Why BitTensor Is the Bitcoin of AI

Transcript

Welcome back to AI Super Cycle, our weekly premier AI show diving into the intersection of AI and blockchain, agentic finance, and agentic commerce. Powered by our friends at Near. Go to near.ai to learn more. Iron Claw, their private confidential intense and more. Near is our partner today on the AI Super Cycle episode 25. The Bittensor bull thesis for 2026 with James Ross. James, welcome to the studio. >> Let's go. Thank you. Thank you so much for having me. I mean, we've done a few shows before over the years, and this is by far the most impressive location I've been.

>> Thank you. Thank you. Well, welcome to the Nest, the Tokenization Tower. Man, we have seen chatter on Twitter the last couple weeks. Bittensor, Bittensor, Bittensor from the likes of Jason Calacanis, as well. >> Yeah. >> I would love just to understand the Bittensor, let's just start with like the architecture. What is Bittensor? >> Yeah, absolutely. So, I mean, the way I would describe Bittensor is it's essentially an incentive mechanism. Um or it's this network that has these subnets, and subnets are an incentivization mechanism, right? They create a competitive environment um to essentially run a task or a competition or a challenge that gets lots of people

from around the world competing to solve some kind of core problem um for some outcomes. So, you have uh subnets like shoots, which is uh compute. You have Templa, which has been going crazy the last couple weeks, which is doing like pre-training. Um you have subnets like Synth, which is like a data science competition. So, subnets are essentially these incentivization mechanisms, and Bittensor is the network which powers all of these subnets. >> Okay, so Bittensor, the blockchain powers the subnets. Some have called it like the Bitcoin of AI.

>> Yeah. >> And kind of like referring it to it that way, Is there Is there a way to articulate the value of the blockchain itself? So, they it provides this incentive mechanism for these subnets. What is the right way to think about the actual base layer blockchain itself? >> Absolutely. So, I think like the reason why people compare Bittensor to Bitcoin is because you're essentially mining intelligence. Like Bittensor's purpose is to create the environment for people to compete to create intelligence. And there's multiple similarities with Bitcoin from very like surface level things like

there's 21 million tokens, um Tao tokens. They run on a inflationary model. They have halvening similar to Bitcoin, right? Um but then the actual network design, it's a substrate chain. Um and it uses a lot of different mechanics from different ecosystems and blockchains to create these subnets which have these like hyper-competitive environments with miners who compete to earn tokens. Um and this is all powered by Tao. So, what we saw last week with Templar getting a lot of like mainstream attention in AI. Um made people want to go and take their Tao and stake it into

Templar's subnet, right? >> Yeah. >> And you know, that immediately creates more demand for Tao and that kind of lifts the value of all subnets and all subnet tokens and you get this crazy flywheel. Um so, yeah, what we're seeing right now is I think what we've seen with Templar is an example of like a subnet being very successful. Um people like Jason talking about it. Apparently, he's just been talking about it um at the Nvidia conference as well. I just saw a tweet before coming on. So, like it's it's like an impressive subnets doing something in the real world and that's driving demand for Tao.

Um Tao is going into the pool of this subnet and yeah, this is increasing the value of kind of all the subnets at once. So, it's a very cool ecosystem. Um there is a lot of complexity to get stuck into, but for anyone who's been in crypto for a little while, um it's a really interesting place and there's it's 128 subnets. There's a lot of interesting experimentation happening um and it's all kind of starting to break through now. >> Yeah, definitely has and I'm glad we're having this conversation at this time. I think it the general context of the rise of AI in the last year or two is also contributing to this. And you mentioned that kind of BitTensor is this is this platform for intelligence and and and

the way that um Jason Calacanis explained it was he was like there's this decentralized intelligence like rise happening and you know, if you have 100 bucks, instead of spending $80 on on tequila, I think he said I think he said put 20 into Tao, but maybe we could understand why this decentralized intelligence even makes sense from the context of this boom in AI. Like what does that even mean to have decentralized intelligence as it relates to like more closed source models and you know, kind of like the big names, OpenAI, Anthropic, these types.

>> Absolutely. So, I'm not like um an expert on LLMs or necessarily this wider aspect, but I think what we've seen, um especially with Templar and with training, is a lot of people thought it wasn't possible um or decentralized training is very expensive. It's ridiculously expensive. You can never compete. Um and this is to do with bandwidth. You know, you need huge servers, uh massive servers, um like the centralized companies have to be able to train cost-effectively. Um what Templar has shown is that actually um they've managed to optimize 70 people using their own like hardware

around the world to train a model. Um and it's had impressive results versus like centralized counterparties. Um and they're continuously improving this with that incentive mechanism, which is all powered by Tao. Um you know, as there's more improvements, the people improving uh the model get more back. Like they're earning money for this. So, essentially what I think we're seeing now is like before it was like, well, decentralized AI can't compete. There's no point. And now it's like, actually it's a lot of impressive stuff is happening and it's beginning to compete and it's becoming like a big topic again in the centralized world, too. Um and that's what we're seeing at like yeah,

Nvidia's conference and we're starting to see big people in like mainstream tech go, wait, actually these experiments are very valid. Um and if they keep progressing like this, then they could be competing. And what that gives us is you know, we can train models without Open AI or without Anthropic owning everything, right? And we have this sense of uh yeah, like intelligence owned by the people, right? Versus being owned by one group, one company. >> Yeah. Okay, so you've mentioned Templar quite a bit. So, I take my Tao, >> Yeah.

>> I deposit or I stake it with the Templar subnet. >> Yeah. >> Okay. And then what does that allow me to do? It allows me to to like earn Templar tokens? >> So, or >> Yeah. >> Yeah, essentially. Like Okay. Essentially what you do is you you take your Tao, you stake into the subnet, um you receive their subnet token back, right? That alpha token. That earns APY. Um, and that's you just like participating, right? If you think that the subnet is going to go on and continue to do great things and rise in value, then that's great. You hold You hold it. Um, some subnets enable you to like pay for services with that token.

But, um, mostly like the incentive mechanism is there to pay the miners on the network. So, that's really why like, you know, the miners get paid in, um, the Templar token, the subnet token, um, and the validators as well on the network. And, yeah, that kind of creates like the flywheel for the the subnet. And then, >> So, okay, so I stake in the Templar and then I'm contributing to Templar's uh, ability to do the service that it wants to do.

>> Exactly. >> And that's how the subnets work. So, effectively, as a TAO holder, you have like this like venture portfolio of subnets, where you're like, all right, which of these are innovative, which team do I think is going to do well? You can take your TAO and then deposit it or stake it with any of these subnets as effectively a way to earn yield in those tokens as a bet on that project. >> Exactly. >> Okay. >> Exactly. >> Yeah. >> And, yeah, there's 128 subnets, I think, right now. Um, and so, I think there's a few interesting features of subnets. The first is that um, subnets have like UniV2 uh, pools, right? So, it's like a like there's not

really any centralized exchanges. So, you put TAO in, um, the slippage is just dependent on how much you put in. It's like a UniV2. There's TAO on one side, there's subnet alpha token on the other. >> Mhm. >> So, you put in your TAO, you get back the alpha token. If you deposit your TAO back, uh, if you do do the other thing, you know, you get the get the TAO back. So, that's quite a nice design. It's very simple. Um and these pools essentially, um depending on how much TAO is going into them, they receive TAO from the network being like drip-fed in every block. So, you also have TAO going into these pools from the network.

Um TAO is distributed on this like inflationary course. >> Mhm. >> And these pools are getting really big. I don't know how much um TAO is in like the biggest pools, like Shoot's pool, but I think it's maybe like $20 either side. So, $20 of alpha token, $20 of TAO. In the synth pool, we have like $4 million of TAO, $4 million of alpha token. So, there's like a very liquid group of subnets right now that people can like go into with size. And this is like fairly unique, like I've spent a lot of time like in the Ethereum community, Solana community.

Um and a big problem is often you get >> like you know, lots of interesting new tokens, but there's no liquidity, right? If you want to go in with like any size, it's it's not working. Um >> And there's obviously lots of risks around like, yeah, things might like not be there tomorrow, right? Um whereas Bittensor subnets, they're intentionally built for the long term. Um and really like the founders of these subnets are only going to get value by building like for the long term because there's this inflationary tokenomics. >> Yeah. So, help me understand why this isn't just a big giant circle of money sloshing around. Like, are any of these subnets delivering any actual value?

Like, are any of these things doing anything innovative and and like it just kind of sounds like a way to take TAO tokens, put them into different subnets, farm different tokens, dump them, move on, buy more TAO, and just kind of do this endless cycle. Like, it doesn't seem very productive to me. >> I think yeah, like that all like the subnets are doing lots of different things, right? There's like a huge variety of things subnets are doing. And like with synth we're focused on running this uh essentially like data science competition.

And uh we have this incentive mechanism. We reward people uh for creating machine learning models that model volatility. And we take all that data and we serve it as an API and we charge people for um using that API, right? So, we're like a revenue generating company. And our goal is to buy back um synth alpha tokens and to make sure the price stays high or high enough that we have consistently high competition within our network, right? And we see uh like projects like shoots doing the same thing. Um they have pretty significant revenues. I don't know the exact numbers, but I

think that was like a few million dollars a year, maybe higher, maybe like near 10 million dollars a year. They're buying back their token and that's increasing the value of yeah, their subnet. Making sure that the miners are getting paid. Another subnet, score, they are doing like computer vision. They have a number of big clients across like um the sports world, like football teams. Uh they're also like doing computer vision within like enterprises. They're revenue generating. Again, like I'm not sure how much revenue, but it's very significant. So, there are many subnets generating revenue and I think 6 months ago it was very much like how how are they driving revenue back to

the ecosystem? Is this just money sloshing around? Now it's definitely not the case. Now it's very much like these companies are businesses. The Bittensor network in some respects has helped them bootstrap. Yeah. Now they have like these very strong foundations and they're going into the world and yeah, doing all this cool stuff. >> Yeah. And so if this BitTensor thesis plays out, like how big are we talking here as far as like the adjustable market for the subnets and the Tao token? Like give us some of like the like your formulated bull thesis here for BitTensor. Like how how big is the TAM and like what are what are some of the comparables that you think about?

>> I think that like BitTensor is recreating a lot of like the early interest that people saw in like DeFi summer um where there's multiple projects launching. Um there's a way to like get access to these projects very easily. Um I think that if we think of like, you know, the like how AI is moving, right? AI is growing at a rapid pace. Um it's hard to get exposure to big AI companies, very difficult. Like we can't just go out and, you know, get exposure from OpenAI or like to Anthropic unless we have a friend in the valley who's

trying to sell us an SPV marked up 20% that's a terrible idea. >> Yeah. >> Um I think that there's a big opportunity uh for yeah, like hundreds of millions of dollars to come into BitTensor. Potentially like Tao will do like a 10 20X 50X from here. I'm not sure, but I think we got like a glimpse of it with what happened um in the last couple weeks with Templar. It was like they did something remarkable. Um a lot of interest came into Templar. This increased the price of Tao by roughly like 30 40%.

And I think we will continue to see this happen over and over again over the coming year. Um and so yeah, I think there's massive upside potential. I think uh Yeah, a good friend of mine, Mark, um has a fund with Jason, um called Still Core, and they were saying like $3,000 TAO by the end of the year. I don't know if I'm that bullish. I'm very bullish. Um but yeah, I think that's a cool target. >> So, if if you're thinking about the, you know, a way to get exposure to the AI trade in the digital asset space, like BitTensor is the way.

>> I think yeah, it's a strong way, a very strong way. Um especially yeah, if you're looking at decentralized AI, you should definitely looking at look at TAO and also the subnets, for sure. >> And then when it comes to some of like the private or privacy features or kind of anything on like the user-owned side, I know you're not like too deep into LLMs and these things, but how does how does BitTensor fare up in in terms of like the or or or the subnets in terms of like privacy or kind of like this user owners- user ownership on data, is is there anything particularly interesting happening there in in the ecosystem?

>> Yeah, so I believe Shoots is doing confidential compute and a couple others. Um but yeah, I'm not like uh an an expert on that. I mean, I've used a few different solutions. I mean, not a subnet, but I used Venice um yeah, before. That's also a cool project. Um yeah, but there's there's a lot happening uh around yeah, decentralized AI, privacy, um confidential compute, things like this. So, I expect that to continue expanding. >> And and is there any sort of like I don't know if if you've heard like any chatter or rumblings about like actually large capital allocators entering into BitTensor ecosystem? Like is there like any institutional adoption? I know that there that there was a data or an ETF.

>> Yeah. >> Is there anything in the community that cuz I know this community is like like rabid. They're very, very It's kind of like early XRP vibes. Is there anything on that front that you've heard about or >> Yeah, so um I believe so. Still core is one large capital allocator. Um DSV is a Bittensor native fund. Uh Crucible Capital and Suitcase Capital. There's quite a few Bittensor native funds that have been uh raising additional capital and raising I think a considerable amount. Um there's also, yeah, a DOT that launched. Um I think that was led by James Ultra, potentially. Don't quote me on that. Um

But yeah, it looks like there's going to be significant capital coming in this year, and it seems like Still core is really the one pushing this with Jason like being at like Nvidia's conference and like really towel-pilling everyone at like that highest level of Silicon Valley. So, yeah, I feel like, you know, it's it's coming in waves, for sure. >> So, yeah, so how do you even get started? Like do you need a So, it's a substrate chain, so you have to get like a a substrate wallet? Like walk me through what it's like to participate in the subnets? Like what does that process look like?

>> Yeah, absolutely. So, it's fairly simple. Um if you've used like MetaMask or something like that before, um you download a wallet. Two I'd recommend there's TAO Stats. Um TAO Stats is kind of like a mix of Etherscan and um yeah, like it also lists all the subnets. So, it's like Etherscan and CoinMarketCap combined. Uh there's tao.com. They have their own wallet, which is really nice. Um so basically, yeah, you get TAO on a centralized exchange. You put it into you move it into your wallet. Um and then you can very quickly like just stake or unstake from subnets. Um yeah, like the resources are very good around the ecosystem. And uh

yeah, I think that like it's going to yeah, it's it's kind of set up for like good growth. There's like a very good foundation. Um the community's been working hard, and yeah, it shows. Like I think there's a lot that is impressive like the founders especially Jacob Const. He uh does uh something called novelty search where he interviews the subnets and he really grows them. Like he doesn't give you an easy time. He like really digs into what you're doing. Um there's a very high bar for being like a successful subnet. Um anyone who's trying to come in and like do any kind of griefing at all they get completely shunned and completely pushed

out straight away like as they should. Um so they're trying to keep Bittensor very high signal. Um and we're continuing to see like yeah the founders supporting projects individually like really like pushing them forward. Um we're continuing to see like yeah the the liquidity in these subnets grow which is derisking them a lot. Uh so yeah it's exciting times. >> And so from the investing kind of lens here is there it so there's obviously subnets that are like the largest by by their market cap and by growth.

>> Yeah. >> And then there's obviously it goes all the way down to like very very small ones like very very small market caps like sub 1 million sub 500k market cap things. And they tend to kind of move together like I saw when Bittensor was running all these subnets were running as well and they kind of act as like a high beta to the to the Tao token. >> Yeah. >> Yeah. >> Yeah exactly. I think it's similar like you know when like ETH was like the only token that people used to do liquidity pools with right? Like ETH would pump and everything would pump around it.

Um good times remember those well. >> DeFi times. >> Doesn't seem to be happening so often recently but yeah. >> Yeah. >> Um but yeah it's the same like these the pools have Tao and the alpha token. >> Right. >> Um and yeah as Tao rises in price it kind of brings you know lifts all boats with it so >> Yeah. >> Yeah it's very interesting. I think people who've been involved in Ethereum or Solana ecosystem um can potentially do very well trading subnets. Yeah. Either trading on like yeah, uh their performance um on narratives. There's like a lot of opportunity there. And I don't think too many people are spending time like really researching subnets and yeah, actively trading them. So, I think

it's early days for that. But yeah, that's something that I'm excited to see more of. >> Yeah. Yep. James, thanks for your time, man. >> Thank you for uh giving us the Bittensor bull thesis. Uh I don't know any Tao. I feel Tao-less after this conversation. >> you some. >> I think our community wanted to kind of hear about it, get in front of it. Um we talk a lot about AI. We talk about a lot about how to get your payments, kind of what's happening. Uh but Bittensor has kind of been like this dark horse in in the space at large. It hasn't really been covered as much by like the broader space as something like Hyperliquid or even any of the majors.

>> Yeah. >> Yeah. >> Yeah, it's uh it's an interesting ecosystem. I think like not many VCs also have exposure. >> Right, that's yeah. >> Yeah. >> There are some big ones like Polychain, um Tao 5 or Tao 5, um DCG of course. >> Yeah. >> Uh there are lots of Oh, yeah, um >> Yeah, and Barry is big on it. Barry Silbert is big on it. >> Yeah, yeah, yeah. >> Yeah. >> Yeah, absolutely. I forgot to say like yeah, if you're thinking of capital allocators, he's been pushing for a long time. Yeah. I think potentially like things like ETFs could come. DCG is driving a lot. They have a Yuma Group and Yuma Group is bringing a huge amounts of capital um and it can help in accelerating subnets as well.

>> Yeah. Ladies and gentlemen, this has been AI Supercycle episode number 25. James Ross on the Bittensor bull thesis. That wraps us up for the day here.