tools

Lambda 1-Click Clusters

Lambda 1-Click Clusters provides dedicated, InfiniBand-connected GPU clusters for distributed AI training, fine-tuning, and inference.

In inglese

Lambda 1-Click Clusters are managed multi-node GPU clusters built with NVIDIA HGX B200 or H100 systems. They include InfiniBand networking, CPU management nodes, local NVMe storage, Lambda Stack software, and access through the Lambda Cloud dashboard, SSH, or JupyterLab.

AI researchers, engineers, and enterprise teams use them for large-scale training, fine-tuning, inference, and proof-of-concept workloads. GPU time is billed per GPU per hour; filesystem storage is billed separately, and applicable sales taxes may apply.

Features

NVIDIA HGX B200 and H100 GPU clusters
16 to 2,000+ NVIDIA GPUs
NVIDIA Quantum-2 InfiniBand networking
Managed Kubernetes or Slurm orchestration
S3-compatible storage
Preinstalled Lambda Stack with PyTorch, TensorFlow, NCCL, and Open MPI
SSH and JupyterLab access
ISO 27001, 27017, 27701, 22301, and SOC 2 Type II compliance

Use cases

Train large language and multimodal models across multiple GPU nodes
Fine-tune models with distributed PyTorch or TensorFlow workloads
Run large-scale inference deployments
Benchmark new GPU architectures and distributed training systems
Prototype production AI workloads with short-term reservations

Pros

Cons

Get it

Documentation

Security

SOC 2 Type II — “SOC 2 Type II attestation.” source
SOC 2 — “SOC 2 Type II attestation.” source
ISO 27001 — “ISO 27001, 27017, 27701, and 22301 certified.” source

Pricing

Prices checked: 2026-09-24

NVIDIA HGX B200 — 16 GPUs

$9.86 per GPU/hr

2 weeks – 1 year
16 GPUs

NVIDIA HGX B200 — 64 GPUs

$9.36 per GPU/hr

2 weeks – 1 year
64 GPUs

NVIDIA HGX B200 — 256+ GPUs

$8.87 per GPU/hr

2 weeks – 1 year
256+ GPUs

NVIDIA H100 — 16 GPUs

$6.16 per GPU/hr

2 weeks – 1 year
16 GPUs

NVIDIA H100 — 64 GPUs

$5.85 per GPU/hr

2 weeks – 1 year
64 GPUs

NVIDIA H100 — 256 GPUs

$5.54 per GPU/hr

2 weeks – 1 year
256 GPUs

Official website