Google to Invest Up to $40 Billion in Anthropic — One Week After Amazon Pledged $5 Billion of Its Own

Google announced on Friday that it will invest up to $40 billion in Anthropic — $10 billion delivered immediately at a $350 billion valuation, with the remaining $30 billion contingent on Anthropic hitting performance targets over the coming years. The deal also commits Google Cloud to supplying 5 gigawatts of TPU capacity to Anthropic over the next five years, locking in compute access at a moment when frontier-model builders are constrained more by silicon than by capital. The announcement comes seven days after Amazon disclosed its own $5 billion follow-on investment and a separate $100 billion cloud-spending commitment from Anthropic to AWS, making this the second nine-figure infusion in eight days from a hyperscaler that also competes with Anthropic at the model layer.

The Compute Cartel Tightens

What Google is buying for $40 billion is not just equity in a leading frontier lab — it is anchor-tenant status for its TPU program at the moment when Nvidia GPUs remain the default substrate for AI training and inference. Anthropic's reported willingness to absorb 5 GW of TPU capacity is one of the largest pre-commitments to non-Nvidia silicon ever made by a frontier lab, and it gives Google a credible reference customer to point to when it sells TPU access to enterprises. The unusual structure of the round — $10 billion in cash plus $30 billion in milestone-conditional follow-ons plus a multi-year compute supply contract — also reduces Anthropic's ability to switch infrastructure providers later, even as it accepts overlapping commitments from Amazon and CoreWeave. The pattern across all three deals is the same: the people who sell Anthropic compute now also own large stakes in its equity.

The IPO Math Has Changed

A $350 billion valuation today, with reports surfacing of a possible IPO as soon as October at $800 billion or higher, would put Anthropic in striking distance of OpenAI's most recent secondary-market valuations within months — a remarkable outcome for a company that was valued at roughly $60 billion barely a year ago. Two factors are driving the re-rating: the commercial traction of Claude on enterprise contracts (especially in regulated sectors where Anthropic's safety positioning helps), and the recent emergence of Mythos, the company's most powerful internal model, which has unexpectedly proved relevant to cybersecurity and government workloads. The $40 billion Google deal is, in effect, a bridge financing that gets Anthropic to the public-market window without requiring it to dilute further on the secondary market — and it locks Google into the cap table as a defensive measure against an OpenAI-Microsoft style entanglement going to a single competitor.