Skip to content

Neo Cloud

Layer 04 · Compute & Inference — Partial or partner-run

Run compute, GPU, and storage as Navon services rather than resold ones, on Navon-owned metal, in-country, at 40 to 60 percent less than hyperscaler equivalents. Managed inference sits one layer up, in the Token Factory.

  • Private cloud, CPU and virtual machines
  • GPU cloud with configurable overcommit
  • Managed Kubernetes and Slurm for training and HPC
  • Serverless jobs, endpoints, and hosted dev environments
  • Committed-capacity pricing, 1 and 3 year
  • Block, object, and NAS storage
  • Unified auth and metering across the cloud