In a massive escalation of the AI arms race, Google is reportedly investing up to $40 billion in Anthropic. This move comes just weeks after Amazon announced its own multi-billion dollar expansion with the AI lab, effectively turning Anthropic into the most heavily capitalized “neutral” player in the industry.
Read: Bandwidth Blog & Smile 90.4FM Tech Tuesday: Is the AI bubble popping?
According to Bloomberg, the investment is structured in two distinct phases:
- Immediate Capital: Google is committing $10 billion at Anthropic’s current valuation.
- Milestone Incentives: An additional $30 billion is on the table, contingent on Anthropic meeting specific performance benchmarks.
As part of the agreement, Anthropic has signed a joint pact with Google and Broadcom to secure 5 gigawatts of next-generation computing capacity by 2027. This infrastructure will rely heavily on Google’s proprietary TPUs (Tensor Processing Units).
The structure of this deal almost perfectly mirrors Anthropic’s recent agreement with Amazon. In early April, Amazon pledged $5 billion upfront (with $20 billion in milestones) in exchange for Anthropic’s commitment to use Amazon’s Trainium chips.
These arrangements highlight a growing trend of “circularity” within the AI economy:
- Big Tech invests billions of dollars into an AI Startup.
- The AI Startup uses that cash to rent servers and buy silicon from Big Tech.
- Big Tech reports increased cloud revenue, while the AI Startup gains the massive compute power needed to train frontier models.
Anthropic’s ability to secure nearly $65 billion in potential funding from two different cloud giants within a single month underscores the staggering cost of developing “General AI.” Having already raised $30 billion in a recent funding round, Anthropic is now positioned as a unique shared asset—and a key customer—for the world’s two largest cloud infrastructure providers.
By securing access to both Amazon’s Trainium and Google’s TPU ecosystems, Anthropic is effectively hedging its bets, ensuring it has the diversified hardware necessary to compete with the Microsoft-OpenAI alliance.
