Google to invest up to $40 billion in Anthropic and escalate the race for AI computing power

Google is ready to invest up to $40 billion in Anthropic in cash and computational resources — and this amount shows how rapidly the AI market is transforming into a race not just for models, but for infrastructure. For Anthropic, this is a chance to secure critical access to computing power after launching Mythos, its most powerful model with a focus on cybersecurity. For Google — an opportunity to embed itself more deeply in the ecosystem of one of the main players in generative AI, even if they formally compete with each other.

Google intends to invest $10 billion immediately at a $350 billion valuation of Anthropic. The company could provide another $30 billion later if Anthropic meets certain performance metrics. This is not just about direct financing: the package also includes support for the startup's growing computational needs.

Against the backdrop of rapidly escalating costs for training and maintaining cutting-edge models, this deal structure looks almost as important as the cash component itself. The trigger for this new wave of interest was the recent limited release of the Mythos model. Anthropic describes it as the most powerful system in its lineup and emphasizes its notable capabilities in cybersecurity.

Due to the risk of misuse, the company did not open broad access to it and initially limited it to work with a small circle of partners. However, reports have already emerged about unauthorized access to Mythos, which only intensified attention to the model and questions about controlling its use. The Google deal fits into a broader reshuffling of the computing power market.

Major AI developers are competing today not just for users and quality of responses, but for data centers, accelerators, energy, and long-term cloud contracts. OpenAI in recent months has also actively expanded its network of agreements with cloud, chip, and energy providers. Anthropic is moving in the same direction: in April, the company reached an agreement with CoreWeave for data center capacity, and then received another $5 billion from Amazon as part of a broader agreement that could involve up to $100 billion in computational infrastructure spending and around 5 gigawatts of power over time.

Google's position here is particularly interesting because the company plays two roles at once. On the one hand, it remains a direct competitor to Anthropic in the AI model market. On the other — it also sells the infrastructure on which these models are trained and operate.

Anthropic is already heavily dependent on Google Cloud, including access to TPUs — Google's specialized accelerators for artificial intelligence tasks, which are considered one of the few large-scale alternatives to the scarce NVIDIA solutions. In early April, Anthropic also announced a partnership with Google and Broadcom to obtain multi-gigawatt TPU capacity starting in 2027; later, Broadcom's disclosures mentioned a figure of 3.5 gigawatts.

The new agreement expands this framework. Now Google Cloud is to provide Anthropic with an additional 5 gigawatts of computing power over the next five years with the possibility of further increase. For the market, this is an important signal: investments are increasingly structured not simply as buying a stake, but as a long-term binding of a model to a specific cloud provider and chip supplier.

In practice, this means that the winners of the AI race will be determined not only by the quality of algorithms, but also by the ability to reserve electricity, racks, networks, and accelerators years in advance. There is also a financial layer to this story. Back in February, Anthropic's valuation was $350 billion, and then investors began discussing levels of $800 billion and higher.

Against this backdrop, the company is reportedly considering an IPO as early as October. If such a scenario materializes, the current support from Google and Amazon could play a role not only in ensuring infrastructure, but also in shaping the market's image of Anthropic as a business capable of scaling rapidly without bottlenecks in computing. What this means: the AI market is entering a phase where money alone is no longer the solution.

Capital must now go hand in hand with guaranteed access to chips and electricity, and the largest cloud players are becoming the actual architects of the industry. Google's investment in Anthropic shows that even competitors are willing to strengthen each other if it helps them secure a place in the supply chain of the next generation of AI.