Dragon 2 is Cere's EU production cluster, built on community GPUs instead of rented cloud. We're recruiting a small founding cohort of EU GPU owners to bond a 3090/4090-class card, run a GPU/inference node, and earn real $CERE for validated work, settled on-chain every cycle. White-glove onboarding: we help you bond, attach, and connect your instance to the cluster.
// a handful of founding spots · EU region · rewards in real $CERE · we walk every provider through onboarding
The DAC attests what your node actually serves each cycle. That validated record is what you earn on, paid in real $CERE and settled on-chain. Not a faucet, not points.
This is not a drop-and-pray faucet. A core engineer walks you through bonding, attaching your node, and connecting your instance to the cluster. You are not on your own.
Dragon 2 swaps rented cloud instances for community GPUs. The inference compute that used to flow to a hyperscaler flows through your idle card instead.
You are among the first external GPU providers on a Cere production cluster. Early validated operation is standing as the provider network grows.
Registration, bonding, admission, and settlement are signed extrinsics on Cere mainnet. Verifiable in the explorer, end to end.
Real $CERE for validated GPU/inference work on the production cluster. Not yield, not an airdrop. You bond a stake and run real hardware; reward criteria are shared in writing before you commit. The EU cluster goes live shortly: you commit and provision now, the on-chain join lands at launch.
Dragon 2 is Cere's EU cluster (Western Germany, Netherlands), built around a GPU/inference core. It runs real agent and inference workloads, so your node serves real traffic from day one. The plan is to swap rented cloud instances out for community-run 3090/4090s: yours is one of them.
Have spare storage or CPU too? The cluster takes standard storage and agent nodes alongside the GPU core. Note what you can contribute in the hardware field and we take it from there.
You run a GPU node (a machine with a 3090/4090-class NVIDIA card) on the Dragon 2 EU cluster: the compute layer that serves inference and agent workloads. Register it on-chain with your wallet, bond your stake, attach to the cluster. The DAC attests what your node actually serves each cycle; attested work is what earns. Spare storage or CPU capacity can join the same node set.
The form below. Your GPU, region, bandwidth, and wallet address double as our screening: EU GPU owners with a 3090/4090-class card. This is the qualification gate.
A human reviews every application against the hardware and region requirements. We line up the providers who can actually serve, not farm a drop.
Accepted providers get a white-glove session: we walk you through provisioning the node and the bonding setup, before you commit hardware time.
Register your node key, bond your stake, attach to the cluster: register_node → bond → add_node. Each step is one signed extrinsic, and we are on the call with you.
The cluster manager validates your node on-chain (validate_node). Once live, your node serves inference and earns $CERE per cycle, settled through the DAC.
Short and screening-grade. We read every one and reply personally.
Two minutes. Your address is where your node is registered and your $CERE rewards are paid.
You will sign node transactions with this later, so start here. Install one, create an account, copy the address.
polkadot{.js} extension → Talisman → SubWallet →Sign in with email, copy your Cere address. Fine for the application; we will help you move to an extension before bonding your node.
wallet.cere.io →// address formats: Cere shows addresses starting with 6; generic Substrate tools show the same key starting with 5. Both are accepted here.
