ViqusViqus
Navigate
Company
Blog
About Us
Contact
System Status
Enter Viqus Hub

Anthropic Secures Massive Compute Power from Google and Broadcom to Fuel Claude's Growth

Anthropic Google Cloud Broadcom Claude AI Compute capacity AI development
April 07, 2026
Source: TechCrunch AI
Viqus Verdict Logo Viqus Verdict Logo 8
Infrastructure Play: Locking Down Compute Supremacy
Media Hype 6/10
Real Impact 8/10

Article Summary

AI lab Anthropic announced a significant computational resource expansion, signing a new agreement with Google and Broadcom. The deal boosts Anthropic's access to Google Cloud's TPUs and includes provisions for a total of 3.5 gigawatts of compute capacity, according to a Broadcom filing. While the full capacity will only be active in 2027, the move is part of Anthropic's larger commitment to U.S. compute infrastructure. This massive investment in raw computing power underscores the intense demand for the Claude models, which are being adopted rapidly by large enterprise customers, despite facing geopolitical labeling as a supply chain risk. The funding round and record revenue figures further highlight the company's accelerated ascent in the market.

Key Points

  • Anthropic signed a major compute capacity expansion deal with Google and Broadcom, leveraging Google Cloud's TPUs.
  • The new compute capacity is vast, estimated at 3.5 gigawatts, with deployment phased out until 2027.
  • The funding and revenue figures—including a $30 billion run rate revenue and a $380 billion valuation—underscore the explosive enterprise adoption of Claude.

Why It Matters

This is not breakthrough technological news, but it is critical infrastructure news. Anthropic's ability to secure such massive, sustained compute capacity from tech giants solidifies its competitive moat. In the current AI landscape, compute supply is the most critical bottleneck, and locking down gigawatt-scale access with multiple industry leaders (Google, Broadcom) drastically improves their operational stability and competitive positioning against rivals. Professionals should pay attention to how quickly this compute translates into model performance gains and market dominance.

You might also be interested in