I Just Bought $5,200 of These 3 Stocks. Here's Why

Hey everyone, and welcome back to another portfolio update for today. It's been a very eventful week. So, over past week, the Couch Investing portfolio was basically flat, 0.17% where the S&P was up 1.1% year-to-date. We're still doing quite well despite all of the volatility, and of course, despite taking a big hit over past month, I am down 11.7% but year-to-date, still up 26.1% where the S&P is up 9.55% and so, in today's video, we will of course talk about the upcoming earnings week because it's still a very important week across the board for my portfolio and maybe for yours as well. I do want to talk more about Meta and the other hyperscalers. We'll look at some stocks that we follow, 12 stocks actually, and look at the technical analysis, and then I do want to spend some time on this a JPMorgan report talking about hyperscalers investments, AI cloud investments, the return on invested capital because it's actually a very, very good report with some solid uh data. And lastly, and maybe we'll start with that to get it out of the way, of course, the big news of the week which was, well, the Leopold story, right? The Leopold story, the fund that blew up, that got saved, maybe was manipulated, etc., etc. So, quickly, by the way, yes, I love these uh dashboards, an interactive dashboard covering the whole uh story here which I made myself. Well, I made with Gemini. Pretty cool. So, the story here, or in short, what do we have? We have one, the manifesto here. So, the 24-year-old Leopold ex-OpenAI researcher published a 165-page viral essay predicting rapid AGI and massive infrastructure requirement, and so, he started this fund, got LPs from Stripe, Meta, uh Jane Street. Great. Returns, the fund absolutely exploded. Incredible year-to-date, he was still up 270% in May, which means that in June he would have been up, I think, year-to-date over 400% because in July the fund did drop 67%. Now, since the inception, it was up 1,000%. So, again, incredibly incredibly returns. Then what happened to mid-July? I mean, we've been covering on the channel what has been happening. So, mid-July, of course, you have the Kimi K3, you have more Chinese news coming out for DUV machines, etc., etc. All of the high-flying names, the SK Hynix, the Micron, the Sandisk, Core Weave Neptus Bloom Energy, you name it. All of them dropped significantly between 30 to 60% or so in a matter of a couple of weeks. And if you're running a normal portfolio, okay, you you would have been down, but you would not have had the issue that he had because he was levered up, okay, between three to four X. So, for every $1 of equity, he got $4 of debt. And of course, when everything goes up, great. But when things go down significantly in a short period of time, well, you are getting screwed. And then and then came Citadel, of course, to the so-called rescue, and Leopold did sell his public portfolio to Citadel at around 10% discount to protect the $10 billion private stakes that he has in Anthropic. Crazy news, market moved, etc., etc. Cool. To me, I just I don't understand this. You are making a hell of a lot of money. Why do you need four X leverage? Just doesn't make like how greedy do you have to be? To me, it it just doesn't make any sense and I know there was this Chris Camillo story as well with Amazon. Here as well, I don't understand what's the point. You've made millions already. Why do you need to risk I don't think he risked it all, but why do you need to risk a huge part purely on on a bet here? It's a pure gamble with Amazon one. Right, weekly options earnings report it doesn't really matter how much hours of research you did. It's earnings. We've seen it happen. Good earnings report, stock drops. It has nothing to do with with research or anything. It has to do with a Degen move and a pure gamble. Now luckily for him, it worked out. But if we had a very bad Friday, if we had I don't know, tweets from Donald Trump, if we had macro announce macro issues, geopolitical issues which I think are happening right now when I'm recording this video, then it doesn't matter that you research Amazon a thousand hours because a good report could still result in the stock being down and if you are making dumb moves for one reason or another, maybe you need to scratch that itch, well you are liquidated done finito. To me, it makes absolute zero sense. Moving on and looking at the 12 names that we're looking at. Some of you find it interesting, some of you don't. That's completely fine. Technical analysis is really just a tiny tiny part of what I'm doing when I'm looking at stocks and investments. So we start off here with Google. Google as you can see has a solid bounce, actually finished the week higher than before the earnings which we covered and we said it didn't make sense. Like why would Google be down after such an great earnings report which goes back to my point. Google good good company, good report, and the stock went down. So, let's say you ran that little gamble with Google. Yeah, the thesis was correct. Incredible quarter. Guess what? Stock market doesn't care. Poof, down, liquidated. C'est fini. Anyways, with Google reclaimed the key area, and now it can continue its way back up. As for a Rocket Lab, the lows here were a bit closer to $60, did rebound a tiny bit to come and end the week closer to $65. We do need to go back above $70 in order to really say that, "Hmm, we now have some momentum with this company." Now, we do have a SpaceX earnings report coming up. Could Could this help this name? Maybe. Could definitely hurt it as well. As for SoFi, SoFi reported here as well. Very good report. Stock was down. Stock didn't really move since the last time we checked it. So, here as well, for us to feel a bit more comfortable, let's say, it needs to go back above 17 and actually go and challenge $18 per share. One that did break out completely is, of course, Amazon after an excellent earnings report, a good good call as well for once. Very good earnings call, and so it blew past resistance and now we're $271 per share. As for Meta, different story. With Meta still struggling here, right now, it is a bit in no man's land, although it's at the same price it was near the bottom at, I believe, in March. Now, with Meta, as you know, and we'll talk a little bit more about Meta when we talk about the hyperscalers, the old AI story here, with Meta, it's just, right now, company is doing very well. Sentiment around the stock is not great. And you can see this with, for example, Jeremy, financial education, big Meta bull, started to criticize the company more and more recently. Why is that? Stock doesn't move. Stock goes down. But the business itself, the business itself is doing very [snorts] well. Right? All of the investments that they're making in AI are already translating into better results for the core business. But of course, as you know, price drives sentiment. And since we are on the topic of Meta, I was also saying Meta extremely undervalued back in 2022. Back in 2022, they were making way worse investments than today. Back in 2022, we also had this whole Apple change that affected a company like Meta. Today, we only have them investing a lot in what they believe is going to be a very, very successful business, which is this whole AI business. But yeah, hey, when the stock goes back up, I'm pretty sure the sentiment will change as well. Moving on here, new holdings didn't really move much. We're still in a very good area. Do expect new to maybe have a bigger move when we have an earnings report, so nothing to say there. With regards to a Netflix, we are literally at a so-called floor right now. That low $70 mark is a floor for Netflix. It can create a base there, and then we'll see if the company can continue to execute in the right way. We can see if these continuous rumors of company might they buy studio, might they not buy a studio, what are the We'll see what happens. But right now, it seems like the stock wants to create a solid base at $70, and then move from there. Moving on to Nebius. Nebius, of course, was down significantly. Then we've had a huge day up almost 30% or so. So we reclaimed a key area above 181, above 187. We closed out the week at 190 dollars. Of course, with Nevius, with AMD, with Coursera, with all of these names yet to report their quarterly figures, so that's going to be quite interesting. Reddit did report, stock dropped 20% on the day. We've talked about that. I'll talk a little bit more in a second, but right now, 140 dollars, that's where we closed out the week. The support floor is a bit higher, a little bit higher at 145, 146 dollars or so. So, we're a bit in no man's land right now. You would like to have a little rebound, a relief so-called rally maybe happening on Monday or Tuesday. As for Axon, still holding support quite cleanly. Axon will report their quarterly figures also on Wednesday, so nothing to say right here. Mercado Libre, same story, earnings report on Wednesday. Over the past week, didn't really move that much, so same position as the last time we covered this. And then we have Uber, also I repeat myself, same story, earnings coming up. Right now, still trading around 70 dollars. Current support area is around that price, but we know with an earnings report coming up, very important for this company, expect a move. Now, as for the portfolio, and then we'll move into the upcoming earnings week. For the Coursera calls, nothing really changed there, a bit in the red. The PayPal bull spreads are in the green, so they basically cover that, but nothing has changed for me. Upcoming earnings week, we have Palantir on Monday, we also have Grab on Monday, I believe. We have Snapchat on Monday. Then on Tuesday, after hours, big one for me, which is AMD. We'll look at what the market expects in a bit. We have SpaceX, Arista Networks, Astera Labs, Zeta, OpenDoor, Bookings, all of that, Tuesday. After the close, we're of course go live. Wednesday is also a big day. Before the open, Eli Lilly, Shopify, Uber, Walt Disney, Unity as well, Novo Nordisk for those interested, all before the open. After the close, for again, for this whole AI semiconductor story, SanDisk, Western Digital. Some people also like AppLovin. Then we have Elf, Mercado Libre, and Axon. So, again, it might not look like a big week, but when you look at the names that are reporting, pretty sure a lot of us own a handful of names. Thursday before the open, we have ConocoPhillips, D-Wave, DataDog, and Celsius. After the close, DraftKings, Airbnb, Rigetti, The Trade Desk. I do wonder if The Trade Desk is actually seeing some momentum finally, or not. Moving on to the rest of the portfolio, NIBUS still a number one here, close to 90% of the portfolio. SoFi is back at the number two spot. Why is that? I did buy a little bit more. We have AMD here at number three. It's at number three also because it did drop under $500. Google at number four. We have some Rocket Lab, Oscar, Meta, DLocal Rubrik Micron Uber Nu Holdings Robinhood Shift4 Axon Reddit Netflix Palantir cash position, and then some shares of Nvidia. Now, as you can see, I did add or I increased my SoFi position by 3.2%, so another 100 shares here. I did increase my Meta position by another 6.2%. That's another three shares. I did buy more Reddit, unfortunately not at the bottom. At $148, stock was already down 16%. Was down a bit more afterwards. It's a 20% increase. It is undervalued. Again, I'll explain in a bit. I did explain already in Friday's video, and I did deposit some cash in the account. That's basically the cash that was from the Core Weave position, slashed that position in half, put some in the options, and the rest is basically now cash for here. Of course, last week I did buy more Reddit. Yeah, the timing there could not be worse. Now, for me, why did I buy more SoFi? Purely again, it's a triple B, it's a very good company. Their assumptions for the year got worse, yet guidance got raised. They're doing excellent. They're doing what they should be doing. It's an extremely undervalued company. Now, why didn't I buy more in Nubank? I didn't buy more Nubank because I was planning on buying Nubank on Thursday, but on Thursday Nubank was up 30% and so, well, what's the point? If on Thursday we had another day where the market was flat or Nubank was up 1%, 2%, then I would have probably picked up shares at $150. Coulda, shoulda. Maybe should have bought on Wednesday, but on Wednesday I was just monitoring the situation. And then again, as you can see, Nubank is already a pretty big position for me, so it is what it is. But as for a Meta and Reddit, Meta for me is such a misunderstood name right now because on the one hand, the market is looking at the name and say, "Oh, you should become a neo cloud." On the other hand, if they did become a neo cloud, the market would say, "Ah, you're over investing." So, they cannot win at the moment despite the business winning. Uh the business is definitely winning. The core business is doing extremely well. Ads are up, ads impression pricing is up. I mean, even the agents that they're releasing for for WhatsApp business or so recently launched and that's already gaining traction, I believe a million or so. So, it's such a good business. It's just the sentiment isn't there right now. Which, okay, we know sentiment can flip very quickly. As for Reddit, which is a position I'm now down 15% Reddit is it's not that I'm saying, "Oh, there is no bear thesis for Reddit right now." And by the way, Reddit, I have 90 shares, 2.3% of your portfolio. I'm okay with that position right now. Now, with Reddit, it's a company worth $27 billion. The trailing PE here is 32.7 times. The forward one is 23.7 times. Margin-wise, you can clearly see it's very good. It's It's a very profitable company. PEG ratio under one, expected to grow rapidly over the coming years. Now, if we want to look at how fast the market is this company to grow, we can obviously look at this. So, they're expected to grow this year, this fiscal year, by 52.8% then 30% next fiscal year, and then 20 close to 24 25% in fiscal '28. Now, the thing here with the analyst expectations is that, well, analysts are wrong. You can clearly see that analysts are wrong, and that Reddit guidance is most likely sandbagged every single time. Now, I will say that the reason why the stock is down is not because of a miss this quarter. This quarter was an excellent quarter, and I've explained this in Friday's video. It's because of what might happen going forward. It's because of what might happen if the traffic [snorts] that they used to get from Google gets reduced even more and more. It's because what might happen if they do not manage to grow their monthly, their weekly, their daily active users, their logged-in users as well. That That's the problem. The problem is not that, "Oh, the users that they have, we can monetize them." No, they're doing an excellent job at that. That's really not the problem. The problem is to grow the user base. In the US, barely grew year over year. Rest of the world, that is growing. But, as you know, ARPU, average revenue per user in the US is much higher than in the rest of the world. But, as of right now, they don't have a problem of running the business in an efficient way today. The market, as you know, is forward-looking, and so the market is saying, "If Google AI Overviews, right, is taking so much traffic from you, well, that will hurt your growth, your user base growth, which, of course, then leads us to maybe Reddit should renegotiate a much better deal with a Google, with an Open AI, with an Anthropic, etc., etc., right? Because if your data is so valuable, is getting used by all of these AI labs or a Google to give users answers, well, you should be paid for it." Right? The The fact of the matter is the data from Reddit is very useful. Otherwise, a Google AI Overview would not put it in their own answer. And so, in my opinion here, it's a matter of, "Yes, it's a cheap company on paper. It's growing very fast. That's great. But, the market is now looking for data licensing deals because it's If they they cannot figure this thing out, then over the long run, it's going to be an issue." Now, moving on to AMD, going into this earnings report, we are experiencing a really tiny drawdown right now of 17.7% after being up 113% year-to-date. This is a $776 billion company, a trailing PE of 154 times, and a forward one of 53.4 times. Of course, as you know, the forward one is based on analyst expectations. Now, this might sound super expensive, but it has a PEG ratio of just 1.3. So, put that into perspective. What are the analyst expecting this company to report? Well, let's have a look, shall we? They're looking for $11.3 billion in revenue, which would represent around 52% year-over-year growth, and that would mean an acceleration in growth. That's not all. If we look at the next one 42.8% is probably going to be higher than that, then 62%, 69.8, 64. So, we're expecting then a very, very nice acceleration in growth. Why is that? MI 500 series. We've been talking about this time and time again that the growth is coming up, which is why the stock has been going up quite a lot. Plus also the server, the CPU story has been picking up more and more for them, especially this year for Intel as well. I'm expecting a very good report. I don't think anybody here is expecting a bad report, but it's all probably about the comments of what's to come. Not what's to come this year, but what's to come beyond 2026, 2027 or so. I'm expecting a very good report. Now, before continuing with the earnings predictions or estimates, it is important to look at this. If we look at, of course, the whole AI story, semiconductor story, there are companies that, of course, are benefiting already today. Right? The semiconductor names, AMD, and Intel Nvidia Micron, SanDisk, you name it. Then on the other side, you have, of course, the hyperscalers. The hyperscalers are spending a ton of money. But the hyperscalers, unlike Meta, can already show the growth in cloud. And so, that's why you're seeing here a huge difference. All of them are spending a lot, but three out of the four have a cloud business. So, they can say, "Look, we are spending a lot. Look, look, the cloud the cloud business is growing, actually accelerating here for all three." And Meta can say, "Look, we're spending a lot, but look at our core business." But apparently, the core business doesn't matter today. Only the cloud. Only when they'll be able to say, "Oh, we have reached an agreement for X amount of gigawatts or whatever with Anthropic or with maybe another company, doesn't matter. The thing here is that for Meta, they said it. Look, there are plenty of people coming, companies coming to us willing to pay a premium. And the thing is, maybe they should. Maybe they should announce a short-term deal, right? We've seen the Anthropic SpaceX deal, right? We've seen that yes, it can run for what is it? A couple of years, but they can cancel it after 3 months. And so I think for Meta, just from a optic standpoint, just announce a short-term deal. And say, "Look, we can do it. Here, take it. But the majority of the compute is going to be used by us." And I do also think that along the way and especially towards 2026, especially in 2027, we are going to see Meta talk more about AI services that they're monetizing, whether it's through WhatsApp Business, Messenger, Instagram, you name it. We're going to hear them talk way more about that. And I think the signal of, "Oh, okay, this is how they're monetizing AI." is going to push this name even higher. Now, some might say, "Oh, you're coping, bro." No, I'm not. Just listen to the earnings call, look at the numbers, look at the business already today, and you'll see that this is a tremendous business. Now, what about the rest, right? What about this whole capex story? Well, for that, we'll have to go and look at the JP Morgan report, which is a very good one. So, of course, we have huge spending. Hyperscaler cumulative capex over 1.4 trillion dollars. That's the expected total capital investment committed by major cloud providers for AI compute and data center expansion. Now, the target capacity for 2028 is apparently at 120 gigawatts. The Wall Street concerns here is of course the continuous capex expansion has raised widespread investor anxiety regarding the timing and magnitude of return on invested capital. Morgan Stanley's thesis, mistake, I said JP Morgan, it's Morgan Stanley's. They say pessimism is overblown. The long-term profitability of the AI inference space is significantly underestimated by uh the market. And so they've had three frameworks. Framework number one is the hyperscaler GPU leasing. Here you have a 31% standard approximately 31% return on invested capital target. That's $22.9 billion in revenue per gigawatt, $7.6 billion of total opex per gigawatt, and then you have $12.1 billion in net NO PAT. This is direct GPU rental to enterprise customers generating an incremental EBIT margin of 67%. Then you have framework number two, the model API where you own the infrastructure. That's here, 46% return on invested capital. $30.4 billion in revenue per gigawatt, $7.6 billion in total opex, and that's $17.9 billion in net NO PAT. So proprietary models, example Gemini, Meta's API, etc. running on own data centers with around 75% EBIT margin. Then you have framework number three, which is the model API, the third-party infrastructure. Here you have around 25% return on invested capital, which is the lowest of the bunch here, but still 25%. Not bad. $40.5 billion in revenue per gigawatt, $27.9 billion in computer rental, but here you're left with $10 billion in net NO PAT. So here it's the model labs leasing, third-party cloud compute, bearing of course intermediary rental cost while prioritizing growth speed. All of these things here, to me, look quite attractive. Now, how did they reach those numbers? For framework number one, these are the base unit assumptions. 1 gigawatt data center facility installation, 410,000 Nvidia GB200 GPUs to around 3.6 billion total annual GPU hours, 75% GPU utilization rate at an average of $8.5 per hour rental price, which leads us to this revenue per year. The operating cost, the $7.6 billion per gigawatt, that's $4.6 billion in IT equipment depreciation, $1.1 billion in non-IT infrastructure depreciation, and $2 billion of energy and operating expenses. Now, of course, when you look at the framework number three, the model API on third-party compute, they put here three cases: bear case, base case, and bull case. Of course, the pricing per token, and this is token a million tokens, so $1 per million, 1.75, $2.50. This, of course, is the story. Cuz if it goes lower, then maybe the bull case becomes the base, or maybe the bull becomes the bear, and then, yeah, okay, EBITDA margin is going to be negative, cuz the compute lease cost here is $27.9 billion across the board. Your revenue is going to change here, right? If the price per token comes down significantly, you're going to be making less revenue, but the cost is going to stay the same. So, this this is still the unknown. Right? This is the token pricing is still the unknown, but here you have the three cases, and if we just focus on the base case, this is going to be a very, very lucrative business. Finishing off here with my predictions for the upcoming earnings week, Palantir on Monday after the close, they will absolutely crush it. There's no doubt about that. Stock is also down quite a lot, so I don't think they're going into this earnings report with a price-to-perfection profile, far from it. Tuesday, AMD, like I said, they will crush it. SpaceX, honestly, no idea cuz it's the first time they're going to report, so that's going to be an interesting one to follow. Zeta, I do think they also have momentum. Let's see if they have already seen some impact some indirect impact from the partnership with Palantir. It's very, very early to tell, but maybe something during a guidance. Shopify, solid quarter here as well. Will the market reward the company or say, "Hmm, we're still paying too much of a premium." We'll have to wait and see. Uber is undervalued. The market will, of course, ask the questions or the analysts will ask questions about this whole Waymo story, their autonomous vehicle strategy, the launching new cities, new partners, etc. I do think that the overall business is doing very [snorts] well. As for SanDisk and Western Digital, I believe that the story here will continue. I don't think there will be showing any weakness. Could they tell us maybe a little bit more already about 27 or so? Maybe. Maybe so. Then, Mercado Libre and Axon, not the cheapest name of the bunch, although with Mercado Libre, I'm expecting a strong quarter. Stock sits at around what is the $1,800, $1,900. Could it make its way back to 2,000? If the quarter, especially on the credit side of things, is better than feared, although it hasn't really been bad. I don't think we can make our way back to 2,000. Otherwise, yes, maybe have a recheck the low 1,600s or so. As for Axon, they have momentum. Stock just isn't cheap, so this one is going into the earnings report with a much higher premium than before. Of course, if it was at $400 per share, would have added more. Luckily, it's also not at 600 plus. So, that's already a big thing. As for the rest here, I don't know. I haven't been following these companies too closely. Airbnb probably had some good quarter with the World Cup, etc. But, yeah, other than that, leave your predictions down below. And so, all in all, that's about it for me in today's video. It's been, of course, a hectic week. I'm still expecting another hectic week or weeks ahead, right? Cuz earning season, we're not done yet. This whole AI story is still so-called shaky, despite all of the companies that are in this AI ecosystem are all reporting very good figures. So, we'll see what happens. Stay tuned for more, and we'll see each other in the next one. Bye-bye. >> [music] [music] [music]