NOVA
SN68Uses AI to accelerate drug discovery, finding new medicines faster
A drug discovery subnet where miners compete to find new medicine. Instead of pharmaceutical companies spending billions behind closed doors, NOVA turns molecular search into an open competition across a chemical space of 65 billion+ compounds.
// Mining for medicine.
NOVA is an AI-powered drug discovery network. Miners compete to identify molecules that could become new medicines by searching through an enormous library of chemical compounds. Validators verify the quality of each candidate using protein-binding simulations, ensuring only genuinely promising molecules are rewarded.
The simple version: Imagine a global treasure hunt where the treasure is a molecule that could cure a disease. Thousands of participants search independently, and the ones who find the most promising candidates get rewarded. The best discoveries are shared openly.
Centralized equivalent: Think Schrödinger (computational drug design) or Recursion Pharmaceuticals (AI drug discovery), but the search is distributed across independent miners competing openly rather than locked inside a corporate lab.
How it works:
- Miners run two competitions. In NOVA Compound, they submit candidate molecules optimized for binding to target proteins while avoiding anti-targets. In NOVA Blueprint, they submit algorithms (code) designed to search ultra-large chemical spaces, making NOVA a source of reusable molecular search engines.
- Validators re-score submitted molecules using Boltz-2 for structural prediction and PSICHIC for binding affinity. They enforce chemical diversity through entropy thresholds, reject duplicates, and verify that molecules meet physical property requirements.
- The problem it solves: Drug discovery takes 10+ years and costs $2.6 billion per approved drug, with a failure rate above 90%. The bottleneck is searching a vast chemical space efficiently while current R&D operates in closed silos, duplicating effort.
- The opportunity: The pharmaceutical industry spends over $200 billion annually on R&D. Even capturing a fraction of early-stage candidate identification represents a massive market.
- The Bittensor advantage: Decentralized competition means thousands of independent search strategies running simultaneously. Miners are incentivized to explore beyond known chemical scaffolds because duplicates are rejected and diversity is rewarded.
- Traction signals: Featured on This Week in Startups with Jason Calacanis. 455 commits, 8 contributors, 101MB codebase. Team led by Micaela Bazo (CEO), Pedro Penna (CSO), and Amanda Casadei (CTO). Dashboard live at metanova-labs.ai/dashboard.
Category: Healthcare and Medical AI | Centralized Competitor: Schrödinger, Recursion Pharmaceuticals, Insilico Medicine
NOVA is one of the most tangible "real-world impact" subnets in Bittensor. While most subnets optimize AI models or trade financial signals, NOVA is searching for actual drug candidates. The incentive design is clever: NOVA Compound rewards molecule discovery, while NOVA Blueprint rewards the tools used to discover them, creating a self-improving search ecosystem.
Mechanism:
The dual competition structure is the key innovation. NOVA Compound is the direct search: miners submit sets of molecules per epoch, scored on target affinity, anti-target avoidance, and chemical novelty. A time-dependent entropy bonus rewards exploration of new chemical scaffolds, preventing miners from clustering around known solutions. Any molecule found earlier in the same target-week is automatically invalidated, forcing continuous discovery.
NOVA Blueprint elevates the competition. Instead of submitting molecules, miners submit search algorithms. These algorithms are run in a standardized sandbox for 30 minutes on an RTX 4090, with randomized target proteins and chemical spaces. The output (100 molecules) is independently scored. This means NOVA doesn't just find molecules, it breeds better molecule-finding tools.
The codebase is substantial: 455 commits across 8 contributors, 101MB repository. Development activity on GitHub has been steady. The team's research background (Micaela Bazo, Pedro Penna, Amanda Casadei) brings pharmaceutical domain expertise that most crypto projects lack entirely.
Market metrics are encouraging. At 88,692 TAO market cap with 3,551 holders, NOVA has built a solid base. Root proportion of 0.172 means it's overwhelmingly organic. Net 7-day inflow of 1,128 TAO is healthy. Emission buy acceleration of 1.34x shows buying outpacing the trend. Gini of 0.660 and HHI of 0.071 indicate moderate concentration.
The critical question is whether NOVA's molecular discoveries have real pharmaceutical value. The subnet has yet to announce any molecule advancing to preclinical trials or partnership with a pharma company. The infrastructure is impressive, but drug discovery ultimately requires wet-lab validation that a decentralized network can't provide alone.
- Validation gap: Computational drug discovery is only the first step. Without wet-lab partnerships to test candidates, NOVA's outputs remain theoretical.
- Regulatory complexity: Drug development is heavily regulated. Moving from in-silico predictions to actual medicine requires navigating FDA/EMA processes that no subnet can automate.
- Niche audience: Pharmaceutical AI is a specialized market. Most crypto investors and miners may not understand the value proposition, potentially limiting growth.
- Competition from well-funded startups: Recursion, Insilico Medicine, and others have raised hundreds of millions for AI drug discovery with established pharma partnerships.
