Bitcoin’s compute power dwarfs top 100 supercomputers by 600k times, says Bittensor co-founder
Global computing is shifting from isolated corporate data centers to open, decentralized networks. At Proof of Talk in Paris, Bittensor co-founder Ala Shaabana argued that Bitcoin shows how distributed incentives can create computing power far beyond traditional supercomputers, with the network’s hash rate exceeding them by more than 600,000 times. Bittensor applies a similar model to AI. It is a Layer 1 protocol with Bitcoin-like tokenomics: a 21 million cap, scheduled halvings, no pre-mine, and no venture backing. Instead of mining hashes, participants run and validate AI across 128 specialized subnets. Each subnet sets its own task, and miners earn TAO by optimizing for that goal. The core claim is that incentive design determines network intelligence: reward speed, storage, or other outputs, and the network will evolve accordingly. This model is presented as a way to source global compute and intelligence without centralized tech control.
