GPU marketplace
NVIDIA DGX Cloud Lepton
NVIDIA managed AI platform that combines partner/cloud GPU discovery with model deployment and inference endpoints; it is broader than a GPU-only marketplace.
Product facts checked October 2, 2026Editorial scores last reviewed September 3, 2026
Some outbound links use a first-party redirect hop for click counting. Commission is only claimed when a partner programme is active for that specific link — most vendor hops here are not paid placements. Affiliate disclosure.
Marketplace referrals (for example RunPod or Vast) may return credits or kickbacks when a programme is active — that is a material connection even when it is not a cash CPA.
Best for
Teams wanting to deploy/customize models to managed endpoints over NVIDIA Cloud Partner/GPU marketplace capacity.
Watch out
Endpoints default to one replica and scale to zero after one hour inactive. The documentation says automatic scale-up from zero is unsupported: an operator must scale the deployment up before it can serve requests again. Partner inventory, rates and SLAs vary.
| Billing | Partner GPU marketplace / DGX Cloud usage |
|---|---|
| GPU choice | NVIDIA and partner GPU classes exposed through DGX Cloud Lepton |
| Cold start | Zero replicas cannot serve requests or automatically wake on traffic under the current documentation. Scale the deployment up again, or configure a nonzero minimum; provisioning/model-load delay depends on capacity. |
| Model access | Deploy models to managed endpoints (including vLLM/SGLang workflows) using selected partner GPU capacity; NVIDIA also links to build.nvidia.com serverless APIs/NIMs as separate surfaces. |
Sources for the dated product and announcement claims
- lepton official product and billing documentation — source checked October 2, 2026
- lepton official product and billing documentation — source checked October 2, 2026
- lepton official product and billing documentation — source checked October 2, 2026
- lepton official product and billing documentation — source checked October 2, 2026
Fit detail
When NVIDIA DGX Cloud Lepton is the right shortlist — and when it is not
Use these lists to kill bad comparisons early, before you compare logos.
Ideal for
- NVIDIA DGX Cloud / partner GPU marketplace access
- Teams that need multi-cloud NVIDIA capacity without building a cluster from scratch
- Workloads already planned around NVIDIA professional SKUs
Usually not ideal for
- A token-only public model API with a fixed published tariff
- Bare-metal SKU shopping across non-NVIDIA ecosystems
- Buyers who only need a closed chat API
Spend & ops
How cost behaves — and what breaks first
No invented $/hour quotes. These notes explain the failure modes that show up on the first real invoice.
Cost mental model
Partner GPU marketplace economics under the NVIDIA DGX Cloud Lepton umbrella. Capacity and rates follow the selected partner inventory — verify the live quote before treating it like a fixed serverless bill.
Ops notes
- Restore replicas explicitly after scale-to-zero; retries alone do not wake a zero-replica endpoint.
- Endpoints default to public access without token authentication. Configure endpoint tokens or an IP allowlist when deploying a protected service.
- Confirm partner GPU capacity, current rates and SLA terms in the live quote; QPM/GPU-utilization scaling does not establish automatic wake from zero.
Fit scores
NVIDIA DGX Cloud Lepton on the decision axes
Same editorial scale as the hub chart — useful for shortlists, not as a price quote.
NVIDIA DGX Cloud Lepton
- GPU choice7/10
- Time-to-serving (editorial)5/10
- Price clarity5/10
- Production ops7/10
- Open-model breadth6/10
Among all 16 providers
GPU choice
Pick exact GPUs
Time-to-serving (editorial)
Warm-path fit
Price clarity
Easy to forecast
Production ops
Less DIY ops
Open-model breadth
Catalog depth
Scores are editorial planning ratings (1–10) for product shape — not published $/hour quotes or vendor SLAs. Verify current pricing on each provider site.
Compare
NVIDIA DGX Cloud Lepton vs alternatives
Open a head-to-head when you are deciding between product shapes.
Related
Keep the decision attached to the rest of the stack
Cloud rental is one path. Local VRAM and open-weight fit still matter.