Subnet (Bittensor)

A specialized sub-network within Bittensor that defines its own incentive mechanism, validator set, and task type — enabling any AI or data workload to compete for TAO emissions.

Bittensor's modular subnet architecture means that different subnets can reward different behaviors: one subnet trains text embeddings, another validates image generation, another hosts LLM inference. TAO emissions are distributed across subnets based on validator votes, creating an internal market for AI compute.