Generating 9M+ images in 24 hours for just $1872, check out the Stable Diffusion inference benchmark!

Where Kubernetes Meets 10K+ Distributed GPUs

With the SaladCloud Virtual Kubelet Provider
Easily deploy K8s pods as container group deployments on the world’s most affordable cloud.

Salad and virtual kubelet

Zero Friction. Maximum Savings.

less cloud cost
Worldwide gpus
starting hourly cost
workflow changes

Why Kubernetes on Salad?

Don’t have enough GPUs in your data center? Overpaying for cloud? With  support for Kubernetes through the Virtual Kubelets (K8s) project, SaladCloud lets you access 10K+ at massive scale and the lowest unit cost in the market.

Pricing Calculator
SaladCloud Portal - GPU pricing
Green dots

Deploy without workflow changes

Schedule deployments directly with SaladCloud API without any changes to development workflow.
See Pricing
Right arrow
No deployment changes
GPU Cloud for Kubernetes

Perfect for GPU-powered applications at scale

With 10K+ GPUs around the world, leave your scaling worries behind and run on SaladCloud easily.
Estimate my Cost
Right arrow

Extend your K8s deployment easily

Deploy the Virtual Kubelet. Deploy to SaladCloud. Start accessing 1000s of GPUs at the lowest price.
Learn More
Right arrow
Easy K8s deployment

How it works

Engineering teams send their apps as a deployment to their existing K8s deployment. On deployment, K8s sends all the pods to SaladCloud Virtual Server in your Kubernetes cluster.

Salad receives the requests, calls the SaladCloud API and starts running your apps on Salad’s  distributed cloud.

SaladCloud Virtual Kubelets architecture