Eight indexes: the Bittensor Classified Subnet Index, a market capitalization-weighted composite of all classified subnets, and seven sector indexes corresponding to each defined Sector in the classification taxonomy. The Composite Index serves as the primary benchmark for the Bittensor subnet ecosystem.

October 27, 2025. All indexes are initialized at a base value of 1000.00 on the Index Inception Date.

Each index is calculated as follows:

Index(t) = Index(t-1) x [sum over i of w(i,t-1) x (P(i,t) / P(i,t-1))]

where i denotes each eligible constituent, w(i,t-1) is the prior-day market capitalization weight of constituent i, and P(i,t) and P(i,t-1) are the USD-denominated prices of constituent i on days t and t-1, respectively.

Market capitalization-weighted. Daily weights are computed as:

w(i,t-1) = MarketCap(i,t-1) / sum over j of MarketCap(j,t-1)

where the denominator sums prior-day market capitalizations across all eligible constituents within the applicable index. Weights are recalculated each calendar day. No constituent weight caps are applied in the current methodology version.

Continuous daily rebalancing. Constituent weights are recalculated each day using prior-day market capitalization data. There is no fixed periodic reconstitution schedule.

A newly eligible constituent receives a weight of zero (0.00) on its first day of inclusion and enters the index at full market capitalization weight beginning on its second day of inclusion. This treatment prevents index level distortion arising from first-day price anomalies.

To be eligible for inclusion, a subnet must satisfy all of the following criteria: (1) the subnet must be actively registered on the Bittensor mainnet (Finney), (2) the subnet must have a verified public description available in the Taostats subnets-infos registry (github.com/taostat/subnets-infos), and (3) the subnet must have a determinable primary function that maps to one of the seven defined Sectors. Subnets that do not satisfy criteria (2) or (3) are designated Unclassified and are excluded from all indexes.

Subnet 0 (the Root Network) is permanently excluded from all indexes. The Root Network is a governance and emission-allocation subnet and does not function as an investable AI application subnet.

Constituents are classified according to a proprietary three-tier taxonomy: Sector, Industry Group, and Industry. Classification is based on the primary economic function of the subnet, defined as the principal output or service the subnet is designed to produce. A subnet is assigned one primary classification. Where a subnet has a material secondary function, an optional secondary classification may be assigned; secondary classifications do not affect index membership or weighting.

USD-denominated price and market capitalization data are sourced from the Taostats Pool API (api.taostats.io/api/dtao/pool). Prices reflect the automated market maker (AMM) price derived from each subnet's on-chain TAO/alpha liquidity pool, converted to USD using the TAO/USD rate at time of data pull. TAO/USD reference rates used in the TAO comparison series are sourced from Kraken (api.kraken.com). Subnet description and registry data are sourced from the Taostats subnets-infos registry (github.com/taostat/subnets-infos).

Index levels are calculated on a daily basis using end-of-day price and market capitalization snapshots. The index series reflects one observation per calendar day per index.

Intelligence and Inference: Subnets whose primary output is a delivered AI model response, including language model inference, multimodal inference, image or video generation, model evaluation, and model training outputs.

Compute and Infrastructure: Subnets that provide foundational computational resources, including distributed GPU compute, model training infrastructure, and network node operation. These subnets do not produce AI outputs directly.

Data and Knowledge: Subnets focused on the collection, storage, indexing, retrieval, or distribution of structured data assets for use by downstream AI systems or subnets.

Finance and DeFi: Subnets generating financial products and services, including quantitative trading signals, price forecasting models, yield optimization strategies, and prediction market infrastructure.

Agents and Applications: Subnets deploying autonomous AI agents that execute real-world tasks end-to-end, or building end-user applications that consume inference and data subnet outputs. This sector has no direct analogue in traditional digital asset classification standards.

Culture and Gaming: Subnets generating creative AI content, including roleplay, 3D asset generation, audio and speech synthesis, and AI-native entertainment experiences.

Security and Verification: Subnets providing AI-powered security and verification services, including cybersecurity testing, AI-generated content detection, code vulnerability analysis, and private network infrastructure. This sector has no direct analogue in traditional digital asset classification standards.