What happened
News reports say Nvidia is expanding its role in Anthropic’s AI plans, in a so-called ninth compute corridor. The discussions reportedly involve about $35 billion in commitments, though Anthropic and Lambda have not officially confirmed details. The idea of a compute corridor is to layer compute capacity through a web of partners. Nvidia would serve as a supplier of GPUs and as a landlord of the infrastructure where the AI models run. The story underscores how the AI build-out relies on a wide network, not just a single vendor, and that big-dollar deals can span many years.
Why it matters
If true, Nvidia could gain more revenue visibility and influence over who can run large AI models. The arrangement shows how the AI cycle now mixes hardware, data-center space, and services across multiple intermediaries. It highlights a shift from buying chips to securing long-term access to both chips and the facilities that host them. This can raise Nvidia’s power in the AI stack, while also tying its revenue to a few big customers and heavy data-center investment.