cloud-hero-image

The best prices for cloud GPUs.

No commitments necessary.

TRUSTED BY FORTUNE 500 COMPANIES
mit-gray-lockup
amazon-2
google-1
raytheon-1
sony-1
NOW AVAILABLE

Lambda On-Demand Cloud powered by NVIDIA H100 GPUs

NOW AVAILABLE

On-demand HGX H100 systems with 8x NVIDIA H100 SXM GPUs are now available on Lambda Cloud for only $2.59/hr/GPU. With H100 SXM you get: 

  • More flexibility for users looking for more compute power to build and fine-tune generative AI models
  • Enhanced scalability
  • High-bandwidth GPU-to-GPU communication
  • Optimal performance density
Lambda Cloud also has 1x NVIDIA H100 PCIe GPU instances at just $1.99/hr/GPU for smaller experiments.
STORAGE

High-speed filesystem for GPU instances

Create filesystems in Lambda On-Demand Cloud to persist files and data with your compute.

  • Scalable performance: Adapts to growing storage needs without compromising speed.
  • Cost-efficient: Only pay for the storage you use, optimizing budget allocation.*
  • No limitations: No ingress, no egress and no hard limit on how much you can store.

*Texas region persistent storage will remain free until the end of 2023.

Storage Rate
Shared filesystems $0.20 / GB / month
LAMBDA DEMOS

Host & share Generative AI apps

Lambda Demos makes it easy to host Gradio-powered Generative AI apps. Simply add your Github repo and host it on an A10 for $0.60/hr. Share publicly with the ML community or privately with individuals.

lambda-demos-examples-sd-t2p-im
SPEND LESS
cost-spot-image

Instant access to cloud GPUs at the best prices

Save over 73% on your cloud bill

Get the latest NVIDIA GPUs for the best prices on the market.

Pay-by-the-second billing

Only pay when your instance is running.

Simple, transparent pricing

No hidden fees like data egress or ingress. 

DISCOVER MORE

Pre-configured for machine learning. Start training in seconds

One-click Jupyter access

Quickly connect to NVIDIA GPU instances directly from your browser. 

Pre-installed with popular ML frameworks

Ubuntu, TensorFlow, PyTorch, CUDA, and cuDNN come ready to use with Lambda Stack.

TensorFlow
Jupyter
PyTorch
Keras
Caffe
Ubuntu
SCALE UP
scale-spot-image

Spin up a variety of GPU instance types, on-demand

NVIDIA GPUs

Access GPUs like NVIDIA H100, A100, RTX A6000, Quadro RTX 6000, and Tesla V100 on-demand.

Multi-GPU instances

Launch instances with 1x, 2x, 4x, or 8x GPUs. 

Automate your workflow

Programmatically spin up instances with Lambda Cloud API. 

TRANSPARENT PRICING

On-demand GPU cloud pricing

Access high power GPUs when you need them and only pay for what you use. There are no charges for egress.

GPUs VRAM per GPU vCPUs RAM Storage Price
1x NVIDIA H100 PCIe New 80 GB 26 200 GiB 1 TiB $1.99 / hr
8x NVIDIA H100 SXM New 80 GB 220 1.8 TB 24.3 TiB $20.72 / hr
1x NVIDIA A100 40 GB 30 200 GiB 512 GiB $1.10 / hr
2x NVIDIA A100 40 GB 60 400 GiB 1 TiB $2.20 / hr
4x NVIDIA A100 40 GB 120 800 GiB 1 TiB $4.40 / hr
8x NVIDIA A100 40 GB 124 1800 GiB 6 TiB $8.80 / hr
8x NVIDIA A100 80 GB 240 1800 GiB 20 TiB $12.00 / hr
1x NVIDIA RTX A6000 48 GB 14 100 GiB 200 GiB $0.80 / hr
2x NVIDIA RTX A6000 48 GB 28 200 GiB 1 TiB $1.60 / hr
4x NVIDIA RTX A6000 48 GB 56 400 GiB 1 TiB $3.20 / hr
1x NVIDIA A10 24 GB 30 200 GiB 1.4 TiB $0.60 / hr
1x NVIDIA Quadro RTX 6000 24 GB 14 46 GiB 512 GiB $0.50 / hr
8x NVIDIA Tesla V100 16 GB 92 448 GiB 5.9 TiB $4.40 / hr

Reserved Cloud Cluster pricing

The best prices and value for NVIDIA H100 clusters in the industry

  Instance type GPU GPU Memory vCPUs Storage Network Bandwidth (Gbps) Per Hour Price Term # of GPUs
Reserved 8x NVIDIA H100 H100 SXM 80 GB 200 20 TB NVMe SSD local storage minimum 3200 $1.89/H100/hour 3-years 64 - 60,000
Sprint 8x NVIDIA H100 H100 SXM 80 GB 224 27 TB NVMe SSD local storage minimum 3200 $4.85/H100/hour 3-months 248