Sources and Nscale's US SEC filings: ByteDance accounted for nearly 75% of Nscale's sales in 2025 and used Nscale's facility in Norway to access Nvidia AI chips
Chinese tech giant was AI cloud provider's largest customer last year but is not prominently named in pitch to stock market investors
Context & Ripple Effects
ByteDance entered 2026 as Nvidia's largest buyer in China, but reporting in 2025 said Chinese regulators had blocked its use of the chips in new domestic data centers. That backdrop makes the reported use of overseas capacity a consequential part of its compute strategy, alongside its stated plan to increase Nvidia-chip spending in 2026.
For Nscale, the customer-dependence report arrives as it has been discussing a major financing round ahead of an IPO. Its planned $3.5B financing and Nvidia participation put revenue concentration, disclosure, and the provenance of demand at the center of investor diligence.
First-order effects
- Nscale's prospective public-market investors must assess whether reported reliance on ByteDance for nearly three quarters of 2025 sales leaves its valuation exposed to one customer's spending and access to Nvidia hardware.
- ByteDance's reported use of Nscale's Norway facility gives it an overseas route to AI compute after the reported domestic restriction, while putting that arrangement under greater scrutiny.
Second-order effects
- Nscale's lenders and equity backers face stronger incentives to test customer concentration and export-control exposure before financing capacity expansion tied to ByteDance demand.
- Nvidia's participation in Nscale financing links its infrastructure partner more tightly to diligence over who ultimately uses the compute built with Nvidia systems.
Third-order effects
- If the reported arrangement withstands scrutiny, neocloud underwriting will increasingly turn on both contracted demand and the jurisdictional path through which customers obtain compute.
- The episode points to AI cloud capacity becoming a venue where export controls, customer disclosure, and infrastructure finance intersect rather than separate decisions.
The trend: AI infrastructure is being financed around concentrated anchor customers, making the geography and disclosure of their compute demand material to capital markets.