This Bittensor Subnet Is Up 400% In 30 Days (ENGY Explained)

[00:00:00]Right now there's a subnet that over the past month has had over 400% gains. It's been leading the chain. A lot of people on Twitter are talking about it and really all the mind share surrounding subnets on Bitensor is pointing back to this uh subnet right here. And so I wanted to talk about it. I wanted to bring it to your attention especially if you have a long-term bag on Bitensor. If you're betting on the AI cryptocurrency space being something you've got to know about the companies out there that are actually moving the needle, the projects that are moving the needle. Now let's get right into it without further ado.

[00:00:31]So today this was a big deal for subnet 53 which is called NNG. They are providing verified inference for Frontier Open models. I'll explain it further but just starting it off you can see kind of the price here back in June uh 24th was at 0.004 and since then it's gone all the way up to 0.022. Big run, a big rally. There has to be a reason for this, right? Is it just hype? Well, there is a a real basis surrounding it and I'll explain it right now. NG is apparently trying to do something that several other subnets are doing but in their own market. They are providing inference. Inference is becoming a bigger and bigger and bigger and bigger market because as it stands training bigger and better AIs is becoming cheaper, more affordable. But the thing that isn't becoming cheaper and more affordable um is inference is providing the compute for people to run models.

[00:01:25]Recently, there was a model by the name of Kimmy K3. It's a Chinese open- source model that is competitive to the giants that the American closed source space has done. I'm talking about the claws.

[00:01:38]I'm talking about the chatbts. Kimmy K3 is competitive against it. And so, a lot of people have been beginning to use Kimmy because at a fraction of the cost, they can access pretty much the same compute. And so what's happening right now is there is a craze inside of who can provide the compute for these big Chinese open- source models. And whoever can do it the best, whoever can do it the most efficient is going to be able to eat really the fruits of their labor.

[00:02:06]One of the biggest criticisms surrounding Bitensor has been that the Bitensor ecosystem, if you don't know, provides rewards for subnets that are performing well. They call it Toflow.

[00:02:18]There's a lot to talk about, but all you really need to know is that there's a contest that happens repeatedly. And a lot of these subnets aren't really making money, but they're getting rewards from Bit Tensor. And so people are kind of asking, is there going to be a day in which these subnets are actually going to be producing value for the network, or are they just going to keep taking taking and subsidizing their models and just call it good tech? And so NG is trying to answer this in their own way and I and I think it's great. So this is led by Ning. I have been talking to you know Mark Jeffrey several people in sort of the towel ecosystem. They say Ning is this very bright guy absolutely great developer and has done a phenomenal job with his product. An adviser for it is also Algot if you guys know him on Twitter. Um he's the trader who called the Luna collapse. Now the pitch is cryptographically verified inference of OpenAI models on decentralized GPUs. This is a big thing.

[00:03:13]So let's break it down. Nothing proves the model you paid for is the one that answered. Now, let's say we are in a world in which for some reason you trust the AIS that you pay for are using the latest and greatest models. Now, I got to ask you, why do you trust them?

[00:03:28][laughter] Why do you trust them? Why do you think that they're providing the service that they're saying they're providing to you? We shouldn't have to trust them. We should just be able to know whether or not they're actually using the models that they portray. Like if I am using Kimmy K3 on this random little website, I should know I'm using Kimmy K3. There's a chance that they're just giving me this watered down version of Kimmy K3. And I won't know that.

[00:03:52]There's no way for me to know that. And so you pay for this big model and then the provider doesn't really need to show you that it's actually that model.

[00:04:00]They'll just say it is. And so is it cheaper? Is it a quantized standin? We don't know. the answer arrives, but there's no proof that it's the model that you're paying for. And so today, every API answer is pretty much a trust exercise. And that's a big gap. That is uh in my opinion kind of unfair for a lot of us. You don't know if you're using an inferior model because they'll never tell you. They'll just say, "Oh yeah, this is totally the one that you're paying for. Just trust me, bro.

[00:04:27]Just trust me, bro. It's it shouldn't work that way." And so, um, they've been able to provide a service that not only allows you to have proof, but also is able to do it at a fraction of the cost.

[00:04:40]Now, um, they're cheaper than even the most right now competitive subnet on Bit Tensor, the highest valued subnet shoots. Uh, their input is 68 cents per million tokens and $1.50 output per million tokens. I believe this is Kimmy K3 here. Pricing, it's really good. It's it's very competitive. very very very very very very very good and I've been looking at the post by you know the their advisers etc and they're saying you know it's not because we're subsidizing it it's because we are more efficiently routing the inference to cheaper places we're just going to cheaper compute and we're not just choosing the most expensive thing out there just to give you something and taking the the money from the Bitensor ecosystem. Now there are quite a few subnetss that are already um providing inference related services inside of Bit Tensor. However, I do want to stress that they are not doing the same thing.

[00:05:36]So let's kind of run down a couple of these inferencebased subnets on Bit Tensor so you can kind of understand. So there's Subet 51 which is a raw compute market. What they're selling is a decentralized GPU rental marketplace.

[00:05:49]And so they're selling the hardware.

[00:05:50]They're not doing any sort of model proof and or doing any of the answers.

[00:05:54]They're just selling uh the actual GPU rentals. We have shoots which like I said is the biggest subnet on Bent right now. They're trying to get cloud serving at scale. They over 50 open models about 120 billion tokens per day. Um do they do proofing? No, they don't. TE keeps prompts private but it does not prove which model ran. So a bit of a different market there. We have Subet28 GM which is doing closed model privacy. They have a confidential onekey gateway to claude GPT and Gemini. The confidentiality is the product. It's not the proof. It's a different market. We have Subnet 95 uh which is I believe trying to do things not on the GPU side but on the CPU side which is a totally untapped market. It's something that a lot of people aren't focusing on but their market is an edge on device. It turns your computer into a private AI inference machine on a shared network. you know, this is local privacy. Not the same thing. We have Targon, which co-authored a paper with Intel surrounding um making sure that your data doesn't get leaked and you can rent out these GPUs. Not the same thing, right? That's confidential. And then there's these smaller ones like Omron and Verithos, which are saying they're kind of doing the same thing. I couldn't find a lot on them. NG is doing proof as a product and they're also doing it at a pretty good cost. And so there's good reason to believe that this is going to be one of the more successful subnets on Bit Tensor. We've seen the bits founder time and time again say, "Hey, this is going to be a subnet that I think is going to do quite well and so a lot of people are excited for what's going to be happening here on uh on Eni. Uh even the biggest inference subnet runs mostly on emissions, not customers. You can see kind of the figures here. There's a lot of subsidies here happening on the biggest subnets on Bit Tensor. Uh the Alpha is right now trading at 0.02, which I don't know the exact market cap.

[00:07:39]I don't think it's Yeah, it's like $21 million. It's still honestly pretty early on. I think the reason it's so low is because uh people haven't really been tapping into the Bitensor ecosystem since the prices started to be depressed. I've been saying this time and time again. If you've been watching the videos, it's time to start looking into the subnets on Bitensor. The money really is going to be made here. And I believe, you know, wholeheartedly that, you know, holding Bit Tensor, you know, might do pretty well, but you're going to do a lot better if you're actually tapped into what the ecosystem is going to grow on. Because let's say Subnet 53 ends up being a big success for Bit Tensor. You're going to get a heck of a lot more upside betting on ENGI because not only are you going to go ahead and get the upside on NG, but you're also going to get the crazy staking APR that is available on NG. You might get like 5% yearly on Bitensor itself, but on NG you're going to get a lot a lot a lot a lot a lot a lot a lot more. So that is my that's my personal perspective on that. Um you know they just joined a big institutional stretch inside of Bitensor. there's it's now easier than ever to go ahead and access Bitensor subnets. I believe that the ecosystem is just getting started. I think that this is yet another big success story that is going to lead uh to gains for everybody inside of the Bit Tensor ecosystem. For too long, people have been kind of ignoring subnets on Bit Tensor ever since the Templar news. And I think they're going to be shocked to find that other than maybe Robin Hood chain and memecoins, Bitensor is the next best place to be to capture upside on on utility coins. Really, it feels like utility coins, especially AI coins, are going to be incubated into the Bitensor ecosystem. And we're already starting to see the crazy amount of gains just off of ENGY. And I suspect this isn't going to be the only one over the next few months. So, I suspect this is going to be a big trend leading into Q4. Uh, Bitensor right now is really just kind of ranging. If you guys enjoy the content, I know a lot of people aren't really providing this type of content.

[00:09:36]They're not providing content surrounding subnets, surrounding what's actually moving the needle inside of the crypto space. Make sure you subscribe so you don't miss out on more content.