Google to Invest Up to $40 Billion in Anthropic via Cash and Compute

Google plans to commit up to $40 billion to Anthropic.

• $10 billion will be deployed first, valuing Anthropic at $350 billion.
• The remaining $30 billion will be deployed additionally upon achieving certain performance targets.
• The investment includes compute infrastructure support, and is also linked to future capacity of 5 gigawatts.

This month, Anthropic released its latest model, Mythos, to a limited set of partners. The company views it as its most powerful model and sees significant utility in cybersecurity, but is not widening access due to misuse risks.

With demand rising, Anthropic is also accelerating its infrastructure expansion.

• Complaints about Claude usage limits have recently grown,
• and in response, the company is expanding data center and cloud contracts with CoreWeave, Amazon, and others.
• It has also secured an additional $5 billion investment from Amazon.

The relationship between Google and Anthropic combines both competition and supply. Google is a direct competitor of Anthropic while also being a key infrastructure supplier providing TPUs and Google Cloud. Anthropic was already set to access multiple gigawatts of TPU-based compute starting in 2027 through Google and Broadcom's TPU partnership, and this investment further expands that scale.

Behind this is the fact that securing compute has become a core resource in the AI race. OpenAI is also aggressively expanding capacity through large-scale deals spanning cloud, chips, and energy.

Anthropic's valuation stood at around $350 billion earlier this year, but since then, investment demand has been discussed even at valuations of over $800 billion. The company is reportedly also considering an IPO as early as October.