Google just committed up to $40 billion to Anthropic, with $10 billion hitting immediately at a $380 billion valuation and the remaining $30 billion tied to performance milestones.
Before Friday’s announcement, Google had put in just over $3 billion since 2023, picking up roughly a 14% stake. Google’s new minimum: $13 billion. This deal doesn’t just add zeros. It cements Google as Anthropic’s primary infrastructure partner, even as Gemini competes directly for the same enterprise AI customers.
The structure mirrors Amazon’s deal from weeks prior: $5 billion upfront with up to $20 billion more tied to commercial milestones. Anthropic CEO Dario Amodei tied both deals to the same pressure point. The company faces what he called “inevitable strain” on infrastructure from surging enterprise, developer, and consumer demand for Claude. Annualized revenue has topped $30 billion, and OpenAI recently criticized Anthropic for failing to secure enough compute.
For operators building on Claude, here’s what changed: earlier this month, Anthropic locked in 5 gigawatts of compute capacity with Google and Broadcom, starting next year. Google also supplies TPUs — its custom chips competing with Nvidia’s GPUs. More compute headroom means fewer degraded API responses when demand spikes.
Monday morning: if you’re building on Claude APIs, your vendor just secured multi-year infrastructure financing from two of the three major cloud providers. The first batch of that 5-gigawatt compute capacity comes online in 2027.
— Nathan Zakhary