Decentralized Physical Infrastructure
Create.
Compute.
Earn.
Caldor is an open, decentralized network where anyone can rent or share computing power — no cloud corporations deciding the price or holding the keys.
GPU provider beta is live — bring your idle cards to the network.
Compute, pooled from everyone
Caldor connects machines around the world into one market. Spare cycles on one side, demanding workloads on the other — rendering, simulation, AI training. The protocol matches them and settles the bill, so you never negotiate with a datacenter.
The CLD token is the unit of exchange between Providers, who offer resources, and Requestors, who consume them. That keeps the network balanced and self-sustaining — payment moves the moment the work is verified.
Get CLD tokens →Requestor side
Rent the power you need, when you need it
Spin up decentralized compute for anything from 3D render farms to large AI jobs. Integrate through the JS API or the Ray interface and pay only for the cycles you actually use — no reserved instances, no idle bills.
Start building →Provider side
Turn idle machines into income
Put the hardware you already own to work. Share spare CPU and GPU with the network and earn CLD as jobs run — your machine stays yours, the protocol just fills the gaps between what you do with it.
Contribute & earn →Community side
Help shape where compute goes next
Caldor is open-source and built in the open. Bring ideas, code, or research and help push what a permissionless compute network can do — the roadmap is decided by the people running it, not a boardroom.
Join the Discord →Stay close to the network
Protocol updates, provider payouts, and new workloads — a short note when something actually changes.
No spam. Unsubscribe whenever.
GPU provider beta opens to the public
Providers can now contribute GPU capacity alongside CPU, opening the network to heavier AI and rendering workloads. Early contributors get priority routing while the pool scales.
Read more →A research team ran protein folding on Caldor
A university lab moved part of its simulation pipeline onto the network, cutting per-run cost against their old cloud provider. We walked through how they split the workload across nodes.
Read more →