The best prices for cloud GPUs. Period.

No negotiation or commitments necessary.

Sign in ›

Training with Lambda

Anthem
Carnegie Mellon
Los Alamos National Lab
MIT
Laptop with deep learning notebook connecting to a cloud GPU server.

Spend less

Instant access to cloud GPUs at the best prices

  • The best prices on the market, no negotiation required

    Save over 73% on compute costs with both on-demand and reserved instances available.

  • Instant access in a few clicks

    Skip the complex setup. Spin up or spin down instances when you need them.

  • Only pay for what you need

    Pay by the second for NVIDIA A100, RTX A6000, Quadro RTX 6000, and Tesla V100 GPUs. Available in 1x, 2x, 4x, and 8x instances. No commitment required.

Sign up for free
Column chart showing Lambda GPU instance at $1.10 an hour vs the competitors at $2.46 and $4.10 an hour.

Discover more

Everything you need for deep learning in the cloud

  • Instant SSH access

    Sign up in a few clicks and get root access with SSH.

  • Pre-installed with an ML Toolkit

    Comes with Lambda stack which includes Ubuntu, PyTorch, Jupyter, & Tensorflow.

  • Easily access storage

    Create and attach persistent filesystems in a few clicks.

Learn more
TensorFlow
Jupyter
PyTorch
Keras
Caffe
Ubuntu

Scale up

Iterate faster, scale effortlessly

  • Optimized for speed

    Instances support up to 100 Gbps of inter-node bandwidth to enable faster hyper-parameter optimization.

  • Easily add more compute resources, ad hoc

    On-demand instances make it easy to add GPU compute as you need it.

  • Distributed training with reserved instances

    Reserve A100 GPUs with high speed RoCE networking (up to 1600 Gbps internode bandwidth) for distributed training.

Explore reserved
Stylized UI showing the ability to spin up many GPU instances.

Transparent Pricing

The best prices on the market for everyone. No negotiation or commitment required.

Lambda offers simple and transparent pricing so you know that you are getting the best price on the market.

On-demand pricing

You get instant access to GPU servers on demand, billed by the second.

GPUs VRAM per GPU Price
1x NVIDIA A100 40 GB $1.10 / hr
2x NVIDIA A100 40 GB $2.20 / hr
4x NVIDIA A100 40 GB $4.40 / hr
8x NVIDIA A100 40 GB $8.80 / hr
1x NVIDIA RTX A6000 48 GB $0.80 / hr
2x NVIDIA RTX A6000 48 GB $1.60 / hr
4x NVIDIA RTX A6000 48 GB $3.20 / hr
1x NVIDIA Quadro RTX 6000 24 GB $0.50 / hr
8x NVIDIA Tesla V100 16 GB $4.40 / hr
Sign up for free
Reserved pricing

Reserve A100 GPUs with high speed RoCE networking (up to 1600 Gbps internode bandwidth) for distributed training.

GPUs VRAM per GPU Price
8x NVIDIA A100 Coming soon 80 GB $11.44 / hr
8x NVIDIA A100 40 GB $8.80 / hr

All options include NVLink™

Get a custom quote