Google Plans to Invest in Anthropic up to $40 Billion

The Story

Google plans to invest in Anthropic with an initial $10 billion commitment at a $380 billion valuation and another $30 billion tied to performance milestones. The deal expands a partnership that dates back to 2023 and comes as Anthropic scales infrastructure for Claude and advanced models, including its restricted Mythos model. Google remains both a competitor through Gemini and a key infrastructure supplier through Google Cloud and TPUs.

The Facts

• Google plans to invest in Anthropic with $10 billion now.

  The investment values Anthropic at $380 billion.

• Another $30 billion may follow.

  The additional funding depends on Anthropic hitting certain performance milestones.

• The deal expands Anthropic’s Google compute access.

  Earlier this month, Anthropic secured 5 gigawatts of computing capacity through a Google and Broadcom partnership that will start coming online next year.

• Anthropic may add more capacity later.

  The Google-Broadcom arrangement leaves room for Anthropic to add additional gigawatts of compute in the future.

• Anthropic already uses Google Cloud infrastructure.

  Google Cloud provides access to Claude models and offers TPUs as an alternative to Nvidia GPUs.

• Google is also a direct Anthropic competitor.

  Google’s Gemini competes with Anthropic’s Claude models and AI services.

• Google’s Anthropic relationship dates back to 2023.

  Google invested $300 million for about a 10% stake, later added another $2 billion, and had invested more than $3 billion before this new agreement.

• Google reportedly owned a 14% Anthropic stake before the latest deal.

• Anthropic recently released Mythos to limited partners.

  Mythos is described by Anthropic as its most powerful model to date, with significant cybersecurity applications.

• Mythos access remains restricted.

  Anthropic is limiting broader access because of potential misuse while selected organizations evaluate the risks.

• Anthropic has also expanded infrastructure ties with Amazon and CoreWeave.

  Amazon invested $5 billion and agreed to invest up to $20 billion more tied to commercial milestones, while Anthropic is also expected to spend up to $100 billion for around 5 gigawatts of compute capacity over time.

• Anthropic’s growth is straining infrastructure.

  Anthropic has described the pressure as “inevitable strain” from growing enterprise, developer and consumer demand for Claude. Its annualized revenue has topped $30 billion.

Numbers that Matter

• Up to $40 billion: total planned Google investment
• $10 billion: immediate commitment
• $30 billion: potential follow-on investment
• $380 billion: Anthropic valuation tied to the Google investment
• 5 gigawatts: Google-Broadcom computing capacity starting next year
• More than $3 billion: Google’s Anthropic investment before this deal
• 14%: Google’s reported Anthropic stake before the latest agreement
• Up to $20 billion: potential future Amazon investment tied to milestones
• Up to $100 billion: expected Anthropic cloud spending under the Amazon agreement
• More than 100,000: customers building on AWS cited in Anthropic’s Amazon statement
• More than $30 billion: Anthropic annualized revenue

Key Risk Factors

Anthropic’s compute expansion is tied to increasingly powerful AI systems that are expensive to run and harder to control at scale. Mythos is already being handled cautiously because of cybersecurity misuse concerns, and reported unsanctioned access adds pressure around model security and restricted deployment.

Market Timing

Google plans to invest in Anthropic as the broader AI infrastructure race makes access to cloud capacity, chips and energy as important as model quality. Anthropic is securing capacity from multiple major providers while Google, Amazon and other tech giants invest tens of billions of dollars in frontier AI labs.

What to Watch Next

The main signals are whether Anthropic meets the milestones needed to unlock the additional $30 billion, how quickly its 5 gigawatts of Google-Broadcom capacity comes online, whether Mythos access expands beyond limited partners, and whether Anthropic moves toward a possible IPO later this year.

Why This Matters

For end users, Google plans to invest in Anthropic is another sign that advanced AI products are increasingly shaped by infrastructure, not just software. Better models, higher usage limits and faster tools depend on massive compute capacity, while cybersecurity-focused systems like Mythos also raise harder questions about who gets access and under what controls.

This article was drafted with the assistance of generative AI. All facts and details were reviewed and confirmed by an editor prior to publication.