The Joint Analysis That Could Reshape Conviction Before Its Next Layer Ships

Conviction’s locking primitive went live on Bittensor mainnet this week, but the version that shipped is deliberately the soft one. There is no subnet takeover capability, no auto-locking, no teeth.

What is live is the ability to lock subnet ‘$ALPHA’ token stake in support of a hotkey, which signals investor confidence and creates the on-chain primitive everything else will eventually be built on. 

Greg Schvey, the Chief Operating Officer (COO) of Yuma Group, featured on Carrot and the Stick for a deep dive into joint analysis conducted by Yuma, TAO.com, and Unsupervised Capital, and the takeaway is sharper than the surface debate suggests.

The locking primitive is broadly seen as valuable. The subnet takeover layer that has not yet shipped is where the real disagreement lives.

The Key Points From the Conversation

What Conviction actually does, what it does not yet do, and the analysis that is shaping the debate around the next version:

a. Conviction lets $ALPHA holders lock their stake to a hotkey with a 90-day half-life on unlocking. After 90 days half the position unlocks, after 180 days 75%, with no fast-exit option.

b. What shipped is the locking primitive only, not the subnet takeover mechanism. The decision logic for takeover is on chain even though the takeover itself is not, which means subnet operators can already monitor who is locking against their hotkey.

c. Conviction is addressing two distinct problems: investor confidence signaling, and the ‘zombie subnet’ problem where useful subnet slots get stuck under absent operators.

d. Around 30 subnets are structurally immune to deregistration today, meaning all stake $ALPHA could sell into the pool and the price would still sit above the dereg threshold.

e. The structural immunity is reinforced by ADR economics, where most of these subnets have ratios below 1, creating a logical actor problem that makes selling economically worse than waiting.

f. Many legitimate subnets rely on emission sales to fund operations, since they have no external funding, no VC raise, and no parallel revenue stream. Auto-locking would structurally hurt this cohort.

g. The takeover logic gates by age and threshold, requiring subnets to be one year old with 10% of $ALPHA supply locked. But the gating effectively forces young subnets to defend from day one if they want to be ready by day 366.

h. Takeover risk varies dramatically by subnet, with some sitting in low-risk positions and others having single non-operator wallets that already hold more than the operator key.

i. The takeover incentive is weaker than it looks, since taking over a subnet alienates the 70%+ of holders who invested under the original operator, and the open-source nature of Bittensor lets dethroned teams rebuild on a different slot.

j. Market structure limits who can even participate in conviction, with roughly 20% of $TAO in subnet pools, ~30% sitting completely unstaked, and around 25% held at exchanges and custodians that have not implemented conviction locking.

k. Fund mandates create another structural barrier, since managed vehicles operate under legal obligations requiring liquidity for rebalancing. US equity markets have around 60% of assets in passive funds with similar constraints.

l. Locking carries a hidden tail risk: conviction follows you through subnet operator change, so if a subnet gets taken over, the locker stays locked under the new operator regardless of preference.

m. The recommendation from the analysis is to keep conviction as an optional signaling tool rather than making it the gateway to automatic subnet takeover. The framing offered was that takeover is “chemotherapy,” powerful but indiscriminate, and that the problem it addresses is too narrow to justify the collateral impact.

n. Near-term gamesmanship is expected, with parties locking against non-owner keys on year-plus subnets just to test how the system reacts, since malicious actors will not actually act until the takeover layer ships.

Where the Analysis Lives

The conversation on this episode lands on a sharper point than the public discourse around Conviction typically does. The locking primitive that shipped this week is broadly seen as a structurally positive addition to Bittensor’s investment surface, giving $ALPHA holders a way to signal long-term commitment that does not exist in traditional equity markets or most other crypto modalities.

The subnet takeover mechanism that has not yet shipped is where the debate gets harder. The argument from the analysis is that the network should prioritize maximizing upside for legitimate builders over minimizing downside risk from a small number of bad actors, and the data assembled suggests the takeover mechanism in its current form would create real friction for productive subnets while only marginally improving the zombie subnet problem.

Whether these critiques are incorporated before the takeover layer ships will be one of the more consequential decisions of the next protocol cycle.