저자: MeetKevin · 외국인
요약 보기
AI 대형주(메타·테슬라·마이크로소프트)의 실적을 분석하며 AI 투자 과잉을 우려함. 메타는 광고 성장으로 밸류 매력이 있으나 R&D 지출 급증이 효율화 주장과 모순된다고 봄. 테슬라는 신차 단종·FSD 구조 변화·대규모 테라팹 투자 계획 등으로 리스크가 크고 PEG 6으로 약 56% 하락 여지를 언급. 마이크로소프트는 AI 추론 비용 상승으로 마진이 압박받지만 PEG 1.77로 약 50% 상승 여지가 있다고 봄. 전반적으로 빅테크의 AI 베팅을 'drunk'하다고 비판함.
다듬은 글
Over $6 trillion of artificial intelligence companies in America are absolutely drunk ducks. I'm sorry to say it, but I went through everything I can get my hands on for Microsoft earnings, Tesla earnings, and Meta earnings, and I have to say I'm starting to grow increasingly concerned that nobody knows how the hell to make money from artificial intelligence, and they're all losing the plot.
I want to be clear about this. I've got exposure to Tesla in retirement accounts. I've got exposure to Meta. It's actually been a stock that I thought has been cheap, and I mean obviously they're doing really well here in the after hours are up like 7%. This is great. Uh we got a course coupon code expiring tonight, but I don't want to talk about that.
It's my birthday. I want to get these bottom lines out of here. I'm feeling lost going through these companies because I feel like all of them have absolutely lost the plot. Let me explain this very carefully. First, I'm going to start with the plot loss at Meta. Then I'm going to hit Tesla, and then we'll hit Microsoft.
So, Meta. The biggest takeaway that I got from the Meta earnings call was Zuck telling us that we are now seeing that one person with artificial intelligence can do what an entire team used to do before that. And as a result, we are massively simplifying the team structures that we have at the company.
Okay, great. So, then I'm like, "Wow, Mark, that sounds fantastic. You must be saving so much money. You must be doing so fantastically on your earnings because you guys are absolutely killing it with actually selling your product. In fairness, quick note here, I don't want to bore you, but not only are they delivering 18% more ads fourth quarter year over year?
But they're selling them at 6% higher prices. So both levers at Meta for income, ads, that's how they make money, right? Skyrocketing. That's why the stock is up because their core business it's a rocket ship. It's amazing. And as an advertising play, they're cheap and it's a great deal. But then I'm like, wait a minute.
Now, we should be seeing way more earnings per share growth because if you could be so much more efficient on your teams, well, then your numbers should be pretty freaking remarkable on the actual earnings pages. And I'm not trying to be a jaded numbnut here. I own the stock. But then I look at this.
I got yes, revenue up 23.7%. Excellent. That's great. I'd love to hear that. Cost of revenue only up 23.7%. That's way better than Microsoft. Microsoft's margins went down here. But wait, what is this? 40.6% more money spent on research and development? Well, wait a second. If we're getting so much more efficient with the teams, why are we blowing all of a sudden compared to last year 40.6% more on research and development?
Something doesn't reconcile here. It's either we are getting more efficient and now we're just deciding to blow our money on the metaverse or glasses or whatever, which is entirely possible, or building super intelligence. Or it's not true that they're getting much more efficient with what one person with AI can do.
Now, I believe that it's likely Mark Zuckerberg is just kind of lost. And this is normal. He did this with the whole metaverse thing, which was a complete flop and a disaster. The advertising business makes so much money though that even though he ends up spending some of the stuff like a drunken sailor while he's looking for a solution of how they're going to make money with these new super intelligence models, he's probably wasting a good chunk of that money.
So, you almost have to I'm picking from these three companies, you almost have to look and go, "Okay, well, maybe Tesla's doing better. Maybe they're not going to sell uh spend money like drunken sailors." And then I heard things on this earnings call that I've never heard them say. Well, frankly, I shouldn't say never, but I haven't heard them say in years.
Tesla now, rather than just hinting at doing a terrafab, is actively suggesting that they're considering raising debt to build a terrafab. That's because even though they have $26 billion of free cash, and we know that their energy margins are 29%, and they're expanding their mega cap development or or megapack development processes, which is fantastic.
They make great money with that. They've got cash. Building a mega or terra factory instead of a gigafactory or megapack factory, building a terra factory, it's probably going to cost Elon, if it's similar to the size of Elon or TSMC, 50 to 100 billion dollars. After the CapEx doubling that they expect to spend that they expect to spend this year, which part of will be rejiggering the Model X and Model S lines to make Optimus robots, and part of will be for cyber cap production, which we'll talk more about that in just a moment, those expenses will lower Tesla's available cash probably to around $20 billion on hand as free cash, net of bills, by the end of the year.
But, $20 billion doesn't buy you a fab. Because the reality check is that when Intel in 2022 started building their two fabs, it took them $28 billion, and it took them almost, well, frankly, I mean, they're still building them. It's taking them about nine years to get them done. Taiwan semi phase one took five years, phase two won't be done until later this year, and Taiwan's complaining that it's four to five or TSMC is complaining it's four to five times more expensive than just building in Taiwan, and it cost them 65 billion dollars.
So, this idea that Elon's going to go for a terrafab that they don't have the money for is to me a little bit of an indicator that Tesla also is trying to find All right, we're in an AI race, what do we want to do? Solar panels in space, that's big. Autonomous vehicles, don't have one yet, but we're going that direction.
Autonomous humanoid robots, don't have them yet, but we're going in that direction. And in order to pay for or like to create our fleet of autonomous human robots, which the God model of Grok is now going to run because we now have a packed or contract agreement to license Grok to run this God army that we don't have yet, uh that God army is going to need so many chips that we're going to need our own fab to build not just logic chips, but memory chips.
And to me, I'm starting to feel like are these CEOs just distracted, not understanding that if they focused on just the core business, they could probably make a lot more money? Instead of focusing on a heads-up display with Reality Labs at Meta or super intelligence models, while at the same time your executives on the earnings call are going, "Hey, um you know, we're actively trying to use and direct workflows to less expensive models so we can cut costs on the more expensive models," which I'm like, "Okay, but here you are trying to blow all this money on super intelligence, which probably isn't going to happen anyway." You know, as as a reasonable, I think, level-headed investor, I'm like, "Man, these people collectively run $6 trillion of companies, and they're freaking lost." Now, Microsoft may be the least lost out of all of them, but the problem is they're kind of held at the b a l l s by other companies and that's why their margins are compressing.
Yeah, total revenue is up 16.7% but their servicing costs for that revenue are up 25%. So, their gross margins actually compressing. So, their costs of goods sold, those inference costs are rising way more than what meta seem, probably cuz they're trying to keep up with Anthropic. You know, there are rumors that the people over at Microsoft are freaking out that they've always been trying to get co-pilot to just keep up with the other AI models and now people are saying Claude is eating their lunch integrating with Excel.
In other words, what do you need to pay a co-pilot subscription for if you could just use Claude? So, they're actively raising prices on their existing customers blowing the rest of the money that they have. Now, in fairness, Microsoft and Meta do have money available to do this. They Microsoft has 89.4 billion dollars of cash.
That's enough money to pay for all of their bills. And you can throw in another 56 billion dollars for receivables which which starts taking off a chunk of some of their long-term debt. Meta is in a great place as well with cash. I've got 81.5 of cash including uh you know, if you include receivables, you're at over 100 billion dollars of of liquid here.
I've got 41.8 billion dollars of bills uh and then I've got long-term debt of 106. So, it it's going to take me about a year of cash flows to actually pay off all that long-term debt if I wanted to and I used my cash to eradicate my debt. They just took on another 29 billion dollars right here of debt, basically 30 billion dollars of debt.
You know, they're they're pre-spending their future cash flows on these ideas that I'm not sure are actually going to go anywhere. And so, then we have to ask ourselves, well, what are the companies going to end up decaying down to when and if some of these big expense plays don't end up working out?
And so, that's where I like looking at valuations. So, let's go ahead and do that. Let's go just pull up the latest valuations that we have. So, I'll pull up the stock tab. You get this if you are a member of the Alpha report. You can get lifetime access still. We're going to move to a monthly recurring model at some point this year.
But anyway, quick look at Microsoft. Last time I looked at Microsoft, they were at Actually, as of today, they're at a 1.77 PEG, which is pretty inexpensive. You know, they probably have about 50% upside for a fair value there. If I look at Meta, I put them at a 1.29, which is basically suggesting they have a 2x upside on their valuation.
And if I look at Tesla, they have a six PEG, which unfortunately a six PEG for I'll give them a service or even though they're substantially lower margin than these other businesses, Tesla has downside of about 56%. So, the valuations are pretty clear that these companies are valuation-wise probably going to you know, Microsoft and Meta, they they both actually have upside.
Even though Microsoft is slipping down, people are concerned that hey, like your margins are compressing, you're falling behind, and those revenues might not sustain. That's a risk with with Microsoft. And if their earnings get written down, then their valuation growth won't be there. Meta on the other hand has this ultimate bailout tool of advertising.
Because of artificial intelligence, they're actually making more money than ever before. Hand over fist more money with advertising. They're killing it with advertising. Elon on the other hand has taken the largest transition and put the company I think into the riskiest position Tesla has ever been in.
I mean, look at this. New car models, dead now. We're not doing new car models. We're just going to focus on autonomy because 90% of all vehicles in the future or miles driven will be autonomy in the future. We're going to focus on killing the SNX. We're not doing those anymore. So, buy one now. You know, going away sale.
And then we're just going to focus on the Optimus. Then we'll focus on a Terafab and Robo-taxi, and we're going to leverage Robo-taxi revenues to uh finance our Terafab. They said this on the earnings call. All this kind of stuff, not worried about it right now. Lifetime FSD, gone, cuz we want the annual recurring revenue.
SNX, gone. New car models, forget about it. This makes me a little nervous because what we're doing it with Tesla is we're coring out the business that's doing really well to focus on things that we don't know will do really well. Look, I love that we got disclosure on how many people are paying monthly for FSD.
1.1 million, that's 330,000 monthly paying customers, that's $32.6 million a month, that's almost $400 million ARR. That's fantastic. However, we're also going to see margins decline as we heard our CFO tell us, because they're not going to get these big upfront sales of FSD anymore where people are paying them 7,000, 6,000 or whatever in a basically 100% margin FSD product.
So, you're not getting those lump sales anymore. Now, of course, they butter those out over time, but still they might recognize that revenue over 2 years or 3 years average lifetime that they think somebody would on average use that car. Average, could make sense, 2 to 3 years. However they justify it, so you take a lot more revenue in.
That's exactly you recognize a lot more revenue. That's exactly why they warned that margins are going to compress on that FSD side. So, a little bit I feel I feel confused, and that's what I want to say here, is that I see artificial intelligence making individual workers more productive, like what we're seeing at Meta, but I don't see this super intelligence bet as a good one.
I don't think Copilot is a good one. I don't think building a Terafab is a good one. These are companies with $6 trillion of American market cap, and I honestly think they're a little drunk right now. I don't know what's going on. Now, look, there's some things I don't Look, some things I do like, some things I don't like about this Tesla thing.
First of all, I hate that we're not going to have new models. But I get it cuz they're going to focus on Optimus. And we had a 50% decline in uh right here, 51% decline in our SX and Cybertrucks deliveries. Cybertruck deliveries, that's terrible. So, I get why they're deprecating it. But the thing is, we're killing off what at least is making some revenue for the Optimus, which right now Elon says there is no Optimus that's doing anything functional.
They're all in R&D project. Not only are they all in R&D project, but every time we make a new one, we deprecate the old one. So, we're kind of just trying to move incrementally one robot at a time. And all these ideas about, oh, we're going to do, you know, a million robots a year or whatever, they're not ready to ramp up the scale anyway because we don't have a finished R&D product.
That's the problem. So, you know, look, I have to say, yes, vehicle margin impressed to the upside. Some of these earnings numbers though it's sort of like the boring kind of giddy giddy giddy mathy stuff, fine. But Elon also made it clear that there's no competition outside of China. But the problem is, China is the cheapest place to mass manufacture robots or cars or whatever.
They're in a deflationary depression basically. They'll manufacture almost for free just to keep their lines working. I'm obviously being a little facetious there. But I don't know if those are the places to compete in, especially by the time Elon, if he goes for making a Terafab, which he sounds convinced in because he wants to hedge for geopolitical risks that haven't happened yet.
They might happen, but we still haven't proved that we have proven we actually need the chips because the Optimus isn't actually fully designed yet. Now, we're going to hedge for these geopolitical risks, build a Terafab. By the time that's done, memory chips from SanDisk or Nvidia will probably be a commodity cuz it's going to take him 6 to 8 years to do this.
All of that makes me nervous at the same time as the CFO says, "Yeah, we're going to we're going to start financing our robo-taxi revenues to pay for the Terafab." Well, again, at the same time Elon says he's thinking about selling shares to pay his fair share of taxes. He's not trying to sell shares to pay his fair share of taxes.
He's trying to sell shares because they're at an all near all-time freaking highs. And he needs to raise money to feed Grok. Stock compensation at Tesla, mind you, was higher than expected at $682 million. That's up from the prior quarters 428, 443, 459. It's easier to pay people in stock comp right now.
Why? A, because the stock is high. B, because you can add it back into your adjusted EPS. Why do you think Tesla beat adjusted EPS by 5 cents? Cuz they just added back in the stock compensation. They paid like 50% more in stock comp the last quarter. End of the year bonuses, I guess. Those are nice bonuses.
682 / 459 is 48.6% more. Energy margins are still great, 28.6%. We're still developing more superchargers, which is great. I'm glad we're doing that. At least we have positive growth of charging infrastructure versus the vehicle growth. But, at the same time, this packed with XAI, to me, is not actually a big beneficial move to Tesla, even though they have the Grok app in the car.
Okay, big deal. Every car now has some assistant in the car. And in the future, other cars will just plug in probably a GPT model or Claude model or whatever. I I I honestly think XAI just needs money to keep staying in existence to pay for the inventory. And you're seeing the same thing at Microsoft.
Their accounts payable are exploding. If you go to Microsoft and you look at their payables, and you can kind of see this drunken drunkenness at all of them. Take a look at this. It says here, they've got 16 Sorry, 18.6 billion in property and equipment still in accounts payable. So, bought and not paid for.
Someone has that inventory. Well, guess who had a spike in inventory the last in the last earnings report? Nvidia. So, you know, maybe that's because Microsoft does isn't a isn't ready yet to install those Nvidia chips, but they're spending money before they have room or the capacity to actually install these chips.
Microsoft is just trying to play catch up. I don't know that they're doing it well and in doing so their costs are going way up. Meta is blowing money even though they got a really good business, and Elon can't find direction. So, I'm nervous. Like I I like I do think these companies Microsoft and Meta will eventually decay to where they go, you know what?
We just need to stop spending on super intelligence and trying to compete. Let's just integrate. So, Microsoft will integrate rather than compete. Meta will stop spending R&D at some point and focus on ads and they'll do well there. Elon, it's it's roulette. This is Elon roulette that we're playing with a massive company, and I don't know how I feel about it because I really love Tesla and I want to see Tesla do really well, but I'm not convinced that the cybercab and Optimus robot are going to be a reality for a lot longer than we think.
Probably the end of the decade would be my guess before we actually have this happening. And maybe we have that time horizon. Hey, Cathie Wood five-year time horizon, great. But between now and then, I think there are a lot of risks, and to me the spending seems drunk. Anyway, speaking of spending, make sure you use my coupon code over at mekevin.com or househack. Have a good n.
Comments 0