Conviction Upgrade Goes Live: Nine Subnet Owners Weigh In. They’re (Mostly) Locking.

The Conviction upgrade goes live on Bittensor mainnet today. It is one of the biggest changes in subnet governance since dTAO, and the first mechanism in Bittensor’s history that meaningfully constrains how subnet owners can move their own emissions.

We reached out to subnet owners across the ecosystem to ask three questions: where they land on the upgrade, whether they plan to lock alpha to defend their subnets, and what risks they’re watching for. Nine responded.

Two patterns emerged across the responses. Opinion is split; bullish, cautious, and skeptical voices are all represented. But the action is near-universal. Every owner we spoke to confirmed they will be locking alpha, even those who think the upgrade itself adds little.

A quick refresher on what Conviction does

For readers catching up: Conviction lets alpha holders lock tokens to a subnet hotkey to accumulate a conviction score over time. The score has a half-life of 62 days, meaning it builds gradually. 100% of subnet owner emissions (the 18% owner share) are automatically locked into conviction on the owner hotkey, generating 1,296 alpha per day per subnet flowing into conviction.

The hotkey with the highest conviction in a subnet becomes the “Subnet King.” Discussions around the upgrade suggest the Subnet King can eventually take over ownership of the subnet. Unlocking is allowed, but unlocked alpha becomes available on a 20.8-day half-life decay, meaning a full exit takes roughly three months.

For the full context on the design rationale and the live community pushback to it, read TaoDaily’s coverage of Const’s appearance on Novelty Search below.

The bulls

Tom from Bitcast (SN93) had the most enthusiastic read, buttressed through his TradFi background.

“The idea that an owner can sell a large proportion of alpha with no warning to investors doesn’t align with my background. I am very happy that it is being introduced from an investor safety point of view but also, Bittensor is once again breaking the mold and introducing a mechanism that if used properly can resolve a problem that has plagued the industry. Rug resistant crypto is bullish!”

Bitcast plans to lock a large proportion of alpha from both its owner wallet and revenue wallets. Tom added that he has already committed publicly, on interviews and on X, to locking and expects the ecosystem to start judging teams by how clearly they communicate their plans.

John from Bitsec (SN60) framed it through the lens of capital formation.

“There are a handful of formal alpha investors and this capital pool needs to increase for subnets to thrive. Investors have few assurances when they buy alpha token. Conviction makes alpha investing more attractive.”

Bitsec will lock alpha but keep some portion liquid as “good treasury practice.” The team is also actively auditing the rollout for what John called “underexplored edge cases”, drawing a comparison to past Uniswap v3 liquidity pool manipulation as the kind of attack vector to watch for.

Jake from Investing88 (SN88) was the most concise: bullish, all owner alpha will be locked.

Mamad from Minos (SN107) framed Conviction as a way for serious teams to signal long-term commitment.

“It allows subnets with a long-term vision to show real commitment to their projects. We are watching the execution closely.”

Vex, who owns SN36, SN70, and SN99, plans to lock most of the owner alpha across all three. “I strongly believe in the long-term vision of all three of my subnets.”

The cautiously optimistic

Egill from Zeus (SN18) described the team’s position as “bullish but cautious”. He is supportive of the upgrade but is still discussing the operational implications internally. The concern he mentioned is interesting: the shift from “100% building to monitoring/capital management.”

James Ross from Synth (SN50) struck a similar note, framing it as “optimistically cautious”, optimistic about the network upgrade, cautious about the technical implementation and rollout. Synth will be locking. His one flagged risk: additional complexity for newcomers entering the ecosystem.

Gareth from Vidaio (SN85) is taking a wait-and-see approach. The team will lock, but “how exactly we do this remains to be seen.” He also surfaced a risk that hasn’t gotten much airtime in public discussion:

“Biggest risk we see is that low value subnets could be taken over. It may be cheaper to do this than buy a new slot.”

If that thesis holds, Conviction could create a secondary market for distressed subnet ownership, which is a dynamic the upgrade’s designers may or may not have priced in.

Youssef from Quasar (SN24) is one of the more nuanced cases. Quasar has been through two attacks, which constrained the team’s alpha position long before Conviction arrived. The Quasar team already had alpha locked with Const and plans to extend that lock further.

“For our case, if we didn’t get the two attacks, we would’ve liked the conviction, as we don’t aim to sell off Quasar alpha. But as an idea, of course it’s important for Bittensor to protect itself from people who see emissions as the only goal in Bittensor.”

His take: scary for teams in his peculiar situation, but important for the network.

The skeptics

Two subnet owners pushed back harder.

Leo from Almanac (SN41) captured the philosophical objection cleanly: Bittensor’s culture rewards exploit-finding, and any new mechanism eventually meets that culture.

“On paper it seems fine but core to Bittensor’s ethos is exploitation; find exploits to build back better. Only time can tell if conviction will be able to do its job or whether individuals will exploit it.”

He also flagged a structural concern that echoes some of the critiques made on the last Novelty Search call: Conviction protects investors specifically by locking owners, but owners aren’t the only large holders. “You plug one hole, another one could appear.” Almanac has committed to locking all of its owner alpha in conviction.

Jose Caldera from Yanez (SN54) was the most pointed skeptic. His position is that Conviction addresses an edge case but creates costs for the majority of subnets that were already behaving well.

“Most subnets are effectively locking most of their alphas without this mechanism. Our company is completely transparent, we have done everything we said we would do. We communicate constantly. We do buybacks, etc. There is no value to us in this, there is no real value to the community, just another thing to keep track.”

Yanez will lock anyway, but for market reasons rather than mechanistic reasons.

“Probably because the market will demand it, not because it adds or detracts anything from our strategy.”

That admission is its own data point: even the skeptics are locking.

What they’re watching

The risks named by owners cluster around a few distinct concerns.

The technical edge cases John from Bitsec mentioned (exploit surfaces that haven’t been stress-tested yet) are the kind of thing that only emerges after live mainnet usage. Bitsec is actively auditing the rollout.

The takeover economics Gareth from Vidaio raised that low-value subnets could be challenged more cheaply than registering a new slot, which is a market dynamic worth tracking once the first “conviction contests” appear.

The unlock-period games Mamad from Minos flagged, coordinated trading against alpha during the 21-day signal window, could become a new concern once the upgrade is live and visible on-chain.

The communication burden that Tom from Bitcast emphasized may end up being the bigger story. He expects the ecosystem to react to locking choices the same way it once reacted to subnet teams selling owner alpha, meaning silence won’t be neutral.

What this tells us

There are genuine bulls, genuine skeptics, and a substantial middle that supports the direction but has specific operational concerns.

But even owners who think Conviction is reactive are committing to lock because the market will price unlocked positions as a red flag, and no team wants to be the one explaining why they aren’t locking.

That’s a meaningful piece of mechanism design feedback in itself. Conviction has already made locking the default expected behavior for a serious subnet team. From here, the question is no longer whether to lock, but how much, for how long, and how publicly.

What to watch next

Watch which subnets publish their lock parameters and how the market prices their actions. The rollout of conviction on mainnet will be the real stress test of the mechanism. And Const’s own framing of a “muted” rollout means the parameters themselves are likely to keep evolving in the months ahead.

We’ll be tracking it.

Read more about Bittensor subnets below:

Want your subnet covered or your perspective in our next piece? Reach out to the TaoDaily editorial team via DM on our X.