Anthropic has reached a landmark agreement to lease approximately $45 billion worth of artificial intelligence computing capacity from Nscale, a British AI infrastructure company, according to a source with direct knowledge of the arrangement. The deal represents one of the largest compute procurement agreements in the AI industry to date.

Nscale, which was only established in 2024, has quickly positioned itself as a significant player in the AI infrastructure space. The company has previously struck agreements with major technology firms, including Microsoft, and will supply Anthropic with computing power through Nvidia's Vera Rubin chip system - the semiconductor giant's most advanced hardware platform currently available.

Nvidia's Vera Rubin System at the Center of the Deal

The Vera Rubin architecture is considered the current frontier of chip engineering. It integrates six distinct chips operating in coordination, making it one of the most sophisticated computing systems available for AI workloads. According to the source, the computing capacity secured through this agreement is expected to begin supporting Anthropic's services and operations in late 2027.

The deal, which was first reported by Bloomberg, spans a period of six years. The computing infrastructure will be housed at Nscale's flagship data center facility located in West Virginia, providing Anthropic with a long-term, dedicated source of high-performance compute.

A Pattern of Aggressive Compute Expansion

This agreement is the latest in a rapid series of compute partnerships that Anthropic has pursued over the past several months. The company has been systematically scaling up its infrastructure capacity in an effort to strengthen its competitive position, particularly against OpenAI, which remains its most prominent rival in the large language model space.

Earlier this month, Anthropic signed a $10 billion agreement with Volta, an AI cloud startup that was itself only founded in January of this year. That deal also covers a six-year period and secures cloud computing resources from a data center based in Norway, further diversifying Anthropic's international infrastructure footprint.

In July, the company entered into a $5 billion compute-related agreement with AMD, signaling that its procurement strategy extends beyond a single hardware vendor. The arrangement with AMD adds another layer to Anthropic's growing network of chip and compute suppliers.

SpaceX, Amazon, Google, and Broadcom Also in the Mix

A few months prior, in May, Anthropic disclosed that it had entered into a substantial computing agreement with SpaceX, the aerospace and technology company led by Elon Musk. The partnership draws capacity from two separate SpaceX data centers and is reported to be delivering approximately $1.25 billion worth of computing capacity to Anthropic on a monthly basis. The arrangement is notable given Musk's well-documented rivalry with OpenAI CEO Sam Altman, a dynamic that has made him an unlikely partner for Anthropic in its competition against that company.

In April, Anthropic also moved to significantly deepen its existing relationship with Amazon Web Services, securing access to an additional five gigawatts of compute capacity through that partnership. Amazon has been one of Anthropic's most prominent backers and infrastructure partners since the early stages of the company's growth.

That same month, Anthropic also extended and expanded its working relationships with both Google and Broadcom, adding further computing power capacity to its growing portfolio of infrastructure agreements.

Industry-Wide Race for Computing Resources

Anthropic's aggressive approach to locking in compute supply is not unique within the broader AI industry. The competition to secure as much high-performance computing capacity as possible has become a defining feature of the current AI landscape. Other major technology companies, including Google, OpenAI, and Meta, are all engaged in similar efforts to build out and reserve substantial infrastructure resources as demand for AI services continues to grow at a rapid pace.

The scale and speed of Anthropic's recent procurement activity - spanning multiple continents, hardware vendors, and infrastructure partners within a period of less than a year - underscores how central computing capacity has become to competitive strategy in the AI sector. As model training and inference workloads become increasingly resource-intensive, securing long-term access to cutting-edge compute has become as strategically important as the research and development work itself.