Anthropic is reportedly in discussions with AI infrastructure provider Nscale for a second $45 billion agreement covering another 460 megawatts of computing capacity at the Monarch campus in West Virginia. The potential deal follows a previously announced six-year arrangement of similar value, highlighting the scale of investment required to meet rising demand for advanced AI models and the growing financial commitments behind the global AI infrastructure buildout.

A Potential $90 Billion Commitment to AI Computing

The proposed agreement would mirror the initial arrangement, under which Anthropic agreed to pay approximately $45 billion for access to 460 megawatts of computing capacity over six years. The first facility is planned to use Nvidia’s Vera Rubin AI chips, with deployment expected to begin in late 2027. If the second agreement is finalized on comparable terms, the two facilities could provide Anthropic with a combined 920 megawatts of capacity and represent commitments approaching $90 billion.

The additional capacity would support the computing-intensive workloads required to train and operate increasingly sophisticated AI systems, including Anthropic’s Claude models. As enterprises integrate generative AI into software development, customer service, research and business operations, providers face mounting pressure to secure reliable computing resources over multiple years. Long-term infrastructure agreements can provide greater visibility into capacity planning, although their economic value depends on utilization, operating costs and the pace at which demand translates into revenue.

The negotiations remain ongoing, and the reported terms could change. The additional facility should therefore be viewed as a potential expansion rather than a completed contract.

Monarch Campus and the Race for Computing Capacity

Located in Mason County, West Virginia, the Monarch campus is designed for large-scale AI infrastructure deployment, with long-term power capacity potential exceeding 8 gigawatts of gross generation capacity. Its scale positions the site to accommodate multiple facilities as demand for high-density computing grows.

For Anthropic, securing additional capacity could reduce the risk of infrastructure constraints limiting model development and commercial expansion. For Nscale, another major customer agreement would strengthen its position in the market for specialized AI cloud services, where providers compete to deliver advanced chips, data center capacity and reliable power.

However, planned capacity is not the same as operational capacity. Facilities require substantial investment in equipment, construction, cooling systems and electricity infrastructure before they can generate revenue. Delays in any of these areas could affect delivery schedules and the economics of long-term contracts.

Financing and Execution Risks Remain Significant

The scale of the reported agreement also illustrates the financial exposure associated with the AI infrastructure boom. Nscale must secure funding for the equipment and facilities needed to fulfill its existing commitments, while ensuring that construction and service delivery meet contractual requirements. Its initial Anthropic agreements are subject to delivery and service-availability conditions, and required financing had not been fully secured in the company’s public offering disclosures.

Concentrated customer exposure creates another risk. Large contracts can provide long-term revenue visibility, but dependence on a limited number of customers can magnify the impact of spending changes, contract disputes or delays. For Anthropic, committing to extensive computing capacity also raises questions about future utilization, technology transitions and whether revenue growth will justify infrastructure costs.

Investors will be watching for confirmation of the second agreement, clarity on financing arrangements and progress toward bringing the first Monarch facility online. Construction milestones, access to electricity, deployment of Nvidia’s Vera Rubin systems and Anthropic’s ability to sustain demand for its AI services will help determine whether the proposed expansion becomes a durable commercial opportunity or adds to the sector’s growing execution and capital risks.