Overview

Contribute GPU resources to the Runpod platform as a hosting partner.

Runpod partners with data centers and compute providers to expand GPU availability on the platform. As a host, you contribute GPU servers that Runpod users can rent for Pods and Serverless workloads.

Hosting tiers

Runpod offers two hosting tiers with different requirements and use cases.

Secure Cloud

Secure Cloud is Runpod’s enterprise hosting tier for established data centers. Partners must meet strict hardware, software, network, and compliance requirements.

Partner requirements

Hardware, software, and compliance specifications.

Requirements overview:

• Minimum 100kW GPU server capacity.
• NVIDIA GPUs (Ampere generation or newer).
• Tier III+ data center with redundant power and network.
• SOC 2, ISO 27001, or PCI DSS compliance.
• 99.99% uptime target.

To become a Secure Cloud partner, contact [e-mail].

Community Cloud

Community Cloud allows individual compute providers to contribute GPU resources. This program is closed to new hosts, but existing hosts can continue operating on the platform.

If you’re an existing Community Cloud host, contact [e-mail] for support.

Host operations

Burn testing

Test hardware stability before listing.

Maintenance and reliability

Schedule maintenance and meet uptime requirements.

Looking for GPU compute?

If you want to use GPUs on Runpod (not host them):

Choose a Pod

Find the right GPU for your workload.

Serverless overview

Run workloads without managing infrastructure.