The agreement covers roughly 460 megawatts of computing capacity at Nscale’s Monarch Compute Campus in Mason County, West Virginia. Nvidia’s next-generation Vera Rubin systems are expected to supply the processing hardware, with capacity scheduled to begin coming online in late 2027.
The scale of the commitment highlights how access to electricity, data centres and specialised processors is becoming a defining competitive factor for developers of frontier AI models. Anthropic is expanding infrastructure at an exceptional pace as usage of Claude, particularly its Claude Code programming service and enterprise tools, drives demand for both model training and day-to-day inference.
The 460MW allocation represents about one-third of the initial 1.35-gigawatt computing capacity planned for the first phase of the Monarch campus. Nscale has designed the West Virginia development to expand substantially beyond that level, with a longer-term pathway exceeding eight gigawatts if demand and power availability support further construction.
Nscale announced plans earlier this year to make Monarch a flagship deployment site for Nvidia’s Vera Rubin architecture. The platform succeeds Nvidia’s Blackwell generation and is being designed for increasingly demanding AI training and inference workloads, combining processors, networking and memory within tightly integrated systems.
Anthropic’s agreement also gives Nscale a major long-term customer as the infrastructure company considers its own public-market plans. Founded in 2024, Nscale has expanded rapidly from a specialist AI infrastructure operator into a major developer of large-scale computing campuses.
The company raised $2 billion in March at a valuation of $14.6 billion. Investors in that financing included Aker, Nvidia, Dell, Lenovo, Nokia and several financial institutions. Nscale has also expanded beyond physical infrastructure, acquiring AI software company Anyscale this year in a transaction valued at about $1.65 billion.
The Monarch project had previously been linked to Microsoft, which signed a letter of intent covering as much as 1.35GW of AI computing capacity using Vera Rubin systems. That proposed commitment did not progress as initially envisaged, leaving Nscale seeking customers for portions of the site. Anthropic’s 460MW agreement now provides a substantial anchor for the development.
The deal forms part of a much wider infrastructure build-out by Anthropic. The company has been securing computing capacity from multiple suppliers rather than relying on a single cloud or chip partner, reflecting both the huge processing requirements of frontier models and the risk that electricity and GPU shortages could constrain growth.
Anthropic has agreements involving Amazon and Google and has expanded its access to Nvidia processors through additional infrastructure providers. It also struck arrangements with SpaceX covering computing capacity at the company’s large AI data-centre clusters and has strengthened its relationship with AMD.
AMD in July unveiled plans to supply Anthropic with tens of billions of dollars of AI servers while potentially investing as much as $5 billion in the company. The strategy gives Anthropic access to an alternative to Nvidia hardware while increasing competition among semiconductor suppliers for the largest AI customers.
Anthropic has also explored ways to increase control over the technology underpinning its models. The company examined a possible multibillion-dollar acquisition of chip start-up MatX before discussions shifted towards a possible commercial partnership. Expanding internal chip-design expertise could eventually reduce dependence on external hardware and help control computing costs.
Those costs have become increasingly important as Anthropic prepares for a possible initial public offering. The company confidentially filed for a listing earlier this year and is expected to publish further details of its flotation plans after the US Labor Day holiday, potentially opening the way for trading later this year.
Expectations surrounding the listing are being supported by exceptional revenue growth. Anthropic has projected revenue of roughly $190 billion to $200 billion in 2028, compared with an annualised revenue rate that had reached tens of billions of dollars during 2026. Such forecasts depend heavily on continued expansion in enterprise AI, coding services and autonomous software agents.
The Nscale contract illustrates the financial risk accompanying that growth. A $45 billion commitment locks Anthropic into infrastructure spending years before the full capacity becomes operational, requiring demand for Claude services to keep rising sufficiently to justify the investment.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
