What is Bittensor? A Peer-to-Peer Intelligence Market
Bittensor is a decentralized protocol that creates an open, permissionless marketplace for artificial intelligence. Founded by Yuma Rao, it incentivizes a global network of machines to produce AI capabilities — from LLM inference and model training to GPU compute and data collection — using its native token TAO. In the post-halving dTAO era (March 2026), each subnet operates with its own alpha token, and emissions are distributed based on market-driven subnet valuations.
This dashboard tracks 41 active subnets across 10 categories, with a combined tracked emission of 67.9%. 3 subnets have active buyback mechanisms, and 12 subnets generate revenue through various models.
Bittensor Network Statistics — March 2026
- Active Subnets
- 41 / 128
- Tracked Emission
- 67.9%
- Active Buybacks
- 3
- Revenue Models
- 12
- Categories
- 10
- Top Emitter
- SN64 Chutes (14.39%)
Complete Bittensor Subnet Directory
All 41 tracked Bittensor subnets sorted by emission allocation. Data sourced from subnetalpha.ai, taostats.io, and Grayscale research — March 2026.
| ID | Name | Owner | Category | Description | Emission | Revenue | Tier |
|---|---|---|---|---|---|---|---|
| SN64 | Chutes | Rayon Labs | Compute | Serverless AI compute — #1 on OpenRouter — auto-stake buyback | 14.39% | Buyback (Buyback) | APEX |
| SN3 | Templar | Covenant AI | Training | Distributed AI training — 70B+ parameter models | 5.62% | — | APEX |
| SN51 | Lium | Datura | Compute | GPU rental platform — ex Celium — 4.44% emission | 4.44% | Fee Share | APEX |
| SN1 | Apex | Macrocosmos | Inference | LLM inference via agentic workflows at scale | 4.2% | — | APEX |
| SN4 | Targon | Manifold Labs | Inference | Deterministic verification & multimodal processing | 3.8% | Fee Share | APEX |
| SN8 | Vanta | Taoshi | DeFi | Prop trading — crowdsourced AI-driven strategies | 3.45% | Fee Share | APEX |
| SN19 | Nineteen | Rayon Labs | Inference | Ultra-low-latency inference — revenue → buybacks | 2.71% | Buyback (Buyback) | APEX |
| SN9 | IOTA | Macrocosmos | Training | Pre-training LLMs with massive datasets | 2.13% | — | APEX |
| SN62 | Ridges AI | Ridges AI | Inference | Coding agents — outperforms Claude 4 on benchmarks | 2.1% | Fee Share | FRONTIER |
| SN6 | Numinous | Numinous Labs | Prediction | Superhuman LLM-based forecasters via agent evolution | 1.8% | — | FRONTIER |
| SN56 | Gradients | Rayon Labs | Training | Simplified AI training — biotech & life sciences | 1.8% | — | FRONTIER |
| SN13 | Data Universe | Macrocosmos | Data | Large-scale decentralized data collection | 1.6% | — | FRONTIER |
| SN5 | Hone | Manifold & Latent | Training | Hierarchical AI training toward AGI reasoning | 1.5% | — | FRONTIER |
| SN21 | Any-to-Any | Omega Labs | Inference | Multimodal AI — text, image, audio generation | 1.3% | — | FRONTIER |
| SN12 | ComputeHorde | Backend Devs | Compute | Trusted GPU compute — on-demand AI validation | 1.2% | Fee Share | GROWTH |
| SN2 | DSperse | Inference Labs | Security | Zero-knowledge proving for verifiable AI inference | 1.1% | — | FRONTIER |
| SN25 | Mainframe | Macrocosmos | Science | Decentralized protein folding research | 1.1% | — | FRONTIER |
| SN68 | NOVA | MetaNova Labs | Science | Drug discovery via decentralized AI models | 1% | — | FRONTIER |
| SN7 | SubVortex | SubVortex | Infrastructure | Distributed subtensor nodes & intelligent routing | 0.9% | — | GROWTH |
| SN14 | TAO Hash | Latent Holdings | Compute | Bitcoin hash rate → alpha tokens via PoW | 0.8% | — | GROWTH |
| SN67 | Tenex | Tenexium | DeFi | Long-only margin trading — automated fee buybacks | 0.8% | Buyback (Buyback) | GROWTH |
| SN120 | Affine | Affine | Training | RL platform bridging subnets — improves AI models | 0.8% | — | FRONTIER |
| SN10 | Swap | TaoFi | DeFi | Cross-chain DEX — Base/ETH into Bittensor tokens | 0.7% | Fee Share | GROWTH |
| SN22 | Desearch | Datura AI | Data | Decentralized Twitter/Reddit/Google analysis | 0.7% | — | GROWTH |
| SN44 | Score | Score Tech | Prediction | Football prediction — 70% avg accuracy via DKING AI | 0.7% | Fee Share | GROWTH |
| SN15 | BitQuant | OpenGradient | DeFi | AI agents for real-time crypto market analysis | 0.6% | — | GROWTH |
| SN34 | BitMind | BitMind | Security | Deepfake & AI image detection for media integrity | 0.6% | — | GROWTH |
| SN42 | Gopher | Gopher Lab | Data | Real-time data scraping into AI-ready datasets | 0.6% | — | GROWTH |
| SN123 | MANTIS | Barbarian | Prediction | Financial forecasting — high-quality alpha predictions | 0.6% | Fee Share | FRONTIER |
| SN16 | BitAds | BitAds | Other | Pay-per-sale decentralized advertising | 0.5% | Fee Share | GROWTH |
| SN32 | It's AI | It's AI | Security | AI-generated text detection & verification | 0.5% | — | GROWTH |
| SN37 | Aurelius | Aurelius Labs | Security | AI alignment — adversarial safety datasets | 0.5% | — | GROWTH |
| SN50 | Synth | Mode Network | DeFi | Synthetic price data — full probability distributions | 0.5% | — | GROWTH |
| SN60 | Bitsec | Bitsec | Security | AI-driven vulnerability detection in smart contracts | 0.5% | — | GROWTH |
| SN70 | Vericore | dFusion AI | Security | Large-scale fact-checking & semantic verification | 0.5% | — | GROWTH |
| SN20 | Bounty Hunter | Team Rizzo | Other | Open bounty competitions for AI benchmarks | 0.4% | — | GROWTH |
| SN52 | Dojo | Tensorplex | Data | Community-driven multimodal training data | 0.4% | — | GROWTH |
| SN61 | RedTeam | Innerworks | Security | Competitive cybersecurity — bot detection | 0.4% | — | GROWTH |
| SN75 | Hippius | The Nerve Lab | Data | Cloud storage — 1st subnet token listed on CEX (MEXC) | 0.4% | — | GROWTH |
| SN96 | FLOCK OFF | FLock.io | Data | Compact datasets for edge AI training | 0.3% | — | GROWTH |
| SN0 | Root | Opentensor | Infrastructure | Governance & emissions distribution | — | — | APEX |
Frequently Asked Questions About Bittensor Subnets
What are Bittensor subnets?
Bittensor subnets are specialized sub-networks within the Bittensor decentralized AI protocol. Each subnet focuses on a specific AI task — such as LLM inference, model training, GPU compute, data collection, or DeFi. As of March 2026, there are 41 tracked active subnets across 10 categories, with a maximum network capacity of 128 subnets.
What is dTAO and how does it affect subnet emissions?
dTAO (dynamic TAO) is Bittensor's post-halving emission system where each subnet has its own alpha token. Emissions are distributed based on market-driven subnet valuations rather than fixed allocations. Currently, the top subnets account for 67.9% of tracked emissions. The highest-emitting subnet is SN64 Chutes (14.39%), operated by Rayon Labs.
Which Bittensor subnets have buyback mechanisms?
Currently 3 subnets have active buyback mechanisms where revenue is used to buy back their alpha tokens: SN19 Nineteen, SN64 Chutes, SN67 Tenex. An additional 9 subnets have other revenue models including Fee Share and Auto-Stake.
What are the top Bittensor subnets by emission?
The top 10 Bittensor subnets by emission allocation (March 2026):
- SN64 Chutes — 14.39% (Compute, by Rayon Labs) — Active buyback
- SN3 Templar — 5.62% (Training, by Covenant AI)
- SN51 Lium — 4.44% (Compute, by Datura)
- SN1 Apex — 4.2% (Inference, by Macrocosmos)
- SN4 Targon — 3.8% (Inference, by Manifold Labs)
- SN8 Vanta — 3.45% (DeFi, by Taoshi)
- SN19 Nineteen — 2.71% (Inference, by Rayon Labs) — Active buyback
- SN9 IOTA — 2.13% (Training, by Macrocosmos)
- SN62 Ridges AI — 2.1% (Inference, by Ridges AI)
- SN6 Numinous — 1.8% (Prediction, by Numinous Labs)
What categories of AI tasks do Bittensor subnets cover?
Bittensor subnets span 10 categories: Infrastructure, Inference, Security, Training, Prediction, DeFi, Compute, Data, Other, Science. The most populated categories are Inference (LLM and multimodal AI), Security (deepfake detection, smart contract auditing, AI alignment), Training (distributed model training), and Data (decentralized data collection and storage).