GeoSphere is a Bittensor subnet that produces a single commodity: real-time, edge-deployable geospatial intelligence for commercial B2B clients and humanitarian disaster response.
Miners compete to provide the freshest, most accurate geospatial detections — new construction, flood extent, fire perimeters, POI changes. Validators maintain curated benchmark datasets drawn from satellite APIs and authoritative ground-truth sources, evaluating miner submissions objectively through Yuma Consensus. Commercial customers (insurance, logistics, retail, construction) pay per query to access top-ranked intelligence; a portion of that revenue cross-subsidises free disaster alerts for at-risk communities.
The result is a decentralised geospatial intelligence engine in which global competition between miners continuously improves data freshness and accuracy, and real B2B demand steers development toward what industry actually needs.
Why it matters: Geospatial intelligence is critical for both commerce (site selection, logistics, competitor tracking) and humanitarian safety (forest fires, floods, hurricanes). Yet existing solutions are either too expensive (Esri: $50k+/year), too slow (annual or quarterly updates), geographically biased toward developed markets, or purely humanitarian-focused with no sustainable funding engine. GeoSphere distributes both the cost and the incentive to produce geospatial intelligence across a global miner network — delivering equivalent capability at a fraction of the cost, and using commercial revenue to fund free disaster prediction as a structural feature, not a promise.
In GeoSphere, the benchmark is the subnet's ground truth. What the subnet optimises is not abstract "data quality" but measurable performance against curated, rotating geospatial datasets that reflect real commercial and disaster conditions. Benchmark design, dataset composition, and evaluation methodology define what intelligence means within the subnet and are treated as first-class protocol components.
GeoSphere receives TAO emissions via Dynamic TAO. Stakers vote by staking into the subnet's alpha token pool; higher inflow produces higher emissions. Yuma Consensus distributes those emissions: approximately 41% to miners, 41% to validators via bonds, and 18% to the subnet owner.
Miner reward calculation: GeoSphere uses a winner-takes-most distribution — emissions concentrate toward top-performing miners, reflecting real market dynamics: B2B clients want the best available intelligence for their use case, not an average across many contributors. The exact reward curve will be calibrated during testnet.
Each evaluation epoch, validators score miners using the following composite formula:
Score = 0.40 × Freshness + 0.30 × Accuracy + 0.15 × Coverage + 0.10 × Latency + (0.05 × DisasterBonus)
Scores are normalised into a weight vector and submitted on-chain via set_weights(). Yuma Consensus distributes emissions accordingly.
Validator reward calculation: GeoSphere launches with a centralised validator operated by the subnet team — a deliberate choice recommended by established subnets including Chutes. The evaluation process is fully transparent: benchmark samples are selected using deterministic public randomness, all scores and weights are published on-chain, and independent verification is straightforward. Validator decentralisation will be phased in as the subnet matures.
Organic revenue (Phase 2+): Beyond emissions, B2B clients pay for query access in TAO or fiat converted to TAO.
This creates a second reward signal: benchmark scores reflect technical quality (emissions), while customer purchases reflect market quality (organic revenue). Over time, market demand becomes the dominant signal steering subnet development.