According to Bloomberg, Google plans to invest up to $40 billion in Anthropic and support the AI company’s growing computing needs. According to Anthropic, this subsidiary of Alphabet has committed to investing $10 billion now, valuing Anthropic at $350 billion, and investing an additional $30 billion if Anthropic meets certain performance targets.

This investment commitment was made after Anthropic released its latest model Mythos to limited partners this month. Anthropic stated that Mythos is the company’s most powerful model to date and has significant value in network security applications. Due to potential abuse, Anthropic has limited broader access to the model while collaborating with selected organizations to assess and address these risks – even though the model has fallen into unauthorized hands. In addition, the cost of running the model on a large scale may also be high.

The AI competition increasingly depends on the ability to acquire the computing power required for training and deploying these systems. OpenAI has actively ensured this capability through a series of multi billion dollar agreements across cloud providers, chip suppliers, and the energy sector, including expanding cooperation with chip manufacturer Cerebras this month.

Anthropic himself has also been working hard. In recent weeks, the company has faced widespread complaints from users about Claude’s usage restrictions and has reached a series of infrastructure agreements in response. Earlier this month, Anthropic reached a data center capacity agreement with cloud computing provider CoreWeave. This week, Anthropic also received an additional $5 billion investment from Amazon as part of a broad agreement, under which Anthropic expects to spend up to $100 billion in the future to acquire approximately 5 gigawatts of computing power.

Although Google is a direct competitor in the field of AI models, it is also a key infrastructure provider for Anthropic. Anthropic heavily relies on Google Cloud’s chips and infrastructure, including the use of Google’s tensor processing units. TPU is a specialized chip designed specifically for AI workloads and is considered one of the best alternatives to Nvidia’s scarce processors.

Anthropic’s relationship with Google was announced earlier than this week’s news. Earlier this month, Anthropic announced a partnership with Google and Broadcom, a chip manufacturer that designs custom AI chips for Google, to acquire multi gigawatt computing power based on TPU starting from 2027; Subsequently, Broadcom’s securities documents showed that the number was 3.5 gigawatts. Google’s new investment expands this arrangement, and Google Cloud will now provide a brand new 5 gigawatts of computing power over the next five years, with further room for expansion.

According to Bloomberg, Anthropic’s valuation was only $350 billion in February; Since then, investors have been eager to invest in the company at a valuation of $800 billion or higher. In addition, it is reported that the company is considering an initial public offering as early as October.