Bittensor Subnet Crashes to Zero After Weeks of Decline

The collapse of a subnet priced like a top-five performer shows how quickly confidence can evaporate when an economic model depends on continuous staking inflows to sustain emissions.

A Design That Rewards Staking, Not Scrutiny

The dTAO update went live on the Bittensor mainnet on February 13, 2025 and rewired how subnets compete for emissions. Allocation is now driven by the amount of TAO staked in each subnet instead of root validator weight assignments, and a subnet’s share of emissions is proportional to the total TAO staked in it. The goal was to democratize governance.

One community observer, writing in a post on X, pointed to basic validator miner gaming patches that have apparently let coordinated actors drain liquidity from low-cap subnets and push their alpha token prices toward zero.

The weakness is structural. Bittensor faces ongoing challenges with weight copying, validator collusion and validator-miner collusion, and these remain serious threats despite Yuma Consensus. The network also has no protocol-level slashing, so no stake is burned for misbehavior.

Without that penalty, the pattern is easy to picture. Miners push minimal work through validators who score it highly, both sides extract emissions, and they sell into thin liquidity pools. The alpha token crashes, and the validators move on to the next target.

Paying Customers Are the Dividing Line

Not every subnet is exposed in the same way. Network data presented at Bittensor’s Exploit Summit on September 28 put annualized revenue from paying customers at an estimated $28 million to $35 million, with twenty-four subnets now earning from actual customers rather than token emissions alone. Score, which does computer vision, and RedTeam, which does cybersecurity, have both landed commercial contracts.

That gap is widening. Established subnets with external revenue face less downward pressure on their alpha tokens because the income creates a floor. Low-cap subnets with no customer base depend entirely on speculative staking inflows and daily emissions.

Sell Pressure Remains Unsolved

The zero-price collapse points to a maturation problem inside dTAO. In April 2025, Covenant AI, the team behind one of the network’s largest subnets, announced it was leaving the Bittensor ecosystem. The price of TAO fell 27% afterward, from $338 to a low of $285 within a two-hour window. Major subnet departures can cascade through the wider subnet token market.

The core tension has not gone away. Subnet staking relies on constant product AMM pools, and miners, validators and Subnet 0 stakers regularly have to sell subnet tokens to recover TAO, which pushes prices down. Until subnets earn enough outside revenue to offset that daily selling, low-cap projects stay open to coordinated extraction.

Whether Bittensor governance introduces slashing and other protections before more subnets hit zero is the next thing to watch.

Hashlytics Take

A subnet going to zero under these rules is less a glitch than the system working as designed. Emissions follow stake, sell pressure is built into every reward cycle, and there is no penalty for gaming the scoring. In that setup, the only real defense a subnet has is revenue that does not come from the token. The Exploit Summit figures suggest the network knows this, which is why the revenue-generating subnets, not the emission-dependent ones, are the ones worth tracking.