Lambda
thecontinuitylayer.com/tools/gpu-hosting/lambdaTeams that want plain, well-run NVIDIA machines for training or serving larger open models
GPU cloud · Lambda
Lambda is a GPU cloud built for machine-learning work. It rents on-demand instances with 1, 2, 4 or 8 GPUs, plus 1-Click Clusters and reserved capacity for larger jobs. Instances come with a preinstalled ML software stack.
For open models it is a straightforward place to run vLLM or similar on an H100 or B200 with no egress fees. Prices vary by instance configuration, and popular GPU types can be out of stock in some regions. There is no cheap consumer-card tier.
Specs
| Price per hour | H100 SXM from $3.99/hr; B200 from $6.69/hr (per GPU, on demand) |
|---|---|
| GPUs | NVIDIA data-center GPUs including B200, H100, A100 and GH200, in 1 to 8 GPU instances and clusters |
| H100 SXM | $3.99 per GPU-hour on demand |
| B200 SXM6 | $6.69 per GPU-hour on demand |
| Instance sizes | 1x, 2x, 4x and 8x GPU |
| Billing | Per minute on instances |
| Egress | No egress fees |
| Larger scale | 1-Click Clusters and reserved contracts |
Good for
- Serving 70B-class and larger open models
- Multi-GPU fine-tuning
- Teams wanting a simple ML-focused cloud
Not for
- Hobbyists on a small budget
- Anyone wanting serverless scale-to-zero
See also
- CoreWeave · Companies running open models at large scale that need whole GPU servers and clusters
- Together AI · Teams that want a dedicated, managed endpoint for an open model without running servers
- Novita AI · Developers who want both an open-model API and rentable GPUs from one account
We last reviewed Lambda on 31 October 2026. Prices and specs change; check lambda.ai before you buy or sign up. Figures we could not confirm on the vendor's own page read “not published” or “varies”.
Vendor source: https://lambda.ai/instances