Meta是强力买入标的，因为其在广告和新模型方面的真实AI投资回报率超过了法律和资本支出风险。

Meta is looking to spend up to $145 billion on AI capital expenditures this year. For the first stock, I want to move to Meta Platforms. I think this is one of my favorite stocks this year. The arrow was punished , and that was justified. I think the market always finds reasons for that, and there are many of them. I mean, many could argue that this company's increased capital expenditure is one of the reasons for its punishment, wouldn't they ? We have seen capital expenditures reach approximately $130 billion to $145 billion for this year. The second thing is that, unfortunately, there are a lot of those legal issues that occur with Meta. We've had a few updates in the past few weeks that suggest things may not be as bad as the market thinks. But the market is still a little scared here and there. I mean, in their most recent second-quarter earnings , legal costs caused the company to lose money or not achieve its expected earnings per share. This is a story that will continue to be repeated over and over again. Therefore, there are those risks related to legal proceedings and capital expenditures.

Now I want to explain why I'm so excited about Meta. Meta is one of the companies that sees a real return on investment in this artificial intelligence story. I believe Mark Zuckerberg mentioned during the call that he expects Meta to be the fastest-growing advertising company of all time, surpassing some other giants in its growth . That's very exciting. Another point Mark mentioned is that the AI-dedicated computing capabilities they invest in greatly help their core business. Isn't that so? If you have a better recommendation system , that system will ensure you stay in the app for a much longer time. It will show you what you like, then it will confront you with an advertisement. He will confront you with an advertisement that he knows in advance you will interact with. Therefore, this will lead to increased advertising efficiency. It will also increase the time you spend on the platform, which increases the likelihood of you being exposed to more ads, and so on. This creates a final impetus for more data. " Oh, this person is watching this to stay here longer." Perhaps if you do this, they will stay here much longer. Or for the next person who has a very similar viewing pattern, I can recommend this because they like similar things. Therefore, it generally helps to support that basic activity. The second thing is that with all this AI revenue , Mark really wants to sell " superintelligence". He says, " I can sell my computing capabilities." "I'm investing in all these data centers, I've built all this computing power, and I can sell it at a great price, just like SpaceX does." "But I do n't just want to sell computing ." "I want to sell intelligence ." "And with intelligence, I sell computing, and I sell code." More recently, what we've seen from the Meta team is their continued release of new "Meta Muse" models. Just yesterday, or earlier this week, they released a new programming solution. Based on rapid testing, they are already outperforming many leading models. This shows that the idea or vision that Mark has is not perfect, but it contains some truth. If they can get these models to a pioneering level, there will be demand for that computing and those tokens, and I don't think the market has fully priced that in yet.

Ah, Rachel, I didn't want to get too caught up in revenue growth numbers because I think there are a lot of places where we can see that. I just wanted to explain this high-level thesis about the position of the data center market and why this computing power is so powerful for Meta and its shareholders.

Yes, there are a few other things I'm keeping an eye on regarding Meta at the moment. As you know, their dedicated MTAA segment already powers recommendation engines that serve billions of users daily, something many investors may not know . But the other thing that just happened is that Meta launched its new "New Spark 1.3" model yesterday for complex programming and multi-step agent tasks, and they are primarily targeting developers who build standalone tools and automated workflows. Therefore, this model deals with multi- file projects, software engineering, a lot of visual interface generation , as well as the workflow of AI agents in general. Meta will offer two main levels of access through the Meta Models API, in addition to third-party aggregators. But what's interesting about this is that it's linked again to advertising activity , isn't it? Because much of what Meta does goes back to feeding the source from which it generates 95% of its revenue, which is advertising. And I think " New Spark" will be a key part of that. As you know, Meta plans to use the capabilities of the agency model to truly improve ad targeting, automated creative production, and conversion rates. Without going into highly technical details, "New Spark" operates under the umbrella of an advertising auction. Therefore, it analyzes user interactions, texts, images, and videos. Okay, so what does this mean for advertising? Essentially, the company can predict user structure with greater accuracy. What does that achieve? This improves ad relevance and enhances engagement metrics for advertisers. Um, as you know, this is just a high-level overview. I'm sure we'll talk more about this update in a future video, but I think this is really important to keep an eye on if you're a Meta contributor or if Meta is on your watchlist.

Um, and another thing I will point out is that they revised their capital expenditure guidance upwards three separate times. Um, as you know, they had their original domain, then they raised it again in April, and then again in July. Hence we got the figure of $145 billion . So, basically , Meta continues to underestimate the size of the expansion required. This means that if they revised their infrastructure spending upwards three times in one year. This is a very strong argument that the physical infrastructure aspect of AI trading , including energy, land, and cooling, may remain a bottleneck for much longer than many investors currently expect. And I think that creates a real buying opportunity , not just for Meta, of course. Meta is the company we are talking about in this case, but the same applies to many other companies as well.

Two completely different angles of the same wave of spending on artificial intelligence. As you know, Meta, the company that Jose covered, is clearly a direct bet on artificial intelligence. It is one of the largest companies in the world.