Loading
Loading
Neuron registration trends
MetaHash Group's Treasury Subnet (SN73) is a decentralized alpha acquisition and settlement system within Bittensor. It operates a deterministic multi-epoch pipeline where miners bid for alpha across multiple subnets, validators conduct auctions and clear winners, and settlements are audited on-chain. The subnet functions as both a capital allocation layer and treasury mechanism for MetaHash Group's portfolio of subnets across compute, robotics, data, agents, and advertising domains.
The subnet implements a client-server architecture where validators (treasury operators) broadcast auction opportunities and miners submit bids for alpha across portfolio subnets. Miners maintain per-coldkey state directories tracking bidding limits, auto-sell configurations, and payment status. The Runtime component in miners handles auction processing with configurable timeouts (3s), while a separate Payments component processes Win notifications (1s timeout) in background tasks.
Validators coordinate a multi-master architecture with metagraph synchronization under RPC locks, deterministic epoch progression, and on-chain commitment publishing via IPFS. AsyncSubtensor instances (shared for reads, separate for payments/settlement) manage parallel chain operations. The StateStore layer persists local state and treasury whitelists across restarts, while the Strategy layer applies YAML-based weights to subnet capital allocation.
The mechanism operates in a deterministic 4-phase epoch cycle: (1) Auction & Clearing (epoch e): Validators broadcast AuctionStart, miners submit bids with subnet_id, alpha_amount, and discount_bps. Bids are valued using slippage-aware pricing and reputation caps. Winners receive Win invoices with payment windows in e+1. (2) Commitments (epoch e+1): Validator publishes auction snapshot as CID on-chain and full JSON to IPFS, maintaining strict publishing order for e−1 only. (3) Settlement (epoch e+2): Validator scans on-chain alpha transfers during payment windows, credits delivered alpha, burns underfilled amounts to UID 0, computes miner incentives based on alpha delivered. (4) Weight Setting: Incentives from epoch e are applied in epoch e+2.
Miners with higher delivered alpha and consistent performance receive higher weights. Reputation caps and budget signaling enforce long-term alignment and prevent capital overallocation to any single miner.
The sections above were auto-generated from the GitHub repository and may not reflect the latest changes.
This information was auto-extracted from the GitHub repository and may be incomplete or out of date. Always refer to the subnet's GitHub repository for the latest instructions.
Holder & total-staked metrics as of 4h ago (hourly snapshot)
Miners maintain per-coldkey state directories and connect to configured Bittensor network endpoints. Async operations require event loop availability.
Validators coordinate multi-epoch pipeline (auction, clearing, commitment, settlement) and maintain metagraph synchronization under RPC locks.
Install core dependencies including bittensor, substrate-interface, and logging utilities
pip install -r requirements.txt
Run the miner entry point with configuration loaded from bittensor_config
python3 -m metahash.scripts.run_miner
Optional: Enable debug mode for async operation tracing and fault handling(optional)
export METAHASH_DEBUG_ASYNC=1
Set bidding control limits and auto-sell parameters in miner configuration(optional)
Configure config parameters: bidding.max_total_alpha, autosell.enabled, autosell.subnet_id
Install core dependencies including bittensor, substrate-interface, and protocol dependencies
pip install -r requirements.txt
Run the validator entry point with configuration loaded from bittensor_config
python3 -m metahash.scripts.run_validator
Set up YAML-based subnet weights and capital allocation parameters via STRATEGY_PATH
Configure weights.yml with subnet allocation strategy
Optional: Enable testing mode for preview-only settlement without on-chain weight setting(optional)
export TESTING=true