BULLISH 📈 : NVIDIA nod and 72B model drive TAO 73% rally this month
Bittensor’s TAO jumped about 73% in 30 days. The move followed a public nod from NVIDIA’s Jensen Huang to decentralized AI training.
Analyst Alex Carchidi links the rally to that recognition. He says the comment validated training LLMs outside centralized data centers and drew capital to Bittensor’s stack according to his note.
The project’s Templar subnet trained Covenant‑72B, a 72B‑parameter LLM. More than 70 contributors used off‑the‑shelf hardware. Large‑scale model training usually needs tightly controlled, costly infrastructure, so this result supports Bittensor’s decentralization thesis as discussed here.
Token design adds a bid. Carchidi highlights supply dynamics partly similar to Bitcoin, which he views as a long‑term tailwind if subnets keep shipping useful services in his analysis.
But revenue lags subsidies. Newly minted TAO sends about $52 million annualized to the top subnet. That subnet brings at most ~$2.4 million in outside revenue. Network‑wide demand revenue is roughly $3–15 million per year versus a ~$3.3 billion token market cap, which Carchidi calls a “valuation mismatch” in his view.
Price levels matter now. TAO trades near $308. Immediate resistance sits around $315 after last week’s failed hold. If $315 flips to support, the move could extend toward $580, per analyst Ali Martinez on X.
Mid‑term resistance at $378 capped the prior leg. TAO remains almost 60% below its $757 all‑time high.
