Bittensor's native token TAO climbed 19.19% in 24 hours, reaching $284.75, driven by renewed interest in decentralized artificial intelligence and a viral moment from one of its most active subnets. The rally highlights how subnet token demand directly influences TAO's price, since investors must hold TAO to access individual subnet tokens within the ecosystem.
The momentum was largely sparked by Templar, operating on Subnet 3, after a single viral social media post triggered a wave of buying activity. Community observers noted that if multiple subnets generated simultaneous buzz, the compounding effect on TAO demand could be even more significant. Three subnets — Templar (SN3), Targon (SN4), and Basilica (SN39) — landed among CoinGecko's top eight daily gainers during the move, pushing TAO above the $280 mark.
The excitement surrounding Templar was backed by a real technical breakthrough. On March 10, the team announced Covenant-72B, a 72-billion-parameter large language model pre-trained entirely on Bittensor's decentralized network. Trained on approximately 1.1 trillion tokens using standard consumer internet connections, the model required no centralized computing cluster or restricted access. Any GPU holder could freely participate in training.
To tackle bandwidth constraints at that scale, Templar deployed a method called SparseLoCo, allowing participants to run local optimization steps before compressing and transmitting updates. The result is a model the team says competes with centralized alternatives like Meta's LLaMA-2-70B.
Crypto trader Michael van de Poppe took partial profits, selling 10.42 TAO at $288 and rotating proceeds into SEI and EIGEN, while keeping roughly half his portfolio in TAO and NEAR. He flagged the move as slightly overextended on the daily timeframe and suggested a short-term pullback could create a better re-entry opportunity. Van de Poppe remains bullish on the broader AI and crypto narrative as a key market theme going forward.
Comment 0