Bittensor’s TAO is up 90% so far this month, and the tokens in its ecosystem are rising even more.
The network’s subnet token category reached a combined market capitalization of $1.47 billion on Monday, with a 24-hour trading volume of $118 million, according to CoinGecko data.
This increase follows TAO’s own run from $180 to over $332 in March, but subnet tokens are where the real action is happening. Templar, the subnet 3 token, gained 444% in 30 days. OMEGA Laboratories grew by 440%. Level 114 added 280%. BitQuant gained 230%. Even the largest subnet tokens saw significant returns, with Chutes up 54% and Targon up 166%.
Bittensor is a decentralized network that creates markets for artificial intelligence. Instead of a single company creating and controlling AI models, Bittensor encourages a global network of participants to contribute computing power, data, and machine learning models in exchange for TAO, the network’s native token.
The network is divided into specialized subnetworks called subnets, each focused on a different AI task, from training language models to running computational infrastructure to analyzing cybersecurity. There are currently 128 active subnets, each with its own token whose value is directly linked to the amount of TAO staked on it.
Several catalysts contributed to these movements of tokens in the Bittensor ecosystem.
Subnet 3 produced Covenant-72B, a large language model trained without permission on Bittensor’s decentralized network by over 70 contributors using standard Internet hardware.
The model was trained on 1.1 trillion tokens and achieved an MMLU score of 67.1, confirmed in a March 2026 arXiv paper. This puts it in a competitive range with Meta’s Llama 2 70B, a model built by one of the most well-resourced AI labs in the world. (MMLU, or Massive Multitask Language Understanding, is a standardized test for AI models that scores them in 57 academic subjects.)
Subnet 3, called Templar, is Bittensor’s decentralized AI training network. Miners contribute computing power to the GPU and compete to produce useful training gradients for large language models, while validators evaluate the quality of their contributions and distribute TAO rewards accordingly.
Think of it as a way to train AI models in the same way that Bitcoin mining blocks do, with globally distributed participants contributing hardware and getting paid for useful work.
Elsewhere, Nvidia CEO Jensen Huang and investor Chamath Palihapitiya endorsed Bittensor’s approach on the All-In podcast on March 20, describing decentralized AI training as complementary to proprietary models. Coming from the CEO whose blog post earlier this month briefly helped reverse a sell-off in tech stocks, the endorsement carried weight beyond the usual crypto echo chamber.
How subnet tokens work
The mechanics of subnet tokens explain why the gains are so outsized compared to the TAO itself.
Since Bittensor launched dynamic TAO in February 2025, each subnet operates its own automated market maker with a native token whose valuation is determined by the TAO staked in that subnet’s reserves. When TAO appreciates, each subnet’s reserve becomes more valuable, inflating token prices and attracting more stakers. The relationship is reflexive and amplifies movements in both directions.
With a market cap of around $3 billion and individual subnet tokens ranging from $1 million to $137 million, subnet tokens function as leveraged bets on the parent protocol.
The network plans to expand from 128 to 256 active subnets later this year, which would lead to a new wave of token launches.
A possible regulatory move to convert the Grayscale TAO Trust into a spot ETF could provide institutional access by the end of 2026. And Yuma, a subsidiary of the Digital Currency Group, already contributes to 14 different subnets, suggesting that smart money is treating this as infrastructure rather than speculation.
The sustainability of the subnet’s rally depends on whether Bittensor continues to produce competitive AI models or whether the Covenant-72B was a one-off that lucked out with Huang’s approval.
