Launched this week
TechNode -Idle GPUs you can rent and run
GPU cloud powered by idle gaming PCs — from $0.30/hr
6 followers
GPU cloud powered by idle gaming PCs — from $0.30/hr
6 followers
TechNode turns idle consumer GPUs into affordable, on-demand AI compute. Unlike traditional GPU clouds, GPU owners can keep using their PCs: workloads yield in about 1 ms when they return. Pre-warmed workers reduce startup delays, while users keep their existing PyTorch, JupyterLab, SSH, and OpenAI-compatible workflows, with per-second billing from $0.30/hour.
Hello Everyone! 👋
We built TechNode because it felt weird that GPUs are so expensive and hard to get, while so many powerful consumer GPUs are just sitting idle most of the day.
Most GPU clouds are built around datacenters. We wanted to try something different: make idle gaming PCs usable for real compute without taking the machine away from its owner.
So with TechNode, you can launch things like JupyterLab or PyTorch on consumer GPUs and pay by the second, starting at $0.30/hour.
The part we cared most about was this: if the PC owner comes back and starts using their machine, your workload pauses in about 1ms and resumes on another node. The owner stays in control, which is what makes this kind of supply possible in the first place.
We’re also trying to be pretty honest about the current limits. This is still a small private beta, the isolation is process-level only for now, and it’s not the right fit for sensitive data yet.
If this sounds interesting, we’re currently invite-only and giving 60 free minutes to invited users.
Would love feedback from anyone working on AI workloads, GPU infra, or just tired of paying too much for compute 😃