AI tokens jump on 72B-parameter model

TLDR

AI-focused tokens are rallying after a decentralized 72-billion-parameter AI model was unveiled on the Bittensor network.

1. Bittensor (TAO) jumped after its Templar subnet released the Covenant-72B decentralized model, drawing attention to AI tokens.
2. The models scale and fully decentralized training strengthen the narrative that AI work can be coordinated and monetized on-chain, boosting demand for AI infrastructure tokens.
3. The move looks overextended on technicals, so the key question is whether real usage of Covenant-72B and other subnets justifies current AI-sector premiums.

Deep Dive

1. What Happened: Covenant-72B And The Rally

Templar, a subnet on Bittensor (TAO), announced Covenant-72B, a 72-billion-parameter language model trained entirely on Bittensors decentralized network using about 1.1 trillion tokens and consumer-grade connections. The team claims the model is competitive with centralized offerings like Metas LLaMA?2?70B and used a technique called SparseLoCo so participants could optimize locally and send compressed updates on limited bandwidth. Following this, TAO spiked roughly 20% in a day, and multiple Bittensor-related subnets appeared among top daily gainers, while broader AI tokens such as Render (RENDER), DeXe (DEXE) and Artificial Superintelligence Alliance (FET) also posted strong weekly gains as the AI sector outperformed the market.

2. Why It Matters For AI Tokens

Covenant-72B is important because it shows that a top-tier model can be trained using a permissionless, token-incentivized network instead of a single data center. Any GPU holder could contribute and earn, tying model quality directly to on-chain incentives on Bittensor. This strengthens the idea that AI infrastructure tokens are not just narratives but economic coordination tools for compute, data, and models, so demand for TAO and similar AI assets rises when credible projects build on them.

3. Sustainability, Risks, And What To Watch

Analyst commentary already flags TAO as slightly overextended on the daily chart, with the possibility of a pullback after the initial spike. The broader AI sector has logged double-digit weekly gains, which often attracts momentum and leverage but can reverse quickly if new usage does not follow. Key things to monitor are: real adoption of Covenant-72B by developers and apps, whether additional Bittensor subnets ship similarly credible models, and if other AI tokens show growing on-chain activity instead of just price moves.

Conclusion

AI tokens rallied because a major decentralized model launch gave traders a clear, concrete catalyst that fits the broader AI narrative. The long-term impact depends on whether Covenant-72B and similar efforts see sustained real-world use, turning todays price momentum into durable demand for AI-focused crypto infrastructure.