In March 2026, Bittensor (TAO) nearly doubled in price, becoming the main driver behind the GMCI AI Index rising 48% since early February. Despite the index still sitting 84% below its Q1 2024 all-time high, the rally highlights renewed market interest in decentralized AI infrastructure.
The GMCI AI Index remains heavily concentrated. TAO (24.89% weight), Render (RNDR), and Artificial Superintelligence Alliance (ASI) together account for over 71% of the index. This means the recent surge was largely powered by TAO’s performance rather than broad AI token sentiment.
The catalyst was the announcement from Bittensor subnet 3 (Covenant) of the Covenant-72B model — a 72-billion-parameter language model trained permissionlessly across more than 70 globally distributed nodes.
The model achieved a 67.1 score on the MMLU benchmark, placing it in direct competition with Meta’s Llama 2 70B. This marked the first credible demonstration that decentralized, distributed training can produce competitive large language models, countering long-standing skepticism about speed and coordination challenges in distributed systems.
The subnet’s native token, τemplar (SN3), exploded over 400% in the past month, reaching a market cap of approximately $130 million.
The breakthrough drew public recognition from notable figures:
Their comments added significant narrative momentum to the Bittensor ecosystem and TAO specifically.
Beyond Covenant-72B, the subnet ecosystem continues to mature. Targon (SN4), a decentralized GPU compute marketplace run by Manifold Labs, recently secured a six-figure deal to power the backend infrastructure for Dippy AI, which serves 8.6 million users.
At the time of writing:
The Covenant-72B success represents a pivotal moment for decentralized AI. It shifts the narrative from theoretical potential to demonstrated capability, potentially opening the door for broader institutional and developer interest in Bittensor’s subnet model.
While the GMCI AI Index remains top-heavy, the underlying progress in distributed training suggests the rally may have more fundamental backing than previous AI token hype cycles.
