Inside Broadcom, Anthropic & Google’s AI Compute Partnership

Broadcom has announced a deal to develop and supply future versions of Google’s AI chips, marking a significant expansion in AI infrastructure capacity.

Previously, the company had expanded its partnership with Anthropic which, beginning in 2027, will access approximately 3.5 gigawatts of AI compute power from Broadcom.

The tech giant will supply tensor processing units (TPUs), which are specialised semiconductors needed for advanced machine learning workloads and neural networks that power today’s most demanding AI systems.

According to a security filing, Broadcom entered into a long-term agreement to develop and supply custom TPUs, as well as a supply assurance agreement to supply networking and other components to be used in Google’s next-Gen AI racks through to 2031.

In what Krishna Rao, Chief Financial Officer (CFO) of Anthropic calls a “ground breaking” partnership, Broadcom, Google and Anthropic have also expanded their current collaboration.

Krisha says: “This ground breaking partnership with Google and Broadcom is a continuation of our disciplined approach to scaling infrastructure: we are building the capacity necessary to serve the exponential growth we have seen in our customer base while also enabling Claude to define the frontier of AI development.

“We are making our most significant commitment to date to keep pace with our unprecedented growth.”

This comes as AI semiconductors have already brought in significant revenue growth for Broadcom. 

Announcing fourth quarter results, Hock Tan, President and Chief Executive Officer (CEO) of Broadcom says: “In Q4, record revenue of US$18.0bn grew 28% year-over-year, driven primarily by AI semiconductor revenue increasing 74% year-over-year.

“We see the momentum continuing in Q1 and expect AI semiconductor revenue to double year-over-year to US$8.2bn, driven by custom AI accelerators and Ethernet AI switches.

“We forecast Q1'26 total revenue of US$19.1bn and adjusted EBITDA of 67%.”

Hock told analysts that his company’s AI chip revenue will be well above US$100bn in 2027, according to reporting from CNBC.

TPUs powering AI applications

Google’s Gemini and all of Google’s AI-powered applications like Search, Photos and Maps – which serve more than a billion users – are powered by TPUs.

They are highly specialised semiconductors used in Google's AI technology, known as application specific integrated circuits.

These chips are designed specifically to handle the computational demands of training and running large language models and other AI systems.

TPUs are used in a variety of AI applications such as agents, code generation, media content generation, synthetic speech, vision services, recommendation engines and personalisation models.

Anthropic’s Claude runs on a range of AI hardware including Amazon Web Services Trainium, NVIDIA GPUs and Google TPUs,.

Scaling AI compute capacity

The new partnership expands earlier collaboration between Anthropic and Google Cloud, which in 2025 expanded its use of Google’s TPUs. 

Worth tens of billions of dollars, the 2025 expansion, is estimated to bring over a gigawatt of capacity online in 2026.

The ever-rising AI demand saw Anthropic’s run-rate revenue surpassing US30bn – up from roughly US$9bn at the end of 2025.

Anthropic also said the deal represents an expansion of its commitment to invest US$50bn in US computing infrastructure, reflecting the scale of investment required to remain competitive in frontier AI development.

US infrastructure investment challenges

According to McKinsey, from 2021 to 2024 semiconductor and electronics companies invested an estimated US$450bn in US semiconductor manufacturing. 

This spike in AI compute is why McKinsey estimates the semiconductor industry valued in the range of US$630-680bn in 2024 will touch US$1tn-1.1tn by 2030.

In an analysis of semiconductor supply chains in the US, McKinsey found that there is not enough domestic supply to meet that level of demand.

Hundreds of different chemicals are needed to manufacture semiconductors in fabs – these include speciality gases such as tungsten hexafluoride, solutions such as ammonium hydroxide and metallic compounds such as aluminium oxide.

McKinsey argues that for these chemicals, supply gaps could be serious by 2030, potentially constraining the AI infrastructure buildout necessary to support continued advancement in AI capabilities.