Anthropic Partners with Google and Broadcom! Securing Multi-Gigawatt Next-Gen TPUs as Revenue Surpasses $30 Billion!
📰 News Overview
- Massive Next-Gen TPU Acquisition: Anthropic has secured a new agreement with Google and Broadcom to lock in several gigawatts of next-gen TPU capacity, set to commence operations in 2027.
- Explosive Revenue Growth: The company’s run-rate revenue has skyrocketed past $30 billion (up from around $9 billion at the end of 2025). The number of large customers spending over $1 million annually has doubled in under two months, soaring from 500 to over 1,000.
- Investment in U.S. Infrastructure: The majority of the secured computing resources will be established within the U.S., further expanding the previously announced $50 billion investment to bolster U.S. computational infrastructure in November 2025.
💡 Key Points
- Multi-Hardware Strategy: Claude leverages a mix of AWS Trainium, Google TPU, and NVIDIA GPUs, optimizing chip allocation for workload demands to enhance performance and resilience.
- Available Across All Major Clouds: Claude maintains its position as the only frontier model available on the three dominant cloud platforms: AWS (Bedrock), Google Cloud (Vertex AI), and Microsoft Azure (Foundry).
- Strong Relationship with AWS: Amazon continues to be a primary cloud provider and training partner, working closely with AWS on “Project Rainier.”
🦈 Shark’s Eye (Curator’s Perspective)
This partnership isn’t just about adding more servers—it’s a whole new ballgame! Securing next-gen TPUs at a gigawatt scale signifies that AI development has turned into a “chip acquisition battle.” What’s especially notable is the flexibility to seamlessly utilize different architectures like AWS, Google, and NVIDIA. The ability to extract the strengths of various hardware without being tied to a specific chip is the secret sauce powering Claude’s muscular infrastructure!
🚀 What’s Next?
Once this massive TPU fleet becomes operational in 2027, Claude will evolve into an even larger and more sophisticated “frontier model.” Given the doubling of large customers, the rollout of specialized AI solutions for enterprises will accelerate, intensifying the competition in AI infrastructure!
💬 A Word from HaruShark
Surpassing $30 billion in revenue is mind-blowing—I’m just a shark, and even I’m amazed by that growth rate! I can’t wait to see a bigger Claude emerge, lured by those next-gen chips! 🦈🔥
📚 Glossary
-
TPU: Stands for Tensor Processing Unit, a specialized processor developed by Google for machine learning.
-
Run-rate: An estimate of annual revenue based on recent revenue data, assuming that the current pace continues for the next year.
-
Frontier Model: A large-scale language model that defines the cutting edge of AI technology, possessing extremely high capabilities.
-
Source: Anthropic expands partnership with Google and Broadcom for next-gen compute