Bittensor's SN68 subnet is now being used to accelerate drug research and development at Metanova Labs. The move brings decentralized artificial intelligence into a field long dominated by big pharma and centralized computing.
What the subnet does
SN68 is a specialized subnet on the Bittensor network. It runs machine learning models that can analyze molecular data, predict drug interactions, and speed up the search for new treatments. Metanova Labs is the first known group to apply this decentralized system to actual drug R&D.
Decentralized AI in drug discovery could lower costs and open the process to smaller labs, universities, and even individual researchers. That would challenge the traditional pharmaceutical model, where a handful of companies control most of the computing power and data. If the approach works, it might make drug development more democratic — and a lot cheaper.
The hurdles ahead
Validation remains a big question. Before any drug candidate found by a decentralized AI can enter clinical trials, regulators will need proof that the models are reliable. No one has yet shown that a drug discovered through a subnet like SN68 can pass muster with the FDA or its equivalents. Metanova Labs hasn't said whether it's already submitted any findings for outside review.
The broader Bittensor network has faced its own challenges. A security breach earlier this year exposed vulnerabilities in the decentralized structure. If SN68 is to be trusted for drug work, it will need to prove it can keep patient data and research secrets safe.
Metanova Labs has not announced a timeline for releasing any drug candidates. The company is likely to face pressure to publish validation data — or to partner with a traditional pharma firm that can run the necessary tests. For now, the subnet is running, the models are learning, and the question of whether decentralized AI can actually deliver a new medicine remains open.
