Nscale's IPO filing reveals $103.4 billion in active and contracted AI infrastructure agreements, providing visibility into committed customer demand.
Nscale has filed for an initial public offering on the New York Stock Exchange, revealing $103.4 billion in active and contracted total contract value (TCV) as of August 31, 2026. The London-based AI infrastructure company disclosed this figure in its September 18 SEC filing, highlighting the scale of agreements with major technology companies and AI developers. Nscale plans to list under ticker NSCL.
The $103.4 billion represents aggregate contracted revenue across the committed terms of signed customer agreements, rather than earned revenue or company valuation. Approximately $2.6 billion is active TCV, reflecting substantial growth from $38 billion in active and contracted TCV at end of 2025. Contracts are generally structured as long-term, take-or-pay arrangements with a weighted average contract life of approximately 5.7 years.
These agreements support Nscale's strategy of developing and operating large-scale AI computing infrastructure for hyperscalers, AI laboratories, enterprises, and government-backed AI initiatives. Services include dedicated GPU clusters, cloud computing, AI model training and inference, and purpose-built data centers.
As of August 31, 2026, Nscale operated approximately 25,000 active GPUs and 461,000 active and contracted GPUs. Its infrastructure portfolio included five active and 12 contracted data center sites, representing approximately 1.37 gigawatts of active and contracted power capacity. The company identified approximately 10 gigawatts of potential development capacity across sites under its ownership or long-term control and related power agreements.
A substantial portion of Nscale's contracted business derives from major AI infrastructure agreements. On August 25, 2026, the company entered into agreements with Anthropic to provide dedicated GPU infrastructure at Monarch Compute Campus, with aggregate payments of up to $44.6 billion and deployment of NVIDIA Vera Rubin NVL72 GPUs.
Nscale is also expanding its relationship with Microsoft. It acquired a 200-megawatt site at Portugal's SINES Data Campus in April 2026 to support deployment of more than 66,000 NVIDIA Vera Rubin NVL72 GPUs beginning in late 2027. This builds on an earlier Microsoft deployment of more than 12,600 NVIDIA Grace Blackwell Ultra GPUs at the campus.
Another significant agreement involves humanoid robotics developer Figure AI. In August 2026, Nscale announced a multiyear strategic partnership with potential deployment of up to 100,000 NVIDIA Vera Rubin GPUs, with initial deployments targeted for the second half of 2027. Nscale is also making a strategic investment in Figure and serving as its preferred AI infrastructure provider for developing Helix models and humanoid robots.
The company's rapid expansion is reflected in its financial performance. Nscale generated $140.6 million in revenue during the first half of 2026 versus $10.4 million in the corresponding 2025 period—representing 1,252% year-over-year growth. Full-year revenue increased from $19.1 million in 2024 to $33 million in 2025.
Despite this growth, Nscale continues to incur substantial losses while developing its infrastructure. The company reported a $1.02 billion net loss for the first half of 2026 versus $368.9 million for the corresponding 2025 period. Its full-year 2025 net loss was $761.8 million. Nscale identified customer concentration, access to financing, and timely infrastructure deployment as key business risks.
To support expansion, Nscale entered into a subscription agreement on September 15, 2026, involving a minimum of $3.1 billion, comprising $2.1 billion in unsecured convertible loan notes and $1 billion in convertible notes or non-voting shares to be issued by NVIDIA.
The company is moving beyond physical infrastructure through its planned acquisition of Anyscale, announced in July 2026. Anyscale provides an AI computing platform built around Ray, the open-source framework for distributed AI. The acquisition is expected to add approximately 200 employees and expand Nscale's software capabilities in distributed computing, model training, and inference, subject to customary closing conditions.
Nscale intends to use IPO proceeds to support data center development, technology investments, working capital, and other corporate purposes. It may also allocate proceeds toward acquisitions or investments in complementary technologies.
Goldman Sachs, JPMorgan, and Morgan Stanley are serving as lead bookrunners for the offering. The preliminary filing does not yet specify the number of shares to be offered or expected price range.