Lambda
GPU cloud for AI training and inference with on-demand and cluster options
Lambda provides GPU cloud infrastructure for AI workloads. On-demand instances are available in 1x to 8x GPU configurations with NVIDIA B300, B200, H200, H100, and A100 GPUs. 1-Click Clusters scale from 16 to 2,000+ interconnected GPUs for large-scale training. Billing is per-minute with no egress fees. GPU pricing starts at $1.10/hr for A100 80GB and $2.99/hr for H100 SXM. Also offers private cloud deployments and physical GPU workstations.
Pricing: Pay-per-minute
What Lambda is
Lambda is a GPU cloud focused on AI training and inference, billed by the minute with no egress fees. It spans single on-demand instances up to reserved multi-node clusters, covering everything from a proof of concept to production.
GPUs and pricing
Per-GPU on-demand rates (per their pricing page, June 2026) run from $0.69/hr for a Quadro RTX 6000 and $0.79/hr for a V100, through A100 at $1.99-$2.79/hr depending on memory and form factor, H100 PCIe at $3.29/hr and H100 SXM at $3.99-$4.29/hr, GH200 at $2.29/hr, up to B200 SXM6 at $6.69-$6.99/hr. There is no free tier.
Who it fits
Lambda suits teams that want straightforward per-minute GPU rental with predictable published rates, from a single card for experiments to reserved clusters for sustained training. Compared to a marketplace like Vast.ai, where prices float with supply and demand, Lambda publishes fixed rate cards. Compared to a high-end cluster provider like CoreWeave, the single-instance on-demand path is simpler to start on, with the trade-off of a narrower top end for very large multi-node jobs.
Lambda Alternatives
Explore 67 products in the Inference APIs category. View all Lambda alternatives.
Genesis Cloud
European GPU cloud for AI training and inference powered by 100% green energy
Nebius
Full-stack AI cloud with GPU infrastructure for training and inference
CoreWeave
GPU cloud infrastructure built for large-scale AI training and inference workloads
Is your product missing?