Nscale’s IPO filing puts AI infrastructure in the public-market spotlight

Nscale has filed a registration statement with the US Securities and Exchange Commission for a proposed initial public offering. The AI cloud provider has applied to list on the New York Stock Exchange under the symbol NSCL. Its announcement says the share count and offer-price range had not yet been set, and the registration statement was not yet effective.

Goldman Sachs, J.P. Morgan and Morgan Stanley are named as lead bookrunners. Nscale describes an integrated business spanning AI software, computing capacity, data centres and power. Its move towards public markets brings further attention to the infrastructure needed to train and run AI systems.

The Financial Times reporting supplied for this article adds commercial context. It describes Nscale pursuing an IPO valuation of around $35 billion and computing agreements worth up to $44 billion with Microsoft and $45 billion with Anthropic over several years. These are reported ambitions and contract values, rather than an achieved listing valuation or recognised revenue. Nscale’s IPO announcement does not confirm a valuation.

The FT also examines customer concentration and disclosure. It reports that ByteDance, through a subsidiary, accounted for 73% of Nscale’s $33 million revenue in the reported year, while its identity appeared in supporting paperwork rather than being named in the main filing. The report describes the arrangement as legal, while discussing the potential impact of tighter export controls and geopolitical scrutiny. Nscale declined to comment to the FT.

For Bolt’s CUDO investment thesis, the relevance is sector-wide. A proposed listing by another neocloud provider illustrates a potential route to capital for businesses combining GPUs, power and operational delivery. It also shows why contract duration, customer concentration, financing requirements and disclosure matter alongside headline growth.

Nscale is a separate business from CUDO. Its proposed IPO does not establish CUDO’s valuation or imply a similar exit. The useful comparison is the business model and the questions public-market investors bring to AI infrastructure: the quality of demand, the economics of delivery and the ability to turn contracted capacity into sustainable cash generation.