The computer that
never sleeps.
ArkNet is a peer-to-peer network of GPU providers that forms a single runtime across thousands of fragmented hardware sources — with verifiable execution and market-driven pricing.
A job is dispatched with explicit constraints. The network selects a provider, executes in isolation, returns results + proof, and finalizes state.
1. Dispatch
You submit a compiled kernel artifact plus constraints (VRAM, region, timeout, strategy).
2. Route
Schedulers match your job to the optimal provider using bids, locality, and reliability scores.
3. Execute
Providers run kernels in a hardened sandbox and return output with cryptographic evidence.
Validators
Validators maintain the integrity of the ledger. They verify execution proofs submitted by providers and finalize state transitions.
- Verify execution proofs (STARKs)
- Consensus on global state
- Slash malicious providers
- Audit routing + settlement
Providers
Providers lend their silicon to the network. Any GPU owner can join and earn yield from idle capacity — from a single RTX 4090 to an H100 cluster.
- Execute WASM/PTX kernels
- Cache model weights for warm starts
- Bid into the marketplace
- Earn rewards per-second
Global footprint
A single runtime spanning regions — optimized for locality and verified correctness.
/api/metrics/summary.Join the network.
Whether you’re consuming compute or supplying it, ArkNet gives you a verifiable path from kernel to silicon.