Broadcom has reached a long-term deal with Google to create and provide upcoming generations of specialized artificial intelligence chips and additional components for the firm's next-gen AI racks until 2031.

The semiconductor company also reached an agreement with Anthropic to grant the AI startup access to approximately 3.5 gigawatts of AI computing power utilizing Google's AI processors, beginning in 2027.

The need for specialized chips like Google's tensor processing units (TPUs), employed for AI tasks, has increased significantly in recent years as companies look for substitutes to Nvidia's expensive graphics processors.

Google was working to establish its TPUs as a competitive option against Nvidia's leading GPUs in the market. TPU sales have turned into a vital growth driver for Google's cloud revenue, as it aims to demonstrate to investors that its AI investments are yielding returns.

Anthropic announced on Monday that the new agreement enhances the company's dedication to investing $50 billion in bolstering U.S. computing infrastructure.

Also Read: AI Demand to Drive Surge in Semiconductor Revenues: Goldman Sachs

The demand for its AI model Claude has surged in 2026, as the startup's run-rate revenue now exceeds $30 billion, rising from roughly $9 billion at the close of 2025.

Anthropic mentioned it trains and operates Claude on various AI hardware, encompassing Amazon Web Services' Trainium, Google TPUs, and Nvidia GPUs.

Broadcom's involvement extends past chip design. The firm will additionally supply networking elements necessary to link numerous AI processors within extensive computing clusters, a progressively essential factor as AI systems expand.

Also Read: Wipro Secures $1 Billion Deal from Olam Group

The agreement highlights a wider shift in the semiconductor sector. Tech giants are increasingly moving away from depending solely on outside chip manufacturers, instead pouring resources into custom processors to secure dominance over performance, supply chains, and AI development timelines.

Also Read: Airtel Expands its 5G Network Footprint across UP East

With the global rise in demand for AI computing, the development of custom silicon is becoming the new arena, moving competition from traditional chips to specialized AI hardware ecosystems.