GPU provider program
Rent out your GPU.
Earn per task.
Install the Provider Suite and the OpenGPU Network sends paid AI workloads to your GPU. Join 325+ providers in 40+ locations. No bidding, no DevOps, no upfront commitment.
Anyone can provide
Who can rent out GPUs.
GPUs, CPUs and AI accelerators all have a place. The routing layer detects your hardware class, memory and stability, and sends you the workloads it is best at, from short tasks to hourly instance rentals.
Datacenters
Fill unused capacity in enterprise fleets with routed demand.
Cloud providers
Monetize spare inventory without building a sales channel.
GPU farms
Put multi-GPU rigs to work on real AI jobs.
Home & hobby rigs
A single GPU at home can earn. No technical knowledge needed.
Edge devices
Compact devices and mini-servers run lighter jobs.
Labs & universities
Monetize research GPUs between projects and semesters.
How providing works
How to earn with your GPU.
- 01
Install and register
Install the Provider Suite and register your hardware. During setup you link the master and provider accounts used for payouts.
- 02
Become a provider
Once registered, your GPU is available to the network. No upfront commitment is required.
- 03
Match with sources
Clients publish sources and routing rules such as hardware type, performance or priority. Your GPU receives tasks whenever it matches a source's criteria.
- 04
Earn in real time
Tasks are executed in isolated containers and earnings stream to your payout account. Monitor everything in the Management App.
Provider Suite
Download the Provider Suite.
Native apps for macOS and Windows. A Docker Compose stack for Linux servers and fleets. Always the latest build.
- Automatic GPU detection and benchmarking
- Isolated, encrypted execution environments
- Graceful shutdown, pause anytime
Platform guides: Provider Suite docs
Linux & Docker
Linux GPU server setup with Docker.
The Provider Suite runs as a Docker Compose stack, built for enterprise deployments with modular services and graceful shutdown.
- Operating system
- Linux, macOS or Windows host with Docker
- Docker
- 20.10.0+ with Compose v2
- Graphics
- NVIDIA GPU with CUDA (CPU-only mode available)
- Memory
- 8 GB minimum, 16 GB recommended
- Storage
- 10 GB for Docker images
- Network
- Broadband connection
1 ยท Download the compose filebash
$ curl -o docker-compose.yml \ https://raw.githubusercontent.com/OpenGPU-Network/provider-suite/refs/heads/main/docker-compose.yml2 ยท Set your keys in .env.env
PROVIDER_PRIVATE_KEY=your_provider_private_key MASTER_ADDRESS=your_master_address
3 ยท Launchbash
$ docker compose up -d $ docker compose logs -f # follow logs
By installing you agree to the Provider Software Agreement. Never share your PROVIDER_PRIVATE_KEY. Use a secret manager in production. For GPU support, install the NVIDIA container runtime.
Track earnings
Earnings per task, per GPU and over time.
Manage sources
Register to the sources that fit your hardware.
Auto or manual
Enable automatic routing or pick jobs yourself.
GPU health
Monitor uptime, utilization and reputation, and scale your fleet.
Management App
Track GPU earnings in one dashboard.
Monitor GPUs, register sources, track earnings and see your provider reputation. From one GPU to hundreds.
FAQ
Provider questions.
What hardware do I need?
An NVIDIA GPU with CUDA on Windows or Linux, or an Apple Silicon Mac, plus a broadband connection. Linux servers need Docker 20.10+ and 8 GB RAM minimum (16 GB recommended). See the Provider Suite docs for the full list.
Do I need to commit anything upfront?
No. Register your hardware and start receiving tasks. You can pause or stop providing at any time.
What happens if my node goes offline during a task?
The network detects it and shifts the job to the next best provider, resuming from the last checkpoint. Reliability feeds into your provider reputation, which affects future routing.
How much can I earn?
Earnings depend on your GPU class, uptime, reputation and network demand. You are paid for every completed task, and you can follow earnings live in the Management App.
Is it safe to run other people's workloads?
Workloads run in isolated, hardened containers with no access to your files.
Start earning with your GPU.
Install the Provider Suite and start receiving tasks today.