Bittensor

Can this project keep operating and funding itself for years?

Dynamic subnet economics and the July 2026 V440 emission gate align miner block rewards with verifiable output across 128 active subnets.12

Can outsiders see and check what the team and the protocol do?

Open-source Subtensor code on GitHub pairs with public tracking of subnet registrations, emission weights, and validator stakes on TaoStats.32

Does the protocol earn real fees or income from use?

Subnet registration burn fees, commercial inference API usage, and institutional trust demand support network economics despite ongoing token emissions.42

Is the technology being built and shipped, and does it matter?

Yuma Consensus coordinates 128 AI subnets, highlighted by the 72-billion-parameter Covenant-72B model and September 2026 Chainlink CCIP deployments.53

Who uses, builds and holds it, and how durable is that base?

Active miner, validator, and AI researcher participation across 128 subnets pairs with institutional access through the Grayscale Bittensor Trust.26

Do supply, issuance and unlocks protect holders from dilution?

TAO follows a capped 21 million supply schedule with periodic halvings, but the 46% circulating ratio applies the high-dilution penalty.46

1. 1docs.bittensor.com/subnets/understanding-subnetsCited for Sustainability
2. 2taostats.io/subnetsCited for Sustainability, Transparency, Revenue, Community
3. 3github.com/opentensor/bittensorCited for Transparency, Innovation
4. 4docs.bittensor.com/emissionsCited for Revenue, Tokenomics
5. 5bittensor.com/whitepaperCited for Innovation
6. 6coingecko.com/en/coins/bittensorCited for Community, Tokenomics

Bittensor provides liquid exposure to decentralized artificial intelligence infrastructure across 128 specialized subnet markets. Instead of relying on one centralized cloud provider, the protocol uses token incentives to coordinate model training, inference, and validation.

• Subnet architecture. The Subtensor blockchain coordinates 128 active subnets covering pre-training, inference, fine-tuning, and synthetic data generation.
• Distributed pre-training. Covenant-72B trained a 72-billion-parameter model across 1.1 trillion tokens using decentralized contributors.
• Emission gating. The July 2026 V440 upgrade directs TAO block rewards toward subnets with verified output utility.
• Cross-chain access. September 2026 Chainlink CCIP integrations brought TAO liquidity to Base and Robinhood Chain.
• Institutional wrapper. The Grayscale Bittensor Trust gives traditional allocators regulated secondary exposure to the network.

• Supply dilution. About 9.7 million TAO circulate out of the 21 million cap, leaving 54% of supply for future emissions.
• Validator concentration. High subnet registration costs and stake weight concentrate governance influence among large validator pools.
• Centralized competition. Well-funded frontier AI labs subsidize inference pricing and challenge open-source subnet adoption.
• Security incidents. Third-party router vulnerabilities and wallet exploits in 2024 and 2026 highlight operational risks around ecosystem tooling.

Bittensor leads the decentralized AI sector by operating a multi-subnet incentive protocol rather than a single compute marketplace. Unlike GPU rental networks such as Render or Akash, Subtensor uses Yuma Consensus to score model output and data quality.

Bittensor leads decentralized AI infrastructure with 128 active subnets, the Covenant-72B pre-training milestone, and the July 2026 V440 emission gate upgrade. Cross-chain deployments on Base and Robinhood Chain via Chainlink CCIP in September 2026 broadened token access alongside the Grayscale Bittensor Trust.