Transcription
Well, it's official. Nvidia is expected now to invest over a hundred billion dollars directly into OpenAI, and OpenAI's nonprofit will maintain control. This is leading a lot of people to question, wait a minute, is this all just the circular bubble, or is this Nvidia doing what they should be doing?
Well, let's take a look at some of the Nvidia numbers here really quickly, just so you can understand how a hundred billion investment into OpenAI looks. Mind you, a hundred billion dollars into OpenAI in order for OpenAI to invest more into data centers and NVIDIA chips is pretty circular, right?
Nvidia takes profits and then they could buy back stock, which they do. They regularly buy back stock. Their outstanding stock keeps going down, which is great for existing shareholders. So then they buy back stock, but then they use that money potentially to go make investments that are strategic by either making promises to companies like Coreweave or making investments into companies like OpenAI, which is then used to go buy more Nvidia chips.
Remember last week we heard that Oracle plans to spend $70 billion over the last, you know, 6 months plus 18 months, so that 2-year period. Uh so over two years they plan to spend $70 billion. They expect that 59% of that $70 billion will go to Nvidia chips. Understand for a moment just those two companies, Oracle spending 59% of $70 billion, that's $41 billion. Plus OpenAI call it $100 billion of circular money. That's $141 billion. Okay. Nvidia has revenue of about $46 billion in three months. So we've got like almost a year of revenue that we are creating at Nvidia simply through OpenAI uh OpenAI 100 bill this is you know spread out over time obviously we'll talk details in just a moment uh and then the Oracle uh 70 * 59% $41 billion right this is almost one year of revenue just from those two it shows you the magnitude of of how large this open AAI deal is.
And this is why Nvidia stock is rocketing right now. It's absolutely taking off because people were going, "My gosh, Jensen is literally taking profits and then getting Nvidia shareholders exposure to OpenAI, but whether or not that exposure matters, they're basically funding the very purchase of their own continued revenue." So yes, it is circular. At some point, it is very bubbly.
But what you find is people are making these investments because of this belief that artificial intelligence will continue to be integrated into our lives in sort of a an irreplaceable manner that it just becomes so core to what we do. And when we look at companies in the United States, they're so much more primed to adopt AI than in other countries around the world. Why? And signal we have more money. I mean, understand this for a moment. China buys some AI chips, but they only buy 10% of the world's artificial intelligence chips. It's not because they don't want AI. In my opinion, it's because they don't have rich people and rich companies. Like, how many Chinese can blow 200 bucks a month on a CH GPT Pro subscription compared to how many Americans can do it and not even notice a difference in in their standard of living. The same is true for Chinese companies versus American companies. You have China that just went through a terrible real estate recession because of the bubble that the the Chinese government popped up and then they rugpulled uh their Chinese real estate investors by changing the very debt and leverage rules that they were encouraging their builders to undertake. This makes for a very hard environment to do business in China. This is selling rich products to rich people. the new iPhone uh you know with the promise of oh we'll finally have some more AI which is the same thing they promised last year people buy it hook line I do too I got it right here I got the new watch I got the new like I get it like we have we have cash and we're like look even if it's incremental we want a little bit more battery we want a little bit more comput power we want that artificial intelligence and I get it more so now than ever before.
Why? Why do I understand that? First of all, it's a big deal that Oracle and open AI. Just those two things, this doesn't even include Meta. This doesn't even include Microsoft, mind you. They're spending tens of billions of dollars on artificial intelligence. Meta has a $20 billion deal in talks right now with Oracle. And on top of that, OpenAI wants to spend another hundred billion on backup servers, you know, to sustain sort of their ability to provide uptime uptime. Microsoft just last week committing another $4 billion to another data center in Wisconsin plus another 3.3 in Mount Pleasant, Wisconsin uh coming online later this year. Right? These these are huge capital investments.
Why uh does this how should I phrase it? Why is this all such a big deal for Nvidia's numbers here? Because remember Nvidia every quarter starts at zero. So you need this continued willingness of companies to keep throwing money at Nvidia and that's how you get Nvidia stock continuing to show EPS growth. If you look at Nvidia's EPS growth at just 20% over the next call it four years on average. They're only trading for a two peg right now. Realistically Nvidia is probably a $240 stock by year end. I really shouldn't give away these things from our alpha reports. Uh it's part of our me Kevin membership. Sometimes that just leaks out from me. But uh you know I've got a big price target for Nvidia by the end of the year and we've had it for for weeks now in the course member live stream. And look freaking rocket ship right now. So you're not part of the Meet Kevin membership yet. I guess there's a a pitch to go check it out.
But uh understand that the American consumer is willing to blow money on just incremental increases in productivity. And I get it. And the reason I was going to say I get it is because what we're doing right now and this isn't supposed to be a pitch for house hack but I want you to see the relatability when we're running models previously on old generation Nvidia chips for house hack we're looking at like 8h hour to 3day training time frames to get certain MLS like smaller MLS that we're running because we we have many different segments we have we have many different MLS that we run for our real estate AI which We plan to start beta testing in Q4 with course members uh and then we'll probably expand to some kind of public sale for real estate AI in Q1. We have many MLS that go into that. Our compute speed 5xed when we installed Blackwell chips. So, we're like, "Oh my gosh, if we're going to spend, you know, x,000 a month on labor for research and development on artificial intelligence, why not spend that same amount on labor, but get five times the output when the chips are operating faster?"
So, yeah, how has has Blackwell chips and and we actually think we're we might be turning into more of an AI play than even real estate. We're real estate backed. It's obviously all real estate. Like we're still like our company stock sells uh you know the the fund raise we're doing right now is just assuming a real estate valuation. So assume this is not an AI sales pitch because I'm not selling you some AI valuation. I'm not selling you any is like don't don't go invest in house. I just think the shares are too cheap. Uh one day we'll raise the valuation maybe you know when we actually start proving our our AI in sales to people. But but you know then the valuation could could 100x or a thousandx if if it works out well. But my point is for bringing this up as a CEO not to make you know projections or bullishness about house hack. I'm very bullish about that. You know that. But it's to show you that just like you and I will buy the new watch to get that extra battery life for a little bit more speed because we could get that email out faster. We could get, you know, multiple tasks running at the same time without the phone black screening and pinwilling. I got that so much on the iPhone 16. It was driving me nuts. Just like we'll pay for that extra 15% A19 processing on our iPhone, you know, 17, companies benefit massively from picking up that latest iteration of a chip. So, like if next year, for example, the Blackwell chips we have right now, we could get another 30% efficiency by buying the next generation chip next year, we'll probably do that because we're getting so much more productivity out of our labor in doing so. And so, that's what keeps this flywheel going right now is the actual results of what's coming out with with that with AI productivity is is good. At least, you know, for our purposes, what we're seeing, we know there are a lot of companies that are just like chat GPT rappers out there, and even they are getting great valuations. I mean, look at this. There's uh Meta spent $14 billion buying a data labeling company that tells you if something's a highway or not. Like, big deal. $14 billion for that kind of data labeling company. We're labeling, you know, interiors and exteriors of of real estate in different rooms and components of real estate. Yeah. And it's like, wow, the valuation of somebody's life highways and they're worth $14 billion. I think it was scale AI. That's crazy. So, the valuations are crazy. And there are some things that I feel like, you know, some of these chat GPT rappers, like I think Nvidia just invested $900 billion poaching the CEO of some chat GPT rapper. Basically, it's $900 million. It's insane some of these valuations. And I agree that a lot of these companies, just like MIT said, are not profitable companies. And there is a lot of froth and it's a bubble. It's a cycle. We know that. We know there's going to be overinvestment.
But for right now, these circular investments from Nvidia, they keep it going, but there's also there's so much hope in why it makes sense to keep spending on these chips. So like, as much as I want to look at it and go, it's a bubble, it it is to some extent. On the other hand, I also as a recipient of the AI chips look and go, "Oh my gosh, but the chips are better. They are good. It's like they make sense. The upgrade cycle makes sense." And the one thing that Nvidia probably doesn't have priced into it right now is people are the people who are not investing in video are worried that Nvidia always starts at zero. People are not pricing in an Nvidia upgrade cycle that happens at an annual cadence. This is what Jensen started trying to tell us about 6 months ago that don't think about this as starting at zero every time. Think about companies saying no no no we want that extra 20% step improvement just like you get that with that iPhone every year that extra 20% battery life and efficiency. The same thing with chips and you can still make use of the old ones. This is what's so valuable here in Nvidia.
And Nvidia, mind you, with this OpenAI partnership, uh, is only committing to make the investment of hundred billion in portions as they invest in the deployment of gigawatt systems with OpenAI. So, let me rephrase that. The strategic partnership is deploying 10 gawatt of Nvidia systems over the next period of time. Nvidia is only going to invest 10% when the first 10% or the first gigawatt is deployed. So they're basically doing for every gigawatt, you'll unlock 1% of an investment for us. So we'll basically give you $10 billion to go build a $10 billion gigawatt facility and then we'll invest $10 billion into your company, right? And and then that is being used to go buy more Nvidia chips. Jensen is the most brilliant circular kind of salesperson here because he's doing exactly what you should be doing. You should be going to these companies doing whatever you can right now to keep the cycle going. And frankly as much like a Ponzi implies that there's no underlying benefit and I get that cuz it's circular but again at house hack I see that underlying benefit firsthand and I get it as a consumer with AI products.
Plus we also have to conf consider remember how artificial intelligence products are getting more and more integrated into our daily lives. People forget that Nvidia is also an FSD play. You know, it's a robo taxi play with their Nvidia drive platform. I can't remember if they're on Thor or Our right now, but VADA infrastructure for coding plus Nvidia infrastructure for basically creating a product that any vehicle company can license. Anyone can license whether it's Neo, whether it's Mercedes or whatever, any of these companies can license Nvidia's AI platform to create FSD and to start catching up to Tesla. You know, they all have a lot of catching up to do, but it's beyond that. We create more and more AI purposes on a daily basis with these products. I know people like they they make fun of the AI goss, oh, I don't want Mark Zuckerberg in my living room or or whatever. Uh, and I get it. But the more we have these AI potential enabled products, which the AI on this is absolute trash right now, okay, trust me, it's like totally trash. But it'll get there. And the more we get to where we're looking at, wow, okay, yeah, smart glasses or smart rings or cars or laptops, PCs, whatever, all integrate inherently AI, the more demand there is for these AI chips. I mean, I think it's incredible that I could have that firsthand experience of of sitting and playing with my kids uh and not have to worry about holding a camera to film my kids, right? And I've got the opportunity to film them with these very glasses I'm wearing. So, I like before I go play with my kids, I put the glasses on on purpose and there's no distraction. Like they don't feel like like here I got a child on me, you know, playing with my watch. That's all she wants to do is play with my watch while at the same time there's chaos over here. Like these are the things that in the future AI just edits together for us. Like who actually goes through uh their you know old footage of of you know things that they filmed or captured with their iPhones or or their glasses or you know whatever which the more we film with glasses the more there is to edit, right? Who goes through all that stuff? Nobody goes through all that stuff. Like most people aren't video editors who are going to go through all of this, right? Uh but what's crazy is over time, the editing will all be outsourced to automatic insights provided by artificial intelligence. And a lot of that is just going to be driven on these AI data centers, whether they're Nvidia or OpenAI or whatever. I'll give you an example. When I went to Japan, you've all heard I went to Japan. News. When I went to Japan, I uh I had a 360 camera while I went on a ride with uh the kiddos. And I go on this ride and AI, you can follow me on Twitter if you haven't seen it yet, real meet Kevin. It was super hot, so I'm sorry. I'm like barely wearing clothing here. But the point is this whole video that you see is AI. Well, like the footage is 360° footage that I filmed by holding the little stick, right? But AI edited this entire 32 second clip. All of the angles, the color grading, the scenes, all of it. AI. Watch this. I'm going to play this for 32 seconds here. I like I go point at something and then it's it's videoing it. comes back to Max, Lauren, Jack, myself. All this people waving at the character over here, the scene cuts, what to cut out, what to keep this zoom in right here. Dude, that was all AI. I didn't tell it to do that. Picking up this scene over here and this ride or this ride. This all AI edited. I did nothing other than film it. And so in the future, you know, we're just gonna have the like as which is already here. We're filming it all on our on our phones or, you know, whatever our glasses or whatever, which I guess is the perfect spot to put up my Meta uh, you know, affiliate code. You get $30 off if you get the Meta Glasses, you know, and all it gets AI edited in the future. We're actually creating more demand. that might not be profitable right now for a lot of these companies, but it's it's so functional uh for as a consumer or as a user. I couldn't be more excited about this kind of stuff. I just flip it through. Oh, by the way, shout out to uh the Cybertruck. By the way, yesterday I took this picture with my meta glasses, by the way. Yesterday I um I took the Cybertruck to Lowe's a few times and I had to get this 12T 2 by6 on my on uh in the car back home. Uh, and so I ended up using uh the Cybert trucks rack that I added on top and just bungee corded it on it. I think it looks pretty good. The black on black. Uh, I think it looks awesome. So, I was really excited about the functionality there of the Cyber Truck. But, but anyway, the point is this is what we're seeing with this AI boom right now.
Now, let's listen for a moment to to what we're seeing the Coreweave CEO say. He had an interview on CNBC this morning with Sarah Eisen and uh you know quick note also on Nvidia because you know me and the balance sheet >> Kevin is much more interested than most people by the way in the balance sheet. >> Uh I am a big big big big big big fan of the balance sheet. There's nothing to complain about at Nvidia. It is so damn good. They have so much money. So it doesn't even matter that they're investing 100 bill into OpenAI. The 100 billion to open AAI is a rounding error basically for Nvidia because they're their profit margins are so high. They're bringing more than 50% to the bottom line. It's it's just it's insane. It's it's incredible. Now, what matters heavily uh is that when we look at cash flows from Nvidia, we have to ask what is the money being spent on? If they're just issuing it all out as dividends, goes back to investors. Great. That's fantastic. But does it keep the cycle going? Maybe not. On the flip side, if people want to sell some of their stock, they can. But what are we getting out of Nvidia? Nvidia is saying, "Hey, hundred billion into OpenAI, not taking a controlling interest. Open AAI worth $500 billion right now. That's like one of the company. Nvidia is paying 38 times sales for this. So they're not getting a discount on uh OpenAI as far as we're aware and you know there are concerns like SoftBank promises to invest $500 billion over four years yet they're laying off people in their venture fund. So that's sort of like all right you know what's actually going on. I don't know but I at the same time which has been killing it since we've been venturing it on the channel 23,000 AI chips now and they say they have capacity for 60,000 black wheel chips. The more I Ren stock goes up, guess what? The more money they have available to go buy more Nvidia chips. So circular in nature without a doubt. Does it make money? Yes. So here's Nvidia, you know, now at 183. But look at Iron for example. Iron's up 7% today. So a lot of these companies, let's look at uh Coreweave. Core is up 5% right now. Look, breaking above the 125 line. A critical line. This line has been so critical. We've been we've been playing this and setting up alpha report trades on this on almost a daily basis. It feels like it feels like maybe probably like twice a week we talk about four weeks and we're like that 125 line. Look this morning we reject it again. We bounce on it again. Now we're breaking out over it probably because of this Nvidia deal.
But let's listen to the core weave CEO and let's listen to see what comments he has and we'll add our own commentary. >> Good to have you here. Fine. Well done. So, so we were in these talks sort of on the sidelines of the UN General Assembly about how governments can work with poor. >> Those lines, by the way, are all Weeble. If you want to get a 2% balance transfer, go to meet.com/weble. Uh you can go to meet Kevin.com/meta for the $30 on the meta glasses. Meet Kevin.com/life. If you want life insurance in as little as 5 minutes. Really good, by the way. Highly recommend you have it, especially like while you're younger. Lock it in. paid promotion, >> right? And they're all sponsorship. >> You also just fresh off of this trip with President Trump to the UK announcing all these deals. >> All right, get to it. I want to hear the comments. Here we go. >> Thank you for having me. And I I do think that uh both of these nibbings uh really highlight um the realization that when you're u building a planetary scale infrastructure that's really required to allow artificial intelligence to to to to reach its potential that you are going to have to plug into really all of the resources uh that are available and those are private sector resources uh and those are public sector resources. the the uh the meeting yesterday, the P3 meeting. Um it was a wonderful opportunity for exchanging of ideas. Uh the opportunity to be with the uh the president and the prime minister in the UK. Once again, uh just an incredible uh opportunity to kind of uh bring together different constituents uh that have different tooling that allows for the mobiliz Oh, Meta killed the $30 off from Oh, Dude, that's lame. Are you serious? Hold on. Let me see if there's a new one. Why would they kill it? It was It was doing so well. So many people were getting it. Hold on. Let me see if I can get you a new one. I I'll I'll look right now if I can get a new one. Uh while we listen to this, okay, I'll try to get a new one. Uh give me, you know, 30 seconds to sort of figure it out. Uh of the order of the magnitude of capital that's required to allow this uh uh buildout to occur. >> What what is that? How big are we? So, so I I think a big uh one of the things, one of the narratives out there is that keeping these incredibly large deals being announced by uh uh um the the the AI labs by by the hyperscalers uh by uh the Neo clouds and you know incredible uh indication of the scaling. >> Uh no, there's a limit to how much you can get on a referral so you have to ask for somebody else. Damn, Beta you guys are so dumb. I show you guys on a daily basis and now you screw up the link because like I don't have like a contact at Meta, you know? I This was just from the app and it's like, "Oh, you've maxed out your referrals." It's kind of like with Tesla, like my Tesla referral link, it's maxed out. And I'm like, you guys are so dumb >> the demand. But what it does not do is it doesn't really speak to all of the things that need to happen and all the investment that needs to occur in order to uh provide an environment where these investments in the data centers and the supercomputers and and I'm talking about this. So far he said nothing. This is so damn s like say something already person. It's everything from energy exploration uh exploration uh through um uh you know uh power generation through >> okay we know we know all this so somebody here is complaining about the coupon code the thing okay look really quickly yes the coupon code expires but the thing is the price goes up so that's the whole point it's like we lock people in at lower prices and then that way if you got the knee Kevin heaven membership for, you know, $400 effectively, you know, in 2017, you know, uh, and now it's $1,400 or whatever. Like in the future it'll be even more. That That's why. So, like when you hear, oh, expiration or whatever, the price just goes up uh transmission uh all the way through the data centers and ultimately up and through the supercomputers to the to the to the to the models and to the application layers. And so, um, >> okay, get to the point. It's it's trillions of dollars and and you really are rebuilding one of the fundamental uh u um components of what the economy I'm surprised this interview so far not a lot of substance like okay we've heard before yeah trillion dollars redo everything I get it like I'm about to turn this guy off although they do have that MSA with Nvidia you got to give them credit they have the material service agreement with Nvidia and that material service agreement the details of which we expect to get on November 12th are going to be huge because it's basically Nvidia's you know conf like Nvidia's claim that they will buy back excess data set or capacity and that's big you know that basically shifts the bag holding from coreweave to Nvidia or generation is different than what we saw in the docom the dotcom was a bolt-on to the existing infrastructure this requires really a a new layer because you're talking about a step function and power consumption because the computer powerful. Um the the uh um you know it is a debate that goes on in many circles like is this as as the represents on the same order of magnitude greater or lesser than uh you know the advent of the internet and and and I believe that you know over the long term you don't get to put these things in in a pecking order. All you get to say is this will be one of the seinal technology buildouts and one of the seinal technologies of our future. That's why it was so interesting to listen to the president address Jensen in that room and say I hope you're right because the commitment to get it made is much different than it was in in the '90s. He so far has said nothing. This is a waste of time. Somebody says, "Yeah, getting Kevin's course. One of the best decisions I could have made a few years back. Very proud to be in the group." Oh, thank you for that. Yeah, it's interesting. You know, the um what's it called? Uh there's somebody else who g who donated $20 or just going like, "Oh, you know, basically, you know, been a course member since 2021. Great decision. That was really nice. >> I I think um you know, there there's there's another narrative out there is like, are we in a bubble?" And that's that's that's come up again and again. And then I'm sure if I you know kind of sat here long enough you guys would ask and and given the sums we're talking about it's not an unrealistic question to ask because is that capital going to actually sign. So so one of the things that's incredibly important to understand here is that the overwhelming majority of the compute that's being built right now is being built to serve inference. Inference represents the monetization of artificial intelligence. It is I don't know. Okay. There's training and there's inference. We get it. Inference, he says, represents the monetization of AI. Do I agree with that? I guess so. Because like for example, with House Act, we're doing all this training. Uh we don't we can't monetize training. We can only monetize inference. you know, when somebody like what we want is we want you to be able to open our software and me tell you if a property you're looking at is a good deal or not. And I think that lets us charge money that that creates value for people, right? And when you create value for people, you could charge money. So I guess he's right. Inference is the monetization component. Okay. All right. All right. All right. because the hyperscalers that are serving clients are being paid to build this infrastructure to serve their current demand and their projection of demand. And so just because a lot of money flows into a particular portion of the economy doesn't imply that it's a bubble. What it is implying is that there is tremendous broad-based demand across the technology space as the society. I mean, yes, I agree that right now there is broad-based demand and and it just keeps getting fueled. So, there's no sign of it hopping, so to speak. This is true. As the uh enterprise businesses begin to integrate that into their uh uh workflows and it's just, you know, I I I think that everybody jumps to the concept of is this a bubble? Um because they they just see these vast sums and they're like, "Oh, it must be a bubble." And I just think you got back to why is the money flowing there and is it being supported by commercial activity, >> right? Well, I don't know that we all do that and in part I mean the companies that are deploying this C. Okay. All right. I I mean like you know not a lot of substance in that CNBC interview have to say I think CNBC did a bad job on that but that's okay. It's CNBC so take it for what it's worth. Uh but I think his his point of view is right on that inference is the monetization cycle. Now the biggest argument that people have especially like MIT is that so much money is being invested in AI but so so little is coming out of it. Does it just take more time? Maybe you know that so so we'll see. uh can can AI be productive enough to uh to to actually be monetizable uh you know more so than than the investments that are being made to where you're actually getting a return on your investment rather than even just some portion of a return of your investment. We'll see. But broadly, I think this is useful because you have to remember that Nvidia right now is only trading for I mean barely a two peg, barely a two peg on this stock. And to me, uh, a company with these margins and the size of the pricing power they have, here's Weeble again, bow transfer 2%. Meet Dear Weeble. uh they are they're not fully valued on just a basic PEG ratio basis. So we'll see that's my take on what's going on with Nvidia knows about this. >> We'll we'll try a little advertising and see how it goes. Congratulations, man. You have done so much. People love you. People look up to you. >> Kevin Pra there, financial analyst and YouTuber. Meet Kevin. Always great to get your take.