Anthropic has signed a six-year, $10 billion deal with Volta, a young AI cloud computing startup, to secure a major new source of computing capacity for training and running its Claude models. The agreement was confirmed this week.

Under the deal, Volta will supply cloud computing services to Anthropic over six years. To deliver that capacity, Volta is working with Bitdeer, a crypto-mining firm turned data center operator, to build a facility in Norway capable of delivering 133 megawatts of computing power.

The Norway facility will run on Nvidia’s Vera Rubin systems, the chipmaker’s newest AI chip architecture, underscoring how closely the current wave of AI infrastructure spending is tied to Nvidia’s product roadmap.

Volta itself is a relatively new player in the AI infrastructure market. The company was founded in January by former executives of Brookfield Asset Management and operates by leasing AI computing capacity while also helping clients finance the purchase of expensive AI chips. Volta recently raised $300 million in venture funding, a round that valued the company at $2.4 billion.

The Anthropic deal is part of a broader pattern of aggressive computing capacity expansion by the company in recent months. Anthropic has also signed new compute agreements with SpaceX and Amazon as it works to keep pace with rising demand for its Claude models and the enormous computing costs that come with training and serving large language models at scale.

The scramble for AI computing capacity has become one of the defining financial stories of 2026, with major AI labs signing multibillion-dollar, multiyear contracts with both established cloud providers and a new wave of specialized infrastructure startups like Volta. For newer entrants such as Volta, deals of this size offer a fast path to scale, while for AI labs like Anthropic, locking in capacity years in advance has become essential to supporting continued growth in usage of their models.