Author: Schatong TechFlow
TAO dropped sharply by 15% today, now around $277, with further downward momentum.
A public letter may have caused the sudden price drop.
The sender is Sam Dare, founder of Covenant AI. Bittensor is currently the highest-market-cap project in the AI + Crypto space, and Covenant AI is the most important subnet team within its ecosystem.
Last month, this team achieved a major milestone, training the 72B-parameter large model Covenant-72B using consumer-grade hardware from over 70 contributors, claiming it as the largest decentralized AI training effort to date.
A month ago, this event made TAO the hottest narrative in the entire crypto space, evoking the feeling of a beacon of AI.
NVIDIA CEO Jensen Huang publicly praised it, and renowned Silicon Valley investor Chamath mentioned it on his podcast; $TAO surged by as much as 90% in a single month, standing out remarkably in the current bear market.
But today, Sam Dare announced he is leaving Bittensor with his team, taking down the foundation of this narrative along the way.
Sam Dare's open letter is long, but its core accusation is just one thing:
Bittensor's founder, Jacob Steeves (known in the community as Const), controls the entire network alone; decentralization is merely a facade.
In his letter, he listed a series of actions, the most severe being: first, directly suspending emissions from the Covenant subnet, effectively reducing the team’s income to zero; second, executing large-scale token sell-offs during the conflict to apply pressure and force compliance through market mechanisms.
Today, the举报 website Tao Papers, which went live today, provided stronger evidence. According to on-chain data disclosed by the website, Bittensor conducted 41 network upgrades between 2023 and 2026, with all 38 upgrades—from proposal to signature to deployment—originating entirely from infrastructure controlled by Const. The other two signers endorsed each upgrade within minutes, with no public discussion records available.
The so-called three-person "governance committee" is in reality a decision made by one person, with two others merely approving it.
The open letter and the reporting website launched on the same day, clearly indicating that Sam did not act on a whim.
Correspondingly, two months ago, Bittensor’s founder, Const, announced his resignation as CEO of the Opentensor Foundation (the development foundation behind Bittensor), citing the promotion of decentralization as the reason.
At the same time, on-chain data showed that Sam sold all 37,000 TAO tokens after announcing his departure, fueling further FUD in the market.
To date, Const’s only response to all of this has been a single statement, essentially saying that Covenant’s departure is actually a positive development, as it could give rise to subnets that operate automatically without reliance on any team.
Regarding the allegations listed by Sam Dare, not a single point has been addressed. However, whether it’s drama or contributions, the market has already priced them in.
In March, $TAO rose from approximately $170 to $337, nearly doubling. According to The Block, TAO increased over 50% in the two weeks following the release of Covenant-72B, making it the largest single catalyst within the overall 90% rally.
CovenantAI is linked to Templar, Basilica, and Grail (subnets SN3, SN39, and SN81), with subnet tokens tied to TAO through a staking mechanism. After the announcement of Covenant-72B, subnet tokens surged as much as 400%, pulling TAO along with them.
So, while buyers of TAO are ostensibly purchasing access to a decentralized AI network with hundreds of subnets, the price structure tells a different story—nearly half of the price increase is tied to a single team.
Covenant made it, and Covenant broke it. This phrase has been mentioned by many today, but most people only see its surface meaning.
More worth considering is why a network claiming to have 125 subnets relies on just a few subnets to drive prices up and others to crash them down. The majority of subnets were virtually invisible during these two market cycles.
Bittensor has been promoting its decentralized narrative for three years, yet its price structure has always appeared centralized. As for who is right or wrong in this conflict, the author believes it’s not the main point.
Every ecosystem has star teams, and any star team may leave.
The challenge is what you can do about it.
It’s not uncommon for crypto project teams to feud during bear markets; when the market lacks hot topics, whether a project rises depends entirely on whether there’s a compelling narrative to hype. Covenant-72B has been one of the best narratives this year—endorsed by Jensen Huang, with TAO doubling in value, and everyone in the community believing that decentralized AI has finally produced a serious contender.
But narrative-driven rallies have an inherent problem: those who build the narrative can also destroy it. When prices rise, they’re the light of crypto; when they fall, it’s the team fleeing. For those who bought TAO, the 90% you gained and the 15% you lost today come from the same source.
This may be the most authentic investment dilemma in crypto today: prices always follow narratives that emerge out of nowhere, and those narratives are tied to certain key individuals.
As long as people are here, the story continues; when they leave, you might not even find a counterparty.