Google to invest up to $40 billion in Anthropic

Google will invest up to $40 billion in Anthropic, providing $10 billion in cash immediately and up to $30 billion tied to performance milestones for Anthropic’s Claude models. The $10 billion is being supplied at a reported $350 billion valuation; the targets that would unlock the remaining funds have not been disclosed.

Neither company responded to requests for comment on the terms of the agreement. Google previously invested $2 billion in Anthropic in 2023 and $1 billion in 2025.

The companies also have a cloud arrangement that could involve access to up to one million of Google’s tensor processing units, or TPUs, and may be worth tens of billions of dollars. In April, Anthropic, Broadcom and Google announced plans to bring multiple gigawatts of next-generation TPU capacity online beginning in 2027, likely referring to Google’s 8t and 8i TPU chips. Anthropic said the majority of that hardware will be built in the United States and will form part of a multi-vendor training platform that includes AWS and Nvidia resources.

Anthropic has secured additional capital and compute from other partners. Last week the company disclosed a deal with Amazon worth up to $25 billion, under which Anthropic will receive $5 billion now and up to $20 billion contingent on commercial milestones, together with an additional 5GW of cloud capacity. Amazon previously invested about $8 billion in Anthropic. The company has also signed infrastructure agreements with cloud provider CoreWeave.

In response to the Amazon announcement, Anthropic CEO and co-founder Dario Amodei wrote, ‘Our users tell us Claude is increasingly essential to how they work, and we need to build the infrastructure to keep pace with rapidly growing demand.’

Anthropic raised $30 billion in a February financing round that valued the company at $380 billion, up from a $183 billion valuation reported in September. Market analysts estimate Anthropic’s annualized revenue at roughly $30 billion, up from about $9 billion last year, and project the company could reach break-even in 2028.

Major cloud and technology firms are planning large capital expenditures for AI infrastructure. Companies have announced plans that together could total hundreds of billions of dollars for compute expansion, and industry forecasts project more than $650 billion in AI capital spending among leading firms in 2026.

Anthropic has said it plans significant infrastructure spending in the United States, including a reported $50 billion allocation for data centers and model training this year. The conditions for the $30 billion in Google funding and the specific performance thresholds remain unspecified.