Lambda provides GPU-backed On-Demand Cloud virtual machines, 1-Click GPU clusters, and larger contracted AI infrastructure. Its cloud documentation describes Linux instances with predefined GPU, CPU, memory, and root-volume combinations. Current capacity and instance pricing should be checked in the provider console or current pricing documentation.
On-Demand Cloud offers Linux GPU virtual machines with predefined configurations; 1-Click Clusters and contracted superclusters address larger multi-GPU deployments. Published instance documentation includes GPU model and count, VRAM, vCPU, system RAM, root volume, and available image families for listed configurations.
Billing and pricing
Lambda bills running On-Demand Cloud instances by hourly usage in one-minute increments, beginning after launch health checks and ending at termination. Filesystems are billed separately per GiB-month in hourly increments while they exist. The billing documentation says ingress and egress are not charged for those filesystems; it does not make a blanket statement about every product charge.
Regions and availability
Each instance is associated with a geographical region. Region and instance availability can vary; verify the current region list and selected instance type in Lambda Cloud before ordering.
Support
Lambda provides documentation and support links for cloud users. Available support and contract terms depend on the account and selected product; no universal response-time promise is recorded here.
Before you buy
Lambda’s public AI Cloud page advertises pay-as-you-go instances with 1–8 NVIDIA GPUs starting at $0.79 per hour. This is a provider-level starting price, not a named GPU configuration; exact hardware, region, and current availability must be confirmed in the live product flow.
Counts describe the records published on HostCritiq, not every offer Lambda sells. Prices can differ by billing term, region, currency and tax. Open a plan for its full terms.