Anthropic has signed a deal to rent approximately $45 billion in AI compute from Nscale, a British AI infrastructure company, according to a source familiar with the matter who spoke to TechCrunch. The deal, first reported by Bloomberg, is the latest in an aggressive, months-long spree of compute partnerships as Anthropic races to keep pace with rival OpenAI.
What the Deal Covers
Nscale, founded in 2024, has already struck deals with Microsoft and will supply Anthropic with computing power via Nvidia’s Vera Rubin chips, the company’s newest chip architecture, which combines six different chips working together and is considered the current state of the art in chip design. Bloomberg reported the deal spans six years, with computing power drawn from Nscale’s flagship data center in West Virginia. The capacity is expected to begin powering Anthropic’s services in late 2027.
Part of a Much Larger Pattern
The Nscale agreement is only the latest entry in a striking run of compute deals Anthropic has signed over the past eight months:
- August 2026: A $10 billion deal with AI cloud startup Volta, founded just in January, securing six years of cloud computing power from a Norway-based data center
- July 2026: A $5 billion compute-related deal with AMD
- May 2026: A large computing agreement with SpaceX, drawing capacity from two SpaceX data centers and reportedly providing Anthropic with $1.25 billion worth of computing capacity every month
- April 2026: An expanded partnership with Amazon adding 5 additional gigawatts of compute capacity, alongside an expanded relationship with Google and Broadcom for additional power capacity via custom TPU-adjacent infrastructure
Notably, the SpaceX deal ties Anthropic to a company run by Elon Musk, whose public rivalry with OpenAI CEO Sam Altman has made him an unlikely but consequential ally for Anthropic’s infrastructure ambitions.
Why Anthropic Is Moving So Aggressively
Anthropic’s compute-buying spree reflects the reality that training and running increasingly capable AI models requires computing capacity at a scale few companies can supply from a single source. By diversifying across Nscale, Volta, AMD, SpaceX, Amazon, and Google/Broadcom, Anthropic is spreading its infrastructure risk across multiple providers and chip architectures rather than depending on any single supplier, while also locking in capacity years in advance in a market where GPU and data center availability remains tight.
Anthropic isn’t alone in this race. Google, OpenAI, and Meta are all pursuing similarly aggressive compute acquisition strategies, part of an industry-wide pattern in which securing future computing capacity has become as strategically important as the underlying AI research itself.
Why It Matters
The scale and pace of these deals, roughly $60 billion in newly disclosed compute commitments from Anthropic alone since April, illustrates just how capital-intensive frontier AI development has become. For competitors and industry watchers, Anthropic’s diversified approach, spanning a mix of established players like Amazon and Google alongside newer entrants like Nscale and Volta, offers a window into how AI labs are hedging against supply constraints as demand for AI computing power continues to outpace available capacity.
For continuing coverage of AI infrastructure deals and the compute race among major AI labs, keep following Tech News Reports for ongoing updates.
