Big Tech Breaks the Bank for AI

0:04Big tech earnings are in and newsflash, they're spending a lot of money on AI. Motley Fool money starts now.

0:25Everybody needs money. That's why they call it money. The best things in life are free. But you can give them to the birds and bees. From Fool Global Headquarters, this is Motley Fool Money. Welcome to Motley Fool Money. I'm Travis Hoyam, joined today by Lou Whiteman and Asit Sharma. We have to start with the big news of the week. That is big tech. And I want to start with, I think, one of the most surprising earnings reports. That is Google. They were nano bananas, if you will, their results. The stock was up a little bit, although not a huge move, 4 % or 5 % after earnings. But revenue was up 15 % in search.

1:06The cloud business grew 13%. Asit, I want to start with you. what was your biggest takeaway from this quarterly report? Because there's so many pieces of Alphabet now, but it seems like they're all kind of moving in the right direction. Yeah, Travis, it's like an Alphabet soup of earnings to piece together what's really driving that engine. For me, the step up in cloud margin was pretty important. So the profit that the cloud business, which is full of generative AI, is making, this was 17 % margin this time last year. This quarter, they came in at 24%. So what does this mean? Well, Alphabet likes to brag that it provides the full stack to its customers in AI.

1:48So from the infrastructure, to agents, to all kinds of machine learning algorithms, etc. And you can price for that. The pricing is favorable. And we're seeing that as this little business grows, it's getting more profitable as it goes along. Yeah, the interesting thing with that is that business just became operating income profit positive in the first quarter of 2023. So we're less than three years into that actually being a profitable business for them. What, Lou, what were the big things that you saw when you looked at earnings from Alphabet? Yeah, well, vibe off of Mark Twain and his code, reports of my death were greatly exaggerated.

2:26Remember when AI was going to swamp Google search and just everything was down from here? But as Asit said, the margin improvement. I mean, overall margins up 200 basis points. But what that shows is this is so much more than search. That wasn't really ad revenue that drew of that. That is all of the premium things they offer as consumers. That is cloud. Cloud grew revenue by 34%, but operating income by 85%. That speaks to the benefits of scale. And that speaks to, I think, a really, really well-positioned company from here. Yeah, I said, Lou mentioned some of these other businesses. YouTube is the other one that I think we often forget is under Alphabet.

3:08And one of the things that stuck out to me in this report was that search and YouTube grew at the same rate, which may sound a little bit strange. But what that's telling me is that they're monetizing at basically an equivalent level on both of those. So your views are in number of searches and number of videos watched and things like that may change a little bit. But if you were seeing search grow at 5%, but YouTube grow at 20%, that would be problematic. Am I thinking about this correctly? That it's actually good that they're growing at about the same rate? I think so, because in some ways, Alphabet is still using search as a way to funnel people to YouTube.

3:46Although those who get a little bit addicted to a certain type of YouTube shorts, like myself, I'll be honest here, we don't need the search to get us there anymore. But this is a business that is, you know, I think it's of it as being tiny, but powerful, tiny in the scale of Alphabet's total earnings, but powerful because of that growth rate that you mentioned, Travis. So the attention economy, which we'll talk about in a bit when we get to meta, is such a big part of the earning stories of much of big tech. And the fact that Alphabet continues to grow this little enterprise is important because over time, the rest of that business inevitably is going to slow a bit.

4:23But I have a belief that YouTube will keep scaling along the lines that it is 15 to 20 % reliably quarter after quarter after quarter. And they continue to take market share even from Netflix, which I think is surprising when you look at some of these results. Let's move to their AI spending because this is, you mentioned, we've mentioned Google Cloud. That's the big growth story within Alphabet. But they have to spend a lot of money to build out the infrastructure to actually grow at 34%. So, Lou, they increased their CapEx guidance. It was increased last quarter to$85 billion. This quarter, I think they said it's actually going to be$91 to$93 billion.

5:02And then they said more next year. We've heard whisper numbers of around$120 billion. This is a lot of money. Is that bullish or bearish for investors long term? Time will tell, won't it? Yeah, to put meat on it, CapEx up 83 % year over year. And I know it's really accelerated this year. But look, AI was around last year. So, we are now doing year-over-year comparisons when this spending was starting to ramp. Look, the good news is, everything we said, they seem to be finding ways to monetize this. And if that can continue, we will be glad that they are investing more. But yes, when you're spending$24 billion every three months on just infrastructure, on CapEx, you darn well better figure out how to monetize it.

5:47So I do think it's the elephant in the room. It could be fine, but it's definitely also what we should be watching. Yeah, I want to chime in here and agree with Lou. There's a little bit of risk in Alphabet's picture. Okay, there's risk in every big tech hyperscaler's plans to build out all these data centers and provide so much inference to the world. But looking at Google's business, they're much smaller than, say, Amazon Web Services, but their spends are approaching AWS spends for their build-out. That means that they're playing a lot of catch-up. And if things go south for all of these, they'll have a bigger hit on their P &L pound for pound than maybe an Amazon Web Services will.

6:28Speaking of CapEx, Alphabet's response, the market's response to Alphabet's earnings, where it was actually pretty positive, it was not as positive for Meta. That was another big one. They didn't actually increase their CapEx guidance. I think it was just actually at the top of the range that they had previously given. But the questions are starting to mount about how are they going to actually turn that spending into more money in the future. So Asit, what should we take away from what the market thought was kind of a flop of a quarter? But you kind of look at the numbers and I didn't see anything that was horrible.

7:02But then again, you start to ask questions about are they going to actually get an ROI on spending tens of billions of dollars a year on it? Yeah, totally, Travis. I mean, what happens when you merge Daddy Warbucks with what's arguably the door's most popular song? Come on, Daddy, light my fire. More specifically, we're going to set our cash on fire. And that's been the mojo of Meta and Mark Zuckerberg for a long time. You've nailed it. This business is strong. There is nothing wrong with it under the surface. And in fact, average price per ad was up 10 % this quarter. Ad impressions up 14%. revenue growth, what, up 26%, billions of dollars to the bottom line in operating income.

7:46They had a little charge against earnings, big charge against earnings. But look at this spend. Will it remain something that investors can sort of whistle past? I think today was the first indication that, no, maybe we're not. And I know, Lou, you've got some thoughts on this as well. Everyone's spending money, right? Right. And yes, this us, it's right. The core cash generation machine is intact and as strong as ever. The difference is now is that we are going off balance sheet at Meta, all of them. But we can no longer just kind of justify all of this spending on AI as, hey, it's coming out of free cash, free cash flow.

8:23And what do you mean by that, Lou? Well, so it's a big difference between them. They make all the money in the world and they're just choosing to spend it. Whether or not it's a good move or a bad move, it's healthy. When you start borrowing for this, so you're actually spending more than what you're making, it can still pay off, but it better pay off. There is just an added element of risk, both to the company and probably to the system. The big thing for me here, just this continued question is, Meta versus Google, Microsoft, Amazon, it's on the distribution side. I think we saw with Alphabet, and I think with all those other companies, their cloud businesses, their consumer businesses, their office tools.

9:05I see how AI will be distributed and infused into the products. Meta stands out to me because the monetization plan, to me anyway, is less clear. And when you are spending so much, borrowing so much, you got to get monetization right. So, I do think that's the difference right now. Yeah, Lou, it's funny. I used to have the same thoughts about Meta on their distribution, But, you know, I'm a WhatsApp user. And so that little AI agent is right at the top. I've never used it, actually, but I know people who do. I've got some cousins in India. They're using that all the time. And just the way that this business has been able to monetize things like Instagram, where I couldn't make that instant connection.

9:46The platform is the distribution, right? It's the users. It is different than the hyperscale businesses. But time over time, they prove that they're able to find a way for folks to fork over some more money. It is one of those companies, this is the one hyperscaler that is spending tens of billions of dollars on AI infrastructure that doesn't have the same third party business that all the other companies has. You know, if Google overbuilds for itself or for third parties, it can use some of those GPUs itself. Same thing with Microsoft. So does that present more risk? I think that's maybe the question that I have.

10:25I'm not a meta shareholder, but I have come to sort of respect what Zuck has built. And then you get to these points where you go, it's really just a trust me. Asit, is that all we're doing is just trust? Do you trust Zuck or not? Yeah, he's got a mini I told you so wrapped up in here, which is a few years ago before everyone really understood the import of Gen AI. He said, well, all this money that we're spending for reality labs, the AI infrastructure, the GPUs, et cetera, it'll be useful. It'll help us really make our ads better. And we can use all this compute to squeeze more ad money out of our users.

11:01And lo and behold, that's what they've done. But I just want to point out here, I'm skeptical of what Zuck said in the earnings call that, hey, if we build extra compute, we'll just use it to make our stuff better. No worries. I don't buy that simply because the magnitude of what they're investing now, as Lou pointed out, going to having to have debt out in the public markets is a magnitude larger than this first iteration that we all saw. So yes, it worked the first time around. Will it work this time? I'm not so sure. The stakes are getting higher. Next up, we're going to talk about one company that the market was very impressed with.

11:37That is Amazon's result. You're listening to Motley Fool money.

11:44Have you ever gazed in wonder at the Great Pyramid? Have you marveled at the golden face of Tutankhamun? Or admired the delicate features of Queen Nefertiti? If you have, you'll probably like the History of Egypt podcast. Every week, we explore tales of this ancient culture. The History of Egypt is available wherever you get your podcasting fix. Come, let me introduce you to the world of ancient Egypt.

12:21Welcome back to Motley Fool Money. Amazon was the other big earnings report this week. Shares were up 10 % after earnings came out, which is a huge move for one of the biggest companies in the world. Asit, what did we learn from Amazon? Is this a story about AWS? Is this a story about retail, advertising? What should we take away from this quarter? Because just like Alphabet, there's a lot going on at Amazon. Yeah, it's a story of all three. And we can start anywhere. Maybe we'll start with AWS simply because the market breathed a little sigh of relief. People felt that its peers like Alphabet were moving faster and maybe grabbing the business ahead of Alphabet.

13:02And what we learned, and I know Lou has some thoughts on this as well, is that Amazon has a business proposition they feel the world needs to understand. So many enterprise businesses have converted from on-premises infrastructure over to the cloud, onto AWS. And those are really ripe customers for that whole AI stack, just like Alphabet that we were talking about. And it turns out that growth is re-accelerating. Margins look very healthy. I'll come back to margins in a sec. But, Lou, I mean, that cloud business really stood out to you, didn't it? It was great. And guys, remember, what was it, three months ago when the cloud business didn't stand out and everyone was sort of leaving Amazon for dead?

13:43You know, maybe we shouldn't read too much into any one three-month period. But yeah, cloud, 20%. You know, growth, it's everything you could have hoped for there. But how about across-the-board performance? Ad sales up 24%. Even retail was up 10%. So much for that weak consumer. It's hard to find a hole in this quarter. This is just a multi... All parts of this business seem to be firing, and it's just really impressive just to see the strength on display. I say one of the things that they talked a lot about was Tranium and the demand that they have for that chip. This is a competitor to NVIDIA's chips.

14:27This is kind of a big can of worms, But it does seem like sort of the alternatives to NVIDIA are starting to gain some traction. Terranium is one. TPUs from Google is another. They apparently can't make those fast enough. And they're, you know, whether it's for use with TPUs or just Google Cloud, they can't. They're basically having to turn away some customers. Are we starting to see an alternative layer of compute in AI start to be built that is maybe getting momentum? Because all of these companies do have an incentive not to be beholden to NVIDIA. That's right, Travis. They are spending a lot of money for these purpose-built chips.

15:07Trainium is a great example. Even Microsoft, which was behind on the game, now has their Maya chip, which is very similar. So what these chips try to do is offer a reasonably similar level of performance for less cost. So think 30 % to 40%. This is the cost savings that Amazon consistently talks about. And yes, lots of businesses want to use what is performant and also cheaper. But the issue here for Amazon more than it is for NVIDIA is that NVIDIA has this really fast cycle of iteration. And Amazon doesn't iterate quite as fast. So the Tranium chip, while it's getting a lot of hype and it's got this huge customer in Anthropic, which is the maker of Claude, the LLM.

15:53they are going to be soon in a place where they're going to have to have a next generation of chips. And this is where I think that maybe training in these other chips just won't live up to the hype. And we also see lots of deep-pocketed customers still want NVIDIA's latest and greatest chips. So they're doing both. They're spreading a little money to Amazon, but putting most of the dollars into NVIDIA's next big thing. And I would say AMD is playing in that space now. So I think you're right. There's a layer now of business that's up for grabs. And some of these hyperscalers are getting it because I already have the customers in house.

16:25But are they going to supplant NVIDIA? No. And do they know that they themselves, Amazon, must buy NVIDIA because their customers want it? Yeah. Andy Jassy talked about that on the call. So is the game for AI compute right now still just who can buy the most NVIDIA chips and then maybe add on some of their own custom ASICs? Is that sort of the idea, Asit? Yeah, it's interesting you should mention ASICS because these sort of custom-built chips that are the ASICS variety, they really help with cutting down the cost of compute. But then here's another issue. GPU is very configurable, so you can reprogram it as use cases change, where an ASICS is more pointed.

17:08So you can save customers money for a certain amount of time. But if the needs change, if we go not nano bananas, but I don't know, like polymorphing pairs, ASICs might not be built to do the same thing that a GPU can do to adapt to what customers want on the inference side. So the way to think about it, B, is if the rate of innovation in AI, the rate of these models being introduced. And in 2023, it seemed like there was a new model every two weeks. Now we've kind of slowed down. So is that slowing of model improvements good for the ASIC business? because you can actually customize your chip to run optimally for Claude, let's say, or Gemini.

17:49And so they're able to actually kind of get to scale before those are obsolete. Is that the right way to think about it? It's a good way to think about it because what the models now are offering really is just more reasoning steps. It's not like we're having huge, huge advances. So ASICs could fill that. In other words, you keep asking chat TPT questions. It keeps asking me, Asit, do you want to drill down on this? After all, I'm like, baby, I'm tired of drilling. I'm going to take a break here. But it's so eager, and all these models are so eager to have you keep reasoning, because now it's a little bit cheaper for them to provide that, and there aren't that great of leaps and improvements on what the models can do.

18:28So, yeah, I think that's an astute point, Travis. Lou, as you look at all of these big tech earnings, what stands out to you, and where do you think the best buys are in the market? I still like Alphabet a lot. And I think as far as Mag7, I still, they just, again, I'm worried about a world where the AI model sort of becomes commoditized. And so then it's going to be who has the resources to really make money off of it. And again, I'd point to probably Amazon, definitely Microsoft and Alphabet as just the ready-made customers. Wild card here is Apple. Apple has gotten so much flack for failing at AI.

19:10And I don't think that's just their Apple waits. I think they really tried and it didn't go well. But with their customer base, with that iPhone, with that as a prize for any one of these models, if we do get to a world where the models are not the big deal, it's who has the consumers, Apple could end up being the biggest winner here at all. When we come back, we are going to play Trick or Treat. You're listening to Motley Fool Money. I was working in the lab at late one night, when my eyes beheld an eerie sight, for my monster from his last began.

19:53Welcome back to Motley Fool Money. It is Halloween, so we thought we would have a little bit of fun and play investing trick or treat. The idea here is that I'm going to give Lou and Asit a stock, and they have to tell me whether it is a trick, so they're bearish on the company, or a treat, and investors should be bullish. Let's start with some of the names that we have been talking about already on this show. Alphabet. Lou, you kind of gave away your answer earlier. But Asit, I want to know, is Alphabet, after these recent gains, is this a trick or treat for investors? It's a treat. It's one of the biggest games in town.

20:28The tech is very solid. If you look at scholarly citations on AI, they're one of the leaders actually in papers. And of course, it was Google engineers who came up with technology along with some others that underlies all of Gen AI that we enjoy today. So they were a little bit late to the commercial side of the game, but knives out now. So for me, a treat. Yeah, stocks up 74%, I think six months. So maybe they've decreased to a fun size treat instead of like just the full size. But it's a dreaded fun size. It's still a treat. I was so surprised it had been a year. OK, really quick, guys. I broke open some Halloween candy in advance and the fun size.

21:09I was like, wait, isn't shouldn't fun size be big? The fun size is small. What a marketing genius thing. Sorry. Sorry to interrupt, Lou. Let's let's go on to Meta, another company that we've talked about a little bit. But price earnings multiple on a forward basis is 23. So it's actually cheaper than Alphabet. But there are maybe more questions. Maybe there's not. Lou, what do you think? Is this a trick or a treat for investors? It's a treat, but it's an apple or something. It's not the treat you want. The one you don't want in your Halloween bag. It's a treat, but you have to wonder, is it really a trick?

21:45As long as that cash machine is working, I can't be too worried about it. But we already talked about it. There are questions here. I'm going to say it is a trick. And I know this is wrong. I came on this show maybe three or four years ago and sort of infamously said that Meta was one of the worst of big tech. And I had so many cogent reasons. It's really hard to bet against this company because of that ad revenue, the profits associated with it, the fact that 3.5 billion people around the planet are active daily users. of one active i think the right way to say it is addicted to instagram i think that's yeah totally uh but but look at the flip side of this you know they continue to sink money first it was vr ar then um i believe it was the metaverse then we went on to a couple of iterations llama was gonna be the next big thing and now super intelligence is the goal so mark zuckerberg believes in this zero-sum game that the internet can be won, AI can be won, all of revenue can be won.

22:54And I just think you take a risk when you do that, even if you have such a beautiful business model underneath. So I'm going to say there's a slight trick in this business. Beware. Let's end our big tech on Microsoft, a stock that we did not really talk about today. But Asit, a lot going on at Microsoft. Is this a trick or a treat? It's a treat for me. I love me some Microsoft because it doesn't get quite the attention of some of its big tech peers, but it chugs along. Office is a franchise. It has so many franchises in the gaming world. It has done an aggressive amount of investment in AI through OpenAI and now is playing good with them again.

23:33So I think it's just a wonderfully managed company. And don't forget, it's got a core of a cloud business that is really all about, again, And that long-term transition of enterprise businesses from premise into the cloud. So I think, for me, it's a treat that doesn't get the attention it deserves long-term. Does the$135 billion, now that we have a reorganization of OpenAI,$135 billion stake at their$500 billion valuation, now we hear that they're looking at an IPO at a trillion-dollar valuation. That could be$250,$300 billion stake. Does that change? Is that something that you meaningfully build into your thoughts on Microsoft?

24:12Yeah, I think increasingly it is. They're in for a penny, in for a pound with OpenAI. We've seen that the management of Microsoft is deadly aggressive and usually deadly right in how they allocate their capital. But it's getting a bit big for comfort. All in all, I think it will be a win for them. But maybe this is the equivalent of what Lou pointed out about the off-balance sheet financing that Meta is undergoing. Each of these companies is getting a little deep in one part of this equation. And that's maybe the pain point for Microsoft as we look at the risk landscape going forward. So let's be clear, that open AI stake is what, less than 4 % of Microsoft's market cap right now, I think.

24:56And if anything, I think, as I agree, they're great capital allocators. And I think this is kind of them saying, it's okay, you can play around. So, I think they're actually de-emphasizing from OpenAI, which I think is probably smart. If Alphabet was a fun size, Microsoft is still the full-size candy bar treat for me. Because I will always favor the enterprise over the consumer in terms of monetization. I am really annoyed with all of the, would you like AI's help every time I open Excel? But I get it as a business thing. They, you know, we have for decades now taken for granted their ability to sell to the enterprise.

25:37And I just think they are the most natural beneficiary of the AI revolution, which I still think is sort of a lot of just back office mundane stuff getting done faster. And I mean, as a consumer, I would love that. But as a business, I pay up for it. Microsoft is the go-to AI play for me. Can I just underscore something Lou said for, for just a moment here. I was in, I think word, maybe prepping for the show this morning and just jiggled my mouse. Hey, can I help you write this? Dude, my cursor is not even blinking yet. I don't need the help today. It's the modern Clippy. Let it's a modern Clippy.