What Nscale does
Nscale is a privately held AI infrastructure company building and operating “full-stack” neocloud capacity—covering power, AI-focused data centers, GPU superclusters, and managed software—to deliver enterprise AI training, fine-tuning, and inference. The company positions its differentiation around vertically integrated delivery of compute, networking, storage, and orchestration software, and around “sovereign” deployments that keep AI workloads under specific national/regulatory control. Nscale also sells managed AI services and platform capabilities through its GPU cloud and fleet-operations tooling (e.g., inference endpoints, prompt workbench, fine-tuning, managed training via Slurm, Kubernetes-based workloads, and operational observability and usage tracking). Recent strategy has been to expand the software layer on top of its infrastructure: in July 2026 it announced a definitive agreement to acquire Anyscale (a Ray-based platform for scaling AI workloads) to add a managed software layer and broaden its customer base. The company’s recent financing and expansion announcements emphasize rapid, multi-region infrastructure build-out, including major customer/partner deployments and large-scale capacity commitments tied to next-generation NVIDIA platforms.