Current Market Opportunities Mercado Libre Amazon Constellation Software Tip808

TITLE: Current Market Opportunities | Mercado Libre, Amazon, & Constellation Software (TIP808) CHANNEL: The Investor’s Podcast DATE: 2026-04-18 URL: https://www.youtube.com/watch?v=VgvTImPYGeU

---TRANSCRIPT---

Andy Jassy shared in an internal all-hands meeting that he expects AI will help fuel AWS to become a $600 billion business in 2036, which is double the previous estimate of how large they expected that business to become. So that would imply that he essentially expects AWS to grow at a 15% CAGR over the next decade.

Before we dive into the video, if you’ve been enjoying the show, be sure to click the subscribe button below so you never miss an episode. It’s a free and easy way to support us and we’d really appreciate it. Thank you so much.

So today’s episode is a bit bittersweet for me to record because it will actually be my last episode here as a host at the Investors Podcast Network, at least for the time being. I just wanted to take the opportunity to thank everyone who has followed along while I’ve been a host here over nearly the past five years. I’ve just met countless amazing people who listen to our show and just really appreciate everyone who has supported us over the years, whether that be just listening to the show, sending me a note, meeting up in person in Omaha during the Berkshire weekend or by joining our Mastermind Community as a member. All of these things, no matter how big or how small, they all go a long way in helping us deliver this content to you. So with that all said, I just really couldn’t be more grateful for your support.

As for next steps for me, in case anyone’s curious, I’ll be making a bit of a career transition and working for an investment management and financial planning firm here in Nebraska. But part of me will still always be with TIP. I’ll still be listening to most of the content the team puts out here, keeping in touch with everyone and remaining a member of our Mastermind Community and likely attending some in-person events as well.

I gotta say it’s a tough pill to swallow for me too. And I count on seeing you at our events, Clay, and obviously staying in touch beyond that. So let me first say that it’s, you know, kind of a surreal feeling sitting here today because I started listening to the Investors Podcast myself in the early days with, you know, Stig and Preston as hosts. But I have to admit, and I told you that privately before, that to me, you are the voice behind We Study Billionaires. And I’ve listened to so many episodes of you. And it has already been a pleasure earlier this year to be part of this show and pitch a stock here. But to sit here today knowing it will be your last episode feels kind of sad to me too. And yet we have talked about your next chapter and I know that it will be just as exciting and a perfect fit for you. So I wish you all the best and I know that you will always be part of TIP. And, you know, I’m confident that seeing you in Omaha in a couple of weeks won’t be the last time.

Yeah, well, thank you for the kind words, Daniel. I know that you’re still relatively new to TIP, so I definitely feel lucky to have had the opportunity to work with you, get to know you, and just look forward to keeping in touch and seeing each other at future TIP events. And I can promise that our listeners will be in good hands as Daniel will actually be stepping in and hosting more episodes here on the show. He consistently surprises me with the quality and the volume of the work that he does, and I know that our listeners are going to benefit greatly from his contributions. He’s already put together several episodes that are some of my favorites over on the Intrinsic Value Podcast, including episodes on Mercado Libre, Hermès, Nintendo, and of course Constellation Software, among several others as well. So stepping back, Daniel, as I’ve been a host here nearly for five years now, the question I’ve probably been asked the most by people is how did you become a host at TIP? So I’d like to turn that question over to you. So please tell the listeners more about your background and what brought you to this conversation here today.

Well, it’s funny because I get that question quite a lot too. But it seems like all of us hosts have come to this job from completely different backgrounds. So you and Shawn have been long-term listeners of the show, just as I was. But you guys actually applied to work here after you heard that TIP was looking for people. And that wasn’t the case for me. I never came across any TIP job advertisements. So for context, some years ago I started sharing my investing journey online. I started an investing blog that has grown to about 25,000 subscribers over time. And basically, I there shared stock research and some general investing philosophy. And I did the same thing on Twitter as well with about 100,000 followers, although I definitely preferred the long-form content on my blog more. And well, one day I got an email from Stig asking if I would be interested in hopping on a call. And of course, I was a bit starstruck, but still, you know, Stig is a great guy. So it went pretty well. And long story short, he pitched me on a concept basically for a new show, which would, you know, allow me to do equity research all week, but instead of focusing on just one niche as I would have to as an investment banker, I could look into every company whenever I want to. And that kind of sounded like a dream job to me. And I got to admit I was a bit hesitant at first because, you know, officially I would then be a podcast host. And it’s not necessarily the same thing as having investment banker on your CV. But I knew Stig, I knew TIP. And I also knew about the quality of their work. So I pretty quickly decided that this is an offer that I just couldn’t refuse. And I’m glad that today I can sit here and say that it has honestly been one of the best decisions of my life. And I really don’t say that lightly.

Yeah, I definitely know the feeling of having the job title of a podcast host. Mentally, I think that people like to put other people into buckets. So let’s say you’re an investment banker, people will naturally assume that you’re probably really smart, you probably make some good money, and you likely have very little free time because you’re working from dawn to dusk. So when I tell people that I do this podcasting thing, they aren’t sure what it is I actually do and if I actually make money from that job. Perhaps I live in my parents’ basement or something like that, just trying to make ends meet, but that’s certainly not the case for us here. But I can certainly say that we live in a world where going to the best school or having the right job title, it’s just not necessary to be successful. So you obviously like what you do here at TIP. What are some of your favorite parts about it?

I would say I love learning new things. And I guess that’s kind of what I like about investing. You just said that people like putting others into buckets. And I think one bucket for people working in finance is that they only look at numbers and at stock charts. And to me, investing is mainly about understanding businesses and to some extent how actually the world around me works. So the great thing at TIP or working for TIP is that you can really just let your curiosity guide you. So whatever I want to study, I can. And of course that mostly means businesses. But honestly, it’s so much more than that. I mean, while it might seem to people on the outside that we are podcast hosts, no one would know better than you that we do so much more. I mean, working at TIP feels pretty much like being an entrepreneur as any job could feel like that. And I’ve learned so much about building a business in the last year that it has just been an amazing journey.

Very well put. But before we get too far off topic, I had a few companies that I wanted to cover with you here today. The first of which would be Mercado Libre. When we did our top stock picks episode that was released on January 1st just a few months ago, I went into the episode pitching Meta and you pitched Mercado Libre. A couple of days after the episode, MELI immediately jumped 15% based mostly on the news from Venezuela. Since then though, the stock has dipped below $1,700 USD, which is a level that we haven’t seen since 2024. And despite, you know, a pretty stellar earnings report that they just put out, MELI is also a company that you and Shawn added to the Intrinsic Value Portfolio a few months back. So how about you share your thesis and some updated thoughts on that?

Yeah, sure. I mean, MELI is a company that I have come to appreciate more the deeper I dived into it. And this last earnings report is actually a great example of why.

I’m cheating a bit here and peeking at your notes. You have down that revenue grew 45% in that recent report. Is that right?

It’s not a typo. I mean, revenue did grow 45% year-over-year. Items sold were up over 40% too. And the credit portfolio actually almost doubled. So that means MELI has now extended its record for the longest ever streak of quarters with over 30% year-over-year revenue growth to 28 consecutive quarters. And at MELI’s size, that’s an insane number. And MELI is actually the only company ever to achieve that. But the stock still reacted negatively because MELI’s margins were down. And it’s mostly because they invested heavily in the credit card portfolio. They lowered free shipping thresholds in Brazil. And also they scaled its cross-border and especially the first-party businesses. So combined, management said that those investments cost 5 to 6 percentage point headwind for operating margins. So the topline was exceptional, but the market didn’t like the profitability picture.

You know, too often we hear that this company is the next Apple, this company is the next Google, the next Amazon. And I can, I think, lead a lot of investors to get a little bit weary when you hear people say that. But I just can’t help but think that this is one of those situations where history doesn’t repeat, but it rhymes. It’s essentially the Amazon playbook that MELI is using here. Amazon looked unprofitable for decades before it finally decided to show its true earnings power. And it seems to me like MELI actually is doing the same thing. Nick Sleep, who most of our listeners will be familiar with, he talks a lot about this model of scaled economies shared. In one of his letters many years ago, he wrote the following. Take Amazon as an example. Last year the company reported free cash flow of just over 500 million. Indeed, it has been around this number for the last few years. What is important is that the 500 million is after all investment spending on growth initiatives such as capital spending, but also research and development, shipping subsidies, marketing and advertising, and price givebacks. The firm has been investing in these items today to grow the business in the future so that free cash flow in the years to come will be meaningfully greater than it would be otherwise. In valuing the business at these prices, as occurred last summer, investors are saying to Amazon management, your gross spending has no value. You may as well turn yourself into a cash cow. This is an odd statement to make for a business growing revenues in excess of 20% per year. So my sense is that this quote is spot on for MELI as well.

Absolutely. I mean, I think the market is completely missing the long-term picture here. However, to defend the market, and I know that sounds a bit weird, there are times where you want to have proof of concept. Earlier this year, I gave a talk in the Mastermind Community about how I always try to understand why markets react in certain ways, and especially when it’s different from how I personally view the situation. And a lot can be explained by the market’s short-term focus. So analysts have to adjust their valuation model for the next quarter when margins are lower than expected. And that will result in downgrades and negative commentary. I mean, you know the game better than anyone else. But there’s also an argument to make that investments can be a negative thing. So that’s usually the case when, you know, they are forced by competitive pressures. So at that point MELI couldn’t just stop investing and margins would suddenly revert to where they’ve been in the past. They would just lose market share to competitors. And that dynamic could mean that it’s structurally a weaker business than one might have thought. And then I think today, in my opinion, though, I don’t think that’s the case here. Just as it hasn’t been the case with Amazon some 10 or 20 years ago.

Another concept that Nick Sleep is well known for is destination analysis. So the idea is basically the opposite of looking ahead to the next quarter. Instead, it’s looking out where is this business likely to be 10 years from now.

This is actually what my main thesis for MELI is based on. So if you invest in emerging markets, you are always taking on more risk than in established North American or European markets. Interest rates are higher, recessions tend to be more common and also more severe, and sometimes markets will vanish completely as it has been the case with Venezuela for MELI about a decade ago. But I always come back to the major tailwinds. So the first one is about the e-commerce adoption. No matter how severe a recession is or, you know, where interest rates are, people won’t suddenly stop buying online forever. And e-commerce penetration in Latin America is still between 14 to 15% today. And if you compare that to roughly 25% in the US, close to 30% in the UK, and well above 30% in China, I think there’s no good reason to believe that Latin America will permanently stay at half the level of the rest of the world. And as long as that gap exists, I do think the entire market is growing enough. And MELI, with its 30 to 35% share in Brazil, which is its biggest market, is capturing a disproportionate share of that growth. And in theory, you would have the same tailwinds for the fintech and the credit business. So this is certainly a riskier business because a recession hits these businesses even harder. But overall, people won’t go back to hoarding cash at home or applying for these complex expensive banking products. So I do believe that Nubank and MELI are very likely to be even more important in 10 years than, you know, they are right now.

It’s such a great point, and I just love the idea of finding a huge industry with these massive secular growth trends and owning the number one beneficiary of that trend. All around the world you see e-commerce being a secular growth industry, even here in the US you see penetration continuing to increase year after year. And your point on being able to weather through the drawdowns while riding a wave of secular growth, it reminded me of my conversation with Francois Rochon the other day. We were talking about LVMH and the headwinds they are facing with what’s happening in China, but if you zoom out and take a five to ten year view, he essentially said that GDP per capita is going to continue to increase at a fast clip in China, and thus more people are going to want to purchase LVMH products. I recently saw a conversation that you had with one of our members of our Mastermind Community where you talked about the credit risk. And I think you mentioned that MELI’s loans, they tend to be very short-term, with an average consumer loan duration of less than four months and an average credit card duration of less than three months. Considering this and the fact that MELI gets its data from the Mercado Libre marketplace, which is probably the part of the market where you would see reduced spending in a struggling consumer fairly quickly, that gives me, you know, a bit of a better feeling regarding how MELI could adjust its loan portfolio if they do see a storm coming.

This is why I’m generally a bit more comfortable with owning MELI compared to Nubank. To me, it just makes sense how MELI gets its data to underwrite its credit and loan portfolio. Nevertheless, things move quickly in bad times. So if there should be a recession or anything like that, I certainly think MELI’s credit portfolio would take a hit. And that’s certainly one of the risks that you have to be aware of and that you want to take when you invest in MELI. But what I like so much kind of going back to the destination analysis point is that MELI built all of its ecosystem because these were essential needs to the people in South America that basically no one else offered. So they built a payments arm to enable more people to shop on Mercado Libre. They built their logistics network because logistics were highly unreliable in large parts of South America. And I think that matters because it shows that MELI is the only competitor of size that is basically South America first. So should there ever be a large recession in Brazil or in any of the other big markets for Mercado Libre, Amazon has a dozen other places to deploy capital. So why would they keep investing in Brazil when it’s burning lots of money and you don’t even know how the market will be in 10 years? Shopee, which is probably MELI’s biggest competitor in Brazil right now, kind of faces a similar problem. So they are very committed to Brazil right now, but its parent company, which is Sea Limited, is primarily focused on Southeast Asia. So if it has to decide where to deploy capital in a worst-case scenario in Brazil, again, I’m not sure they would choose to invest through the storm and stay in Brazil. And at least not to the same extent that MELI would. And in fact, Shopee used to be in many more South American markets, and it was quite successful in expanding its presence in those markets. But when Sea Limited stock plummeted after COVID, I think it went down almost 90%, they kind of had to reconsider where to invest their money. So they closed down shop in all of their South American markets except for Brazil. MELI on the other hand, pretty much has no alternative. So MELI is fully focused on that part of the world.

To play a bit of devil’s advocate on the MELI pick, let’s compare MELI to Amazon again, which is another company we’ll be discussing here shortly. Amazon stock has just been a bit of a dog ever since it got a big boost post-COVID. Even with AWS crossing $140 billion in revenue run rate and the advertising business being an $85 billion business, it seems to me that the market is just discounting Amazon due to how capital intensive it is. They’re caught up in the AI arms race, purchasing chips and investing in data centers, and the e-commerce and logistics part of their business may have been more capital intensive than the Amazon bulls may have assumed originally. So in 2022, when the stock was heavily beaten down, there was this narrative that Amazon was essentially being bailed out by AWS due to the e-commerce division having such low margins. So in light of that, how do you think about where margins will land longer term in valuing MELI and looking at their e-commerce business?

I believe we will talk about some of the major tailwinds I see for e-commerce companies generally in a minute when we talk about Amazon. So I will generally say it’s not a coincidence that I only recently invested in both Amazon and MELI. I do believe that we are on the verge of seeing innovation that will make the e-commerce industry a lot more profitable. So just for context, MELI never used its marketplace as a loss leader. Amazon did do that for a long time and to some extent they’re still doing it. So to a large extent, just because Amazon could do it is why they do it. AWS was so profitable that Amazon used it to basically subsidize the marketplace business and become the undisputed market leader in both the US and also in large parts of Western Europe. So Amazon’s e-commerce operation didn’t look profitable mostly because this was the strategy to basically price below competitors, offer free shipping through Prime, and reinvest every dollar of margin back into the flywheel. And that’s also why Amazon earns half of its GMV with first-party products. So that means Amazon is basically buying inventory, it’s holding inventory, it’s taking pricing risk and it sells it directly. And that’s just a structurally low margin business or activity to be in. But you do improve your product variety, which is why Amazon decided to do that. So you’re basically running a retail operation with all the associated costs. You have procurement, warehousing, markdowns, returns. The 3P marketplace where basically third-party sellers list products on Amazon and Amazon just takes a commission, that one is actually the very profitable business. Amazon’s 3P take rate is now in the low to mid 30% range when you include FBA fees, advertising, and referral fees. And Mercado Libre shows a different strategy from the beginning. It has always focused on the 3P market. So by now it’s over 90% of MELI’s GMV that is coming from third-party sellers. And that means that the vast majority of their commerce revenue is high margin intermediation. So you’re talking take rates, shipping fees, and advertising. And MELI had to play that game because they didn’t have an AWS equivalent to subsidize the marketplace business. Also, kind of growing up in South America, you needed to be more profitable and more resilient earlier than Amazon needed to be. But to me, all of this just kind of shows that e-commerce generally is not an inherently bad business to be in.

With that let’s talk a little bit more about Amazon. You sort of alluded to this e-commerce being a better business than it previously has been. And part of that thesis may be due to robotics. And you and Shawn covered Amazon months back and you initially decided against owning Amazon. But after the recent drawdown below $200 you decided to add it to the portfolio. Talk more about what made you consider Amazon and talk a little bit more about the robotics piece as well.

Well, Amazon is obviously one of the best businesses out there. So we probably knew that at some point it would likely make its way into our portfolio. And when Amazon dipped below $200 it just felt like we should take that opportunity. It can always go lower and it’s not fully rational to buy just because the arbitrary price of $200 has been undercut. But it did feel like a good moment to start a position and it was also way below the fair value estimate that we considered back in our episode a couple of months ago. So the market currently is certainly scared because of what you mentioned, spending a lot of capital. There is a massive 200 billion dollar investment that Amazon announced and it very well may be that this will cause some overcapacity in the next year and maybe in the next two years. But getting back to the destination analysis, do we really think AWS won’t grow into that capacity? I mean cloud demand is still way outpacing supply and there are only three companies that can supply this at scale. And those are Amazon, Microsoft, and Google. So this isn’t a huge investment with a highly unlikely outcome. It’s more or less clear that the demand is there and the only question is whether Amazon will have too much capacity maybe next year or for the next two years. But not so much whether they have overcapacity for the next five years.

Yeah, I mentioned that I pitched Meta during that Top Stocks episode back in January and that stock is also going through the same issue. The business results they’re publishing are largely very good but the market is much more concerned about these AI investments that they’re making and the returns those investments will generate in the future. Meta is also in a similar boat in the sense that they’re seeing much more internal demand for the AI compute from their family of apps than they’re able to supply, which is why they provided that jaw-dropping guidance for AI spend in 2026.

Well Shawn and I always try to make sure that our portfolio won’t become a Mag 7 proxy. And it really took us about a year until we only covered I think three Mag 7 companies. And for the longest time Google was the only Mag 7 position. And funnily enough it actually doubled for us. So while the small cap investor in me kind of dislikes the idea of owning Google, Amazon and potentially at some point Meta, if they are good investments it would just be foolish not to buy them. But getting back to your point about robotics, I believe that’s probably one of the biggest tailwinds for Amazon. Amazon has used these little orange disk-shaped machines that slide under shelving units and carry them to human workers for over a decade now. But the new generation of robots will actually be able to do the entire process. So we are talking picking and packaging as well. And if you consider that Amazon employs over a million people in its warehouses globally, you can imagine how much they would save if they could automate most of that work. And it’s not only about the money you would save, it’s also about the efficiency gains. Right, so robots don’t need breaks and they can work 24/7. And they already store incoming inventory about 75% faster and reduce order processing time by up to 25%. So to me it always feels inevitable that at some point we will see more robots and less people working in their warehouses.

Yeah, Amazon, it’s just been one of those companies that’s continually redefined the whole shopping experience. I remember back when I was in college ordering online was something that just a lot of people didn’t do. But then, you know, you go online, you figure it out, you hope that the package is gonna show up to your place, and it tended to work most of the time. But then now, you know, I’ll order things on Amazon all the time and commonly have one-day delivery options. And I’m sure Amazon’s gonna figure out same-day delivery at some point as well. For the robotics piece, do you have any idea what that could mean in dollar terms for their economics? You mentioned, you know, a million people working in these warehouses globally. How do you think that’s going to impact their financials?

It’s hard to put a precise number on it, but Amazon’s cost to fulfill, ship, and deliver is supposed to be around $90 billion a year. It costs all shipping and fulfillment costs. And even a 10 to 15% reduction, which is likely to be very conservative, would add about $9 to $14 billion to the bottom line every single year. And that’s something that I would expect to see very soon. And I actually wouldn’t be surprised if the majority of those $90 billion in costs can be saved in let’s say five years from now. And mind you, these are tens of billions of dollars in profits without selling any additional items. It’s really just about the cost savings that you can now take advantage of. And just getting back to the scale economies shared model that you mentioned with MELI, if Amazon’s fulfillment costs come down, there are basically two things that they could do. They either improve their margins or they drop prices further and basically reinforce their market position. And that’s what Amazon has done for decades. And in reality it’s likely a mix of both that will happen. I do personally feel though that Amazon by now is so strong in the markets that it’s operating in, that I feel like in the future they will likely take more margin than they have done in the past.

Yeah, and I would assume that MELI has similar plans to automate their warehouses as well.

Absolutely. MELI is deploying robots in its warehouses too. Generally this will come to all big e-commerce retailers with the logistics footprint that Amazon and MELI have. So it should be a tailwind for the entire industry. But if you imagine that they start turning these segments into meaningfully profitable businesses, I think that will be a huge boost to profits. And Amazon is the company with the biggest revenue in the world. So most of that comes from the marketplace. And if they turn around those economics, I think the earnings power is beyond anything that we have seen with Amazon in the last 10 years. And margins have already inflected in 2025. So they have doubled their North America margin, which is basically where you have the entire e-commerce operation, from 5% in 2021 to 11% today. And I do believe that the true earnings power is still vastly underappreciated.

Yeah, and tying back to jumping on these massive secular trends, I can’t help but also mention Amazon’s ads business alone. I’m personally a big fan of digital advertising, you know, this is a market that has just been growing at an incredible clip and has a fantastic margin profile. And Amazon now has one of the largest ad platforms in the world with more than $70 billion in ad revenue. And as we all know, that’s some of the highest margin revenue one can get. And it shouldn’t be a surprise that ads are a growth driver for MELI as well. In your pitch, you also talked about their highly successful ad network, which is only about 2% of gross merchandise value, while Amazon’s is closer to 7%. So there’s a lot of room for that to grow for MELI as well. And turning back to Meta, once again, one thing that I love about Meta is that there’s just so much internal demand for that compute and so many ways to utilize AI, that it seems unlikely that these investments won’t pay out at least for the time being. You can arguably make an even stronger argument for Amazon on that front, since the ecosystem is even wider. As a matter of fact, just this week, Andy Jassy shared in an internal all-hands meeting that he expects AI will help fuel AWS to become a $600 billion business in 2036, which is double the previous estimate of how large they expected that business to become. So that would imply that he essentially expects AWS to grow at a 15% CAGR over the next decade. It’s also worth highlighting that AWS growth has been accelerating. As in the recent quarter, revenue grew 24% year-over-year. We’ve already talked a bit about Amazon’s investments and how they can utilize it via AWS, but what other ways do you see Amazon making use of AI? And I know that outside of this recording, you also said that you like Amazon right now because it feels like one of the few bets where the upside through AI is clearly more dominant than a possible downside.

Yeah, maybe that’s why I like Amazon so much because it does feel like, also talking to people in the industry and talking to people in our communities, that AI makes things so uncertain. There are companies that we thought would be huge and structural winners for the next 10 or 20 years, and Constellation is actually one of those companies. And then AI came into the picture and you just don’t really know how the company will look 5 or 10 years out. And to me, Amazon is one of those companies where I feel like I have a pretty clear picture or at least idea of how they could actually benefit from the entire AI investment cycle. So Amazon’s AI strategy is basically stretching over their entire ecosystem really. So what you think about first is probably AWS Bedrock and selling AI compute to enterprise customers. And what Bedrock does is that it basically gives customers access to hundreds of foundational models, including Amazon’s own Nova models, but also Anthropic’s Claude where Amazon holds a substantial stake in, and now through a recently announced partnership, OpenAI’s models as well. Amazon actually invested 50 billion dollars in OpenAI as part of that deal. And I’m not a huge fan of that, but to some extent, I mean they’re hedging their bets. And considering how fast Claude seemed to have outpaced ChatGPT, who is saying that it can’t change quickly again? I mean, I’m humble enough fortunately to know that I don’t even know one percent of what the decision makers at Amazon know. So I’m better off letting them make those decisions for me. And what I like about this strategy is that Amazon is not really competing in what to me seems like a very commoditized LLM playing field. But I’m probably even more excited about how Amazon can utilize AI internally to improve its own business compared to the investments that they make into OpenAI or Claude.

Yeah, that’s exactly my thesis with Meta. They’re already seeing these AI investments generate strong returns internally, and I can imagine some ways in which Amazon could also utilize AI in their own business. For example, they might provide more customized search results based on individual preferences, or better ad targeting, similar to Meta. What are some other examples of this for Amazon?

Yeah, some pretty exciting stuff you basically just mentioned where they could, you know, sell even more to customers because the recommendations are getting better. One thing that we also talked about already is robotics. The reason the next generation of robots can pick 65% of Amazon’s product catalog is because of advances in computer vision and AI. I mean, every robot arm in Amazon warehouses is essentially a huge AI system that has to identify a product or an object, figure out how to grasp it, and then place it. And that’s a quote unquote simple machine learning problem. But that is getting better with the more data you have. And there’s pretty much no company that has more data on that than Amazon. And then you also have the customer-facing side of the business, which, you know, is mostly about improving the algorithms and getting products to customers easier. One of those things that they invested in is Rufus, which is their AI shopping assistant. They actually invested in this a while ago, but it keeps getting new upgrades. And instead of, for example, typing keywords into a search bar, you can now have a conversation with Rufus about what you’re looking for, and it will basically guide you to the right product. You can also, and this is in my opinion a better feature, get information on the price history of a product and then set price alerts and automated buys. So for example, I want to buy, you know, something for my house and I know I buy that every single month. Then I can say, okay, well at this price, let’s say $20 for coffee or whatever, this is when I will always buy my coffee. And I won’t buy it if the price is increased. And you can also see where it traded in the last let’s say four weeks. So this is a feature that I personally like. And I would say that generally all of this will feed even more data into Amazon’s AI capabilities. So you could also imagine that in the Rufus conversation itself, over time you will, for example, have promoted products. There are just so many options to monetize this further. Although Rufus itself is, I would say, still far from being perfect. And I think it makes a lot of the typical LLM mistakes that we see across the board. And I would probably personally not yet use it for my my shopping on Amazon.

Actually have a little bit of experience selling products on Amazon. And through that experience, it was clear to me just how dominant of a position Amazon is in. And, you know, I read several of the books that are about their history and Jeff Bezos, as I’m sure you’ve read as well. And, you know, you just read how so many companies try and fight Amazon. And at the end of the day, they still end up listing their products on Amazon, even though Amazon’s taking a 30% cut and getting a massive slice of each transaction through their FBA fees, their fulfillment, and everything else that they provide. So the set of customers that they have is just insanely valuable, and it’s going to be extremely difficult for anybody to disrupt. But if there’s anything I don’t like about Amazon, it’s probably the name that they chose for the AI shopping assistant. Surely they could have come up with something better, but maybe that’s just me. Part of me feels like Amazon is following everyone else’s lead and just putting a chatbot embedded into everything, even if most people don’t necessarily need it. Admittedly, I’ve probably used it just a couple times over the past six months or so, but I could imagine it being quite useful in asking me to tell me the difference between two of these products, help me generate gift ideas for my six-year-old nephew, or help me get everything I need to put together for my home gym, just a few examples. And shifting gears here, as we were prepping for this conversation, I was happy to see that you wanted to discuss Constellation Software and their two spinoffs. I first covered Constellation back in February 2023 back on episode 531. And I think it’s safe to say that that was one of my most popular episodes I’ve ever put together. As I was prepping for that episode, I read Mark Leonard’s shareholder letters, and essentially was compelled to purchase shares on the spot. The letters were just that good. And I was just chatting with Francois Rochon here on the show, and he heard about Constellation on Christmas Eve, I believe it was 2013, at a Christmas party. And he was pretty upset when someone mentioned the company and he hadn’t heard of it before, he wasn’t familiar with it. And he had a very similar experience in reading, you know, going home, reading the letters that night. Now today, pretty much everyone knows about the company, given the stock’s incredible run, and recently somewhat dramatic fall.

I guess getting information on Constellation in 2013 qualifies as probably being the best Christmas gift that you could have gotten. I actually think it’s incredible and it speaks for Mark Leonard and his skills also as an author that everybody feels the same way about his letters. I must admit I only read them recently, maybe a couple of months ago, for my research on Constellation. And I couldn’t motivate myself before that to actually research Constellation or even read the letters of Mark Leonard when it was still trading at, you know, 40 times cash flow. Since I just knew I wouldn’t actually buy the company. So I am thankful for the drawdown, although I probably shouldn’t say that here on this show and speaking to you as a Constellation shareholder, a longer-term Constellation shareholder. But I finally had an excuse to read the letters and, boy, Mark Leonard is really an outstanding guy. I felt the exact same way. I read them and I felt like he understands business so well. What he has built is exceptional anyway. So I really felt like I wanted to buy his shares immediately.

Yeah, and anyone following Constellation knows that the share price is down due to AI taking down the stock prices of so many AI companies. And AI, it could come for Constellation’s business in some ways. I know that most of our listeners will be aware of Constellation, but for those who aren’t, Constellation owns more than a thousand small, niche vertical market software companies. Things like cemetery management software, court administration systems, or public transportation scheduling. So the market’s fears around AI is twofold. The main concern is that creating such software with the help of AI is now much easier, much cheaper, which is why competition will enter these verticals where Constellation is active. So that would mean that their current portfolio of over a thousand companies could be disrupted at least partially, and that the pipeline for future acquisitions might be shrinking. So if AI tools flood the market, there won’t be any high-moat VMS companies up for grabs anymore. So that is the simplified bear thesis. Talk about your take on this.

I think there’s a lot to say about this. So perhaps we should start by just quickly explaining the nature of VMS businesses once more. So VMS stands for Vertical Market Software. And those are the types of niche businesses that you basically mentioned. So the reason why Mark Leonard, the founder and former CEO of Constellation, liked these businesses in the first place is that they are as close to everlasting as possible. In fact, this is why the market for 20 years was certain this is an undisruptable business that should trade at 40 times cash flow. So we are talking about industries here where the leading edge of tech is simply irrelevant. So the assumption of tech guys out there and also investors is that everyone would adopt new technology as soon as it is there. But everyone who has worked in the corporate world knows that’s not the reality. We actually just talked about Kinsale a couple of weeks ago and you talked about the insurance industry and how just changing software at such a company takes a whole lot of time and is probably the more inefficient choice than just keeping your old software. And if employees are used to a certain software or system, it can be way more efficient to stick with what works even though you might save a couple bucks on implementing the new software or it can technically do more stuff than the old one. There is a rule of thumb that generally something has to be 10 times better than what you use right now to actually be willing to change it. And this is obviously not true for you and me. I mean we can change the software that we use in an instant. But it is true for companies that handle critical data. I actually switched from ChatGPT to Claude in a matter of a day. I felt like I was some kind of software engineer when I did that. Up until I figured out that pretty much the entire world changed to Claude at exactly the same time. But that is just an example that there’s basically no switching cost involved for me if I switch from one LLM to the other. But imagine you’re a company or you’re a public institution which is having decades worth of data and extensive knowledge in your current software. Well at that point you can’t just change software overnight. A good example is court administration software. So court records have a legal chain of custody and that means there must be a clear, documented history of who created a record, who accessed it, who changed it, when, and why. And if you switch software, you have to prove that nothing was altered during the migration process. The fact that, you know, timestamps and access logs remain intact, that permissions stayed correct, and that all records are still admissible in legal proceedings. And the amount of extra work required makes switching almost impossible in practice. And every AI tool involved would have to be tested and certified to handle that level of data sensitivity. So in this case, you would really only change the product if the alternative is actually 10 times better or significantly cheaper.

[Full transcript saved — see lines above for complete content]

Yeah, and again, that’s one of the common arguments, right? The idea is that AI now allows others to create new software solutions at a significantly lower cost. And the fact that AI now enables the creation of such software much more affordably means that it can then be sold much cheaper to the current client base of Constellation.

That’s one of the main arguments, but I do think it’s missing one important point as well. So the software itself can certainly be produced cheaper, but that is not really what clients are paying for. So over three quarters of Constellation’s revenues actually come from maintenance. So customers pay for ongoing support with the systems, making sure that everything is running and updated. And that support team is where the main cost base for Constellation lies. So you would need such a team even if you created a software tool using AI. And except for the fact that it might be easier to build a leaner company structure if you start from scratch, which has always been true, not only since the age of AI. I don’t see why Constellation shouldn’t be able to reap the benefits from the fact that you now only need let’s say two instead of five software engineers thanks to AI.

And then there’s the cost structure piece, where CSI’s software is rarely a big cost factor for their clients. So a number that I always see floating around is 0.1% to 1% of revenue, but it can be lower for companies with lower margin profiles. So to help put this into perspective, if a golf course is making $2 million per year in revenue, perhaps they’re paying $10,000 on the software that handles tee time reservations, membership dues, member data, payments, and provides solutions for managing tournaments. It’s not entirely known how much it costs and it depends on the company at hand, but given that Constellation has retention rates far above 90%, it seems that the value proposition matches the pricing pretty closely. One example I actually came across is that there’s something like 30 golf courses in Indonesia, and almost all of them use a software within Constellation’s ecosystem. So to me, this is a perfect example that if anyone knows the needs of the customers in that niche market, it’s going to be Constellation. So they are best positioned to potentially utilize AI to improve their offerings and add even more value to customers. And as these software solutions only become more embedded into the company’s workflows, this would naturally increase the switching costs over time as well. And I also just think back to the argument of, you know, is a manager of a golf course going to pull up Claude and just put together their own software solution to save $10,000? It just doesn’t seem to be likely to me at all. If someone manages a golf course, they probably just like to golf and they like the camaraderie around golf and it provides them a good living. And if they try and do that and the software fails, then they are out significantly more than the $10,000 they could have just paid Constellation. Another part of the Constellation story I think back to, the episode I did in 2023, and sort of my thesis back then, what really appealed to me was that these businesses are relatively resistant to poor economic conditions. So when the economy inevitably enters a recession, since these solutions are mission critical, the money just keeps coming in because these are relatively fixed payments and they’re recurring in nature. And better yet, Constellation can actually come out thriving on the other side of a recession because they run into desperate buyers that need to exit their software company. Such as a struggling conglomerate who wants to carve out a niche part of their business to help them weather through a crisis. You just run into these sellers that need to sell and don’t have to sell, which means Constellation can get some really attractive prices.

I like the golf example because it often shows that you don’t need just mission critical industries in general, it’s also enough to just be mission critical to one company. Even if there’s a drawdown, even if there’s a recession, and you know, less people are golfing, even though you could make the argument that the people who do golf probably don’t care that much about the recession. But anyway, the $10,000 that this business has to pay for the software, they will always have to pay it. If the business is still in business itself, then it has to pay for that software because otherwise the entire business would just not keep living on basically. And I gotta say, another thing that is probably underappreciated is the factor of distribution. So let’s say I have extensive knowledge in the gym sector. So I decide to build software for gyms. Now it does not only need to be 10 times better, which we have discussed, it does not only need to be cheaper with the same level of support and service, but I also need the distribution to actually reach out to gyms and make them want to adopt my product. And as investors who likely visited a business school, it’s easy to sit at home and think about the efficiency of capitalism and be like, your product is better and cheaper, then people will also use it. In reality, that’s not the case. It’s not always the best product that wins. I would first need to figure out how to even reach gym owners and then pitch them on the idea to switch their entire software system. And I’m not sure how many gym owners will be open to that. After all, their business is already doing well, and if anything it just introduces risk that they didn’t have before. That’s when switching costs generally come to play. I think everyone would agree that highly regulated industries might be difficult to disrupt. But there are switching costs in other industries that have low to no regulation. We just talked about the golf course, but the same is true for gyms. Gyms make a large portion of their profits from members who pay regularly but rarely ever show up. And those members don’t cancel proactively due to a mix of convenience and also the thinking that they just like to pay for a gym because technically they could go, and of course they are fit people. But the moment you switch software and perhaps something goes wrong with the payment data, you have to ask customers to actually re-enter their credit card information. And that most of the time will be the trigger point where cancellation decisions actually happen. And suddenly you’ve lost a good chunk of your most profitable customer base.

And since Constellation is known for its low single digit organic growth rates, much of the value creation is happening through the acquisition front. Constellation is historically known for buying these really tiny VMS companies that can actually move the needle because they just do so many of them. We’re talking about 100 plus acquisitions a year at around a $5 to $10 million purchase price. Leading up to this discussion, I was taking a look at some of the most recent deals that have been announced. So they purchased a company called Comprose. It’s based in Missouri. The company was founded in 1987 and it had an employee count of roughly 20. Another company they purchased is called SK Trend, which also has 20 employees, but it’s based in Budapest, Hungary. So Constellation clearly has this web of companies all over the world. So SK Trend, this is a company that is in the insuretech space. They provide a comprehensive platform for insurance brokers and partners. And there’s lots to nerd out on in terms of just digging into what some of these companies do that Constellation purchases. Recently though, they’ve been dipping their toes into public markets as they announced a new strategy called PMS, permanently engaged minority shareholders. So instead of buying businesses outright, they target these larger public companies, take a stake in them, and try to get seats on the board and push through some operational changes to improve the business and improve its efficiencies. I’m curious to get your take on this. Whether they’re sort of running out of these smaller deals or they just need to find other ways to get capital deployed.

Yeah, I think it’s really much not about AI, which has also been a theory that people say, you know, now you have AI coming into the market and they immediately don’t find any good opportunities in the private market anymore. I don’t think it’s that, it’s mostly just the law of large numbers. So Constellation has said previously that it’s just difficult to go way beyond 100 acquisitions, which is currently what they average per year, without diluting the quality and still hitting their target returns of about 20 to 25%. So I do think it was a matter of time before we would see this strategic expansion. And if anything, I think AI pushed forward the timing. There’s an argument to make that CSI should actually see more M&A activity right now because of the depressed valuations in the sector. However, prices in the private market move much slower than in the public market. So if you are the owner of, you know, a private software company whose business is not touched by AI at all, why would you accept a lower purchasing price for your company simply because investors in the public market are spooked. So it’s only natural for CSI to look for opportunities in the public market where, again, valuations are much lower due to these AI-related fears. And this naturally comes with some risks that you didn’t have in the private market. I mean as a value investor we don’t necessarily believe that markets are always rational and efficient. But the public markets are more rational and efficient than private ones, just because there’s a lot more liquidity that’s pushing, you know, money to the stocks that do well and money away from the stocks that don’t do well. So when you buy a company that is down 90% in the last five years, as it has been the case with the acquisition of Sabre, the company that Constellation just built a 10% stake in, there is a good chance that there are real problems within the business. And also turning a business around as a minority shareholder is a much more difficult game than in the private market when they buy a business outright.

Yeah, it was interesting to see Sabre’s stock price jump on the CSI announcement. To me, it makes a lot of sense. If a business, you know, there is a core product that is maybe a decent business for the next, say, five years or so, I mean Constellation, they aren’t always looking out, you know, what business is still going to be around in 50 years. Like if a business is generating cash and it’s likely going to be generating cash for the next five years, and they believe that it’s significantly undervalued, then, you know, it might make sense for them. And what’s also interesting is that for me, as an investor, if I’m looking at Sabre, I might look at the balance sheet for example and say this company has way too much debt. But Constellation might be able to make the numbers work for them much better if they’re a strategic partner in helping them work through the debt issue. And I think you’re right about private market valuations just being much slower to react. Constellation’s Chief Investment Officer, he actually shared that they haven’t really seen valuations change at all in the private markets. And that’s competition to purchase those businesses is still very strong. So it’s interesting to see this, these conflicting viewpoints where public markets are obviously very spooked, but in the private markets, demand for these businesses is still quite strong. One of the things I admire about Constellation too is just how adaptable they are and deploying capital in the best way possible. Whether it’s doing the smaller deals, going into the public markets, or venturing into new markets, or at least considering new markets, they will pursue what is in the best interest of their shareholders. So the public company you mentioned is Sabre, ticker SABR. This is a software solution for the global travel industry. And what stood out to me about this deal was that it sounds like Mark Leonard of all people is leading the charge on it. So it does seem that his health has improved in recent months, and he’s still doing what he loves at Constellation, following the critical procedure that he had last year. Sabre’s management team, they actually threatened what’s referred to as a poison pill, which I had no idea what this was before they mentioned it. So they threatened a poison pill when CSI informed them about their stake. A poison pill, to put it simply, it’s a way for the current management to maintain control by diluting shareholders’ holdings, thereby reducing CSI’s stake and their influence. So obviously, this is a suboptimal outcome for every shareholder. Sabre’s management, they didn’t end up using the poison pill option, but it shows that these things just can get messy, especially since there might be selection bias. So the management teams of bad performing stocks might not be the most exceptional management teams out there. In the end, I personally don’t see AI as a huge risk for Constellation more broadly. On the margins, of course, it could be impacted. I’ve gotten to know one of our community members pretty well who runs a VMS acquirer, used to work at Constellation, and he shared with the group that these verticals will always have the need for a professional vendor that is not the end customer to take all of the tools that are available, AI being one of those tools, and piecing together a solution that adds value in the best way possible for the customer. So to go back to my golf course example, could the course hire a software engineer to create a software that helps them book tee times? Of course! But first, there’s so much that can go wrong in adopting a new solution. And second, even if there are some cost savings, that golf course can still be missing out on some additional features that the vendor could be offering them. Which, in the end, will likely help the company make even more money. It’s not like the end customer is getting screwed and price gouged. This is a low amount of their overall cost structure and critical to their operations. So instead, their size and the shift in how they approach M&A as they scale, it’s going to be really interesting to see how it plays out. They are very strict on what returns they’re targeting in all of these deals. So I’ll be continuing to keep a close eye on how much capital they’re able to deploy each year. I think largely for me, this is a sidecar type investment where I believe these are some of the best managers in the world, and they’re not going to pursue opportunities that have even a fairly significant chance of not creating shareholder value. So I know that you spend also some time looking at the spinoffs as well, and other public companies that they’ve gotten involved with in Europe. I know you and I are pretty interested in one of those today especially, but let’s talk about Topicus and Lumine. So these were both spinoffs over the past few years, and I’m also excited about the spinoffs due to their incredible track records paired with, again, exceptional operators, and these are coming from a smaller starting base from which to compound from. So interestingly, it seems that the market puts a bit of a premium on Constellation’s reputation and track record, because it seems that the multiples of Constellation relative to the spinoffs is fairly similar. So for me, I think it’s going to be much more difficult for Constellation to continue to grow at the rate that they have historically. You know, again, they’re going to need to purchase 100, 150 plus companies per year to keep growing at double digits. While the spinoffs, for them to compound, you know, look at Lumine for example. They could do two to four deals a year and still compound at 20% plus, which these also tend to be bigger deals than Constellation’s bread and butter, but still, two to four relative to 150 plus is a huge, huge difference. And Lumine, they only own around 34 companies relative to Constellation’s a thousand plus. So in theory, it’ll be much easier for them to move the needle.

Yeah, looking at all these companies, the spinoffs but also the companies that the spinoffs buy, has been really an incredibly interesting look into how the entire CSI machine actually works. And why also you have to pay a premium for those companies, because even though they don’t have a proven track record to the same extent that Constellation has them, you see the exact same way of operating. So you have the same incentive systems. You have oftentimes people from either Topicus or CSI at the helm at these companies. So you kind of know what you’re getting and you know that this has worked so well in the past that as soon as they acquire a company and it’s a small base that company is starting from, it’s incredibly attractive for investors to look at those companies. And especially today where the same trust that CSI and those companies have gotten in the last years is not there anymore due to AI. And you probably know all of these companies even better than I do since you’ve owned them for a while now. I looked at a lot of companies in that universe including the ones you alluded to in Europe which are mostly Polish companies. But I think I’ll focus on Topicus and Lumine here. We will have an episode coming out soon where we cover those other companies as well. Topicus and Lumine differ in that Lumine is focused on a specific vertical but not bound to any geography, while Topicus is bound to a geography but operates across many different verticals. So Topicus was spun off in 2021 and it sits on top of TSS, which stands for Total Specific Solutions, which is basically the actual operating unit that goes out and then acquires other VMS businesses for Topicus. So Mark Leonard, the founder and former CEO of Constellation, is sitting on the board of Topicus and Constellation owns about 30% of the shares. The interesting thing about Topicus and TSS is that they operate in Europe. And more specifically in the Northern and Western parts of Europe. And I talked a bit about the European VMS market in an episode I once did with Shawn on chapters group, which is a German VMS acquirer. And generally speaking, it’s just a lot more difficult to consolidate European companies because you know the cultures and most importantly the languages are so different. Those are problems you don’t face in the North American market where everyone speaks English. And Topicus is one of the few companies in the European market with enough scale to actually consolidate companies on this continent. And another huge tailwind is the succession problem, which is even more acute in Europe than it is in North America. There are tens of thousands of highly successful software companies that don’t have a successor in place and will either be sold or just have to close down shop. And Topicus is undoubtedly the closest copy of Constellation among all spinoffs and subsidiaries, which obviously makes it pretty interesting.

Yeah, looking at Topicus’ web of companies, it looks similarly complex to Constellation, and the incentive system is practically identical too. One interesting nuance though is how the incentives differ slightly depending on what each part of the business is actually supposed to do. So Topicus is made up of two main arms. You have the original Topicus business, which has historically been more focused on growing its existing products organically. And TSS, the acquisition engine, that goes out, buys more VMS companies, and each arm does have its own CEO, and their bonuses reflect that difference. The TSS CEOs are measured on overall revenue growth, including acquisitions, because deploying capital into new deals really is their job. The Topicus CEO, on the other hand, is only measured on organic revenue growth. So acquisitions don’t count toward his bonus, because his unit is expected to compound what it already owns, not grow through acquisitions. So it’s a small but deliberate distinction that shows how detailed companies in the CSI universe are with these types of questions.

And to me that’s what we talk about when we say that Constellation is just operationally excellent. And that’s something that you don’t often see in most other companies that you know try to copy what Constellation does. And to me Topicus is probably an option to invest in a somewhat younger and less mature CSI, with the option to benefit from the European market tailwinds that I’ve just mentioned. But if you’re worried about the law of large numbers with Constellation, I think there’s an argument to make that Topicus will run into similar problems soon.

Yeah, and I actually covered Topicus on the show shortly after that Constellation episode back in 2023 because I just couldn’t help myself. And as you mentioned, Topicus being based in Europe, that really attracted me. The thesis for me at least was that there’s less private equity activity in Europe, which would give Topicus less competitive pressure in making acquisitions. And Europe is also a highly fragmented market, as you mentioned. There might be two different companies in a very small niche that do the exact same thing, but one may be based in the Netherlands and the other is in Hungary, which have different sets of laws and regulations and different languages. You just don’t have that same dynamic when you’re looking at Texas and Nebraska, for example. It’s largely going to be the same. So Lumine was also quite an interesting company to look into. And since I liked Constellation, Topicus, and Lumine for their own reasons, I actually decided to take an ownership stake in all three. Lumine is not bound by geography, but solely focuses on companies in the media and communications industry. Again, their track record of growth under David Nyland’s leadership is just phenomenal. Like you look at what they did when they were under the CSI umbrella entirely and it’s just unbelievable to look at what they’ve done to date. So the whole Constellation ecosystem is also interesting to me is that these guys just totally avoid the press and they just fly under the radar. They are just solely focused on creating shareholder value. I actually invited Nyland to join me here on the podcast. I wasn’t able to get him on, which is probably best for me as a shareholder that he continues to focus on his business. Again, Lumine is not bound by geography. They focus on the media and communications industry. And it was a company that was started in 2014 and then spun off of Constellation in 2023. And I really just couldn’t believe the price drop below $20 just last month, which would have been unthinkable just six months prior when the shares were above 50 and then they drop below 20. It’s it’s quite amazing what markets can do in swinging in both directions.

What’s attractive about Lumine too is the type of companies they go for. So they focus on so-called carve-outs, so orphan parts of larger companies. And one of the advantages of carve-outs is that you don’t have a lot of competition for those deals. And the parent company is also often happy to get rid of that part of the business. So the price that you have to pay is relatively lower compared to other parts of M&A that you could do. And another advantage is that there’s lots of potential for improvements at the acquired companies because they have been neglected for a long time, which means there’s a lot of potential to increase margins just by getting rid of unprofitable contracts, repricing products, and also integrate Lumine’s best practices into those companies. The downside though is that it’s quite difficult to predict reinvestment rates in any given year. I mean in some years Lumine acquires a handful of companies, and in others they only buy one or two companies. Carve-outs though tend to be larger than the typical Topicus or CSI acquisition, which is why it doesn’t need as much volume to move the needle for Lumine, which to some extent is shielding you from the problem of, you know, the law of large numbers. Adding to that that Lumine is also just a significantly smaller company today.

Yeah, it’s uh, what’s also interesting about these spinoffs is that, to understand Topicus and Lumine, you clearly need to understand Constellation. To understand Constellation, you need to read the letters, get acquainted with this being a Canadian company that buys companies all over the world. So there’s sort of multiple layers to this. And with Lumine, the carve-out dynamic is also the reason for the high volatility in their organic growth. So with Topicus and Constellation, it’s just a slow and steady, low single digit organic growth, consistent M&A engine. But Lumine, you even see some quarters where organic growth is negative because they’re doing a lot of work on these businesses and optimizing them upon making that acquisition. And it really takes some time to see those positive organic numbers start to show up in the financials.

Yes, one of those things where it’s not enough to just look at the numbers because if you do so, you have no idea what’s going on or why it’s so volatile. But when you just think about it, I mean these companies that they acquire, they were part from a parent company and Lumine basically has to rebuild all of it from scratch. I mean that means you restructure processes and again, you often cut some contracts which of course puts pressure on the topline in the short term. So as you just said, it’s really about understanding the model and then also trusting the process. Because if you buy any of these companies, you have to trust the people at the helm and then, you know, kind of sit in the bandwagon and hope they do what they’ve done in the last two or three decades.

Yeah, one of my favorite lines from Mark Leonard was in the press release regarding the Lumine spinoff. He stated, I look forward to having CSI’s long-term shareholders become long-term shareholders of Lumine Group. I hope my grandkids are still holding Lumine shares 50 years from now. And if you know Mark Leonard, you know that is quite a bold statement from him. And certainly a stamp of approval for who David Nyland is and what he’s doing in building Lumine. I know we’re jumping a bit here, but I also quickly wanted to ask you about your opinion on SaaS in general. I know the Intrinsic Value portfolio has some SaaS companies such as Adobe. You’ve looked at companies like Salesforce. Talk about how AI plays into SaaS more broadly because I think a lot of investors have come to the conclusion that VMS and CSI will be pretty resistant to the AI threat, at least in the near term. Who knows what happens, say 10 years from now or more. So talk more about software more broadly and how AI could impact that industry.

I should probably use that question as an opportunity to say that I’m not naive about the AI threat. Even for CSI, we’ve spent dozens of hours in our community talking about the AI threat, and many of our members understand the topic much better than I do because they work in the space. And you just said that, you know, a lot of people kind of consider VMS to be more secure in terms of the AI threat. I think if we look at the market, that’s not necessarily the case yet. I think to some extent we’re living in a bubble where we feel like people understand these companies pretty well. And I think there are good arguments to make that they will survive the AI threat. But I apologize for making it about Constellation again for a minute. But I do think that in Constellation’s case, the more detailed risks I see, and to some extent they also apply to all of these SaaS companies, are that agents will take over the majority of workflows. And then the underlying software loses pricing power. And besides its data storage advantage, also switching costs. Staff wouldn’t be trained on the software to the same extent, because basically the agent does most of the navigating and actually operating the software. And that could be the case for Constellation’s companies. But that’s also something that you can think of for Salesforce. I mean we’ve discussed this in detail in our community, but I think it would go beyond the scope of this episode. Just so much: there’s obviously the opportunity for Constellation to build its own agents, but in this case, CSI’s decentralization, which is usually a huge advantage to how the company operates, could actually become a major competitive disadvantage. Because technically, horizontal agents, like the ones from the big LLMs, they are trained on more data points than any other model could ever be. And Constellation subsidiaries operate in silos. So they don’t share databases, infrastructure, or even engineering talent. So each subsidiary’s data is kind of locked within its own walled garden. And then a competitor building a vertical AI for let’s say utilities, for example, can consolidate data across multiple utility clients to then better train models. And that could lead to the fact that they simply have better models than Constellation can build by itself. It’s also a large investment. If you can amortize the cost of building an agent or a specialized LLM across basically thousands of products. So ServiceNow for example, one of the SaaS companies that has come under a lot of scrutiny in the last couple of weeks and months, they could technically build one AI workflow engine for the entire platform. And then if you think about what Constellation could do, well they have a thousand plus businesses and each would have to build or integrate their own AI independently. And at a per subsidiary cost structure, that’s maybe not even an economically viable choice for the average 5 to 10 million EBITDA subsidiary that Constellation owns. And then arguably even larger problem though is the ability for companies to actually build their own software solutions from scratch. I actually just listened to Bill Gurley, who is a venture capitalist who I like to listen to from time to time. And he talked about Salesforce recently, and how most companies only use a fraction of the whole product suite that Salesforce actually offers. And I analyzed as you said Salesforce a while ago. And while I believe it will be very hard for any other company to disrupt Salesforce. I think insourcing might actually be a problem because if you think about the customer base. Salesforce’s customers are the biggest companies in the US. About 90% of the Fortune 500 companies are customers of Salesforce. So those are the companies that certainly have the resources to insource this stuff and basically tailor it to their personal needs without paying for software that they never use. You know, we talked about the golf course for Constellation. If I’m the founder of that golf course, I will not spend my time vibe coding through Claude my own software solution. But if I’m a huge Fortune 500 company, I can hire software engineers that through Claude might have the ability to now basically program their own system. And then I can also employ them to basically do all the maintenance that Salesforce would currently do. That said, and this is the tough thing about AI right now. The opinions of people who use Salesforce daily in their jobs compared to those who use AI daily just couldn’t be more different. If you talk to people who use Salesforce, they always say I don’t like it, but it is impossible to remove it from our company, we just couldn’t function anymore. And then you talk to people who use AI daily and they say, it’s almost certain that this stuff can be insourced and it’s probably happening rather sooner than later. I think ultimately I believe the software companies that will survive are the ones with you know these massive ecosystems that basically combine dozens of critical software solutions or software that’s actually hard to replace by AI.

Yeah, I mean it just reminds me of your comment earlier that just because there’s a better product out there that might be cheaper doesn’t mean people will necessarily use it. So similar to how the stock market isn’t always rational, I guess people aren’t always rational as well, which makes our job even more difficult as investors. You know, I could certainly see the case where some larger companies find it worthwhile to insource some of these solutions, especially if they have a lot of resources to do it and take on that sort of undertaking. But then I could also see why even bother with a lot of these software solutions insourcing if you know it’s a relatively low cost spend and it’s just been a great partner you’ve worked with for potentially decades. So it’ll be certainly interesting to continue to watch it play out. And you know, as with Constellation, if you knew nothing about AI, you would think it was just business as usual. If you were only looking at the financials of how much revenues increasing, how much their acquisition spend has been in recent years. So yeah, it’ll be interesting to watch continue to play out. I’d like to shift away from AI for the final segment of the show here and discuss one more company. A few weeks back I put together an episode based on the principle of inversion. So instead of asking what company is well positioned to succeed in this rapidly changing world of AI. I asked, what company is almost certain not to fail? And that can be a tough question to answer given how fast technology is advancing and how many industries technology is getting its arms into. The company I covered in that episode was Linde plc. This is an industrial gases company, that episode number was 796 for those interested. And that sparked me to want to chat with you about Hermès. Which as to the time of recording is down more than 40% from its all-time high, and it’s undergoing a sharp drawdown due to I believe the conflicts happening in the Middle East. At least that’s the narrative. A couple of months ago you covered Hermès on the Intrinsic Value podcast. It’s a company I’ve dived into on the show as well. And it’s one of the most fascinating businesses I’ve ever looked at. I don’t think there’s any other brand that embodies the luxury playbook better than Hermès. Because of that you’d always have to pay up for the shares. When a company trades at 50, 60 times earnings it’s easy to sort of have it fall off your radar. But now the valuation might be getting to a little bit slightly more interesting levels.

Talking about falling off the radar, I actually didn’t even notice the massive sell-off until a member of our community posted about it. And I guess the problem right now is that there are so many other high-quality stocks selling off that it’s kind of difficult to make a decision regarding where to add. But I do believe that Hermès is one of the best opportunities in the market right now for a relatively smooth let’s say 10% annual return for the years to come. If I compare Hermès to LVMH for example, I would always prefer Hermès. Technically LVMH is a more diversified company and one could argue that’s an advantage. But I think in my mind there are only a handful of high-end luxury brands, and LVMH just doesn’t own that many of them. They own a lot of brands that basically target the aspirational luxury market. But those markets are much more bound to macroeconomics and they carry a lot more brand risk due to possible overexposure of the brand. And Hermès is really targeting the top 1%. You could probably argue it’s the top 0.1% of the population. And that’s the fastest growing segment of luxury buyers, growing at a CAGR of almost 10% per year compared to only 1% for the aspirational buyers. And it’s barely dependent on macroeconomics. It’s almost impossible to replace a brand in this segment. I think the biggest risk is that a brand of Hermès’ caliber is kind of losing its exclusivity by trying to sell more volume and then invite these aspirational buyers into the ecosystem.

Yeah but that seems pretty unlikely in the case of Hermès. The company is family run, you know it’s been family run since it was founded all the way back in 1837. It’s on their sixth generation of family members running the business. And to this day every heir of Hermès has to begin his or her career as an apprentice in production. They have to spend a decade there also before they can even get to an executive position. So this shows just how much pride the family takes in what they do and how much they value what the Hermès brand stands for. So there’s zero risk that as a shareholder you would wake up one day and hear that Hermès has moved their production to Asia to save on costs or anything like that. And as long as you avoid these drastic mistakes, which I have no doubt that the Hermès family will be good stewards of the brand, it’s hard to replace a brand with just that much prestige and history as Hermès. It reminds me of Buffett talking about Coca-Cola. If you had a billion dollars or ten billion dollars and tried to create the brand that Coke has created, you just couldn’t do it. It’s the same thing with Hermès. So the management truly views the shares in this company as a generational asset and it’s looked down upon to try and optimize the next quarter’s EPS figures. So given that the luxury players are betting on strong growth in China in the years ahead. China has really put a drag on the luxury market overall in recent years. And I was surprised to see that even the Chinese government is cracking down on citizens who flaunt their luxury goods on social media, with the thinking that you know it’s just not good for young people to flaunt things like leather handbags. So yeah the Chinese government is trying to discourage sort of the purchase of these high-end luxury goods. But similar to the macro conditions I would expect this to be just a temporary headwind for the luxury players.

I think it’s one of those news that you read and you kind of get scared over it but I don’t think it’s long term too. I would also say that this is probably yet another piece of news that feels like it could come for LVMH rather than Hermès, because the top 0.1% of luxury buyers they tend to be more understated about their luxury consumption and not necessarily post it on social media. So I wouldn’t expect Hermès to take a huge hit because of these measures. A very long-term risk that I could potentially see for Hermès in China, if you wanna look for one, is a movement which is basically toward more Chinese luxury brands. So there’s a general trend that we’ve seen in China and especially with Chinese consumers in recent years. However, once again I feel like Hermès is somewhat shielded from this because you know the top 0.1%, and to be honest it’s probably even okay to say the top 0.01% that shop at Hermès, are more likely to be global citizens anyway. So the trend of choosing Chinese brands likely won’t be that strong among this customer base. Generally though I just want to point out again that you still have to pay 40 times cash flow for Hermès. So it’s not a bargain at all. And Shawn and I looked at Ferrari a while ago when it was still trading at similar multiples. And that’s a company that is applying a very similar playbook in the auto industry. And nobody would have thought that we would see a cash flow multiple of just 25 times anytime soon. And yet if you look at the chart and the valuation, we are there now. So that doesn’t mean that Hermès isn’t a good buy here. But looking just at you know the relative valuation can be dangerous for companies that have traded at these massive premiums for such a long time.

Yeah, I certainly tend to agree with you. If you’re paying 40 times earnings for any company, you need to have a very long time horizon and you know you have to be fairly certain just that earnings are going to continue to grow. Because the market is implying that double digit earnings growth for at least the next decade, and that nothing will negatively impact the business over that time period. And the other thing is that if shares of Hermès are down significantly, odds are that shares in other companies are down as well, which leads to many value investors just never managing to get a position in. So it is an interesting opportunity though given that the stock is down, but I don’t think anything fundamentally has really changed with the business from a long term perspective. And that’s the key point is that that long term perspective while the market is looking ahead at what’s happening over the next one to two years. But Daniel was actually kind enough to put together a valuation model for Hermès that our listeners can check out. It shouldn’t be used as what the intrinsic value necessarily is, but the listeners can pull up the model, plug in their own assumptions and come up with their own valuation assessment. So I’ll be sure to get that model linked in the show notes for those interested. With that, I think we can close out the episode there, Daniel. Thank you as always for joining me here for this wonderful conversation today, and I look forward to continuing to follow along in your journey here on the show.

Well it was an honor for me to be part of your last episode on TIP, Clay, and I think the great thing is that here on TIP I can really say that we are all friends, so I’m very sure that we will still keep very close contact. And yeah, with that, thanks for everything you’ve done for TIP and all the best for your new journey. And to all of our listeners today, thanks for tuning in and I will see you soon.

Take care.