Why I’m Doubling Down on This Hated Stock

Hey everyone and welcome back to another video for today. In today's video, we're going to talk about the hot topics of the market right now. Going to have a look at the current state of the market. As you can see, this was basically how we ended Monday. So, yes, some of our favorite names were down. But then again, we also have a micron that was up 2.7%. Was a lot of news, macro related news that shake the market a little bit. We've had some more legal issues with regards to an Amazon, an investment from Google, and then Meta, still the ugly duckling here. Meta continues to execute. Meta actually came out with some great news. We'll touch on that as well. I also want to touch on SoFi in this video because it does feel like people have completely left this name behind. Now, some say for a good reason because the stock doesn't move. And so, why should we own a name where the stock doesn't move, right? We should only own stocks that go up regardless of the company's execution. Now, I want to start off with this to open up a a small discussion here when we always talk about we need more compute because well, people are saying, I've used Grogbot, I've used this, I've used that and I've reached my limit in 20 minutes or so. Well, yes, we would need more compute, but don't forget don't forget that models, especially newer ones, right, new release ones, they're not super efficient. Okay, so you're burning tokens in a way you should not be burning tokens. Second of all, for the majority of things that people are using right now, maybe you shouldn't be using a a frontier model. Third, look at this. For example, Nvidia Vera Rubin and 72 that will deliver up to 30% better throughput per megawatt. This is of course a huge jump from Blackwell and of course Blackwell was a huge jump from the hopper. So GPUs are becoming more efficient. Models should becoming more efficient as well. So yes, I agree that we need more compute, but a lot of people just don't understand that the existing megawatts that we have today, the existing gigawatts that we have today are going to become better utilized as well. If you can get 30x better throughput per megawws, same amount of megawws, just 30 times better throughput, of course, we are going to get better results as the end consumer. Yes, we would still need more compute, but don't forget that yes, you can have more compute, but you can also have just way more efficient GPUs and models to help with that. Continuing here on the same trend and this is an updated uh image from what we've covered I think couple of weeks ago. This was at 15 million I believe or the expectations were 15 million. This is codeex growth from 6 million to 25 in roughly 25 weeks and according to the person that made this image this is weekly. Okay, this is about active weekly users. Now are those increases real? real meaning. Is this happening because codeex is just so good or is this happening because we've all experienced that popup, right? Use codeex, try codeex out. So, we try it out once or twice. We might like it, we might lot and then we go away. I don't know. Might be a combination of both. But this is definitely something good because yes, more recently we have seen a lot of negative headlines surrounding OpenAI's growth. But it seems like OpenAI is reacelerating. We've also heard OpenAI well reports that say that OpenAI's advertising business is close to a billion dollars ARR and that's real ARR not AR R IN style. So of course all of this is great. This is also not me saying oh they're taking share from entropic. They're probably taking some share back from the overall pie but entropy is also growing quite fast. Open source model open way are also growing. So, seems like the pie is just becoming bigger and bigger. Which then moves me to this Mark Zuckerberg tweeting on X is the signal you need to get when you are maybe thinking about selling Meta or thinking about giving up on this name. Well, Muse code apparently is out of beta and now built to handle bigger, more complex engineering tasks. I'm not going to bore you with all of the details here, but fact that Zuck is tweeting this out. He's very excited about this. That's not the only thing that's exciting here. If you look at the pricing of Muse codes everyday usage under what is it? $5 per month. Let's call it $5 per month. Yes. Okay. This is not great usage, but then again, it's $5 per month. High usage, what is it? probably $15 or so. Power usage close to 50 bucks per month. The pricing here is insanely cheap. Now, of course, you might say, but maybe it's cheap because it's not a good model. Well, it seems like from what we're seeing online, it seems like this is getting very, very good. And we're not talking about their next model, right? Watermelon, which should be maybe released in the next couple of weeks or so. We're not even talking about the reported hatch product. talking about AI agents operating in WhatsApp or Instagram and ways to monetize that. No, we're purely talking about Muse code as of right now. And this this is basically the power move from Meta because if they believe that their model is so good, we know that the core business is insanely profitable. We know that they want to grow the adoption of Muse Spark, Muse code and the next models. Best way to do that is to release a very good model at a super attractive price. Now I know what people are going to say. Yeah, but couldn't this hurt then an entropic or an open AI? Don't we need these companies to succeed, right? For the whole AI ecosystem to be healthy to survive. Well, let's think about this both ways. On the one hand, Meta is a huge player already. Capex wise, it's it's huge. I would rather have a successful meta with already a very very profitable core business than an unprofitable frontier AI lab that might always have big question marks surrounding that company is that business model profitable they need to raise money how much money does need to come from Nvidia etc etc because when you think about this when you think about this if one of these companies go away AI doesn't go away okay this whole revolution does not go away because a frontier AI lab is going to go out of business or something like it, right? This whole move will continue to exist. The demand is just going to flow to different types of players. And if these different types of players are a Meta, a Google, an Antropic, an Nvidia, you name it, so be it. Now, that's not all of the bullish reason around Meta. I know I've been talking about Meta so many times. One might say I've been paid for it. I wish I wish a way for me to get paid is if the stock actually goes up. I think that would be the best case scenario for everybody else. But more recently, we have gotten more bullish, let's say, commentary from analysts. So, here we have one from Bernstein saying, "Digital ads, when does Meta overtake a Google search?" If you're still skeptical on AI investment returns with digital ads heading into September 2026, a quick look back at analyst estimates for this year's ad revenue growth just two years ago was less than half the growth rate Meta and Google search are expected to deliver this year. In 2026, Meta captured nearly half of every incremental digital ad dollar. Put these two points together and there's a path to Meta overtaking Google search as the largest destination of ad dollars before the year is out and yet it's the e-commerce names that are winning. And it's really not that crazy to think about Meta being well Meta's advertising business being bigger than Google here. If you look at last quarter for example, it is already very close. It has been growing faster than Google, right? over the past couple of years has been growing at a company and a growth rate of 20.4%. Smaller base, I get it. But still, it is catching up and it seems like it's going to catch up and become bigger sooner rather than later. So, that part is great. We already know that Meta Score business is becoming better and better and that will fund the AI buildout for them. But what comes next, the added services. So, we have personal AI agents that can autonomously complete tasks. That's the expected launch that could happen pretty soon. Now, them just launching doesn't mean it's going to be successful. Of course, it helps. It helps to have over three billion users across the ecosystem. It definitely helps, but it needs to be very good. It needs to work well with maybe outside apps, maybe with WhatsApp and Instagram together, Messenger, etc., etc. A Jenclick offering will need to be differentiated. How much better is this going to be from what Antropic came out with, from what OpenAI has, from what Grock has a Gemini and all the rest? Because in this world of AI, we are switching models on a daily basis. So unless you create something that's so good, so sticky, the ecosystem is great, I will just switch to whatever works best and whatever is at the best price. And so, yes, to reiterate what I said a couple of days ago, I do think that Meta is very similar to what happened to Google in 2025. It is just taking a little bit more time than what I expected. But hey, if we're going to have an amazing second half of 2026 and a very good 2027, fine by me. If at the end of the 2027, you tell me Meta is going to be at $1,000 per share, I'll say, "Great. Where do I sign?" just a question of if you have the patience to own these types of names, which which goes back to what we're going to talk about when covering SoFi again. But before that, a quick uh news item here. Entropic signs a $35 billion cloud deal with Nvidia backed Lambda. This is the second huge deal that they've announced in the matter of of a week or two, I believe. So Nvidia will supply the chips and is leasing the underlying 700 megawatt Texas data center campus from HUT 8 while Lambda will provide the compute to Entropic. Hut 8 had already disclosed a 15-year leases covering the entire campus which he said could represent a $20 billion facility but hadn't named the tenant. And Tropic also signed a separate $45 billion dealer earlier this month with Nvidia backed nscale for Nvidia compute capacity in West Virginia. So of course this goes against every comment that was made after Ian, right? Because Iran said, "Oh, we we've had an agreement with the Frontier AI model, but we cannot disclose who it is." And some people are already trying to throw out excuses. It's probably anthropic because well they're now in the quiet period because they're going IPO and so they cannot say it. Well, guess what? A couple of days later this came out. So clearly they can. So we might now speculate that it's not an entropic. It's probably not an open AI. So who is it? I don't know. And for everyone that said, "Oh, you don't know what you're talking about." I specifically mentioned Hut 8 during the stream. The fact that Hut 8 can announce big deals. Okay, even if some of these deals are collocation deals, who cares? They are getting more and more traction. And you know what? Traction is what you need in this buildout, especially when you are so small and it still is super early. So I truly hope iron catches a big big fish because yes, a lot of investors are in this name and I would love all investors to make money and iron. Yes, they do have a lot of potential. It's just execution has not been there in my opinion. The fact that the coco had to go on X and reexlain himself because he was there on the earnings call to reexlain himself and say, "Oh, people might have uh tangled some of these comments. So, let's untangle them." Mate, you were the one that was speaking out of a script. Maybe next time read it before you start speaking during an earnings call and explain a little bit more how your business is doing and stop throwing around that ARR number that really is not a real ARR number. Quickly touching on SoFi here and SoFi, yes, the stock if you own it, it's horrible, right? It's down 35% year to date. You would say opportunity cost, it sucks because, well, I could have bought a memory name. I could have bought this name. I could have bought that name. Would have done much much better than owning a SoFi. And you know what? If we could go back in time, okay, I would agree, right? I would say yes. I know it's going to be down 35% this year until September. So, let's not own this name. Let's only buy the names that are going up. But guess what? It it doesn't work like this. Now, what I will say time and time again is yes, there are plenty of people that talk about SoFi purely because it gets views. Gets views on YouTube, gets views on X. They don't really care about what happens with shareholders. They don't really care what happens with the business because guess what? The business has been doing extremely well. The business has been doing extremely well. guidance has been raised. Guidance has actually first stayed the same while the assumptions became worse. And now the assumptions that became worse, right? They're now expecting one to two rate hikes. While they did increase revenue guidance, still the same profitability. Now it does seem like we are heading towards one rate hike at least if not two for this year. So Sofa is again prepared for these let's say badcase scenarios which is good and despite all that the business is still performing extremely well. I still have the same DCF available to all of you out there. You just have to change of course the current stock price maybe some of the metrics that you want to change you can change those as well. Those are usually in blue. But you can clearly see that I'm still expecting this name to be much higher. It is undervalued today. Even with even with the bad expectations, we still think well we still expect this company to grow 30% through 2028 revenue-wise. We still think that EBIT margins are going to go up over the long run. Although right now we might still be sitting around 34 35%. That's fine. But the growth is still there. The growth is still there. When you look at the base case, yes, 30% growth is what we're expecting. Could they exceed it? It's very possible. EPS- wise, EPS is still expected to grow faster than revenue despite all of this environment that we're in. That's the thing with these types of names. And it's the same with a meta. Meta to the core is a very good business. The stock might not reflect it. free cash flow margin might not reflect it but to the core if you actually dig a little bit deeper it is an amazing business so far look at the stock might not reflect how good the business has been performing despite all of the macro headwinds but to the core SoFi has been executing and doing extremely well I don't know what some investors want management to do maybe they want management to turn into scam pumpers and make up metrics that don't really exist just to pump up the stock. That's not how SoFi rolls. And yes, SoFi does have a bank, so they cannot really play around with some weird shenanigans. To the core, SoFi is doing extremely well. I'm not going to spend more time explaining here. Just a quick reminder that the stock does not always reflect the business. Broken stock, not a broken business. And yes, I'll continue to cover this name even if we're at $18 towards the end of the year and the business has performed extremely well. The moment the business does not perform well, I will call it out just like I've called out every other name that I own or used to own because execution has not been uh there. And so all in all, that's about it for me in today's video. This is the pre-market action as of right now. Seems like another reddish day. It's September. You know what they say about September. It's not always the greatest month out there. Well, fine. Can we have a little pullback or a month where names stay flat or are down a little bit? Yes. And that's to the advantage of the long-term investors. And then we'll see what if we go through the next FOMC meeting and there is no rate hikes. H, that could be a big surprise. Anyways, that's all I've got for you in today's video. See you all in the next one. Bye-bye. [music] [music] >> [music]