Google Just Bet $40 Billion on Anthropic, and the Most Popular Agent Framework Has Admin Access Wide Open

Google will invest up to $40 billion in Anthropic, the largest single investment in an AI company to date. The deal, reported by Bloomberg and confirmed by multiple outlets on April 24, structures as $10 billion now at a $350 billion valuation, with $30 billion more contingent on Anthropic hitting performance targets. Separately, Anthropic secured 5 gigawatts of next-generation TPU capacity from Google and Broadcom, starting in 2027; enough to power roughly 3.5 million homes.

The numbers tell a story about where compute leverage is concentrating. Anthropic's annual revenue run rate has surged past $30 billion, up from $9 billion at the end of 2025. Google is simultaneously Anthropic's investor, compute provider, cloud reseller, and competitor. That structural tension is the defining feature of frontier AI's capital stack: the companies funding the race are also running it. For builders, the takeaway is that Anthropic's infrastructure runway just extended significantly, which means Claude's model cadence and API reliability are unlikely to be constrained by compute in the near term. For investors, Google's willingness to invest at $350 billion signals that the market for frontier model providers has consolidated into a two-to-three player oligopoly where the entry cost is measured in tens of billions.

(Reuters · TechCrunch · Anthropic)