A market of startups that moves less than the S&P 500
Bittensor subnets are young startups with small revenue and huge speculation, yet their markets swing less than the S&P 500, one of the most stable markets on earth.
That is not a flex. Const, Bittensor's co-founder, agreed on a recent community call that it is a problem, and that the fix is already being built.
The same call covered four more changes that will shape every TAO and alpha holder's next year: permissionless lending, null consensus, eTAO, and gamma. Most people saw the headlines and the angry replies. Very few saw the full reasoning behind them.
This article walks through all of it in plain English, including the parts Const admitted the team got wrong.
The problem: too much liquidity
Bittensor's subnet markets are too calm because the chain itself, not traders, supplies most of their liquidity.
The argument came from Yousef of Yuma, who presented it at the Exploit conference. Const called the talk great and the direction correct.
Here is how the markets work today. Every subnet token trades in an automatic pool against TAO. The chain injects liquidity into those pools as part of emissions, so a new team gets a working market on day one without hunting for buyers, sellers, and market makers.
That is a big reason teams launch on Bittensor. Const contrasted it with launching elsewhere, where the first big sell can send a token with no buyers straight toward zero.
But the cushion has a cost. On Uniswap, liquidity providers stay when they earn and leave when they do not, so liquidity itself is a market. On Bittensor, liquidity is mostly hardcoded through emissions, so it never shrinks to match real demand.
Why does that matter? Bittensor uses subnet prices to decide where emissions flow. If prices barely move, the ranking barely moves, and strong new teams struggle to climb past established ones.
Yuma proposed a new pool design, nicknamed the ASMR pool, tied to each subnet's 21 million cap and its outstanding supply, plus removing some alpha from pools and burning it. Const said he cannot judge it until the full paper and parameters are released, but he agrees with the core point: the pools need more volatility.
The fix: lending built into the chain
The Opentensor Foundation's answer is permissionless lending against TAO and alpha, built directly into Bittensor like Maker or Morpho.
Const was clear about the name. People call it shorting, but he said that is the wrong frame. It is lending.
Here is his own example. You post 1,000 TAO as collateral. With a 25% loan to value ratio, you can borrow 250 TAO from a subnet pool, plus the matching amount of that subnet's alpha.
From there you can sell the alpha if you think it is overpriced, or do the reverse and borrow TAO against alpha to buy more if you think it is undervalued.
Either way, borrowing pulls liquidity out of the pool. Less liquidity means prices respond more to real demand, which is exactly the volatility the Yuma talk asked for.
It also solves real problems for people already in the ecosystem.
• An owner sitting on alpha emissions can raise money without selling it.
• A miner on a subnet like Lium can borrow against their alpha instead of dumping it.
• A believer in a fast growing subnet can lever up, taking the risk themselves.
Const's preference is that the market sets the volatility, not a fixed hyperparameter. The team would pick the borrowing limits and let pool behavior adjust around real demand.
He put the difficulty on the level of the dTAO upgrade and said it may end up merged with Yuma's proposal. No launch date was given.
Null consensus: the change that made people angry
Null consensus lets a subnet officially run without validator consensus, and Const argues it only labels honestly what about half of subnets were already doing.
Some history helps. Bittensor's first consensus system had no agreement at all; anyone with stake could set weights. Later versions forced validators to agree with each other, then added tools to fight validators who simply copied others.
Over time, subnets split into two camps.
• One camp kept on chain consensus, where many validators must agree. Trust comes from the math.
• The other camp chose auditability. One main validator scores the work, publishes the data, and anyone can check it. Trust comes from transparency.
Chutes took the second path about two years ago. Lium, Targon, early Templar, and others followed, usually by routing nearly all stake to a single validator.
Those subnets were still paying the chain to run a consensus algorithm that was not really doing anything. Null consensus lets them switch it off.
Critics read it as Bittensor giving up on decentralization. Const's reply was that nothing changed on those subnets. The same single validator setup existed before; it is just stated openly now.
He added two details holders should know. Root validators can still police a subnet only with a supermajority of stake, exactly as before. And on subnets that switched, pure weight copying validators now earn less, because copying no longer works.
His bigger point is that null consensus is the first of many options. The core structure of stake, weights, and registration stays fixed, while subnets choose the rule that turns weights into rewards.
The decentralization that actually matters in 2026
Const's view is that the decentralization that matters most today is who decides where new TAO goes, and on Bittensor that budget is spread across thousands of holders.
He split decentralization into two kinds.
The first is protection. Bitcoin is decentralized so no government can shut it down. Bittensor has every tool for that, Const said, and any subnet could throw away its keys tomorrow. But in 2026 that is not what most teams are asking for.
The second is fairness in the budget. Every day, holders decide which subnets to back by buying and selling. That decides where emissions flow, and Const said about 99% of that decision sits with the market, not the foundation.
His proof was blunt. If he controlled emissions, he said, a subnet he likes would not be falling in price right now.
He was also openly critical of Ethereum. He pointed out that around 75% of its early token distribution came from one centralized sale, that a small group still steers its upgrades, and that it moves slowly. In his words, it ossified too early.
His rough take on revenue made the same point. He guessed Ethereum earns somewhere between about $20 million and $150 million a year against a much larger market cap, and asked the chat to correct him. Treat those figures as his off the cuff estimate, not data.
That is why Bittensor is moving fast now. Once a chain ossifies, missing features take years, not weeks, to add.
eTAO and gamma: from fixed rules to real businesses
eTAO lets each subnet choose how its new tokens are split, and Const described it as the first step toward gamma, which could let subnets pay each other for services.
Until now, almost every subnet split emissions the same way: 18% to the owner, 41% to validators and stakers, and 41% to miners. eTAO lets owners change that, within limits, to fit their business.
• A compute subnet can pay miners more to bring in more GPUs.
• A subnet on null consensus can pay validators less, since it needs fewer of them.
• A research team can raise the owner share to fund its work.
Soon, owners should also be able to send part of their emissions to other subnets. That is gamma.
Const used the example of Ditto, a subnet building a product where maybe 15% of the work is the kind of optimization miners do. With gamma, it could direct emissions toward buying compute or inference from other subnets instead of paying for mining it does not need.
Mark, the call's co-host, shared an insight from Mashier that may be the most underrated part. With gamma, one Bittensor wallet becomes your account on every subnet that opts in. No sign up, no API keys, no credit card, no KYC.
Const called it a supermarket for services. An AI agent could ask which subnet offers the cheapest inference and pay for it in a single chain call.
There is also a built in honesty check. Gamma tokens are effectively a claim on a subnet's commodity. If a subnet refuses to deliver, other subnets stop sending it emissions, and its alpha suffers.
Are stakers losing out? Const's answer was no. Every change is opt in, and stakers can push back by moving their stake if a subnet's choice hurts them. He also said gamma could arrive as early as next week, while warning people not to hold their breath.
Why speed is also a security feature
Const's strongest defense of constant upgrades is security: AI has made serious chain level attacks far more common, and a chain that can change quickly can patch them quickly.
He was specific about what he meant. Not miners bending rules, but attackers finding a cryptographic flaw and stealing funds. He said the rate of those attacks is going exponential, and the team has already had to respond to several reports from white hat hackers on short notice.
If the chain could not move fast, he said, much worse things could have happened.
He also revealed that the team has finished, but not yet launched, wallets designed to resist two kinds of future attack.
• Quantum computers breaking today's encryption.
• AI cracking the signature scheme most wallets rely on.
His point was that features some chains build their entire identity around, like quantum resistance or privacy, can simply be added to Bittensor because it can still upgrade.
The honest part: mistakes, founders, and what investors ask
Const admitted the team got TAO Flow wrong, and a community member raised the hardest open question: whether constant change is keeping new founders away.
On TAO Flow, he did not dodge it. The team was overconfident, he said, and the market was not as efficient as they expected. People gamed it, so the foundation had to step back in, add some central control over emissions, and return to price based emissions. Lending is the long term fix that lets that control be removed.
Then Sammy, a community member, made a sharp point. Most new subnet founders in 2026 came from inside the Bittensor community. Very few top tier founders have arrived from outside, and he believes it is hard to bet a company on a protocol that keeps changing.
Const's answer was that outcry follows every upgrade and he has heard it for five years. He expects a few more big changes ahead, including a possible hard fork, a move toward a different validator model for the chain itself, and a more formal governance process. He called those coming fights, not finished ones.
Mark offered a different view from the investor side. At Stillcore, he said, not one prospective investor has cited complexity or change as a reason to pass. What they ask instead is simpler: if these products are so good, why is none of them worth a billion dollars yet, and why does nobody they know use them?
That question is the real scoreboard. It points back to the same metric every serious TAO holder should watch: products people use and pay for.
What to watch next
Every change on this call points the same way: Bittensor is handing teams the tools to run real businesses while it still has the speed to build them.
Here is what I would track from here.
1. Gamma's launch. Watch whether subnets actually opt in and start buying services from each other.
2. The lending proposal. Watch the final borrowing limits and whether it merges with Yuma's pool redesign.
3. Subnet price movement. If lending works, rankings should start moving faster, and strong new subnets should climb sooner.
4. eTAO splits. Watch which subnets shift emissions toward miners and whether that shows up as more output and revenue.
5. Outside founders. The clearest sign of maturity will be top teams arriving from outside Bittensor.
None of these upgrades matter on their own. They matter if they turn into products people use without knowing Bittensor runs underneath.
Const compared markets to democracy: a mess, full of pumps, dumps, and scams, and still the best tool we have for finding the truth.
Bittensor is betting everything on that idea. The calm markets of today are the proof that the bet is not finished yet.