Google's $40 Billion Stake Makes It Anthropic's Investor, Chip Supplier, and Rival

Google will invest up to $40 billion in Anthropic — $10 billion immediately at a $350 billion valuation, with up to $30 billion more contingent on performance targets — making Alphabet's subsidiary the startup's single largest financial backer while remaining its most direct model-layer competitor.

The deal bundles a fresh 5-gigawatt Google Cloud compute commitment over five years, expanding an arrangement that predates this week's news. Earlier this month, Anthropic, Google, and chipmaker Broadcom announced a partnership to access multiple gigawatts of TPU-based capacity starting in 2027; a subsequent Broadcom securities filing put that figure at 3.5 gigawatts. The new investment layers an additional 5-gigawatt TPU allocation on top, with room to scale further. Google's tensor processing units, purpose-built for AI workloads, are among the credible alternatives to Nvidia's processors for frontier inference at scale.

The timing reflects acute compute pressure on Anthropic. The company faced widespread user complaints about Claude usage limits in recent weeks, and has since signed a separate datacenter capacity deal with cloud provider CoreWeave. Amazon this week added a further $5 billion to its own Anthropic position, part of a broader agreement under which Anthropic is expected to commit up to $100 billion for roughly 5 gigawatts of compute over time. The Google investment closes the loop: Anthropic now holds multi-gigawatt supply commitments from two of the three hyperscalers simultaneously.

The capital structure that emerges has no direct precedent in the AI industry. Google competes with Anthropic at the model layer via Gemini, supplies the TPUs that underpin Claude's inference workloads, and now holds the largest single investor position in the company. That three-way role — model rival, infrastructure landlord, and majority financial backer — gives Google both competitive visibility into Anthropic's technical roadmap and pricing leverage over its competitor's cost structure. For enterprise architects evaluating vendor dependency risk, the exposure is no longer a cloud spend question alone; it is a governance one.

The deals arrive alongside Anthropic's release of Mythos, described by the company as its most powerful model to date, with significant cybersecurity applications. Access is restricted to a limited set of partners due to misuse risk — though the model has already reached unsanctioned hands. Mythos's compute demands at scale almost certainly factor into the urgency of locking in capacity ahead of broader deployment.

The company was valued at $350 billion as recently as February; investors have since expressed interest at $800 billion or more, according to Bloomberg. Anthropic is reportedly weighing an IPO as soon as October. A Google-backed public offering would carry structural complexity: underwriters will need to price a business whose largest shareholder is simultaneously its most capable competitor and its primary chip supplier.

For enterprise buyers with active Anthropic deployments, the near-term signal is clear: Claude-based workloads have committed infrastructure backing them at scale. The structural read is starker. When a single hyperscaler can sell you the chip, supply the model that competes with the one running on it, and hold equity in that competing startup — "multi-vendor AI strategy" graduates from preference to procurement policy.