AI Compute Is Still Constrained — These 2 Stocks Could Win Big!!

Welcome back to the channel. My name's Hussein Naharro. This is What the Chip Happened, what I believe is the greatest semiconductor and AI channel out right now. Now, today's episode I want to take a closer look at two stocks that I believe are super undervalued in the AI semiconductor space, and also are some of the top players leading this market right now. But, before we take a closer look at these two stocks, I do have some interesting news about something the CEO of Iran, a new cloud player, just said. So, let's take a closer look in today's episode. But, before we begin, if you're serious about semiconductor and AI investing, I break down major earnings and conferences inside my community. Institutional quality research built for retail investors. Get 33% off at what the chip happened.com. All right, guys. So, it's been a pretty interesting week. I don't know how many videos I've actually done on my channel this week. It's just been crazy. I've been doing a lot of live streams. We've been doing a lot of earnings coverage. I've been hopping into a lot of other channels. So, if you if this is your first time here, welcome. My name's Hussein Naharro. I focus pretty much in the semiconductor and AI space. I have a master's degree in electrical engineering. I used to work at some pretty cool engineer places before I stopped and kind of do semiconductor and AI research as my full-time gig. Now, that's what you're going to get in this channel, right? You're not just going to get financial numbers. You're not just going to get very top level of hey, this is a company that's doing this much in revenue growth, and this is the margins right now, and that's what it takes to see if this is a good or bad stock. Well, I don't feel like things like that are important. I do believe in the tech industry, especially with an innovative market like AI, you need to understand at least somewhat of what's going on. What are some of the tools evolving? How can these tools expand into other avenues? And that is what I bring in this channel. Now, the first thing I want to bring up to is a post that Daniel, the CEO of Iran, just posted. He said, "Every new AI data center creates more demand for AI, not less. That's what people referring to past bulls keep getting wrong about this one." Now, obviously, obviously, the easiest rebuttal to this is you are the CEO of a new cloud company. Anything you say is going to be bullish for your business. But, I also just want to hear his reasoning for it, right? He says that the biggest thing of bears complaining is 2027, 2028 we're going to see this huge demand of data centers. So, there's huge growth of data centers turning up because of all the capex that's being spent right now is going to be spent next year. And then, bears are saying that that's going to lead to the over supply in AI tokens, the over supply of AI infrastructure. And when you have an over supply, that creates the bubble or that's when the bubble pops. But, his argument is that better AI needs more computing power to run. And cheaper AI gets used more, not less. And every step change in capabilities creates new cases that didn't exist a year ago. So, now when you have more AI data centers, the first thing you're going to do is you're going to be able to run better models. And if you're able to run better models, you're going to be able to create more work cases, right? You have now a model that can do something it couldn't do before. The second thing is cheaper AI gets used more, not less. So, maybe this cheaper is AI is not as strong, but now someone's going to be able to run this AI solution a lot more, again increasing the demand that need for AI infrastructure. Uh they go on to say that demand grows at the speed of software, but supply cannot. Supply grows at the speed of the real world. Power connection strength, mission lies, concrete, steel, and the thousands of people it takes to build. That is also something I did not think about. When an AI model comes out, developers can start creating something instantaneously, nearly instantaneously, right? And then if something gets created that needs even more and more demand or creates a workload a market thanks to AI that wasn't existent before, it's not like you could turn on a new data center tomorrow to be able to keep up with that demand. Right? There's always going a lag. In past booms, technology helped supply catch up. This time, technology is on the other side of the equation. It's fueling exponential demand and the real world can't build fast enough to keep up. I really like it. Obviously, he like I mentioned, he's extremely bullish in this AI and space. So, he's going to say something bullish, but it does make sense to me. I would love to hear your thoughts before we look at some stock, guys. If this is something you agree with or disagree with, right? I mean, the first thing I talked about was there right? It was maybe in the agriculture, right? You created an agriculture machine that helped you farm create crap crop or or collect your crop a lot faster. Now you're able to collect your crop a lot faster. The demand was I guess people feeding. And then it's not like the that solution or that technology made humans more uh hungry. Some might argue it didn't that's a moral fact, but it didn't make them more hungry. Where AI is doing that, right? It's like it's Yes, you're able to get more food AI tokens, but those AI tokens are creating more demand because they create a new use case that then that demand increases dramatically that now you have to wait for another technology to innovate to grow and grow even more crop or collect more crop. Hopefully, that makes sense. Hopefully, that didn't sound like gibberish, but it makes sense in my app. All right, so now let's take a closer look at the two stocks that I have. But if you are curious about the semiconductor industry and my overall thoughts, I do a lot of deep dives, make sure to check out my community at whatthechiphappened.com. This is a community that I share deep dives on earnings. I share my thoughts on the overall market. I share my thoughts on deep dives 101. We actually do exclusive live streams. And right now we're in prime earning season. A lot of great information gets shared here. So, check it out at whatthechiphappened.com. All right, now let's take a look at the two stocks. Right, the first stock I want to jump into is Meta Platforms. I think Meta is extremely cheap. And in today's episode, I really want to take a closer look at two AI leaders. Right, two AI leaders and two stocks that have high that have I want to say low risks. But it have a high probability of success and and have other opportunities that can help them continue to be a safe play. Remember, there's never a safe play in the overall market, but I believe the next two stocks are going to outperform the overall market in the next few. So, Meta Platform is the first one. The first is just valuation. I mean, this is a company trading at forward PE ratio of 18.4. The market is definitely punishing it for a nice scenario of CapEx is frowned. But the CapEx is giving real returns for this business. I mean, this is a company that did 16.8 billion dollars in the most recent order, up 28% year over year. When you continue to see massive growth, right? When you have a June quarter, a June quarter do better than a December quarter, that speaks volume of AI innovation. It is very rare for a June quarter to outperform the December quarter, which has all the ads for all these different types of holidays, right? So, that that right there can indicate just the overall ROI AI is created for is creating for their advertisement platform. Now, one of the opportunities that Mark Zuckerberg has, right, is to sell compute. For those that are not familiar, Meta uses its AI tools to make the advertisement business a lot better. Uh with AI, they're able to recommend better videos. If you have better videos or better posts recommended to you, you get stuck in the platform or the app a lot longer, which allows you to get hit with more ads. And the more ads, they are able to monetize more and more, right? So, it's this this great flywheel that they created that they show content, increase your time in the market, then they're able to get show you ads, and then when they show you ads, they get paid, so then there's more content in their platform. Um but now there's also another opportunity because they're building a lot and a lot of AI infrastructure. And there is the opportunity that they can sell AI compute in the future. Uh we see companies like SpaceX say that you can get so much money out of that AI infrastructure in the short term. So, there is the opportunity that Mark will start selling that. Even though I don't think it's his main plan, right? He says, "Look at the high-level observations that there's just nowhere near enough compute for all of the demand." That is why we see that basically we are getting large number of offers for the compute that we have. We have a lot of internal uses that we think are going to be quite valuable. Now, in terms of running the business, obviously a common trade-off that we need to make is around how much do you monetize something today versus develop future assets for the future. So, pretty much it's like if I know AI is going to be worth a lot two three years from now. Do I sell compute at a premium right now or do I hold that compute to be able to develop that tool in the future, which I believe will be even better returns for investors? The other thing that we hear from them is that they continue to be the main constraint, right? The main constraint under compute and they see that for the foreseeable future. That really includes their core business. So, many people always wonder why AI Meta is spending in AI infrastructure. Why are they doing this? Their core business is getting massive returns and they still have numerous ROI positive places that we would put compute towards if we added. Right? So, this really is limiting how much computing would sell if they had to uh because it does seem like they have massive opportunity. They also think beyond 2027 and when we look into 2028 and further than that, the world is going to evolve a lot. Our own internal demand will evolve. We'll have turned over a lot of cards by then. Finally, Mark Zuckerberg, the last kind of quote that I will share is he believes that there will continue to be a significant higher margin on selling intelligence rather than selling compute directly. So, there's three ways to make money in AI infrastructure. The first way is selling the compute itself. The second is being an AI model company, renting out compute and selling that intelligence for a great price. The third thing The third thing is owning is owning the intelligence, is owning the model and owning the infrastructure cuz you get to sell that compute and you get to sell the margins on the token, which is extremely extremely beneficial. And that's where they think there's a big opportunity. Obviously, to sell compute as well, but I think where he sees one of the areas that there is going to be massive growth opportunities is actually both, right? Both these areas, compute and selling AI intelligence. And as an AI investor myself, this is what I love to see, and this is what I love to read. Because to me, it's giving me an indication that we are going to be compute constrained for a long time. So, whoever has compute will be able to benefit. And second is AI is going to continue to innovate at crazy levels, that companies that are working in AI models are going to do extremely well as well. Uh so, that's the first stock, right? Meta, extremely undervalued, in my opinion. And I really do see Mark's vision of AI. I think his vision of AI is definitely pretty strong. I think he needs to continue to invest in AI infrastructure, and I think they will be a big winner. Now, it's not without risk, obviously. Their AI models used to suck. They're getting a lot better. But it is important to understand that there is risk that they are spending a nice amount of money. All right. So, now I'm jumping to stock number two. Stock number two shouldn't be a uh a shocker. And some of you might be saying, "Jose, Meta and Nvidia, what's going on with you?" Well, welcome back to the boring channel, right? This is this channel that there are times when there's opportunities to buy a high beta name. And we did, right? We bought Ford we bought $60. We bought Air at $60. We bought so many different companies in high beta when there was this downturn. Now, if people get scared, that's a whole different reason, but now it's not the time I'm going to be chasing high beta. So, if you're out there saying, "Jose, why are you talking about these high beta names anymore?" There was a time for that. Now, I go back to my normal status of now I just sit and wait. Patience gets opportunity. Now, sit and wait. I'm not going to hold cash, cuz I don't want to hold cash. So, what's the next big best thing for me? Well, Twitter big players they have the opportunity to grow and and if we ever get another high beta sell-off, sure my Nvidia might not I might sell off a bit, but it's not going to sell off to the same types of levels as those other high betas. So even with a drop I can still take out that money and still have better opportunities when that high beta turns around. So that's usually one of the ways I play the market is work extremely well for me numerous times and again, right? I'm back into the boring story for now. So if you hate the boring story, that's okay. But Nvidia mean forward PE ratio of 22.59. This that's a crime. That's a crime within itself. I I I I have to say it. That's pretty much a crime. The median for the past decade or so is roughly 34.59. This is trading extremely cheap and it's because the market believes Jose Jose and Nvidia's revenue can't grow that much longer. Nvidia's revenue is going to be stopped by so many things, competition, people deciding their own ASIC, AMD. And while I am bullish in all those other companies, I don't think it's a bearish point for Nvidia. We have SpaceX, one company committed to Nvidia GPUs only. And Nvidia and and they even mentioned they are expected to build at least somewhere in the midpoint of 5 gigawatts next year. 5 gigawatts times 15 billion dollars, which is what Jensen says he can collect per gigawatt of revenue, that's roughly 250 billion dollars from one customer. If you're telling me you think Nvidia is going to stop growing anytime soon, and that's saying extreme even if it stops growing, right? Let's say it stops growing. It stops growing revenue. Maybe 5%. You're still already trading buying it at a forward PE ratio of 22.59, which to me is like, okay, so even if you stop growing, I still have an amazing company growing at single digits, which I don't think we're going to see. Low double-digits, I don't think we're going to see. But you're telling me I can get that for 22 4 P ratio 22.59, and then you have customers like Elon Musk saying they're going to commit and and potentially spend over $200 billion on your chips next year. And you have other big cloud service providers saying the same thing, that they expect CapEx to be much higher in 2027 because their AI infrastructure constrained. All right. All right, you got it, bear. Buyer bears, you got it. That's completely right. You you made so much sense for me. Now, the other thing that we are saying that I thought was pretty cool for me was this. Nvidia partnered up with multiple financial companies, right? The The who's of the who when it comes to the banking and the lending. And they kind of created this way to build allow If you're building an AI infrastructure, you're going to be able to get favorable rate. And the reason that this happened is Nvidia is the goat is the goat at making hurdles get eliminated, right? They've done a lot of hurdle eliminations I've talked about in the past. There were were a few to kind of show They first came out with the GX Cloud, a way to show AI accelerating cloud platform. And then all the hyperscalers started to fall behind it. And they started to show how and they started to build their own AI cloud solution or AI accelerated accelerated cloud computing, not even AI accelerated cloud computing. Then you had Nvidia show the AI server makers how to build AI servers. How to do it in a time where it's faster. Because the customers, they don't want their servers late. They want it as quickly as possible. So Nvidia makes sure to reduce that hurdle of time. After that, what did they do? That one of the biggest hurdles was building data centers. So they created this platform the SX to showcase how to build data centers, eliminating the time to do it. What is the next kind of hurdle in AI infrastructure? The next hurdle in AI infrastructure is the financing side. And then Nvidia is working to make sure that is blown up. And the reason Nvidia is able to do this is because their argument is that in prior financial models, in prior financial models, a GPU was seen as a depreciating asset. So, if you're lending money to somebody who's buying a product that's going to depreciate over time, obviously they're going to have higher risk. If for that reason you're going to charge them higher interest rates to be able for you to make money on taking that risk. Now, Nvidia started to argue that that's not the case. Because of AI, GPUs are not typically the depreciating asset that they once were. Where instead of maybe depreciating in a few years, you have Core Weave saying that the A100 is running and has contracts for 9 years extension life. So, now financial bankings can go with this data point and go back and say, "Hmm, now I don't have the risk of this asset being invaluable or or lack of what is it of of having no value in 3 4 years. I actually have data points to show that why it can still hold value 7 8 years from now." And that completely changes the risk model for financial and includes the financing of those that go with Nvidia chips. And at the moment this is mainly for Nvidia chips because Nvidia is the only one that can go say, "Look, I've had this chip since 2000. We have massive AI infrastructure." Even though I am the AMD board, AMD can't say that, right? The MI300. There wasn't much adoption of that and it didn't come out to what, 2022 2023 potentially? When did the AI uh when did the when did the MI300 It it it was December of 2023, right? That's when they officially launched, but I think it really didn't start shipping till like early 2024. So, you only have about 2 years on this product. They don't have enough data points. Nvidia is truly the only one. The CPUs are all internal use cases. You don't know what external users can really benefit from that. Because Nvidia can showcase that not only have I had these chips since 2000, it's been the chip that it didn't matter what workload you ran. Maybe the older generations of CPUs, they only ran really well with certain types of workloads with certain types of models. That adds risk. So, Nvidia right now I think this was an extremely smart move. And now if you are in AI infrastructure player, you go and say, I'll take A and B chip and take out at higher interest rates. Obviously, Nvidia is working with their customers to be able to work something else. Or do you go with hmm, if I build on Nvidia AI infrastructure, I know I'm going to get better financing and based on the data that's out there, I believe I can make money with this solution. So, I thought that was pretty pretty interesting and it shows where how Nvidia continues to eliminate hurdles impacting the AI market. I mean, another hurdle that they're trying to release right now or or kind of in remove is power, right? Or certain power solutions and that's what what they're working on A100 V100 or E100. Now, Nvidia did also release their 13F recently. There weren't any major changes. The biggest change this time was they did they do have a new position of SpaceX stock, but they bought SpaceX individually. They had a prior investment in XAI which became SpaceX. They roughly own 21% there. So, SpaceX is roughly 33% of Nvidia's portfolio. Their biggest position is still Intel with nearly 47.3 position. third place is Coreweave, sitting at roughly 20%. Now, the next 13F, I think Coreweave will actually be number four because Nvidia will finally join join in with the recent investment that they made, I believe, in July. Right? I don't think if this didn't count the July era. But, I think Nvidia would be number three next order, just based on the stock price going up and the amount of money that they bought just a few uh few months ago. Uh but, if Nvidia's investment public portfolio has grown dramatically. In quarter four, it was 13 billion. In quarter one, last quarter was 18.4. Now, it's sitting at roughly 63.4 billion dollars. Obviously, it comes from a bit of a bit of AI invest private investments going public, but still doing extremely well and Intel is doing amazing for Nvidia position as a whole. Uh there wasn't anything else notable in things selling or buying for this quarter. Nothing that I personally would really complain or talk about. Now, there are also the final things I want to talk about is Nvidia is reporting that they are revising the plans to support back funding on Ohio Open AI data center. It instead of 250 billion dollars, it's 125 billion, but it also is only for the first phase right now. So, for other further phases, things can change and and they might implement. But, I don't think this is a bearish flag anywhere. It's just Nvidia deciding, "Look, I I'm going to I don't know I'm going to back you up. I have a guaranteed of 120 billion. I don't think that you need the 250. Right? You're going to have other opportunities with what I'm building, especially with that partnership that I'm working with all these big financial companies on how you will be able to get better rates." So, nothing crazy there and yeah, I'm super excited. I mean, earnings are going to be in a few weeks. This is going to be the best place you're going to get any information on that. Make sure to leave it to me. Make sure to leave it a thumbs up. Make sure to leave it a subscribe button. And like always, make sure to check out whatthechip.com for exclusive live streams, exclusive dives, and so much more. Peace out, and see you all next time.