Bitroot Nodes

The Bitroot network runs on two complementary node roles: validator nodes, which handle consensus and network security, and compute nodes, which supply the GPU power behind the distributed AI training and inference network. This page covers what each role does, how to participate, and the mechanics worth understanding before you commit hardware.

1. Two roles, two jobs

• Validator nodes take part in consensus, block production, and transaction verification — the foundation of the network's security and decentralization. The main requirement is stable uptime and reliable connectivity.
• Compute nodes contribute GPU capacity to the distributed AI compute network, taking on training and inference workloads. The main requirement is hardware that matches the tasks you want to serve.

You can run either role on its own, or both at once, depending on your hardware and how you want to participate.

2. Staking and slashing (conceptually)

Both roles stake BRT as collateral for participating in the network. The point of staking is to make misbehavior cost clearly more than it could ever gain:

• Staked tokens stay locked while a node participates normally.
• If a node misbehaves — a validator double-signing, a compute node returning bad results or refusing service — some or all of its stake is slashed, scaled to the severity of the violation.
• Nodes that participate honestly and reliably build reputation over time, which affects how often they're selected for tasks and what they earn.

Exact staking minimums, lock-up periods, and reward formulas evolve as the network's software evolves, so this page deliberately doesn't hardcode numbers that could go stale. Check the protocol repository (linked below) and official announcements for current parameters, rather than planning around a number that may no longer be accurate.

3. Running a validator node

Validator nodes take part in block production and consensus voting — the direct source of network security. Running one typically requires:

• Stable connectivity and a high uptime rate (frequent disconnects hurt your reputation score).
• Hardware (CPU/memory/bandwidth) that meets the current version's requirements — see the protocol repository's deployment docs, since requirements shift as network throughput scales up.
• Completing the staking process to join the active validator set.

For deployment steps, sample configuration, and startup commands, see the protocol repository: https://github.com/brt-chain/bitroot.

4. Running a compute node

Compute nodes contribute GPU capacity to the distributed AI training and inference network, earning rewards based on compute contributed, task quality, and accumulated reputation. Running one typically requires:

• GPU hardware suited to the task tier you want to serve — home GPUs, enterprise-grade hardware, and idle mining rigs can all participate at their respective capacity levels.
• Stable connectivity for receiving task assignments and uploading results.
• Completing the staking process to join the compute node pool.

Task assignment uses a verifiable random function (VRF) rather than manual allocation, so the selection process can't be gamed. For critical tasks, the network also distributes the same work to multiple independent nodes and checks for agreement before accepting a result, adding another layer of trust.

5. Where rewards come from

Rewards scale with a task's computational complexity, how quickly and accurately a node completes it, and the node's accumulated reputation — nodes that participate consistently and honestly tend to get better task allocation and higher reputation weight over time. See the protocol documentation for the current reward formula and payout schedule.

6. FAQ

Can an ordinary user's idle GPU participate? Yes — letting idle home, enterprise, and mining-rig GPUs participate and earn at whatever capacity tier they fit is one of the explicit design goals of the compute node network.

Can the same machine run both a validator and a compute node? Technically yes, but the two roles have different hardware and uptime requirements, so it's worth evaluating which role your hardware is actually suited for first.

Can staked tokens get slashed arbitrarily? No. Slashing only applies to verifiable misbehavior (returning incorrect results, double-signing); normal participation and occasional network hiccups don't trigger a penalty.

7. Resources

If you run into something this page doesn't cover while deploying a node, reach out via the Media Contact page or our community channels.