At first glance it sounds like a paradox: Google has just committed up to $40 billion to Anthropic, the company that makes Claude, the AI assistant that competes directly with its own Gemini. The deal starts with an immediate $10 billion, at a valuation of Anthropic of $350 billion, and the rest arrives if the company meets certain targets. In addition, Google Cloud commits to providing five gigawatts of computing capacity over the next five years.
The logic behind this move is clearer than it seems. The real war isn’t between Gemini and Claude, but between Google Cloud, Amazon Web Services and Microsoft Azure. Whoever hosts Anthropic’s infrastructure takes billions in computing revenue. Google isn’t betting on Anthropic’s models; it’s betting on being the one to run them. And Anthropic’s growth figures justify the investment: from $1 billion in annualised revenue at the end of 2024, to $9 billion at the end of 2025, and $30 billion in April 2026.
There’s an additional detail worth attention: Anthropic has internally developed a model called Mythos, described as the most powerful they have created so far, with significant capabilities in cybersecurity. They only showed it to a small group of organisations because of the risks of misuse, and according to reports it has already fallen into unintended hands. This is the company Google has just deposited $40 billion in. The AI sector doesn’t have a boring moment.
Article published on 29 April 2026 | Technology blog
