AI
Anthropic locks in $45B compute deal with Nscale as AI giants race for infrastructure
The AI lab joins a competitive scramble among major players for massive computing resources to power advanced model development.
TechCrunch — AI
Anthropic has entered into a $45 billion agreement with Nscale, a UK-based AI infrastructure provider, to access significant computational resources over a six-year period, according to reporting from Bloomberg.
The arrangement represents part of Anthropic's broader expansion of its compute footprint as it seeks to keep pace with rivals in the AI sector. The hardware will run Nvidia's Vera Rubin chips—a new-generation architecture combining six processors in a coordinated system—and will be delivered from Nscale's West Virginia data center. Services powered by these resources are expected to come online by late 2027.
Nscale's rapid rise in AI infrastructure#
Despite being founded just two years ago, Nscale has already established itself as a significant player in the compute-supply landscape, having previously partnered with major tech firms including Microsoft. The company's expansion into high-capacity AI infrastructure reflects growing demand for computational power among leading AI developers.
Anthropic's aggressive scaling campaign#
Anthropic has pursued a rapid series of infrastructure agreements over recent months to bolster its computational capacity:
- August: $10 billion six-year deal with AI cloud startup Volta (sourced from a Norwegian data center)
- July: $5 billion compute arrangement with AMD
- May: Major computing partnership with SpaceX drawing from two data centers, valued at approximately $1.25 billion monthly
- April: Expanded partnership with Amazon adding 5 gigawatts of capacity, plus expanded arrangement with Google and Broadcom
The broader AI infrastructure arms race#
Anthropic's infrastructure push reflects an industry-wide dynamic. Competing organizations—including Google, OpenAI, and Meta—are simultaneously pursuing aggressive compute-acquisition strategies to maintain their competitive standing in AI development. The escalating demand for computational resources has become a defining feature of the current AI landscape.
