Google confirmed last Friday an investment of up to $40 billion in Anthropic, the creator of the Claude models, with $10 billion deployed immediately and the remaining $30 billion contingent on the company hitting certain performance milestones. Anthropic's valuation was set at $350 billion, the same level recorded in a funding round back in February, not counting the capital recently raised.
For those following the sector, this deal doesn't come out of nowhere. Anthropic CEO Dario Amodei previously worked as an AI researcher at Google, and the company has been receiving investment from the tech giant since 2022. Anthropic has since established itself as a large-scale strategic customer of Google's TPU servers and cloud services. We are therefore looking at a relationship that is simultaneously a partnership and direct competition, something increasingly common in this industry.
The operational context is equally telling. The deal comes shortly after Anthropic released Mythos, described as its most powerful model to date, with significant cybersecurity applications. Due to the potential for misuse, access was restricted to a select group of partners while risks are assessed, though the model has already reached unauthorized users. At the same time, the company has faced widespread complaints about Claude usage limits in recent weeks and responded with a series of infrastructure agreements, including a deal with CoreWeave for data center capacity and an additional $5 billion investment from Amazon, as part of a broader agreement that plans up to $100 billion in computing capacity spending over time.
Investors have shown appetite to value Anthropic at $800 billion or more, and the company is also reportedly considering an IPO as early as October.
What strikes me as most significant here is not the amount itself, impressive as it is. It's the conditional structure of the investment: $30 billion tied to performance milestones. This means Google is not simply betting on Anthropic — it is establishing a financial accountability mechanism over its development. In an industry where valuations frequently outpace the reality of products, this kind of structure is, paradoxically, a sign of market maturity. The AI race is moving away from being purely about who announces the most and toward who actually delivers measurable results.
Google confirmed last Friday an investment of up to $40 billion in Anthropic, the creator of the Claude models, with $10 billion deployed immediately and the remaining $30 billion contingent on the company hitting certain performance milestones. Anthropic's valuation was set at $350 billion, the same level recorded in a funding round back in February, not counting the capital recently raised. For those following the sector, this deal doesn't come out of nowhere. Anthropic CEO Dario Amodei previously worked as an AI researcher at Google, and the company has been receiving investment from the tech giant since 2022. Anthropic has since established itself as a large-scale strategic customer of Google's TPU servers and cloud services. We are therefore looking at a relationship that is simultaneously a partnership and direct competition, something increasingly common in this industry. The operational context is equally telling. The deal comes shortly after Anthropic released Mythos, described as its most powerful model to date, with significant cybersecurity applications. Due to the potential for misuse, access was restricted to a select group of partners while risks are assessed, though the model has already reached unauthorized users. At the same time, the company has faced widespread complaints about Claude usage limits in recent weeks and responded with a series of infrastructure agreements, including a deal with CoreWeave for data center capacity and an additional $5 billion investment from Amazon, as part of a broader agreement that plans up to $100 billion in computing capacity spending over time. Investors have shown appetite to value Anthropic at $800 billion or more, and the company is also reportedly considering an IPO as early as October. What strikes me as most significant here is not the amount itself, impressive as it is. It's the conditional structure of the investment: $30 billion tied to performance milestones. This means Google is not simply betting on Anthropic — it is establishing a financial accountability mechanism over its development. In an industry where valuations frequently outpace the reality of products, this kind of structure is, paradoxically, a sign of market maturity. The AI race is moving away from being purely about who announces the most and toward who actually delivers measurable results.