GPU Compute
Coming soonGPU jobs at peer-network prices.
What it does
Train models, run inference, render frames — bring your container and a GPU type, and we schedule it on a peer machine with the right hardware. Common consumer GPUs (RTX 3090, 4090, A6000) are available on demand. Each job is metered per GPU-second, so you only pay while the work is actually running.
Most idle GPU capacity in the world is on personal machines. Chess Cloud lets people contribute those GPUs and lets you rent them at a fraction of cloud GPU prices. There's no commitment, no reservation, no API quota tier — if a peer with the right GPU is online, you can run.
Why Chess Cloud
- Access consumer GPUs (RTX 3090, 4090, A6000) on demand
- Billed per GPU-second — only pay while work is running
- No commitment, reservation, or API quota tiers
- Rent peer GPUs at a fraction of cloud GPU prices
Coming soon
GPU Compute is not available yet. Leave your email and we'll notify you when it launches.