TL;DR
Nscale, the London-based AI infrastructure company backed by Nvidia, updated its US listing papers on 12 November with preliminary results for the nine months to 30 September, Reuters reports. Revenue came to about £231m ($307.5m), against about £16m ($21.3m) a year earlier, a rise of 1,344%. The company expects a net loss of about £1.3bn ($1.7bn), up from roughly £325m ($433.3m).
The company intends to float on the New York Stock Exchange under the ticker “NSCL”. It has still not set out the size or price of the offering.
What the amendment adds
When Nscale first filed in September, it published figures only for the first half. The amended filing adds the July to September quarter, taking the reported period to nine months.
The numbers point in two directions. Revenue is now around 14 times the year-earlier level, as the company sells compute, power and data centre capacity plus the software customers use to train and run models. But losses are growing faster in absolute terms: the expected net loss is roughly four times last year’s.
The lead underwriters are Morgan Stanley, Goldman Sachs and J.P. Morgan. Reuters also notes a Telegraph report on 10 November that Nscale was weighing whether to push the listing into next year.
The wider backdrop
The filing lands amid coverage we have tracked closely: the AI share sell-off after the US midterms, investor doubts over Anthropic’s IPO timing and Microsoft passing on a larger Nscale deal in Texas. Nscale has not said whether it will delay.
Sterling figures here are our conversions, at roughly 75p to the dollar, and are approximate.
Looking Forward
The filing leaves the key decision, whether to proceed this year, unresolved. In our view, the revenue line supports the case that demand for UK-headquartered AI compute is real, while the loss line shows how much capital that growth consumes. For UK investors and policymakers, the more pointed fact may be the venue: one of Britain’s most prominent AI infrastructure firms is choosing New York for its listing. If Nscale does proceed, its pricing will be a direct read on how public markets value loss-making AI infrastructure in current conditions.