
Overview
Registration Activity
Neuron registration trends

Neuron registration trends
The alpha price in τ, day by day, and what moved it each way.
Log change over the last 30 days −4.02%: protocol buys +0.04%, root sells −0.66%, users net −6.59%, cross-subnet +3.18%, unattributed −0.0006%.
The chart follows the alpha price in τ. Each period starts at its opening price and steps through the mechanisms below, in a fixed order, to its closing price: a step that rises pushed the price up, one that falls pushed it down, and its height is how far it moved the price. A thin line carries each close into the next period’s open; it is dashed where the two differ. The faint candle behind each period is the market’s own open, high, low and close, for context.
With many periods on screen (90 days, or hours), the steps would be too thin to read, so the chart draws the price as a line instead and colours each period by the mechanism that moved it most. Hover a period for the full step-by-step breakdown.
Not here, because they do not move the spot price: the per-block TAO injection into the pool, and burns.
All figures are log changes, not simple percentage changes, so the parts add up exactly to the total; on big moves the two differ (a 10% log fall is a 9.5% price fall).
2026-11-06 UTC, as of ≈15:47 UTC (block 9,448,028)
The day isn’t over yet, so these figures keep changing. They are in no daily total, average or rank.
Environmental forecasting powers energy markets, trading desks, and grid operators—yet it remains expensive and slow. Traditional physics-based numerical weather prediction (NWP) requires massive computing infrastructure and hours of simulation time per forecast, limiting frequency, resolution, and accessibility. Data-driven machine learning alternatives promise speed and cost savings, but they need competitive incentive structures to attract diverse modeling talent and drive continuous improvement.
A decentralized, incentive-driven network of forecasters solves this by allowing miners to compete on prediction accuracy under the same challenge rules using publicly available ERA5 data. Validators verify forecasts against ground truth using standardized metrics, rank miners transparently, and allocate emissions based on verified performance—creating a market where the best models earn sustainable rewards and weaker approaches naturally exit.
Zeus (SN18) implements this on Bittensor.
Miners precompute and commit SHA256 hashes of compressed global weather predictions to the blockchain four times daily, then reveal full forecasts for validators to score. Validators run hourly 49-hour and 15-day predictions across four ERA5 variables (2m temperature, 100m wind, surface solar radiation) on a global 0.25° grid (721 × 1440 points). Scoring applies latitude weighting and geographic scalars (Europe 1.5×, Germany 2.5×) to reflect market demand, stores rank history in SQLite, and emits weights from rolling rank averages via a background thread. On-chain commitments prevent relay mining; dynamic burn from a performance API deters weight-copying validators.
Auto-generated from the GitHub repository and may not reflect the latest changes.
Zeus (SN18) is a decentralized environmental forecasting subnet that incentivizes miners to produce hourly weather predictions for ERA5 variables (2m temperature, 100m wind components, surface solar radiation) across a global 0.25° grid for 49-hour and 15-day horizons. Validators score predictions against ground truth using RMSE and MAE, applying latitude weighting and geographic scalars (Europe 1.5x, Germany 2.5x), and set subnet weights from rolling rank averages with dynamic burn to prevent weight copying.
The subnet operates on a 6-hour challenge schedule (00:30, 06:30, 12:30, 18:30 UTC). Miners precompute and commit SHA256 hashes of compressed float16 predictions to the blockchain for 4 ERA5 variables and 2 horizons (49 and 361 hourly steps) covering the full Earth (721×1440 grid at 0.25° resolution). Validators run three phases: (1) hash commit read from Subtensor Commitments pallet; (2) query best miners for top-k long-horizon reveals 1+ hours post-commitment; (3) final scoring when ground truth is available (after ERA5 data is complete, typically 5–7 days).
Miners return base64-encoded blosc2-compressed predictions. Validators verify hashes, compute RMSE/MAE with area and regional weights, maintain per-challenge rank history in SQLite, and emit weights from rolling rank averages via a dedicated background WeightSetter thread. Dynamic burn fetched from a performance API discourages weight-copying validators.
Predictions are scored using (RMSE + MAE) / 2 after normalization by latitude-based area weights and regional scalars. Miners are ranked competitively; ties on RMSE share the same rank. Weights derive from rolling averages of per-challenge ranks, with the best-ranked miner receiving the majority of a challenge's weight and remainder distributed logarithmically.
Variable weights (2m_temperature 0.2, 100m_u_component 0.3, 100m_v_component 0.3, surface_solar_radiation 0.2) and time-window weights (short-term 0.2, long-term 0.8) are combined. Shape errors, non-finite values, missing responses, hash mismatches, and collusion result in penalties. On-chain commit-reveal prevents relay mining.
The sections above were auto-generated from the GitHub repository and may not reflect the latest changes. View repository (github.com, opens in a new tab, external site)
Pushing weather forecasts beyond state-of-the-art
Powered by Ørpheus AI
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 (github.com, opens in a new tab, external site) for the latest instructions.
GPU not required for inference; GPU recommended for training. A GPU is required for training (unless you want to wait weeks), but is not required for inference.
Not yet verified: Notes
Machine-extracted from the subnet's repository docs; may be wrong or out of date. Check the repository before buying or renting hardware. Open the repository (github.com, opens in a new tab, external site) · Extracted 28 Sep 2026
Clone the Zeus repository and navigate to the folder
git clone https://github.com/Orpheus-AI/Zeus.git && cd Zeus
Create a Conda virtual environment with Python 3.11
conda create -y -n zeus python=3.11
Activate the virtual environment
conda activate zeus
Run the setup script to install dependencies
chmod +x setup.sh && ./setup.sh
Register your hotkey on mainnet (use netuid 301 for testnet)
btcli s register --netuid 18 --wallet.name [wallet_name] --wallet.hotkey [wallet.hotkey] --subtensor.network finney
Create and configure miner.env with NETUID, SUBTENSOR_NETWORK, WALLET_NAME, WALLET_HOTKEY, AXON_PORT
cp miner.env.example miner.env && nano miner.env
Launch the miner
conda activate zeus && ./start_miner.sh
Clone the Zeus repository
git clone https://github.com/Orpheus-AI/Zeus.git && cd Zeus
Install Python 3.11 (or use system package manager for your OS)
sudo dnf install python3.11 -y
Create and activate a Python virtual environment
python3.11 -m venv zeus-venv && source zeus-venv/bin/activate
Run the setup script to install dependencies
./setup.sh
Register your validator hotkey on mainnet (use netuid 301 for testnet)
btcli s register --netuid 18 --wallet.name [wallet_name] --wallet.hotkey [wallet.hotkey] --subtensor.network finney
Create and configure validator.env with CDS_API_KEY, WALLET_NAME, WALLET_HOTKEY, AXON_PORT, PROXY_PORT, PROXY_API_KEY, optional DISCORD_WEBHOOK
cp validator.env.example validator.env && nano validator.env
Launch the validator
source zeus-venv/bin/activate && cd Zeus && ./start_validator.sh
Run a local Subtensor node to minimize outages and improve performance (see Subtensor documentation)
Ensure local Subtensor node is running
| Day (UTC) | Share |
|---|---|
| 2026-11-05 | 0.03% |
| 2026-11-04 | 0.03% |
| 2026-11-03 | 0.04% |
| 2026-11-02 | 0.03% |
| 2026-11-01 | 0.03% |
| 2026-10-31 | 0.03% |
| 2026-10-30 | 0.03% |
| 2026-10-29 | 0.03% |
| 2026-10-28 | 0.03% |
| 2026-10-27 | 0.03% |
| 2026-10-26 | 0.03% |
| 2026-10-25 | 0.04% |
| 2026-10-24 | 0.03% |
| 2026-10-23 | 0.04% |
| 2026-10-22 | 0.04% |
| 2026-10-21 | 0.03% |
| 2026-10-20 | 0.04% |
| 2026-10-19 | 0.04% |
| 2026-10-18 | 0.04% |
| 2026-10-17 | 0.04% |
| 2026-10-16 | 0.04% |
| 2026-10-15 | 0.04% |
| 2026-10-14 | 0.04% |
| 2026-10-13 | 0.04% |
| 2026-10-12 | 0.04% |
| 2026-10-11 | 0.04% |
| 2026-10-10 | 0.04% |
| 2026-10-09 | 0.04% |
| 2026-10-08 | 0.04% |
| 2026-10-07 | 0.04% |
| 2026-10-06 | 0.04% |
| 2026-10-05 | 0.04% |
| 2026-10-04 | 0.04% |
| 2026-10-03 | 0.05% |
| 2026-10-02 | 0.05% |
| 2026-10-01 | 0.05% |
| 2026-09-30 | 0.04% |
| 2026-09-29 | 0.04% |
| 2026-09-28 | 0.04% |
| 2026-09-27 | 0.03% |
| 2026-09-26 | 0.05% |
| 2026-09-25 | 0.04% |
| 2026-09-24 | 0.03% |
| 2026-09-23 | 0.04% |
| 2026-09-22 | 0.04% |
| 2026-09-21 | 0.04% |
| 2026-09-20 | 0.05% |
| 2026-09-19 | 0.04% |
| 2026-09-18 | 0.03% |
| 2026-09-17 | 0.04% |
| 2026-09-16 | 0.04% |
| 2026-09-15 | 0.03% |
| 2026-09-14 | 0.04% |
| 2026-09-13 | 0.03% |
| 2026-09-12 | 0.04% |
| 2026-09-11 | 0.03% |
| 2026-09-10 | 0.03% |
| 2026-09-09 | 0.05% |
| 2026-09-08 | 0.05% |
| 2026-09-07 | 0.04% |
| 2026-09-06 | 0.05% |
| 2026-09-05 | 0.05% |
| 2026-09-04 | 0.08% |
| 2026-09-03 | 0.06% |
| 2026-09-02 | 0.08% |
| 2026-09-01 | 0.07% |
| 2026-08-31 | 0.05% |
| 2026-08-30 | 0.03% |
| 2026-08-29 | 0.04% |
| 2026-08-28 | 0.05% |
| 2026-08-27 | 0.04% |
| 2026-08-26 | 0.03% |
| 2026-08-25 | 0.06% |
| 2026-08-24 | 0.06% |
| 2026-08-23 | 0.06% |
| 2026-08-22 | 0.07% |
| 2026-08-21 | 0.04% |
| 2026-08-20 | 0.06% |
| 2026-08-19 | 0.04% |
| 2026-08-18 | 0.10% |
| 2026-08-17 | 0.05% |
| 2026-08-16 | 0.10% |
| 2026-08-15 | 0.06% |
| 2026-08-14 | 0.03% |
| 2026-08-13 | 0.08% |
| 2026-08-12 | 0.11% |
| 2026-08-11 | 0.09% |
| 2026-08-10 | 0.12% |
| 2026-08-09 | 0.09% |
| 2026-08-08 | 0.13% |