The editorial argues that every dollar Anthropic spends on TPUs flows back to Google's cloud revenue line, making the investment partly a customer acquisition cost dressed up as a venture bet. The circular nature of the deal — Google invests billions, Anthropic commits to spending billions on Google Cloud infrastructure — means the money largely stays within Google's ecosystem.
The editorial highlights that Anthropic signed two major cloud commitments in rapid succession — with AWS and Google — suggesting the company was severely capacity constrained. The price of relieving that compute scarcity was deeper entanglement with both major cloud providers through 'somewhat adverse contracts,' and model quality reportedly recovered once capacity was secured.
Bloomberg frames the deal as strengthening the relationship between two companies that are 'at once partners and rivals' in the AI race. The $40 billion potential total investment makes Google by far the largest financial backer of a company whose Claude models compete directly with Google's Gemini across enterprise, consumer, and developer markets.
Google is investing $10 billion in Anthropic PBC, with an additional $30 billion potentially to follow, according to Bloomberg. This brings Google's total investment in the Claude maker to roughly $42 billion when including prior rounds — a staggering figure that dwarfs most tech acquisitions, let alone minority stakes.
The deal makes Google by far the largest financial backer of its own most dangerous competitor in AI. Anthropic's Claude models compete directly with Google's Gemini across enterprise, consumer, and developer markets. Yet here's Google writing checks that could total $40 billion to keep Anthropic running on Google Cloud infrastructure.
The timing matters. Just weeks ago, Anthropic signed a deal to purchase "multiple gigawatts of next-generation TPU capacity" from Google and Broadcom. Before that, Anthropic deepened its relationship with Amazon Web Services through a separate multi-billion-dollar commitment. Two major cloud commitments in rapid succession.
The surface reading is straightforward: Google wants to keep Anthropic on Google Cloud Platform. Every dollar Anthropic spends on TPUs flows back to Google's cloud revenue line. The investment is partly a customer acquisition cost dressed up as a venture bet.
But the deeper story is about compute scarcity and what it forces. As one Hacker News commenter observed, "the subtext of the last few weeks is that Anthropic was becoming severely capacity constrained. They seem to have had to sign two somewhat adverse contracts with Amazon and Google in short succession. Suddenly model quality is back up again." Anthropic appears to have been compute-starved, and the price of relief was deeper entanglement with both major cloud providers.
The "circular deal" criticism is worth addressing head-on. Google invests $10B in Anthropic. Anthropic commits to spending billions on Google TPUs. Google's cloud division books the revenue. Some of that revenue supports more TPU production. Anthropic buys more TPUs. The cash goes in a circle, but the assets created — trained models, cloud infrastructure, customer relationships — are real. It's not fundamentally different from a bank lending money to a developer who hires contractors who deposit their paychecks at the same bank. The question isn't whether it's circular; it's whether the underlying businesses generate real value.
The competitive dynamics are genuinely unusual, though. Google is simultaneously Anthropic's largest investor, biggest infrastructure vendor, and primary competitor in the foundation model market. A Google Chrome engineer on Hacker News put it bluntly: "I work at Google for Chrome, I can assure you nobody in our team is using Gemini over Claude." When your own employees prefer the competitor's product, investing in that competitor starts to look less like strategy and more like hedging.
One commenter raised the most important long-term question: "All progress points to a commodification of foundation models — Google first named it as 'we have no moat, neither does anyone else.' So there must be some secondary play."
This is the right frame. If foundation models commoditize — and the trajectory of open-weight models from Meta, Mistral, and others suggests they will — then Anthropic's value isn't primarily in having the best model. It's in the ecosystem built around the model: the API platform, the developer tooling (Claude Code, the agent SDK), the enterprise contracts, and the safety research brand.
Google's bet may be less about Claude-the-model and more about controlling the infrastructure layer beneath whoever wins the model race. If Anthropic succeeds, Google profits from both the equity stake and the cloud spend. If Anthropic's models get commoditized but the company still needs massive compute, Google still profits from the cloud spend. The only scenario where Google loses is one where Anthropic migrates entirely to AWS or builds its own chips — both of which this deal makes less likely.
For Anthropic, the calculus is different. They're a public benefit corporation with stated goals around AI safety. Taking $40 billion from Google doesn't technically change their governance structure, but it creates gravitational pull. When your largest investor is also your largest vendor and your largest competitor, the phrase "platform independence" starts to sound aspirational rather than descriptive.
If you're building on Claude APIs, the immediate signal is positive. Capacity constraints have been a real problem — rate limits, degraded performance during peak hours, and model availability issues have all been reported by developers in recent months. A massive influx of TPU capacity should ease these bottlenecks. Expect Claude to feel snappier and more available in the coming months.
If you're choosing between AI providers for a new project, this deal slightly increases lock-in risk for Claude-dependent architectures. Anthropic now has deep financial ties to both Google Cloud and AWS, but the Google relationship is by far the larger one. If you're running on Azure or on-prem, keep an eye on whether Claude API performance varies by cloud provider — it shouldn't, but infrastructure dependencies have a way of creating subtle advantages.
For teams evaluating foundation models, the practical takeaway is to build abstraction layers now. Whether through a library like LiteLLM or your own provider-agnostic wrapper, the ability to swap models without rewriting your application code is increasingly valuable. The corporate chess game between Google, Amazon, Microsoft, and the model labs will create unpredictable shifts in pricing, availability, and capability. Your code shouldn't care.
The enterprise pricing implications are also worth watching. With $40 billion in potential backing, Anthropic can afford to compete aggressively on API pricing. Google and OpenAI will likely respond. The next 12 months could see meaningful price compression for frontier model inference — good news for anyone with a large AI spend.
The AI industry is entering a phase where the capital requirements are so extreme that even well-funded startups can't remain truly independent. Anthropic, OpenAI, Mistral — all of them are now deeply intertwined with hyperscaler economics. The question isn't whether this is good or bad; it's whether the intertwining produces better models and more competition, or whether it slowly collapses the field into three vertically integrated stacks. For now, developers benefit from the competition. The $40 billion question is how long that lasts.
Context: a few weeks ago, Anthropic signed a deal to buy "multiple gigawatts of next-generation TPU capacity" from Google and Broadcom [1]. There have been several previous deals, too.Some people call this sort of thing a "circular deal", but perhaps a better way to think of it i
I think the subtext of the last few weeks is the Anthropic was becoming severely capacity constrained (or approaching that). They seem to have had to sign two somewhat adverse contracts with Amazon and Google in short succession. suddenly model quality is back up again.
It feels like the market is full Wiley Coyote on frontier model makers, and I like Anthropic's B2B business model.But all progress points to a commodification of foundation models--Google first named it as "we have no moat, neither does anyone else." So there must be some secondary pl
It feels like Anthropic is everybody's insurance policy against someone else winning the AI race. So you have Amazon, Google, Microsoft basically every major tech company pushing their own tech hard but simultaneously ensuring they have a survival level stake in Anthropic if they can't bui
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.
https://archive.ph/u274V