# Inference capacity

> Steady inference GPUs with a city and a term. L40S, RTX PRO 6000, H100, and other serving nodes.

Inference buyers on this board want nodes that stay up in a city, for a year, under one tenant. The hourly spot price is a different product.

## What to publish

GPU model, count, whether you can swap in a named substitute, the metro, and the month you can start. Mention egress only if it is unusual. A request that says “L40S or RTX PRO 6000, Singapore, 12 months, single tenant” is complete enough to answer.

Post an offer: https://privatecompute.com/post?kind=offer&workload=inference
Post a request: https://privatecompute.com/post?kind=request&workload=inference