Bittensor's TAO token delivered the most impressive performance among crypto assets in March 2026, surging over 100% in a single month, from approximately $180 in early March to above $350. The rally was catalyzed by several converging events: the completion of the largest decentralized language model Covenant-72B, an endorsement from NVIDIA CEO Jensen Huang, and Grayscale's spot ETF filing.
Covenant-72B - the Largest Decentralized Model in History
Bittensor's Templar subnet (SN3) completed the training of Covenant-72B on March 10 - a 72-billion-parameter language model trained by over 70 independent contributors from around the world. All participants used standard consumer hardware and regular internet connections, without access to corporate data centers.
The model processed 1.1 trillion tokens and achieved a score of 67.1 on the MMLU benchmark (zero-shot), outperforming Meta's LLaMA-2-70B and LLM360 K2. The results were confirmed in a March 2026 arXiv paper, making Covenant-72B the largest decentralized LLM pre-training run on record. All weights and checkpoints have been published under the Apache license.
NVIDIA CEO Compares Bittensor to Folding@home
On March 20, the TAO token jumped 17% after NVIDIA CEO Jensen Huang discussed the project on the All-In Podcast. Investor Chamath Palihapitiya raised the topic of decentralized AI training, and Huang called Bittensor "a modern version of Folding@home," endorsing the idea that open-source and proprietary AI models are complementary.
This endorsement from one of the tech industry's most influential leaders significantly boosted confidence in the project. Unlike Bitcoin, which traded in a narrow range around $71,000 throughout March, TAO delivered explosive growth of over 100% in a single month.
Subnet Market Reaches $1.5 Billion
TAO's rise triggered a cascading effect across the Bittensor environment. The combined market capitalization of subnet tokens grew from $550 million to $1.5 billion, with nearly every token posting double- or triple-digit gains over 30 days. TAO staking volume on subnets surged by 833,000%.
- Templar (SN3): the subnet responsible for training Covenant-72B became the sphere's flagship
- Lightning Protocol: a new Rust-based communication system between validators and miners launched on March 4
- Scaling: the 2026 roadmap includes doubling active subnet capacity from 128 to 256
Grayscale Files for Spot TAO ETF
Institutional interest in Bittensor is confirmed by Grayscale Investments' filing for a spot ETF under the ticker GTAO on NYSE Arca. The company submitted an S-1 registration form to the SEC on December 30, 2025, to convert its existing over-the-counter Grayscale Bittensor Trust into a full exchange-traded fund.
The SEC's decision is expected within 240 days, approximately by the end of 2026. If approved, GTAO would become the first spot TAO ETF in the United States, opening regulated access for institutional investors to the decentralized AI token, similar to how spot Ethereum ETFs expanded access to the second-largest cryptocurrency by market cap.
Outlook for the Market
Bittensor's success demonstrates growing demand at the intersection of two of the hottest technological narratives, artificial intelligence and blockchain. If a 72-billion-parameter model can be trained without corporate infrastructure, it challenges the monopoly of big tech companies over AI development.
However, investors should consider the risks: a rapid 100%+ surge in a single month is often followed by corrections, and the SEC's ETF decision remains uncertain. At the same time, deep factors, from technological achievements to institutional interest, suggest that Bittensor has established itself among the leading AI crypto projects.
Comments
Your email address will not be published. Required fields are marked *