NVIDIA DGX Cloud Lepton Review 2026

NVIDIA-run marketplace-and-runtime platform that connects developers to multi-cloud GPU compute with a single API, formed after NVIDIA's acquisition of Lepton AI.

NVIDIA DGX Cloud Lepton is a focused option for multi-cloud GPU capacity, training, and inference deployment. Its clearest differentiator is a unified NVIDIA-operated layer across participating GPU cloud providers, with serverless inference, region choice, and NVIDIA-optimized containers. The staging record identifies these concrete capabilities: Multi-cloud GPU marketplace, Serverless inference with autoscaling, Per-second billing across partners, NVIDIA-optimized container images, Region-diverse GPU access, Managed training with checkpointing. Those details make the product worth evaluating for a defined workflow, but they are not a substitute for a hands-on pilot with the files, prompts, permissions, latency requirements, and failure cases your team actually has.

Pricing needs special care. The existing record lists Marketplace: Per-second GPU billing (Varies by cloud partner); Pro platform: Paid monthly (Team management + higher limits); Enterprise: Custom (SLAs, dedicated capacity, private routing). During this scheduled run, both the vendor page and the expected pricing route were requested directly with curl, but neither returned usable HTML. Every figure and plan label therefore needs manual verification before it appears in a budget or purchasing decision. Do not assume that an open-source component makes hosting, support, storage, model inference, or implementation free. Ask for a written breakdown of subscription or usage charges, minimum commitments, overages, support, data egress, implementation, renewal terms, and cancellation rights. Model a 12-month total cost that includes engineering, security review, monitoring, and human quality control.