TIP825: Meta, Adobe, Booking Holdings w/ Stig Brodersen, Tobias Carlisle & Hari Ramachandra

Hari Ramachandra: My pick for this time is Meta. When I was looking at the recent shuffle in the market, I see many names falling down and Meta was one of them. Its share price from its peak has fallen down by 20%. And when I looked at the company, the business is pretty strong. They're one of the two best advertising machines ever built. In fact, they are on track to beat Google in terms of ad revenues. They are forecasted ad revenue for 2026 is $243 billion, which will be $3 billion more than Google's. And they're growing, their operating margin is very healthy at 41% with a 46 billion free cash flow in 2025, a 30% net margin. Their revenue has been growing pretty healthily for last five years with a 18.5% CAGR revenue growth. So what's the problem? And the problem is something that is not new to Meta. They are very bold and very swift in making serious bets. And they put serious dollars behind those bets. Metaverse was one of them, which market got spooked when they didn't see much returns and they saw it as a money pit. VR Labs was another one, the Reality Labs. And this time, what has spooked market is their investments of their projected CapEx, especially of $135 billion into building their data centers and infrastructure for their AI, their big bet that they're making. Their first LLM llama was not a big success, but recently their super intelligence group came up with their latest model, which has performed really well compared to other foundational models out in the market, which gives me confidence that one, they have the ability to come up with a good model. Two, as we are seeing that models are pretty much getting commoditized. That means the incremental difference between models is kind of getting saturated. Distribution becomes more advantageous. It's the distribution that matters, whether it is Grog with XAI, Gemini of Google. Meta has a solid distribution. The second thing with Google and Meta is they have a lot of use for AI to make their products better, their ad targeting better. So they don't have to look for subscription model immediately. They can actually improve their profitability, their revenue streams for their existing products with AI. So my base case is based on these, they're able to recover their FCF margin and also the growth stabilizes without any re-rating of the price to earnings. I see a 46% upside from here. If they really hit the ball out of the park with their AI monetization, then it can be much more.

Hari Ramachandra: When I was looking at the recent shuffle in the market, I see many names falling down and Meta was one of them. Its share price from its peak has fallen down by 20%. And when I looked at the company, the business is pretty strong. They're one of the two best advertising machines ever built. In fact, they are on track to beat Google in terms of ad revenues. They are forecasted ad revenue for 2026 is $243 billion, which will be $3 billion more than Google's. And they're growing, their operating margin is very healthy at 41% with a 46 billion free cash flow in 2025, a 30% net margin.

Hari Ramachandra: As we are seeing that models are pretty much getting commoditized. That means the incremental difference between models is kind of getting saturated. Distribution becomes more advantageous. It's the distribution that matters, whether it is Grok with XAI, Gemini of Google. Meta has a solid distribution. The second thing with Google and Meta is they have a lot of use for AI to make their products better, their ad targeting better. So they don't have to look for subscription model immediately. They can actually improve their profitability, their revenue streams for their existing products with AI.

Hari Ramachandra: If you look at AI in general, I think it is constrained by power. It is constrained by real estate or space because you got to build a data center somewhere. And it is constrained by chips. are the GPUs currently. Of course, now there are competition from Google and Amazon and others with TPUs and there are specialized chips for inference versus training coming in and then the models are also being optimized. So the arc is towards that. So right now we are heavily constrained by chips today. But as you mentioned, the arc is towards a place where we might no longer be constrained by that, by a factor of 1, competition catching up, so supply coming into the market, 2, the model themselves becoming more efficient. And there is a lot of papers now being published on that area, and I'm following that, where so far it was all about features in the model, if you will. Now it's all about how to optimize the model for energy, for cost. In fact, Google Flash, Gemini Flash, recently in their Google IO, they talked about how if you use Flash, you will save billions of dollars because it's much more efficient. So that conversations have already started. So I think with that arc continuing, Facebook has the advantage that it has the walled garden. Not everybody has access to what the data Facebook has, and it can leverage that data whether it is for better ad targeting or for suggesting products to its customers, or even coming up with new features and product and subscription model for its customers based on the treasure trove of data they have. So that can prove to be an advantage for Facebook because their cost of delivering AI goes down as chips gets commoditized or model becomes more efficient. Now they have the advantage of data.

Tobias Carlisle: The big risk to booking is that folks think that they can somehow they will figure out how to get AI to control. You'll just type into your AI that you want to book a trip somewhere and the AI will do all of that for you without you having to interact with any of these sites and it will either do it directly and therefore cut out booking.com. So booking.com, their thesis though is that doesn't happen. So they're not disintermediated because there is a large database of all these sites and there are relationships that need to be maintained. It's not a simple matter of just calling them and paying them. They have these specialized relationships. And so in order for ChatGPT or Claude or one of the other LLMs to access these travel agents that'll need to go through booking.com in order to do it. ...The rest of the business is, it's very well managed. Capital allocation is excellent. As a result, it tends to be very high return on invested capital. It's grown very steadily for years and years. And has all the things that make it a great business, very asset light, great sort of network. Once people get used to the site, there's high switching costs, lots of free cash flow.

Tobias Carlisle: Mine is also a AI impacted name. I like booking. Booking.com. If you're in the States, you'll see any ads, right ads. Booking.com really sticks in the mind. It's a big company. It's much bigger than I usually pitch. And it's more expensive than I usually pitch. But I think it's interesting because it is so dominant and I think it's one of the unusual chances that you get to buy one of these companies cheaply, just so for folks who don't know, Booking is a business that allows you to book travel. They have Booking.com Priceline, which has been a great business forever. Agoda, I don't know particularly well, but Kayak is also a name that folks will recognize and OpenTable. Over time, there's secular growth in travel. As people become more wealthy, they tend to travel more. and they're able to control the experience. They're able to upsell and control various parts of the experience. They have a business model that doesn't. So through 2020, some of the other sites in an effort to compete with Booking, they buy rooms up front. They get caught when the rooms weren't taken. Booking.com doesn't do that. They don't buy their rooms up front, so they're very capital light. They don't own the underlying assets, they just hook people up. The big risk to booking is that folks think that they can somehow they will figure out how to get AI to control. You'll just type into your AI that you want to book a trip somewhere and the AI will do all of that for you without you having to interact with any of these sites and it will either do it directly and therefore cut out booking.com. So booking.com, their thesis though is that doesn't happen. So they're not disintermediated because there is a large database of all these sites and there are relationships that need to be maintained. It's not a simple matter of just calling them and paying them. They have these specialized relationships. And so in order for ChatGPT or Claude or one of the other LLMs to access these travel agents that'll need to go through booking.com in order to do it. I don't know what the likelihood of that actually happening is, but that's their thesis. They think that they won't be disintermediated, that they will continue to be part of that acquisition, that purchase journey. The rest of the business is, it's very well managed. Capital allocation is excellent. As a result, it tends to be very high return on invested capital. It's grown very steadily for years and years. and has all the things that make it a great business, very asset light, great sort of network. Once people get used to the site, there's high switching costs, lots of free cash flow. So the competitive advantages I think are durable. The risks to booking are travel is still sort of somewhat cyclical with if the economy goes through a weaker period, then folks just tend to travel less. And because of the way that booking is priced, really is priced, assuming some future growth or assuming that it continues to grow into the future somewhat like it has in the past, if that sort of revenue growth slows down, that's the sort of return we would likely expect. I think the most likely outcome is that travel just sort of generally grows as it has historically. I like companies that buy back stock at opportune times and I think that they're doing a good job buying back here. The bull case is that AI helps them and they become this sort of channel for all of these other LLMs and they just continue to grow. They interconnect with them easily seamlessly. You don't even know that it happens. And so they do a lot better than they have historically. you can handicap the bull and the bear and maybe they cancel out. And so the base case is the most likely, which is just that they keep on sort of muddling along. I think that booking is a reasonable risk adjusted bet at these levels because it's a little bit depressed with the fear around LLMs, but there's a reasonable chance that they are beneficiaries of that.

Stig Brodersen: My pick is Adobe. And if anyone is unfamiliar with Adobe, it's a software company. They're known for Photoshop in particular. But now, they have a ton of different apps you would typically use if you're a designer, creative person. They also have something called Digital Experience that's more about analytics. But my pick is Adobe. I guess that's what I'm trying to say. Market cap roughly $100 billion. And it has been growing double digits for the longest time. And it's not only trading at a 52 week low or near that, but near a seven-year low. It traded at 224, not too long ago. At the time of recording, it's 270, so it had a small bump. But the market, like all software or SaaS companies these days, is just unloved. Because at first glance, there is a lot to love. 96% of the $23 billion in revenue comes from subscription revenue. That is exactly what you want to see. Diversified base of more than 41 million paying users. And I mean, if I had pitched this just a few years ago before the era of the LLMs, we'd be all over it. The most important moat, I would say, is switching costs. And I was hinting at before, you can think of Adobe as having two segments. They have digital media, which is 76% of their business. And that's where you have the creative cloud, you have the document cloud where that resides. And then you have digital experiences. That's the enterprise software business that help companies manage marketing, customer data, e-commerce, digital content, and so on. And we're all creature of habits. And as uninspiring as it sounds, inertia is just very good whenever you're thinking about it in terms of a business model. And you also have to consider that people follow incentives. And of course, in this day and age with AI, it seems like it should be a tailwind. Shareholders, management, they want the employees of whatever kind of company to embrace AI. For example, to become more productive or to save on costs, guess what? Most employees do not have the same incentive. They think, and perhaps rightly so, that any efficiency gains doesn't really benefit their paycheck. And worst case, they can lose their own or the coworkers' jobs. And so management and shareholders just are way more excited about embracing AI to cut costs, which very often mean salaries. Even if we are looking away from the financial incentives and job risk and so on and so forth, there is a professional pride in knowing how to use Adobe. If you're a designer and you're taught those tools, you've used it for decades, it's really difficult to embrace a new technology where there is a level playing field for I wouldn't say everyone, but more or less everyone and certainly someone who's entry level who might be making a third of what you're making. Like you don't really have that incentive to stop playing that game. And I'm going to talk a bit more about incentives here, but I think anyone who's been running a business would tell you that whenever you roll out a new initiative, everyone would constantly or subconsciously be thinking, what's in it for me? One of the things I often think about whenever I would make an investment is how does this align with human nature? in this day and age, we see disruption from everywhere, but it's very difficult to disrupt human nature. And so one of the things that I wanted to highlight is like for 99% of people have an easier time spending other people's money than their own. It sounds so passive aggressive whenever I'm saying this. But I think a lot of the AI fears are probably overblown in the sense of how rapid things are going to change. And I'm not questioning at all that AI is going to change a bunch of different things. But I think we sometimes underestimate how much of these changes that need to be almost brute force and how difficult it is to brute force something in an organization. And it is very difficult to get people to understand something whenever your livelihood is depending on not understanding it. And of course, even that there's a limit to. If you don't understand the car and you swore by faster horses, you would eventually have to wake up to the reality. And so I'm not saying that if AI isn't 10 times better and despite the switching cars, then we would eventually have to use something else than say Adobe's products. And of course, whenever you ask the CEO about that, he's saying, oh, no, AI is not going to disrupt us. It's going to be a tailwind. We're going to be so much better because of AI. So I think that's a pretty weak signal. Any CEO that's worth his salt would probably say that today. But I would also be the first to say that a company like the Investors Podcast Network, we are so dependent on Adobe. And we're 20 people on the team. 15 of them are using Adobe. Some of them are just using one app, but most are using the entire suite. and the overall cost is significantly less than 1% of our total cost. So as much as it sounds great, let's use something that's cheaper and it's sort of like a different discussion if it's better, but let's say that this is going to be a lot cheaper. You know, even me as one of the owners of the company, as much as I would like to save on cost, it's just not really a cost saver. And whenever you run a company, and of course it depends on the company, but It's not always the creative people that are paid the most. They are typically one team inside a larger organization. It's the same with TIP. Then you have the salaries. So the people, that's the biggest cost. And then the second biggest cost is typically the equipment. And then a very small part of the cost is the software. So you're looking at something that's very small. So whenever you're doing that, I would say that Adobe is an incredible cheap product for the value that it provides. And I think it's very important that you also, whenever you hear people talking about the cost of Adobe, you would have a lot of people who are saying, it's ridiculously expensive. You know, the freelancer who has a side hustle.

Stig Brodersen: We're all creature of habits. And as uninspiring as it sounds, inertia is just very good whenever you're thinking about it in terms of a business model. And you also have to consider that people follow incentives. And of course, in this day and age with AI, it seems like it should be a tailwind. Shareholders, management, they want the employees of whatever kind of company to embrace AI. For example, to become more productive or to save on costs, guess what? Most employees do not have the same incentive. They think, and perhaps rightly so, that any efficiency gains doesn't really benefit their paycheck. And worst case, they can lose their own or the coworkers' jobs. ...Even if we are looking away from the financial incentives and job risk and so on and so forth, there is a professional pride in knowing how to use Adobe. If you're a designer and you're taught those tools, you've used it for decades, it's really difficult to embrace a new technology where there is a level playing field for I wouldn't say everyone, but more or less everyone and certainly someone who's entry level who might be making a third of what you're making. ...the overall cost is significantly less than 1% of our total cost. So as much as it sounds great, let's use something that's cheaper and it's sort of like a different discussion if it's better, but let's say that this is going to be a lot cheaper. You know, even me as one of the owners of the company, as much as I would like to save on cost, it's just not really a cost saver.