collaboration is set to provide Anthropic with “multiple gigawatts” of computing power via Google-designed Tensor Processing Units (TPUs), promising to bolster the AI startup’s ability to scale its AI models effectively and meet growing demand.

Anthropic, known for building advanced AI models, announced that its revenue run rate has surged to over $30 billion recently, reflecting soaring customer engagement and expanding commercial success. This growth underscores the escalating demand for AI compute capacity, a critical factor driving Anthropic’s strategic moves.

The partnership primarily hinges on utilizing Google’s TPUs, which are specialized chips designed to accelerate machine learning workloads. Broadcom plays a crucial role as the supplier and manufacturer, working in close coordination with Google. They have finalized a long-term supply agreement that extends through 2031, ensuring a steady and reliable chip supply for Anthropic’s AI infrastructure.

This expanded deal also builds upon Anthropic’s previous commitment made in 2025 to invest around $50 billion in U.S. computing infrastructure focused on AI development. Notably, most of the new TPU capacity will be deployed in the United States, reflecting ongoing efforts to strengthen domestic AI technology resources.

From Broadcom’s perspective, this partnership comes at a pivotal time when the market for AI chips is expected to grow exponentially. Broadcom’s CEO, Hock Tan, highlighted that AI chip sales could surpass $100 billion in revenue next year, demonstrating how critical these custom AI chips are becoming in powering next-generation AI applications.

Google’s TPUs have been gaining traction not only with Anthropic but across several AI-focused ventures, positioning themselves as a formidable alternative to the dominant NVIDIA GPUs in the AI hardware market. Google is not limiting itself to a single cloud partner either, integrating TPU technology alongside Amazon’s AWS Trainium and NVIDIA GPUs to create a more diversified and robust AI compute ecosystem.

For investors and market watchers, this partnership signals several important trends. First, the rapid rise in Anthropic’s revenue reflects the broadening commercial adoption of AI technologies. Second, the extended supply agreement through 2031 underscores a long-term belief in the continued demand for AI chips, with companies aiming to secure their hardware needs well into the future.

Moreover, this move highlights a growing shift toward a multi-hardware approach in AI development. Instead of relying solely on one chip type or supplier, leading companies are diversifying their compute infrastructure to maximize efficiency, reliability, and innovation.

As global markets continue to show volatility with mixed sector performances, the AI chip segment remains a bright spot and a key area for growth. Anthropic’s collaboration with Broadcom and Google not only strengthens its production capabilities but also reinforces the importance of strategic partnerships in the evolving AI ecosystem.

In conclusion, Anthropic’s expanded partnership with Broadcom and Google marks a pivotal advancement in the AI chip arena. With substantial computing power secured and a long-term supply plan in place, Anthropic is well-positioned to capitalize on the accelerating demand for AI services and technological innovation. This development is a clear indicator that the AI infrastructure market is rapidly maturing and becoming a critical investment space for the future.