Latest τemplar (SN3) News Update

Deep Dive

1. Next Research Initiative TPLR-006 (TBD)

Overview: The project's published research outlines TPLR-006 as the next open problem (Templar). The initiative will focus on advancing "optimizers, incentives, and systems for training frontier models over the open internet." This follows the successful completion of the Covenant-72B model, which demonstrated large-scale, decentralized pre-training. The research is framed as collaborative, with grants offered to external labs.

What this means: This is neutral for τemplar because it signals continued intellectual development, but the lack of a specific timeline and team adds uncertainty. The value hinges on attracting new researchers to advance the core decentralized AI mission.

2. Transition to New Team & Governance (Late 2026)

Overview: Following founder Sam Dare's exit and allegations of centralization in April 2026, the subnet was left without an owner (The Defiant). Bittensor founder Jacob Steeves ("Const") has proposed a lock-based subnet ownership system to prevent similar abrupt exits. This system would require team owners to lock alpha/TAO tokens with visible unlock schedules, tying economic commitment to long-term stewardship (Oliwier).

What this means: This is bullish for τemplar because it addresses a critical governance flaw and aims to restore investor confidence by making ownership transitions predictable. The success of this transition is the single most important factor for the subnet's medium-term viability.

3. Continued Operation of Covenant-72B (Ongoing)

Overview: The Covenant-72B model, a 72-billion-parameter language model, was completed in March 2026 and remains a live, permissionless resource on Bittensor (BlockBeats). Its weights are open-source under an Apache license, ensuring the project's flagship technical achievement continues to provide utility and attract miners to the network, independent of the founding team's departure.

What this means: This is neutral for τemplar because it provides a solid foundation of proven utility and a reason for the subnet slot to persist. However, without active development and innovation, the model risks becoming a static asset rather than a growth driver.

Conclusion

τemplar's immediate roadmap is a governance rebuild and an open call for research, transitioning from a single-team project to a more resilient, community-steered subnet. Will the proposed lock-based ownership successfully attract a new team to build upon the Covenant-72B foundation?

Deep Dive

1. Founder Exit Sparks Major Sell-Off (10 April 2026)

Overview: On April 10, Sam Dare, founder of major subnet operator Covenant AI, announced a full exit from the Bittensor ecosystem, accusing Bittensor founder Jacob Steeves of "decentralization theatre." Following the announcement, Dare allegedly sold a large amount of subnet alpha tokens—community reports suggest 37,000 TAO worth around $10 million. This triggered a cascade: τemplar's price fell 63% in 24 hours, and nearly $10 million in TAO long positions were liquidated. Steeves denied the allegations, stating his own sales were minimal and that he cannot unilaterally control subnet emissions. What this means: This is bearish for τemplar in the short term because it creates significant sell-side pressure, destroys holder trust, and leaves its primary subnet (SN3) without an active development team. However, it has prompted a proposed governance upgrade, with Steeves announcing a new lock-based subnet ownership model to better align team incentives. (The Defiant)

2. Historic Decentralized AI Model Launched (10 March 2026)

Overview: In early March, the τemplar team completed and deployed Covenant-72B, a 72-billion-parameter large language model, on Bittensor's Subnet 3. This was hailed as the largest decentralized LLM pre-training run in history, involving over 70 contributors using commodity hardware. The model achieved a competitive score of 67.1 on the MMLU benchmark. The operational launch was a major utility milestone for the Bittensor network, directly increasing demand for TAO staking and validator services on SN3. What this means: This was profoundly bullish for τemplar, as the successful deployment validated its core utility and catalyzed a 194% token rally within days. It demonstrated that high-performance, decentralized AI is feasible on Bittensor, attracting institutional interest and setting a new standard for the subnet ecosystem. (BlockBeats)

Conclusion

τemplar's narrative has rapidly pivoted from a showcase of decentralized AI's potential to a stress test of its governance and founder dependency. The key question now is whether the proposed lock-based ownership mechanics can successfully attract a new team to revive SN3 and restore holder confidence.