Nscale is placing an unusually large contracted-revenue figure at the center of its pitch to prospective investors as the Nvidia-backed U.K. AI infrastructure company prepares for a potential initial public offering. The approximately $103 billion figure illustrates the scale of capital being committed to AI computing, but it also highlights an important distinction for investors between signed future capacity commitments and revenue already recognized by a rapidly expanding infrastructure provider.

$103 Billion Backlog Changes the Scale of Nscale’s IPO Story

Nscale is telling prospective investors that it has approximately $103 billion in total contracted revenue, more than double the roughly $51 billion figure it had been presenting before the latest wave of customer commitments. According to documents reviewed by The Information and cited by Reuters, Nscale’s contracts have an average duration of 5.7 years, translating to approximately $18 billion in annualized contracted revenue across the portfolio.

The distinction between contracted revenue and recognized revenue is critical. The $103 billion figure represents future revenue associated with signed lease and compute agreements rather than cash already generated by the company. A person familiar with Nscale’s discussions with investors cautioned that the figure is illustrative and not formal revenue guidance.

Nscale’s reported operating scale remains considerably smaller. The Information reported that the company generated more than $100 million of revenue in the second quarter of 2026, compared with approximately $37 million in the first quarter and roughly $33 million for all of 2025. That sharp acceleration demonstrates genuine growth, but it also shows the distance between Nscale’s current revenue base and the headline value of its contracted commitments.

Anthropic Deal Becomes a Major AI Infrastructure Anchor

The biggest catalyst behind the expansion in Nscale’s contracted backlog is a reported $45 billion, six-year agreement with Anthropic for AI computing capacity at Nscale’s West Virginia data-center campus. Nscale is expected to deploy Nvidia’s Vera Rubin architecture to support Anthropic’s computing requirements, with the facility expected to become operational toward the end of 2027.

The agreement reflects a broader change in the AI industry. Frontier-model developers increasingly require enormous amounts of computing capacity not only for training but also for inference as usage expands. Anthropic has committed to multiple large infrastructure arrangements, demonstrating how AI model companies are securing capacity years in advance to avoid shortages as demand grows.

For Nscale, the strategic importance extends beyond the size of a single contract. The company operates an AI-native infrastructure platform spanning data centers, GPU compute, networking, storage and software. It has also announced a strategic partnership with humanoid-robotics company Figure involving up to 100,000 Nvidia GPUs, with an initial $3.5 billion compute commitment and an intention to scale beyond $6 billion.

Nvidia’s Role Extends Beyond Supplying Chips

Nscale’s development is also significant for Nvidia because it illustrates how the chipmaker’s influence in AI infrastructure is expanding beyond GPU sales. Nvidia participated in Nscale’s $2 billion Series C announced in March, which valued Nscale at $14.6 billion. Nvidia was also among the participants in Nscale’s earlier $1.1 billion Series B financing.

Nscale has simultaneously been building the financing infrastructure required to deploy expensive AI computing clusters. In February, the company announced a $1.4 billion GPU-backed delayed-draw term loan to finance cluster deployments across Europe. That financing structure illustrates the capital intensity of the neocloud model: large customer contracts require substantial upfront spending on GPUs, data centers, power and networking before the associated revenue can be recognized over time.

IPO Investors Will Need to Look Beyond the Headline Number

The potential IPO therefore offers a test of how public markets value AI infrastructure companies whose future economics depend on long-duration contracts and substantial capital expenditure. Nscale was valued at $14.6 billion in its March Series C financing, while Reuters reported on September 4 that the company was seeking approximately $3.5 billion in additional pre-IPO financing, including $1.5 billion of convertible notes and a potential $2 billion investment from Nvidia. Discussions remained subject to change.

The key question will be whether Nscale can convert its contracted capacity into revenue and cash flow at attractive economics. Investors will need to monitor data-center construction, GPU deployment, power availability, customer concentration, financing costs, contract duration and recognized revenue rather than treating the $103 billion figure as equivalent to current sales.

Nscale’s backlog nevertheless provides an important window into the scale of the AI infrastructure cycle. If customers continue signing multibillion-dollar compute commitments, the demand outlook for Nvidia-powered data centers, specialized cloud providers and related power infrastructure could remain substantial. But the company’s next phase will be defined by execution: turning enormous future commitments into operational capacity, recognized revenue and sustainable cash generation while managing the heavy capital requirements of the AI infrastructure business.