EcoPredict is a Bittensor subnet for decentralized environmental forecasting.
EcoPredict is a Bittensor subnet that builds a competitive, decentralized marketplace for AI-powered environmental forecasting. In regions like Port Harcourt, Rivers State, Nigeria, centralized weather and pollution services often lack granularity due to sparse data, leading to devastating crop losses, flood damage, and health risks from events like oil-related contamination. Traditional forecasts miss hyper-local factors (e.g., river levels + rainfall + pollution runoff), costing communities billions annually.
EcoPredict solves this by incentivizing a global network of miners to run specialized AI models that generate accurate, location-specific predictions (rainfall, flood risk, air quality, disaster alerts). Users query the subnet (e.g., via a simple app: "Flood risk for my farm in Buguma next 7 days?"), miners compete to provide the best forecasts using diverse data sources (satellites, IoT sensors, local feeds), and validators score them objectively against real-world outcomes post-event.
Key strengths:
Incentivized Proof of Intelligence — Top-performing models earn 80% of TAO rewards per epoch via accuracy-based scoring, driving continuous improvement.
Decentralized & Resilient — No single point of failure; miners in data-scarce areas (like Africa) can contribute local insights for better relevance.
Real-World Impact — Empowers farmers, governments, and insurers with actionable intel, enhancing climate resilience in vulnerable regions.
This subnet complements existing Bittensor efforts (e.g., Zeus for general weather) by focusing on environmental + pollution/disaster use cases, attracting non-crypto users (agri, NGOs) and boosting TAO utility through practical adoption.
Progress During Hackathon (as of Feb 21, 2026):
Completed full subnet design proposal (Notion doc with incentive mechanics, miner/validator roles, business rationale, and GTM).
Recorded 5-10 minute explanation video walking through the architecture, farmer use case in Rivers State, and reward flow.
Created public introduction post on X (Twitter) to build visibility and community feedback.
Prepared 10-page pitch deck highlighting problem-solution fit, competitive edge, and Nigeria-specific impact.
Set up GitHub repository with proposal docs, diagrams, and README overview.
Currently bootstrapped (solo contributor from Port Harcourt, Nigeria). No external funding raised yet. Seeking:
Ideathon prizes ($18K+ cash pool + up to 𝞃1000 TAO discretionary funding from partners like Unsupervised Capital).
Compute credits (e.g., from Basilica AI) for Round 2 prototyping.
Potential incubation/accelerator support if selected as a top idea.
No prior grants, VC raises, or token sales. Open to partnerships for data sources (e.g., satellite/IoT providers) or early pilot funding to bootstrap miner participation.