The FERMI Paradox - Will Project Matador AMD Build Out topple NVIDIA ? #amd #nvda #fermi

Okay. So, um >> this is about Fermy. >> Um that's uh Google talking in the background. This is about Fermy, but this is TensorWave. This is their new tenant. And that's what uh this video is going to start about. I may split this up into two parts, but this is TensorWave. Um let's move on to more information. Okay, this is MXUX. You know, I'm going to just going to say this to start. I planned on making this video two or three weeks ago and I thought Fermy had a chance and they were an interesting company, but they had so much they have so much debt I mothalled the project. Some of the pro video you're going to see later in this uh this report here was made a couple couple weeks ago. Here's the what changed everything. Uh, Fermy landed its first customer and uh, you know, it's like a bar. When there's nobody in the bar, nobody wants to go in. You get a couple people in there, all of a sudden people start walking in. They got a lease agreement with TensorWave. This is TensorWave right here. TensorWave is a uh, Neocloud, a data center data. They're going to I guess um uh project matador is going to supply the data center. So they're a uh an AI model company. I'm not sure. Uh but it doesn't matter. The point is we're going to gloss over that. Uh they got a uh this is the final agreement. The partnership intensive 650 megawws over three phases. First phase deliver turnkey data center matador fermy will deliver 222 megawatts beginning in the second half of 2027. I think they got everything there to do that. Can they get the financing to put the buildings up and everything? I I don't know. I I I do believe I I I'm not certain. You got to look at this contract yourself. I'm not an attorney. I don't I I I I do believe they're going to supply the whole data center and and they're going to supply the building, the racks, the cooling, the power, and then I think TensorWave might be supplying these new AMD GPUs. Dallas, Texas, Fermy uh announced that it has executed its first binding customer lease at Project Manador uh in Texas. TensorWave is the cloud provider. Okay. TensorWave LL text TensorWave one a subsidiary of AI cloud provider TensorWave. So they're they're a cloud provider. Uh uh you know um Cororeweave uh I think that's how they're positioning themselves of that type. Okay. And what uh Matador is really, it's kind of a REIT. They are going to own the land and supply all the power. And I think they're going to build the buildings. They own them too. Construction at manager uh Matador is well underway with approximately 6 gawatt of the plans 17 gigat gigawatts already per permitted. More than 15 billion invested in the buildout to date and the first power targeted for 2026. They do have they're not bluffing. They They have all the permits. They have the water. They have the air. They have the They got a I think like a dozen turbines that were ordered way in advance of even knowing if they'd have a customer. Uh and they got a major gas line, two of them running right next to this place. Uh turnkey data center lease. So, okay, Fermy is going to do the whole thing. They're going to throw up uh the AMD units, the buildings, the power, the cooling, the whole thing. Uh they're going to do they're going to rent it out. Rent the capability out for I believe a fixed fee uh and a fixed power fee. So these guys, I believe Tensor, you got to look at this contract yourself. TensorWave got a deal I think uh 220 megawws. So a um a power plant atomic power plant is 1 gawatt terms initial term of 15 years following commencement of final delivery phase. So they got some benchmarks they got to meet with two renewal options of 5 years each. So, they got a tenant uh for 15 years and they're going to build out for this tenant and they're going to pay a a fixed price blah blah blah. You got to read the contract yourself [snorts] approximately 6.5 billion over the initial 15-year term for phase one. I think Elon Musk is getting hourly rates for his GPUs. This is not an hourly rate kind of thing. Delivered phase beginning the second half of 2027. And I do believe you know I I you know the lease contains customary conditions blah blah blah. Power is the critical constraint now and the vision and scope of Fermy with our own they have their own uh Fermy is they're big on supplying power and that's this is the bottleneck that everybody's facing and it's I think it's going to shut down construction next year uh for other people hundred uh our customers need hundreds of thousands of next generation AMD Instinct GPUs in aggregate on timeline lines they can count on. Firmeny has assembled the powered land and permits to deliver that price and we are proud to be the first customer at a campus built for the decades ahead. So, um it's all execution now. Um this is all here tense. Here's the thing. Uh, no CEO at Fermy. They're being run by a guy who was exchilan struggles can get him to stay. I think he's retired. He's on the board. I think he might be chairman of the board. what a fantastic candidate he would be for CEO. Um, but he's running he's a interim CEO. They don't they don't even have a CEO yet, right? They're in debt. Okay. But they got the gas, they got the land, they got the permits, they got the turbines, they got the contract. Now they can borrow against this rent due. This is the uh let's just go back here. This guy is uh talks too much on [laughter] he's really he's really something though. This is the se this is one of the co-founders of TensorWave and TensorWave is in uh venture funding. I don't know what round they're in, but they've been working with AMD either as TensorWave or with other companies for, you know, I don't I forget what the guy said, a long time, 10 20 years. Uh so they're all AMD all the way. Uh their rub is what they have done at TensorWave is uh they're working very uh very closely with AMD to to field test these GPUs and do all this other stuff. what they have at TensorWave which uh is their mcguffin is they have a ex Nvidia guy there who has written with the help of AI coding a utility as we all know you got Nvidia GPUs and then you have CUDA which is the real moat for Nvidia it is the coding language that's behind everything and you write in CUDA and that's how they program the CPU chips chip That's what makes everything works. And what happens is [snorts] the racks of course are proprietary for Nvidia and the power and everything else. But the CUDA language is the is the problem because people get millions of lines of code or whatever and CUDA and then they can't leave Nvidia. These guys have an ex Nvidia guy with the help of AI wrote a utility to transfer that CUDA code over to their own proprietary kind of CUDA code copy. So no more mode for Nvidia. Uh and this guy says it works already and they and the AI coding is making it better and better and better. So if you want to switch and you're working with uh somebody else that's that's all Nvidia um you can say hey I you can come over here we'll just uh run you through our utility and you'll be fine. So that's their rup now. Um I think that's really something. These guys seem very very well they're all the all these this cat is inside. This is the cube. This is a very good YouTube channel. Um, they really seem to know the market and they really seem to know what they're doing. This is uh AMD Helios. You see this Helios here? This is the rack. It weighs uh three tons, three and a half tons. Okay, it's gigantic. I'm just going to let this play in the background. This is their they've been working on this for years in the background and they finally got it up and running and this Matador installation which is going to happen next year is I think going to be their first full scale run at this thing. So you see here they're they're testing this is a testing facility and this is the same old story with the GPU 7 5 million Iraq. Okay, which is a lot. However, how much you know I think that's that's beating uh what does a rack of uh Nvidia GPUs go for? I don't know. They're 30,000. Uh I think got 16 at 30,000. I don't know. Anyway, the point is this is a way as you can see here. There you go. Blackwell Vera Rubin and his debut system better, more efficiently, and customizable. So, uh, the point is I think they're going for cheaper, better, faster, and um, you know, project manager Fermy is going to be their, you know, their Halo site, their their first big site, I I believe. But anyway, this is all happened while I was planning this while I mothball the other video. So, let me [snorts] um bring up some of that old footage and I'll just give you some background so you know what's going on here. But I'll tell you, very interesting. Uh, of course, uh, Google and, um, Amazon and Meta, they're all coming up with their own inference chips. And, uh, I I believe this is an inference and training chip. I'm not sure. Um but uh this is a thirdparty competitor to uh Nvidia. Very exciting, very interesting. And of course uh single digits firmy stock. Got about a 25% pop on this. It's still sub $10. So it's not getting any real coverage. It won't get any coverage from any players until it's over $10, but I'm not recommending anybody buy this stock. This is not a not a financial advisor, but I would say put it on your watch list. Take a close look at it. All right, let's go into these next sections. Okay, just to close out here, this is the Fermy website and um their last CEO got got in a argument with uh Letic dragged away by security. He resigned. They wanted to take it private. There's a big scrap going on. A lot of insiders have sold this stock, a lot of debt. But they got this tensor wave firmy deal going here. And uh I think it's uh I think it's worth a