Nscale's IPO filing reveals unit economics and profitability models of neocloud AI computing providers competing with hyperscalers.
Nscale, a UK-based AI infrastructure provider that emerged from a crypto mining operation around two years ago, submitted its S-1 registration to the SEC on September 18, seeking NYSE listing under ticker NSCL. The company has a $103 billion contracted revenue backlog.
First-half 2026 revenue reached $140.6 million, up 1,252% from $10.4 million in the prior year. Net loss: $1.02 billion.
**The numbers behind the neocloud gold rush**
Nscale operates a vertically integrated model, owning or contracting data centers across the US and Europe and managing power supply, GPU fleets, and the software layer connecting them. As of late August, the company had approximately 461,000 GPUs active or under contract, with a power pipeline exceeding 10 gigawatts across 14 regions.
The contracted revenue backlog stands at $103 billion, with an average contract duration of 5.7 years. Anthropic, the AI safety company behind Claude, accounts for the largest share with a contract valued at approximately $45 billion. Microsoft is also among key customers.
**From crypto mining to AI powerhouse**
Nscale was spun out of Arkon Energy, a crypto mining infrastructure firm, in 2024. In March 2026, the company raised $2 billion in Series C funding at a $14.6 billion valuation. Nvidia contributed $1 billion in convertible notes as part of a broader $3.1 billion notes package.
Nscale is targeting a valuation of roughly $30 billion. Goldman Sachs, JPMorgan, and Morgan Stanley are leading the underwriting.
**What this means for the AI infrastructure market**
Nscale's net loss expanded from $368.9 million to $1.02 billion year-over-year. With roughly $45 billion of the revenue backlog concentrated with a single customer, the relationship's stability represents a significant risk factor. The IPO will establish a benchmark for how public markets value contracted GPU capacity, long-term AI compute agreements, and vertically integrated infrastructure models.