Lambda Hyperplane

Same DGX-1 Hardware, 30% Less

Tesla V100 Server: 4-GPU / 8-GPU Options

This is required

Trusted by Top Research Groups

Apple is a customer Microsoft is a customer Amazon is a customer GE is a customer MIT is a customer Raytheon is a customer

Choose Your Server

Not seeing what you want? Customize your own

Basic

4x Tesla V100 Server

4-way NVLink
2x Xeon Silver 4116 (12 Cores, 2.1 GHz)
4x Tesla V100 GPUs (32 GB VRAM)
256 GB RAM
Fully Customizable Storage
10 Gbps Ethernet
Starting at
$ 49,555
Customize
Premium

8x Tesla V100 Server

8-way NVLink
2x Xeon Gold 6148 (20 Cores, 2.40 GHz)
8x Tesla V100 GPUs (32 GB VRAM)
512 GB RAM
Fully Customizable Storage
100 Gbps InfiniBand
Starting at
$ 109,425
Customize
Max

8x Tesla V100 Server

8-way NVLink
2x Xeon Platinum 8168 (24 Cores, 2.7 GHz)
8x Tesla V100 GPUs (32 GB VRAM)
768 GB RAM
Fully Customizable Storage
100 Gbps InfiniBand
Starting at
$ 120,987
Customize
custom

4x or 8x Tesla V100

NVLink
Any Processor
4x or 8x Tesla V100 GPUs (32 GB VRAM)
Up to 768 GB RAM
Fully Customizable Storage
100 Gbps InfiniBand
Starting at
$
Customize

Easy Upgrades with Lambda Stack

Easily upgrade TensorFlow, Keras, PyTorch, CUDA, and cuDNN with Lambda Stack. When new versions are released, simply "apt-get upgrade." Learn More