I'm Still Selling

That's right, I'm still selling one of my longest held companies and most profitable investments, which is Apple. It's moved down to only a $6,000 position with $34,000 in gains. I'll be going over why I'm selling Apple and where I'm putting this money, as well as some thoughts on some of the biggest companies that reported earnings this week, ones like ASML. ASML stock moved up 5% after earnings, it was strong across the board. Then we have Microsoft that also beat on their earnings, but it sold off. We'll be looking at what investors are concerned about with Microsoft. Meta continued its win streak of posting strong earnings. Tesla stock missed on almost everything. What is causing Tesla stock to go up? We'll be discussing in this episode. We have Mastercard and Visa that both reported earnings. Both of these companies are on a roll, posting strong results. And we can't forget about Apple. This is after all the company that I've been trimming all year long. We'll be looking at this one as well. And then finally, while all this busy news was going on of the earnings week, we had semi-analysis looking more in-depth on the deep-seek narratives. All these different narratives that have taken the world by storm, that have caused trillions of dollars of market cap loss in the US markets. And they released an in-depth report debunking many of the claims. We're going to take a look in this episode. So we have a very full episode to get to. Let's go ahead and start off with the first companies to report earnings this week. Out of all the many companies that reported, one of the first major ones was ASML. And this is a key position in my portfolio. It's the newest addition to the passive income portfolio. ASML reported very strong earnings. They beat on the top and bottom line. They posted record high free cash flow in a single quarter. And overall, the biggest highlight this quarter was one of these KPIs that investors track specific to ASML that is particularly important. It is the bookings. And we have this metric here in Qualtrant. It's called net bookings. This is basically how many companies have actually put orders in with ASML. What is the amount of business that they're locking down every quarter? The expectations for Q4 were about $4 billion worth of bookings. So right around there. They came in with $7.09 billion of bookings. So they almost doubled it. Obviously, that was way above investors' expectations. That caused the stock to jump around 5% to 7%. So this was a huge quarter in terms of net bookings. But again, this wasn't the only metric that was good. Across the board, ASML showed a lot of strength. This was a great quarter for the company. And they also showed a lot of optimism for the future. When the CEO was asked about the impact of deep seek or demand in the future of AI, he seemed very unbothered. It didn't seem to affect or really impact the CEO's view at all, nor do I think it should. As they say here, the company remains optimistic about its product pipeline and market position as it sees progress in 2025. They continued on with their strong guidance that they had prior to the deep seek news. The CEO of a very similar company, KLA Corp, said similar things, but more directly. He says, We are aware of recent revelations of deep seek and the implications that it portends a diminished demand for advanced semiconductors in support of AI infrastructure buildout. As it pertains to the demand environment for advanced semiconductors, we see no reason to believe that increased compute efficiency in AI will have an impact on the advanced demand environment in the foreseeable future. He does not buy it. He doesn't buy the idea that having better efficiency in these chips will reduce demand. Neither does the CEO of ASML. Neither does Microsoft. Now, you may say that these CEOs will just talk positively about their companies, but in most cases, if these CEOs actually see, and they model out, impacts on their future, they have to relay that to investors. CEOs frequently express concerns on impacts of demand, but none of them are doing that for the deep seek news. ASML's margins went up, their profits went up, their revenue went up, their booking went up, and they reiterated their strong guidance. As a result, the stock jumped from around $670 to $750. But that's where things get interesting. The pricing of this company. When I look at ASML today, it jumped up around $70, so it could look like it's getting a little expensive, like now is not a great time to buy it. But again, if we look at this, we see something interesting here. First of all, ASML is a little bit slower than most companies in filing their numbers with the SEC. Every company gets around two weeks after their report to submit their numbers to the SEC. When we look at it right now, the numbers are still on Q3, so they'll be updated within a week to Q4. ASML will be updated to have $9 billion in free cash flow in Q4, bringing the 1% free cash flow yield that looks expensive today to 3.4%. So all of a sudden, ASML will get cheaper overnight. With this massive growth in free cash flow, this company is cheaper than it was a quarter ago even though the stock price has gone up. I still believe the company right now is a great risk and reward. And that's why during my trade this morning, I sold $5,000 out of Apple and bought $5,000 of ASML. Now we've gone over ASML, and I'll go over Apple later in this episode. Next we get to Microsoft. This is another large holding in my portfolio. It's been a great one so far, but Microsoft was a bit of a disappointment with the after hours reaction. If we look at what they did, Microsoft beat on their revenue, they beat on their earnings per share. When we look at the most recent quarter, we can see that the productivity, intelligence cloud, personal computing, all three segments have grown to record highs. Microsoft continues to grow at a brisk pace. When we look at this on a trailing basis, this is what the earnings per share looked like. Zooming in over just the past five years, you can see it's growing at 12% per year plus they're paying a dividend on top of that. That's a company that should be compounding around 15% per year. So why did Microsoft trade down? Well, when the numbers for the most recent quarter looks great, it has to be with their guidance or their commentary. The commentary is that the guidance was a little bit on the softer side, a little bit on the low end of expectations, but overall, I thought the commentary and the guidance was still fine. Satya Nadella said that he's extremely bullish on AI scaling and infrastructure. He cited doubling the data center capacity in three years. He's making massive investments into AI. Nadella discussed the expansion of Microsoft's tech stack, particularly in the areas of Azure, data services, and AI tools like GitHub Copilot. He emphasized the extensive use of AI enhancing employee productivity, streamlining workflows with a Microsoft 365 Copilot. So he's talking about AI being mixed into everything, growing out the data center, growing out the cloud business. Overall, Microsoft's earnings are good. The company is at a high valuation, but not an extreme one. Microsoft only trades at a 30 PE ratio over the next 12 months. And based on 2026 estimates, it trades at a 27. This isn't super cheap, but for a company like Microsoft, it's also not that expensive. I think right now, Microsoft's at a decent valuation. For me, I consider it a hold. At the same time that Microsoft reported, we also had Meta, and this was possibly the highlight of the week. This was really an incredible earnings report. Let's look at a couple things here. First of all, the top line revenue grew at roughly 21% year over year. Last quarter, doing $48.3 billion. Incredible growth from this company. They're still maintaining that super growth. Growth above 20% is in a category of its own. That's in an elite category, especially for a company this scaled. So Meta is doing something right. They're making their products way more efficient, and they're driving extremely strong revenue growth. Another key performance indicator that was equally impressive was the family daily active people. This is a metric that tracks how many people are using one of Meta apps at least once a day. And this grew 5% year over year. The growth rate is starting to accelerate from the previous quarter, going up from 3.29 billion to 3.35. This is incredible how they continue to grow this metric. The free cash flow per share grew 25% year over year. The net income increased 59% year over year on a trailing 12 month basis. The earnings per share increased 60% year over year on a trailing basis. Mark Zuckerberg was also enthusiastic on the call, expressing all the AI developments, personalized AI experiences. He particularly focused on the Llama4 model, saying that the DeepSeek news illustrated why it's so important for us to have our open source models. He said it confirmed that his approach to being open source and developing the common standard is the correct approach. In terms of outlook for Meta, they anticipate revenue growth driven primarily by user enhancement, engagement, monetization. This is the same playbook they've been using for the past couple of years. They're using all of this AI to enhance their product, make it more addicting, make the ads better, make people want to spend more time on their apps. And it's working with an incredible level of effectiveness. He announced that he was going to spend a fortune on CapEx. Now in most cases, when a CEO comes out and says, I'm going to spend all of our income and roll it back into CapEx, investors don't like that. They want the money themselves. But in this case, I thought that investors would be okay with it. I thought they would tolerate it because they're on board with his vision of the future. And that's exactly what happened. Even though he announced incredible levels of CapEx in the future, they just shrugged it off. They're excited about the future. I don't own Meta, and it's one of the companies I wish I stayed in. I sold out of Meta early 2021. As soon as Mark Zuckerberg started talking about the Metaverse, as soon as he started doing the demos of him in this three-dimensional world with like six monitors on his screen, I didn't buy that vision at all. And neither did Wall Street. The stock subsequently fell, but I didn't buy the Meta dip. I didn't buy back in towards the bottom. I was busy at the time buying Netflix, Amazon, and Google. And that's the problem when a lot of great companies sell off at the exact same time. Either you can buy a little bit of all of them, or you can pick and choose which ones you want to go into. But I've been so impressed with the fundamentals of Meta over the past two years. In terms of the valuation of this company, when we look at the PE ratios, on a trailing basis, it trades at a 28. On the next 12 months, it trades at a 27, and then on 2020 success estimates, it trades at a 24. The company's not the cheapest it's been throughout its history. It's actually trading at an elevated valuation for Meta's history. But I believe that's deserved. Meta is a better company today than it was five years ago. It's less risky. It's more established. It's more profitable. It has more applications with more users. It's more ubiquitous. Its network effects are greater. So it makes sense that the PE ratio expands a little bit to coincide with the progress the company's made. I think it trades at a very reasonable price, given its fundamentals and situation. Now next we have Tesla, which is perhaps one of the more confusing earnings report of the week, because typically when a company posts the earnings, and every single metric is lower than what the expectations were, that typically under normal circumstances coincides with the stock dropping a little bit to reflect this new information where they did not meet anyone's expectations. But that's not what happened with Tesla. Tesla is a different breed of a company that has a different breed of a shareholder. Tesla investors will be quick to tell you that that's old news, even though it was reported just minutes ago, it's old news. And what really matters with Tesla is, of course, the future. And that was the case this quarter as well. Tesla reported their earnings and their revenue number did not meet expectations. When we look at it over just the past three years, you notice a trend that it's a little bit flattish. Tesla, in fact, is having difficulty growing revenue as of right now. Now, in this case, I'm a little bit forgiving because when we zoom out five years, Tesla's grown revenue a lot. I get more concerned when it's been like two or three years that a company hasn't grown revenue. And that's not the case with Tesla. So it's going through a little bit of cyclicality. The revenue is flat right now. But either way, Tesla did miss on their revenue number. If we zoom out and smooth this out on a trailing basis, this is what the revenue looks like. It looks a bit similar to Apple's revenue, where it's been smooth for a couple of years. But that's not the biggest problem in the world if it's followed by increases in revenue. And we've seen that before back in 2019. They also missed on the bottom line. The earnings per share and net income estimates were a miss, but they also missed while making $600 million in market to market gains on digital assets, meaning even with the benefit of their crypto going up in price, in valuation, this was still a miss on the earnings per share. In terms of free cash flow, they posted $2.1 billion in the quarter. That is a 1% change over Q4 of 2023. None of these numbers are impressive at all, and Tesla investors know that, and the stock rose anyways. So why did Tesla go up 4.4% the day after earnings? Well, like you may point out, Tesla's priced for the future, not for the past. Even though this quarter was a day ago, it's old news in the minds of investors, and they're looking to the future, specifically Elon Musk's comments. Now when I look at the AI summary in Qualtrim of what he talked about in his long opening statements, the summary is a bit interesting. It says Elon Musk expressed significant optimism about Tesla's future, emphasizing autonomy and humanoid robotics as key growth areas. He highlighted the exponential progress in full self-driving, and foresees Tesla becoming the most valuable company globally. Musk stressed the transformative potential of autonomous vehicles and humanoid robots. We have significant optimism, the most valuable company in the world, and humanoid robots. When we actually look at the transcript itself, this isn't far off. In fact, it's an understatement of the level of optimism he expressed. Elon Musk said, quote, I mean, there is a path where Tesla is worth more than the next top five companies combined. If you didn't think that was enough, he also mentioned on the call that Optimus has the potential to be north of $10 trillion in revenue. No, that's not a typo. That's not a mistake. And he didn't change that statement. That is a real statement by a CEO on an earnings call for a company that just posted a run rate of revenue less than $100 billion, that one of their product lines that has yet to exist yet could be worth $10 trillion in revenue. He says, quote, it's really bananas. Okay, it's really starting to come into focus now. When you have a CEO as big and influential as Elon Musk saying these things, it's going to have a positive effect on the stock. When Elon Musk was asked about their roadmap and who they're facing off with, they say that there is no company in the world that is good at real world AI as Tesla. I don't even know who's supposed to be in second place. Like you say, who's in second place for real world AI? I would need a very big telescope to see them. So he's basically saying that Tesla is the only game in town, the only one doing these real world AI things like autonomous driving. These are the type of statements that Elon Musk was saying on the call, that Tesla is going to be worth more than the top five companies in the world. The top five companies in the world combined did $335 billion of free cash flow in the past 12 months. Tesla's done four. He says that he can't see the nearest competitor with a telescope while Waymo's actively doing 200,000 rides per week, testing their technology in 10 different cities. Zooks from Amazon is going to be tested and launched in Vegas this year. There are multiple competitors that Tesla faces in the full self-driving game. The only explanation I have seen from Tesla investors compared to Waymo is that Tesla works with a neural net and vision instead of using LIDAR along with vision. But when customers decide whether or not they want to ride, they're not thinking about the technology behind it. They're just thinking of whether or not it'll get them from point A to point B safely, and Waymo's accomplishing that. So what Tesla's stock really comes down to is just seeing this narrative, believing this vision of the future. These type of stocks partition people into two groups, the group that believes and is willing to pay the price today for the future vision, and the group that believes it's too risky or too unlikely. But right now, in terms of the valuation, the Ford PE ratios, the free cash flow yields, none of that matters to Tesla investors. The prize is whether or not this company can establish global dominance in transportation through full self-driving. And whether or not this company can release millions of humanoid robots and generate trillions of dollars in revenue. As the investor, you have to make up your mind of whether or not you think that's going to happen. JP Morgan released a memo on their analysis following Tesla's stock's move, and they say it's not clear to us as to why Tesla's stock traded 5% higher after the market closed on Wednesday. Although we do have some leading theories. Perhaps it was management's statement that it had identified achievable paths to becoming worth more than the world's five most valuable companies, taken together, more than $14.8 trillion combined. That's the market capitalization of Apple, Microsoft, Nvidia, Amazon, and Google. Or maybe it was management's belief that just one of its products has by itself the potential to generate north of $10 trillion in revenue. It may have even been related to management's guidance for 2026. Of course, there's no financial targets provided, but it was said to be, quote unquote, epic. And for 2027 and 2028, quote, ridiculously good. What does seem clear is that the move higher in Tesla shares bore no relation whatsoever to the company's financial performance in the quarter just completed, or its outlook for the growth in the coming year. Now moving away from Tesla, we get into MasterCard and Visa, which have far less spectacular of earnings. There's not really any bells and whistles. And shockingly, the CEOs of MasterCard and Visa, which did post very strong earnings, both of them beating their guidance, both of them beating their revenue and earnings estimates, both of them are compounding machines executing on their strategy with precision. But in either case, neither of the CEO used the terms epic. They just continue to execute on very solid business fundamentals. Both of them face all different levels of competition across the globe, but they continue to adapt over time. And the huge tailwind of trillions of dollars of cash still in existence being used is slowly making its way to digital payments. I'm up around 50% on MasterCard. It's a $35,000 gain. And it's still my second largest position in the passive income portfolio. Now we get to the most recent of the big tech, which just reported yesterday, which of course is Apple. Apple's report was a mixed bag. They bid on their top and bottom line, which they almost always do. But there is a little bit more to the report. Apple gave okay forecast. Investors were enthused about that, and they bid the stock up 3% after hours. So the initial reaction was very strong. Investors were a little bit enthused. They came in today with the stock up 3%, but Apple has since sold down to being up only half a percent now. Some problems have been highlighted in the report, specifically in China. Revenue from China plunged 11% to $18.5 billion in the last fiscal first quarter. Analysts had anticipated $21.6 billion, so that's a pretty big miss from their China revenue. Tim Cook said during the conference call that more than half of the decline in China stemmed from inventory issues. So we have different commentary on here giving a little bit of context, but it does raise some questions. Are consumers in China preferencing their own phone companies like Huawei over Apple? Is this a trend that's going to continue on in the future? Apple like many companies is playing catch up with AI. We were promised Apple intelligence that would renew the iPhone, give it an entirely new feel and make it feel super smart. Most of the feedback that I've seen online is that people are a little bit underwhelmed with Apple intelligence. And that's a problem because other competitor phones like Huawei in China already have intelligence built into them. They are AI enabled and Apple intelligence hasn't even rolled out in China. So part of the issue is Apple simply being late to the game. Right now, Apple's being super slow and cautious, meticulous. The next few quarters will test whether it can balance its cautious approach with the market's hunger for AI innovation. Overall, year over year, Apple just grew their revenue by around 4%. Now based on the 3% bump this morning, I sold another $5,000 of Apple. And I'm doing this for a specific reason. I don't think that Apple's a bad company and I don't think it's unreasonable to include Apple in many different portfolios. If you have a more conservative portfolio where you're okay having a little bit slower growing companies that are highly profitable and stable, Apple fits really well. It fits along with Coca-Cola, McDonald's and many other great companies. But in my portfolio, I like growth that is above the rate of the market. The rate of the market is 8% and I'm looking for companies that can grow higher than the market at 12 to 15%. So the growth just isn't there for my portfolio. When you combine that with the fact that Apple's not on a dip, it's not trading very cheap right now, it's trading at valuations and forward multiples that are elevated based on its history, I believe it's best for my portfolio to transition that money into companies that have a little bit more upside. And I did that with another $5,000 this morning. Now as we got through this busy week of earnings, we can finally take a look back on the news that rattled the markets on Monday. The DeepSeek news, the release of this new innovative model from China that only cost $6 million to train. It caused an eruption of stories and articles, of theories and conspiracies. All these different things happened on Monday, causing US equities to lose combined worth of trillions of dollars. And now that we've had a little bit of time to investigate the DeepSeek news, we're getting a lot of new information. SemiAnalysis, for example, did an entire investigation on the DeepSeek narrative that took the world by storm. And in this article, there is a lot of debunking. Now before getting into any of the debunking, they do acknowledge the skill involved behind what DeepSeek accomplished and the extravagance they've gone to in putting together this team. DeepSeek has sourced talent exclusively from China, with no regard to previous credentials placing a heavy focus on capability and curiosity. DeepSeek regularly runs recruitment events at top universities, where many of the staff graduate from. Roles are not necessarily predetermined, and hires are given flexibility with job ads even boasting of access to 10,000 GPUs with no usage limits. They are extremely competitive, and allegedly offer salaries of over $1.3 million USD for promising candidates, well over the competing big tech Chinese companies and AI labs like Moonshot. So there is a lot of skill involved in what they're doing, but still some of the numbers don't add up. The first of course was the major claim that they trained with only $6 million. We believe the pre-training number is nowhere the actual spent amount on the model. We are confident that their hardware spend is well higher than $500 million over the company history. To develop the architecture innovations during the model developments, there is considerable spend on testing new ideas. Multi-head latent attention, the key innovation of DeepSeek, took several months to develop and cost the whole team in man hours and GPU hours. The $6 million cost in the paper is attributed to just the GPU cost of the pre-training run, which is only a portion of the total cost. Excluded are important pieces of the puzzle, like R&D, TCO, the hardware itself. For reference, Cloud 3.5 Sonic cost tens of millions to train, and if that was the total cost Anthropic needed, then they would not raise billions of dollars from Google and from Amazon. It's because they have to experiment, come up with new architectures, gather and clean data, pay employees, and much more. So basically, to arrive at this $6 million number, which is what caused the sell-off in all these companies, they had to focus on one part of the bill. They just ignored everything else involved, they went to the one part of the pre-training cost and said that's how much it cost to train it. In reality, it cost DeepSeek closer to $1.6 billion to train their model. That's an incredible difference between the quoted number and reality. So this is a long way of saying that that $6 million number is bogus. It's a number that was pulled out of the entire process to try to make it look like this was far more effective and cheap than it actually was. But that's not the only thing that was highly misleading about the reports that we saw earlier this week. The next one is the simple performance of DeepSeek. You heard article and article after it was groundbreaking, it was so much better than what the US companies have come out with, it was far and away better. And that's not necessarily the case. DeepSeek's v3 is no doubt an impressive model, but it's worth highlighting impressive relative to what? Many have compared v3 to GPT-4.0 and highlight now that v3 beats the performance of 4.0. That is true, but GPT-4.0 was released in May of 2024, AI moves quickly and May of 2024 is another lifetime ago in this algorithmic improvement. Further, we're not surprised to see less compute to achieve comparable or stronger capabilities after a given amount of time. Inference cost is collapsing, and it's the hallmark of AI improvement. An example is small models can run on laptops and have comparable performance to GPT-3, which required a supercomputer to train multiple GPUs to inference. Put differently, algorithmic improvements allow for smaller amounts of compute to train and inference models of the same capacity. And this pattern plays out over and over again. This time the world took notice because it was from a lab in China. So this story of models becoming cheaper to run and being more powerful than their predecessors is not new. This has been happening since the beginning of LLMs, and it's expected to happen in the future. To be clear, DeepSeek is unique in that they achieve this level of cost and capability first. They are unique in having released open weights. But prior, Mistral and LLAMA models have done this in the past too. DeepSeek has achieved this level of cost, but by the end of the year, do not be shocked if costs fall another 5 times. We are expecting the same thing to happen all throughout 2025. So what happened was impressive, but not nearly as impressive as what it was talked to be. Further along, we look at the actual performance of DeepSeek, and we compare it not against but against newer leading models like what Google has recently released, and it's not any better. While there is a frenzy of hype for R1, the $2.5 trillion US company released a reasoning model a month before for cheaper. Google's Gemini, Flash 2.0 Thinking, this model is available for use and is considerably cheaper than DeepSeek's R1, even with a much larger context length for the model through the API. On the reported benchmark, Flash 2.0, which is Google's latest model, beats DeepSeek's R1. Google only released three benchmarks, so it's an incomplete picture. Still, we think Google's model is robust, standing up to R1 in many ways while receiving none of the hype. So when you stack up DeepSeek against an old US model, which is ChachiPT4, it makes it look comparatively better, but when you pin it against the latest model of a leading company like Google, it's not comparatively better, and Google's cheaper. Overall, this is the problem with shock narratives and shock stories in the market. The market festers on them, they sell out of stocks in a rush, and what we see here is that virtually every part of the story, every single aspect of it, has been overstated and overblown. DeepSeek spent orders of magnitude more money than what was reported on training their AI. DeepSeek's model is an iteration that fits well within the norms of AI development. DeepSeek is better than ChachiPT4, but comparatively the same as Google Flash. So it's not a shock at all that stocks are back to where they were before DeepSeek. That's going to be it for this episode, hope you enjoyed, see you in the next one.