Bittensor Is Having Its 2013 Bitcoin Moment RIGHT NOW

[00:00:00]One of the largest funds on BitTensor is calling what's happening right now with BitTensor Bitcoin's 2013 moment. That's the moment in which really everything started to take hold for Bitcoin before it had its meteoric rise and really made history in our financial system.

[00:00:18]Recently, Mark Jeffrey, he is the one of the heads of Stillcore Capital, he went ahead and published a state of subnets report. And I think that he's likely one of the smartest people in the ecosystem and definitely something someone to watch if you're interested maybe not even in just how but in AI cryptocurrency in general. A very bright mind. I went ahead and broke down his entire state of subnets report and we're going to be going over it. So I hope you enjoy that. Let's get right into it.

[00:00:49]Major Talon says the entire subnet class trades at roughly about 1% of its centralized rivals. He's stating that he believes that subnets, if they prove to be as useful as the economics are slating them to be, are very undervalued. They're very undervalued. A lot of them are doing the same exact thing that some of these multi-billion dollar companies are doing. And so, these are the receipts surrounding it.

[00:01:15]The first thing is nothing has really happened yet from a subnet standpoint.

[00:01:19]Every subnet on Earth, these are Bidtensor subnets combined is worth about $700 million right now. The all-time high back in March was about $1.5 billion.

[00:01:32]The biggest one shoots uh is an $85 million project with a $140 million all-time high. And for context, Tao itself trades near $200 with a market cap around $2 billion. So Still Core's frame is that Bitcoin in 2013 and Ethereum in 2016 look a heck of a lot like what BitSensor looks right now. No subnet has crossed that chasm yet to mainstream adoption. We haven't seen new money arrive. It's just a small pool of wallets that are all trading these subnets against each other. And this is directly from the mouth himself. All this motion you're seeing in the subnets, this bubbling up, bubbling down, it does not matter because nothing has happened yet. Every number in the rest of this report happened before the new money showed up. There's a big problem in crypto right now where really people do these big announcements and they see all these things are going to happen and their coins are pumping and really in the end they get nothing.

[00:02:37]There is no value that ends up being delivered. they don't deliver on the promises that they make in the past. And what Mark believes is much more important than having these big ups and downs, he actually uh talked about virtual protocol as a great example of an AI protocol that went up a little bit too fast and then went down is that once you have this big rise, people's expectations are absolutely insane. But the second that it goes right back down, well, your reputation's kind of damaged for the rest of time. And so he'd much rather have this ecosystem that kind of just like slowly goes up and slowly goes up and slowly goes up, maybe uh underdelivers on some of that price performance in the beginning, but overdelivers on what they're actually trying to achieve. Because personally, if we're trying to be taken seriously, I would much rather have an ecosystem that actually delivers first and then has the price action that we expect to have uh in an asset class like AI. That's something that I think we can all agree on and something that Mark makes uh really a great point on in terms of BitTensor. Now, with this being said, we can talk all about what they've delivered so far, but really what we're all looking for is revenue this year.

[00:03:47]Year one is over for subnets and year two is the revenue year. 2025 was the first year subnets and subnet tokens existed. The open question was whether this would even form as a functioning marketplace. And it did. And the part Still Core Capital highlights is what did not happen. No subnet ended up mooning to a $5 billion tulip valuation and then collapse the way these agent tokens did on virtuals. They've had a slow restrained healthy growth instead.

[00:04:19]Even when it fell unnaturally quiet, which is great. We love them. And so at the two big Bit Tensorheavy conferences in Europe this summer, Silkor heard the same refrain from everyone. Year one was the development year and from here the rubber meets the road. And so year two is going to be the big year for revenue.

[00:04:40]These subnets on Bit Tensor have to prove that the money they're getting from the Bitensor chain is going to translate into realworld value. And a lot of that comes down to revenue. And I think he's 100% right on this. We have a couple of these subnets that are revenue positive, but we need to see more and more and more and more of this. This is all happening during a time in which centralized AI is cracking at both ends.

[00:05:07]I don't know if you saw the news recently, you could watch my video I made yesterday, but there is a model out there right now. It's an open- source model that is uh not just competitive, but in some benchmarks it's better than Fable. And Fable right now is the best centralized AI LLM out there. It's created by Claude. It's better than GPT 5.6 in some respects. It's like a really good model. And uh that is a really big deal if you're someone that's getting into opensource cryptocurrencies because well, I think that centralized AI isn't going to have the victory a lot of people were thinking it was going to have. Open source is coming for their throats. it's coming for their throats. And while OpenAI made a lot of money, you know, OpenAI generated $13 billion in 2025, they also posted enormous losses with leaked figures suggesting burn in the tens of billions of dollars. Anthropic's been doing better. They touched a roughly $30 billion annualized run rate with profitability still years out. And so their costs are big. It runs 2 to 3x revenue across the Frontier Labs.

[00:06:16]They're not making money right now.

[00:06:18]Meanwhile, the openweight side keeps closing the gap. We had Deepseek, then we had GLM 5.2, and in the last week since his report came Kimmy K3, which per the report now beats the strongest closed model on coding and agentic task at about, check this out, 16th the cost.

[00:06:35]It's way cheaper under a permissive license. And the off switch turned out to be real. A lot of people were saying that these closed source models were just going to magically leave. And the US government had an 18-day ban around Anthropics Fable release. They banned Fable for 18 days because they were saying uh that it was some kind of threat. That's a really big deal if you're trying to access the latest and greatest AI model. It's a very big deal and that's why open source is only going to get stronger. They had direct proof from the US government that centralized AI can be shut off by decree. It's never been more bullish for these big openw weight models and a lot of them are going to want to compete. They're all going to want to compete. And so Stillcore's conclusion is that intelligence is commoditizing fast and when the model itself stops being the moat, value flows to whoever distributes and specializes intelligence the cheapest. Now what does he mean by this?

[00:07:34]He means that you can think of these big LLM models as PCs and you could also you can also think of these different bit tensor subnets as essentially hypers specialized mainframes. What is the difference? A PC can specialize in a lot of things. It can do a lot of things.

[00:07:54]Your computer can do a ton of different things. But does it really specialize on on anything? No, it doesn't specialize on anything. And so all these different Bitensor subnets, all these subnets that are all competing for a slice of the pie, they all specialize in one thing.

[00:08:11]And some of them do one thing very well.

[00:08:13]Some of them have made history doing just that one thing. And so the world is going to be moving from these massive big models to these hyper specialized mainframes that are going to be actually useful because you can have the most capable model in the world, but if it doesn't specialize in the thing you actually need it for, is it actually useful? And so his argument is that the moat, the long-term advantage here is not going to be the model anymore. It's going to be exactly what Bit Tensor is built for, which is building these really hyper specialized models. I want to talk to you now about really Bit Tensor's economics, but point it over to Bitcoin because Bitcoin is the, you know, biggest coin out there. Everybody knows Bitcoin. So, here's the entire mechanism and it is the same trick that Bitcoin pulled. Remember, Bitcoin is one of the greatest experiments humanity has ever made. Nobody pays Bitcoin's security bill. And if they tried to imagine there was one company out there that tried to sustain all of the hash rate on Bitcoin, all of it. Can you guess the cost? Well, if you took a peek at the video, you would know. If you weren't, guys, lock in. Lock in. But uh if the someone tried to sustain the network's full hash rate, it would cost a single company, check this out, 8 to 15 billion a year. It's it's not happening. Like there's no single company that's going to be able to do this. and it's just going to get more and more expensive. Plus, uh, a midsized country's worth of electricity.

[00:09:46][laughter] And so instead, miners end up paying these costs themselves, competing for newly minted coins. So imagine there's 8 to 15 billion worth of electricity that needs to be sustained by something. No single company can do it. But if we distribute it all over the world and give them a reward for doing it, we get it for free essentially. That is a really big deal. The network gets that product for free. And so that was the big technical innovation that Bitcoin did. And so now we're trying to do this with talent. Bitensor's role here is not just AI, right? AI is a cool little buzzword and most of the innovation happening on Bitensor is AI related, but ultimately it is talent. Talent is what they're looking for. If you are the best at doing whatever these bits subnets need, you are going to make money. And so the idea is that they're able to source talent, really the best talent that you really wouldn't be able to get normally. it'd be impossible to get normally, but now you're able to get it on Bit Tensor and the network's paying for it. So, here is kind of the process behind how it works. The chain mint towel and subnet tokens every block.

[00:11:04]These miners, which are the people that are producing whatever results these Bitensor subnets need, they compete and they pay whatever it costs to do that.

[00:11:17]Now, only the best win the tokens. So the best miners in the building the tokens. The miners that aren't good don't win anything. And the operator sells the product for revenue. And so imagine we have this really good product that's made because we have the best miners from around the world all competing to try to make it a better product. Well, that operator comes and says, "Well, now that I have a great product, because I've got the best miners in the world, the the best talent in the world, I'm going to go ahead and sell it." which normally I wouldn't be able to do but thanks to Bit Tensor I am able to do because anybody can be a part of this. We don't need to be geoblocked.

[00:11:52]We can do it all in a decentralized manner. So anyone from around the world can contribute to this thing. They say now that we have this great product let's sell it for some revenue. Let's sell it for some revenue. And instead of booking all the profits, what they decide is due to the fact that we are getting such good talent from this Bit Tensor network and it seems like they're really helping us with costs, we buy back this token with the revenue that we make and we burn them and that creates this loop in which this subnet is going to go up in price over time if it really is making revenue and it continues to get the best talent from around the world all competing to try to make this product better. And we all know the better the product, the more it's going to sell. And so you can kind of see this cool flywheel that's happening inside of the Bit Tensor ecosystem. And I just I I think this is really one of the best explanations for Bit Tensor. The subnet operator never pays the minor, the chain does. And so the operator essentially runs a product department with zero production costs. The pattern that Stillcore keeps seeing is that building a product subnet style costs about onetenth of hiring the same capability in house. They save a lot of money. And instead of local hiring pool, you get 256 of the best minds on Earth competing in what Mark calls a crucible of flash evolution. And this also answers the acquisition question that investors keep asking. Would someone that acquires um shut a successful subnet down and bring the work in-house? Like let's say someone acquires a subnet, would they go ahead and bring all that in-house? And the answer is no. That'd be insane. For the same reason Google would keep Bitcoin's miners if it could somehow buy Bitcoin. The subnet is the product department. The subnet is the entire reason all of this works. If they did not have the subnet, they would not have the best miners all over the world competing to go ahead and solve this thing. They would not have the best talent. They would lose the talent that made their product as strong as it is.

[00:14:03]And so there's no world they would do that. It's a product department that's free and it improves itself. The operator's only real job is keeping the token attractive, which is what the buy and burn loop does. And so the point is that a subnet is a product department that pays for itself and never stops improving. That's really what this big unlock is for subnets. He talked about 12 subnetss and I would like to talk about them. I know you guys are so interested in what these big funds are looking into. So these are the big ones that were highlighted. Uh one of the big ones was was score and you know I've talked to about SCORE in the past. I've actually had Max on the channel. They are uh doing enterprise AI vision sold through PWC France as one of their partners. They're projecting $40 million in annualized revenue this year. So if anybody is bullish on year two being a big year for revenue score is one of those subnets that's doing it right. We have bits which which is uh subnet 60.

[00:14:57]They're doing AI code security. They found over 160 vulnerabilities on a live codebase where a jailbroken frontier model found 60 and caught everything it caught. Very big deal. We have shoots largest subnet out there. They are doing open- source model inference. So imagine we have these big open-source LLMs.

[00:15:18]These are some of the best in the world.

[00:15:19]Example is Kimmy. Well, they're going to need to get inference from somewhere.

[00:15:22]they're going to need to get compute from somewhere if they're going to want to use these big AI models. And so what better place as an open source model than an open source model inference marketplace. So that's what shoots is trying to do. They have a you know over a million dollars monthly revenue. Uh you know they were the open number one open source provider on open rider and they have you know tons of tokens monthly. We have Targon which is doing confidential compute and they also co-authored a paper with Intel very big.

[00:15:51]It's a big unlock because for a very long time these decentralized compute marketplaces were leaking data. There was no way to know whether or not the compute that you were getting was actually keeping the data that you were prompting it secure. And now through Targon and this co-authored paper with Intel, they cracked the code on that.

[00:16:11]And that was only done on Subnet 4.

[00:16:13]Really is a big unlock for the decentralized compute marketplace. Now we can actually trust that our data isn't getting leaked through Targon.

[00:16:20]very big unlock especially if you're someone that's interested in uh scaling your enterprise compute. A great place to to go ahead and go. We have Bitcast which you know I've talked about many times on the channel. I've actually done many of their campaigns. They have an AI managed creator ad network and they're also the first subnet that is profitable. They aren't leaking money.

[00:16:41]Their revenue fully offsets their minor emissions. We have Synth which is doing price predictions. They're also the second subnet that is profitable. They announced a $100,000 buyback and burn.

[00:16:51]We have IOTA, which has been doing a lot of really good things lately. They are working on decentralized training. In fact, I've actually gotten my MacBook in the past and done a little bit of mining myself. They completed Orion, which is a 100 billion parameter globally distributed run up to 65% of data center efficiency. So, they're doing these big models. They're training them to get better and better and better. We can assume over the course of the next year if this subnet continues to be uh successful that this research subnet is going to produce some great things for the ecosystem. We have Zeus which was doing weather prediction Vid AIO uh which is doing video compression. You know you can get up to 80% smaller files with better perceptual quality. We have 404 gen which is doing the largest 3D object library on Earth on Earth. We have Minos which is doing genomic variant calling. They have 37,000 evaluations in 11 days, which is about 5x faster than a conventional lab. And we got a lot more here, but we got 795, which is also running a test inference engine on home PCs and Macs, even a Windows. 95 bucks. Okay, what's the point? The point is that these are all verticals. All of these are verticals.

[00:18:02]This is 12 different verticals with one pattern. Thanks to Bit Tensor, thanks to the talent that they're able to get from these miners.

[00:18:11]They're able to do things that are exceptional and they would not be able to do this without Bit Tensor's network.

[00:18:20]That's the truth of the matter, guys.

[00:18:22]It's real output and a lot of them have real revenue that's starting to land and their market caps are still in the singledigit millions. They're so small.

[00:18:30]And so, this is the big gap and the thumbnail finally explains itself.

[00:18:35]They're all doing the same job at 1% of the price. So let's talk about some comparisons in terms of the inference category. Shootes is worth about $84 million right now. In terms of its centralized peer together AI had an $ 8.3 billion valuation. 8.3 billion valuation. That's a 98x gap from the training perspective. We have IOTA at $32 million and Deepseek had a $4 billion raise. That's a thousandx gap, 385x. We have Hippius, which is doing storage, which quite frankly is much better than Filecoin and Arweave. I don't know anyone that uses Filecoin or Arweave. If you do, then you're uh probably in crypto since like 2017 because there's no way you'd want to use Filecoin and Arweave. You're going to have a better time using Hippius.

[00:19:25]There's a 45x gap just there, just in that one alone. There's no reason these cryptos should be worth as much as they are. Hippius is much more competitive.

[00:19:32]We have BitSec at 6.2 2 million and we have Sync which had a $7.4 billion peak valuation in code security. That's a thousandx gap. And so either the market is wrong about the subnetss or they're wrong about the giants. Gaps this wide do not survive contact with new money.

[00:19:53]When crypto comes back, when Bitcoin comes back and people see the work that's being done on Bitensor, this gap is not going to last forever. It's not to say that IOTA is going to hit a $45 billion valuation. I think that's a, you know, a little bit far-fetched to be completely honest. But it is to say that there is a gap there. Somewhere in there, there is a gap in valuations. A lot of it due to the fact that we're inside of the crypto space and people aren't really looking into what's happening inside of this niche. But once they do, something here is going to change.

[00:20:24]Something here is going to change.

[00:20:25]either shoots is overvalued um or or we're going to have a a big moment for these subnets. And so now I guess the question is Jesus, it's great that you're talking about the fact that Bit Tensor subnets might be undervalued when the new money comes in, but like what are the things happening to suggest that the new money might be coming in at some point or another? Well, I'm glad you asked that. the doors for new money just open. And this is still Cor's explanation on this for why subnets have stayed cheap for as long as they have.

[00:21:02]Even cryptonative people had no easy way to buy them. And I I actually agree with you. When I first joined Bitensor back in was it it's been a while. I think it was May. [snorts] Back in May, I was having a lot of trouble buying BitStensor subnets. And actually, Bitsensor subnets by far have been my most profitable investment of the year.

[00:21:20]They literally have been um you know I I kind of flip-flopped in and out of them but I'm very happy that I did. Bits of subnets kind of peaked temporarily. I think we're finally getting to a point where these subnets are looking like very attractive buys once again and have started allocating into a subnet ecosystems cuz I I think it's an absolute no-brainer. Now this is an ecosystem that at one point was very hard to invest into and now thanks to a couple rails it's becoming very easy. So the first rail is is Kraken. So Kraken did a listing of seven subnet alpha tokens. Shoots, Targon, Score, Liam, Ridges, Hippius, and Vanta. The first ones are already live. This is a tier one exchange where normal crypto money can just buy subnet tokens directly.

[00:22:05]They don't have to go through any of this fluff. They just buy them. We also have OKX, which had their TOAO USDC USDT spot, which went live on June 30th. plus in OKX AI marketplace with the open tensor foundation as a launch partner.

[00:22:19]So that's a second tier one door and it's an exchange actively building within the ecosystem. We have wormhole and salana which built out I have a podcast on this with tao.com surrounding their canonical towbridge that's been live since May. That means that people from Salana can finally go ahead and hold Tao. And we have Thor Chain, which means they did a big integration with mainet TA native Bitcoin and ETH swaps into TA with no centralized exchange in loop. For the first time since these subnets have existed, outside money has real doors and this is the first time they really have it in subnet history. They really haven't had access to this uh in any time. Who knows what the market cap of these bits subnets would have been if back in May all of this was live. I mean, a lot of people were excited about when subnets peaked at $ 1.5 billion.

[00:23:08]They would have been much higher if they had easier access. I don't know how many comments I got asking me about how to buy these subnet tokens. And now it's really easy. It's much easier to do them on Kraken on on wormhole and salon. It's much easier. The next thing to talk about is the big chain upgrade that happened very recently. This is a big upgrade for Bit Tensor. The biggest seller of Subnet tokens just switch sides. It's a very big change. This is the deepest change in the report and it's about 48 hours old as of Mark's recording. Under the old design, the chain sold Subnet tokens every block to generate yield for root stakers. So root stakers are the people that are just staking cow normally. The way that they would get their yield is through these submet tokens that get automatically sold. That actually changed. It changed in a big way. And the big problem surrounding that is that it was like a permanent seller that just pressed on every single subnet's price. And so the subnets that are doing really well would really be bogged down by this system.

[00:24:13]There's a new system. It's called root reborn.

[00:24:17]And the chain now buys subnet tokens with newly minted tow every block with validators directing which ones. So validators get to choose these root validators get to choose which subnets they end up actually grabbing. They end up grabbing and holding. Very big unlock because it means that towel is not automatically sold. It's now automatically it's not automatically selling anymore. The chain's buying now offsets roughly 80% of what it emits.

[00:24:46]And before the change, the chain one was a net seller. And the other rule change all points in the same direction.

[00:24:52]Talflow which had become gameable was rescended. Uh we have conviction which now locks subnet owners tokens with a visible unlock process. So the take money and run rug pattern from earlier this year structurally harder. They're talking about the Templar rug where the Templar subnet owner decided to just sell everything in one go. They can't really do that anymore under conviction.

[00:25:14]Subnet Tempo has become programmable and the approved shorting design which is a brand new thing coming for Bit Tensor requires shorts to be collateralized in Tao itself. No lending anywhere in the system which means even the bears become Tao buyers. You need to have Tao in order to short TA [laughter] in order to shout the short these TO subnets. Now these are definitely a risk. So let's be honest, these constant rule changes are definitely a risk.

[00:25:42]Every one of them forces investors to rethink assumptions. Still's view, which I also share, is that these particular changes were good ones made fast in what's essentially the Elon style. Find the biggest problem, fix it, and repeat.

[00:25:59]That's a really big deal. The single biggest structural seller of Subnet tokens is now a structural buyer. And the market has had only two days to d to digest that. This has only been two days. This is a fresh thing. So here's my take now that we've listened to Stilcore Capital. Here's my take on this whole thing. Here's what I buy. I think that the mechanism surrounding Bit Tensor is very real. There are two subnets that are already profitable against their own emissions. There's a third that's projecting $40 million.

[00:26:29]It's score through a big four adjacent channel. And I mean the receipts run across 12 of these unrelated verticals.

[00:26:35]There's a lot of them out there. The rails are very real. The chain flip is real. And you know, I believe that the data surrounding everything I've covered says Bit Tensor story works on narrative, not on price pumps. So none of this needs a green candle to work, which I think is exceptionally unique regarding Bit Tensor. We don't need to have these massive pumps right now in order for the network to be validated and for innovation to come in. It could just happen. Now, what keeps me honest surrounding what Silcore has said, because I I don't believe literally everything Silkor has said. Store is at the end of the day a fund that is talk in its own book. They're they're they're a fund at the end of the day that is incentivized to have investment inside of subnets. [snorts] I believe that comparing a liquid tokens market cap to a private company's equity valuation ignores a lot of things regarding float liquidity and rights. So the gaps might be a little bit overstated.

[00:27:30]I still think that even if it's a 10x gap or 100x gap against a thousandx gap, it's still a big thing. And uh another thing that I am very keen on monitoring is new money. New money hasn't come in yet. It really hasn't come in. Until this new money actually walks through Kraken and Salana, these new doors don't really matter. They really don't matter.

[00:27:52]You can have more doors open, but if nobody comes through them, it doesn't really matter. And so I really want to see some of that new fresh money come in due to innovation surrounding TA which has happened in the past and I'm sure will continue to happen uh if the rate of change in which the Bitensor ecosystem suggests is going to happen.

[00:28:12]My verdict is I agree with the direction of the report. This is 100% a positioning mark and I've said it time and time again. I do not think you can time this market. It's impossible. You will not be able to time this market.

[00:28:24]you can position into the market and benefit from it, but you can't time this market. You're going to miss it. The move is watching these revenue subnetss.

[00:28:32]Are they going to be making money? Are they going to be making money or is this big experiment a whole lot of nothing? I think it'll be worth something. And so that's why I put my money on tow. But I believe there has to be an invalidation at the end of the day. And I believe if this year two closes, if in 365 days, no subnet really shows any level of growth with no meaningful volume coming through new rails, then we're going to run into the same situation most of these big dino coins run into, which is maybe the experiment was a failure. Maybe we do need to change some of the things that are happening here.

[00:29:08]But I would rather be early and bored than late and chasing. But early only pays if year two delivers the revenue.

[00:29:19]And so let me know your thoughts in the comments section surrounding everything.

[00:29:22]Thank you so much for watching this video. Share it around with your friends. I don't think there are enough people actually talking about the inner workings of Bit Tensor. Till the next time, stay classy and I'll see you all