![Transcrição [EN]: 1º Encontro Anual de Investidores](https://notion.so/icons/document_gray.svg){kind=icon}
Presentation:
{DAVID K} [00:00:00]: Good morning! First of all, thank you everyone for being here. For us, this is a very important day for TB. Those who know us know that we don’t seek much exposure; we prefer to spend our time working, studying, getting to know businesses. But after three years, with an event to close out a first cycle aligned with our performance and looking at the long term, we’re taking a new step in the direction of increasing our communication with our investors.
At first, we began with monthly newsletters, which later became quarterly. We’ve already had two annual letters. And today we’re holding a broader event that leaves room for less linear conversations. So the plan is an approximately one‑hour linear presentation.
Then, right after, we’ll open for questions and answers. The goal is not for this to be a monologue, but a dialogue—with plenty of actual Q&A. Then we’ll pause for a break and come back for another 45 minutes to an hour. Noon is not set in stone.
If there’s interest, if people feel like going a bit longer, we can keep going until 12:30.
{GABRIEL S} [00:01:46]: So let’s get started here. Beginning with this first slide. This will basically set the tone of the event, before anything else, by recalling our first foundational presentation. A quote by the self-taught philosopher Eric Hoffer, which says the following: “In a time of drastic change it is the learners who inherit the future. The learned usually find themselves equipped to live in a world that no longer exists.”
That quote was very relevant for us at the beginning of TB, as a sort of philosophical correlation with many of the ideas David and I already had internally, and which would, curiously, become even more relevant as we got the business going. From the start, our impression was that investing in today’s world in a disconnected way makes very little sense. In a connected world, by definition, one should invest in a connected way. And interestingly, a month after TB began, in November, started what we consider to be an industrial revolution, driven by AI, and this quote became even more prescient.
So this is just to set the tone, and I’ll now show some charts more as a way of raising questions. Starting with this first slide, which shows the returns of three different indices. I won’t say which ones just yet—let’s see if you can guess. At least for me, it was counterintuitive to see this result.
There’s a very common idea out there, often repeated in the media, about American exceptionalism. It’s an idea that’s easy to package, but as we see time and again, reality is often very different from what people imagine and this chart symbolizes that.
The United States, if not for the seven companies that really set the future, would have been just a conventional index, an industrial index for all intents and purposes, one that simply tracked GDP plus some yield. In fact, the most relevant European economy index, the DAX, among major continental economies, outperformed the S&P 493—that is, the S&P excluding those seven companies—over the same period. An even more interesting piece of evidence is the late surge in the DAX, which actually comes from the weight of a specific technology company in the index: SAP. So what do we want to put forth here as backdrop? There’s no such thing as geographic exceptionalism, but there is corporate exceptionalism. Leading-edge companies truly shape the future.
But that alone is not enough. For us at TB, another idea that goes hand in hand with our worldview is something Buffett said at the last Berkshire shareholder meeting—that opportunities become very attractive very occasionally. The corollary is that things also become extremely unattractive at certain points in time. And that’s exactly what this picture shows: the torture of being invested in certain indices. In this case, the most shocking is investing in the Nasdaq from the peak of the Dot-Com and the difficulty of carrying the future, even when you get the future right.
I think this chart is self-evident. An investor who had invested at the Dot-Com peak, in March 2000, would have had to wait 14 years just to get back to breakeven. Along the way, they would have spent 9 years, with the lumping of crises, 75% below the watermark.
That’s the kind of tragedy that makes intertemporal compounding nearly impossible for any investor. It leads to crystallized losses, which is what we want to avoid. So another message here, again, just to set the tone for what we’ll discuss later, is that investing isn’t just about picking the right companies. It’s also crucial to have enough balance to actually reach the future.
If you can’t get there, the corollary is—you don’t get there. So this is a way of giving you a sense of how we think. One way or another, the future arrives. And this same mythical investor, this “poetic muse,” so to speak, if he had managed to stay invested for 25 years in the Nasdaq, even starting at an absurd peak, investing in companies that had nothing to do with today’s Nasdaq—just a quick parenthesis, the Nasdaq of the late ’90s and early 2000s was essentially a telecom index, with massive leverage.
Some companies were outright frauds. WorldCom alone was 4% of the index, and everyone knows it was the company that led to Sarbanes–Oxley, the regulation that imposed a whole new level of punishment for fraudsters. So that index had little to do with what we see today. And even so, anyone who stayed long enough would have recovered most of their capital and achieved a very respectable nominal return, 8% in dollars. Curiously, if swapped into reais (BRL currency), that was even higher than Brazil’s real (inflation-adjusted) interest rate which, by the way, was very high during that period.
So those who can stay in the game do manage to get that capital compounding. And even more importantly, those who could rebalance their portfolio throughout that cycle would have achieved an internal rate of return far superior to the simple start-to-finish return. That’s what matters. Just more evidence worth considering: when the Nasdaq imploded, Google wasn’t even listed yet, the iPod hadn’t been launched—let alone the iPhone, which came in 2007— even Facebook would only go public in 2012, much later.
In other words, the index was totally recomposed, and a recomposition that ultimately means that the companies that actually created this future had a much higher compounding rate than what this snapshot shows. But again, it's crucial to understand that without rebalancing, it's very difficult to arrive at this nice-looking outcome. So, the final result is nice, but the journey is very tough. And now I’ll start on the official agenda, with the idea of why have an annual meeting.
Talking with some of you here, we got praise for the idea, and we wanted to share where it came from. David and I have discussed the idea of holding an annual meeting basically since TB’s beginning. Obviously, at the start it made no sense—we had nothing to show, nothing to say, so it was just an idea. But more recently, at a company off-site, we decided the time had come—especially because now we’ve reached our third year, and the first performance cycle is an important marker to share some things with you.
And where does this come from? We believe that, particularly from what we might call post-impeachment onward, Brazil has developed —more than other regions, though also elsewhere— an excessive cult of performance. Many structures that appeared in that period began to sell short-term performance as if it were truly long-term performance. At the same time, communication via YouTube became more relevant—which is great.
At the same time, the traditional gatekeepers started losing power, and a new form of communication arose that heavily emphasized this “number go up” mindset. Which, while important, isn’t necessarily true historically, even though it is really important. So, we feel investors became less true allocators of capital and more like investment bankers or sell-sides. Selling the future became almost a sine qua non condition for raising money in the local industry, especially for new players. At TB, we naturally started after that wave, in 2022, when there was already some correction of that dynamic—because the future arrives. Selling the future but not delivering—it eventually catches up.
So from the start we questioned how we might share our way of thinking, which is diametrically opposed to selling the short term as long term. We’ve just completed three years, and we know that’s nothing on a long-term journey. So the idea of making this meeting started, as David said, with newsletters, then letters, and now this meeting—our idea is to give you another chance to interact, to have a frank debate, to put ourselves out there and hear your questions. And, maybe most importantly—and this is our wish, however naive it may sound—to move away from the cult of performance.
We’re in an industry that worships performance. Without false humility, our first three years delivered very good returns. But internally, we don’t see that as a marker of the firm’s success. We see much more evidence of success in what we’ve been able to do analytically, in building ideas. So our goal here is to bring you into the mindset that we will go through downsides—it’s inevitable. And through a relationship based much more on trust, we want to help you make better decisions over time.
{DAVID K} [00:14:14]: Just a quick comment, because I think our communication should always focus on explaining the investment horizon we look at. Everyone who knows us knows we look at relatively long investment horizons. So, the 3-year period for the performance fee is the minimum, which is why we waited at least that long before starting to communicate more. It’s very important to understand that the vehicle itself, in a way, smooths the investor’s journey. We invest in businesses that are uncorrelated—more than in terms of price swings day to day, but in the type of businesses we choose. Still, it’s an equity investment vehicle. Roughly speaking, we’re 80 to 100% invested, and there will be moments when the correlation goes to 1, and there will be drawdowns. So, it’s very important to understand this is a vehicle that requires patience; the long term does arrive, but it’s not linear, and in the end, it’s critical to understand our alignments.
So this presentation will go over these three pillars in more depth, with specific points and it needs to be very clear what we do—because if an investor doesn’t understand what we do, they shouldn’t invest with us. It’s not because the fund performed well that people should invest with us. It’s because they understand and believe in what we do. Because bad times will come, and everyone knows that bad times in investing create discomfort—sometimes even pain, eventually even physical. So the ability to understand that price volatility is one thing, but the volatility of a real-world business is another, is extremely important.
We’re going to go through nine pillars of our culture and investment philosophy. It’s important to understand what we do and how we do it. To say that equity investing is a poor way of putting it—because those who buy a stock buy a piece of paper, a ticker and I think the big differentiator of long-term investors who’ve succeeded is the ability to invest in businesses. That’s maybe the main misunderstanding about Berkshire and Buffett. People look at Buffett as a stock investor. He isn’t a stock investor. He was at the start of his career, when he had the partnership. At a certain point he flipped the switch and decided to become an entrepreneur. Berkshire is his instrument as an entrepreneur to be a business investor.
So, understanding that we invest in companies—and that through investing in companies you earn carry, through revenue growth, through the ability of these businesses to grow profits, that’s how you earn returns. Returns don’t happen in the stock price, stock prices in the short term are driven much more by noise than reality, but in the long term, it’s reality that drives businesses in the same direction as their profits and returns are going.
We invest in businesses with the understanding that they are the best instrument to carry our savings intertemporally, earning a premium over the real interest rate. In Brazil, the real interest rate is represented by the NTN-B and we believe this is the best way, over decades, to carry your savings— it is not in instruments without duration or in nominal instruments. Brazilians, curiously, compared to the U.S., tend to only look at CDI (Brazilian Interbank Interest Rate) as the reference, in the U.S., everyone looks at the S&P, because the S&P is in fact a picture of the companies that exist in the American substrate, so we invest to earn that intertemporal return. And how do we do that?
An investor in stocks—in businesses—has a natural, obvious choice to make at the beginning of their journey, and we've already made our very obvious choice. They have to choose between being a specialist or a generalist. There are many ways to specialize, but we, obviously, chose to be generalists. And from the moment you choose to be a generalist, in our view, it made little sense to remain stuck in a market with so few businesses, so few companies. The Brazilian market seems to have depth, but in practice, the types of businesses that exist here don't reflect the reality of the world, especially the world of the last 25 years. So, being generalists almost forced us in the direction of going abroad. At a certain point, we found ourselves compelled to look for the best businesses, and the world's best businesses are not in Brazil, so we decided to look abroad. And how do we do it on a day-to-day basis?
We invest in good businesses, analyzing and understanding their microeconomic characteristics, the competitive behavior of these businesses, we look long term, aiming for returns over at least 5 to 7 years, we build a concentrated portfolio —We’ll show in the presentation how our top 5 and top 10 positions represent a significant level of concentration — and we try to do this in a very aligned way, therefore, these four characteristics are extremely important to us.
The first pillar is economic profit. Most of you have already spoken with us over TB’s history. When you set out to be a global investor, you move from a universe of maybe 100 to 200 companies — the Brazilian universe — to a universe that, in the U.S. alone, has roughly 5,000, so a key part of the investment process is how you filter—how you choose which companies to focus on, historically, we talked a lot about our filter, the microeconomic filter, in which we basically look at three major characteristics: increasing marginal returns, significant capital allocated in many cases, and network effect, and when we find these three characteristics in a well-established, stable business, we can be comfortable micro economically that profits will remain stable and grow over time, so that has been our filter. Today we decided to bring another definition of what a good business is, this one is not ours; it’s Buffett’s, from 1991. Investing is great, you don’t need to invent much, for most things, you just need to copy, be able to understand, connect the pieces, and be disciplined. The thing is, being disciplined is much harder than it sounds. This definition is called an “economic franchise.” The translation into Portuguese— “franquia”—is terrible, but even in English it isn’t great. Buffett thought of this by looking at his original business, maybe his most emblematic one: Coca-Cola. Coca-Cola had the holding company and the franchises, the bottlers. The holding produced the syrup and managed the brand, while the bottlers did the CAPEX, bottling, and distribution. A very important share of the economic profit—the ability to distribute and grow profit over time was in the holding. That’s what he defined as an economic franchise. In that 1991 letter, he was actually talking about the newspaper business, very presciently, understanding that at that time magazines and TV news channels were starting to displace newspapers, capturing consumers’ attention.
{GABRIEL S} [00:22:19]: Pre-internet.
{DAVID K} [00:22:19]: Pre-internet, yes—later crushed by the internet. But even then, Buffett was right. He managed to call it 20 years ahead consistently, because he was once again looking at the business’s microeconomic characteristics. At the end of the day, it’s always important to invest in businesses—and to invest with a microeconomic focus. So, what is an economic franchise?
It’s a product or service that is necessary or desired by consumers, where the consumer sees no clear substitute, and where prices are not regulated, so three very clear characteristics. In the modern era, this is a 1991 definition, new kinds of businesses have emerged, and we think they are important to complement this definition of attractive businesses; these are platform companies, in Bill Gates’s definition—companies that generate more value for the ecosystem than they can capture themselves— and demand-aggregation companies, which only became viable or relevant with the internet. These two types of businesses have a very important network effect component.
And in that way, franchise businesses, platform companies, and demand aggregators—much like monopolies—while everyone likes to define monopolies as the businesses that manage to escape, these three types of businesses are the ones that manage to escape one of the most important forces in the economy, which is the force of competition in capitalism.
Competition is what brings players into any market with excess returns, drives prices down, and leaves the returns with the consumer, not the investor. So what we try to do is look for businesses that can escape this gravitational force of capitalism.
{GABRIEL S} [00:24:04]: That first definition, in ’91, there really wasn’t yet a clear understanding of this second half, so to speak, of businesses. I mean, in theoretical terms these businesses already correlated with 19th-century businesses—if you go back to the great American industrialists, many of their businesses at the time were monopolies with very significant scale advantages, later broken up by the Sherman Act—and then we spent a long time without clarity about what a platform or a demand-aggregation business was, and in the ’80s it maybe started to become more evident that some software businesses had this characteristic.
Interestingly, at the very moment Buffett makes that definition—again, Buffett was already 60 then, so it’s natural that his worldview was anchored in his historical perspective.
Around that same time one of the leading academics of modern economic theory as applied to business, Brian Arthur, was writing a foundational essay, “Increasing Returns and the New World of Business,” which basically cemented the idea of the different economics that accrue to platform businesses, it’s the foundational text, and around that same time, Clayton Christensen wrote The Innovator’s Dilemma, which also adds theory around these new businesses, and we would only really see this theory in practice with the rise of demand aggregators at scale and platform companies becoming firms with tremendous power and bundling—roughly from the mid-’90s into the 2000s. So today, discarding this as a pillar—as a major driver of durable economic profits—is simply untenable.
{DAVID K} [00:26:20]: This might be the most important slide, alongside alignment, because, as I said, patience is essential, but it’s hard to maintain along the way, it’s “easier said than done”. Everyone says they like to be patient, but the path isn’t linear and time feels slow. You get drawdowns along the way; you face the usual setbacks every investor faces, but the only way to earn the returns from equity investing is to let time pass. The only way to earn the carry is to wait for the carry to happen. You don’t make money in equities by watching the price tick up or down.
{GABRIEL S} [00:27:12]: And again, I think what David said is exactly right, time is non-linear. It's important to have evidence that time doesn't just unfold on a day-to-day basis. We don't live in a 'one day after the next' world; things aren't a traditional definite integral. Therefore, being out of the game can be very problematic; it can be very consequential.
{DAVID K} [00:27:42]: There’s a problem for us, for Brazilians: businesses in Brazil, for the most part, don’t compound over time—which is not entirely true. We ran an exercise which, I admit, I hadn’t done in a long time, looking at the main businesses, the best businesses. If everyone here voted on the ten best businesses in Brazil, I think very few would miss the list.
And Brazil’s ten best businesses did compound in real terms over the last ten years, there was just one issue: they started from a very high multiple back then, which is why the returns weren’t great. But the businesses themselves, in practice, weren’t that bad.
{GABRIEL S} [00:28:18]: Earnings growth.
{DAVID K} [00:28:19]: Right— in terms of earnings growth. The problem is either you start from a very high multiple, or you don’t have enough businesses to build a portfolio. When you have a substrate—a large base—of companies (not 5,000 U.S. names; 5,000 is an exaggerated number), but when you have 100–150 companies at hand as your repertoire, at any given time you can find 5 or 10 that seem to have a very attractive return and where it makes sense for you to be allocated.
The problem is when you’re in such a limited market that the investor’s challenge is almost impossible, because at many points in time there simply aren’t opportunities. And when you don’t have opportunities, there isn’t much to do. Either you bend reality, which is a common human trait, human beings have this flaw; humans weren’t biologically made to invest, they were made for many other things, but not to invest, so either you bend reality or you stay invested and, unfortunately, go through rough periods of deep drawdowns or mediocre/middling returns.
So what matters isn’t just “time in the market” —that original Berkshire/Buffett phrase — what matters is “time in the right market”. It’s finding the actual substrate of companies — businesses — that will give you that carry, and once you find it, building a portfolio and having the patience to earn that carry.
{GABRIEL S} [00:29:46]: And just one more quick aside: we have a strong view that, post-2008, there was a kind of glorification of a set of ideas that goes very much against what we’re saying here, they can even work, but they still run counter to what we’re discussing, this notion of time — if you read the basic literature from the ’50s and ’60s — it’s always been the case. 2008 is a marker that makes us wonder whether those who lived through it and thought it would scar them for life didn’t actually go through a “curse in disguise” — the opposite of a “blessing in disguise”. People who lived that moment tended to overweight “timing the market” versus “time in the market”. The idea that options-like structures or holding more cash at certain times, predicting the next crisis, without giving enough weight to the fact that fundamentals compound over time has become commonplace among many investors. We think there are some figures that illustrate this a bit—Roubini, Taleb—figures who defined an era and somewhat distorted what had been a very positive “genetic code” among certain investors.
So, after 2008—we’re already fairly far from that period — this idea of chasing black swans and believing you’ll protect yourself from black swans suffers from a basic problem: by definition, you can’t hedge a black swan — It’s a black swan. And that affected people a lot. Anyway, that was just a brief aside.
{DAVID K} [00:31:32]: This attempt to avoid drawdowns at any cost, it’s impossible in practice, it’s nearly impossible. And maybe Buffett—again, often misread by not understanding who he is—doesn’t help much here, you’ll remember his rules of investing: the first rule is “don’t lose money”; the second rule is “remember the first rule.” But then, in the next breath — and he says this in his first interview; by the way, this week we decided, as we do regularly, to revisit classic investing content, philosophical investing classics. I rewatched his first interview when he already had more than 15 years with Berkshire: he says the first rule is not to lose money, and then immediately says he ignores market price volatility, and that for him what matters is the health of the businesses. He’s an entrepreneur. Entrepreneurs don’t think about what’s happening to the asset’s market price, they think about whether their asset is getting better, about its qualities, competitive positioning, pricing power. Is the asset better over time or not? If it’s better, he’s comfortable over time, he’ll make money. That’s how an entrepreneur thinks.
So when he says “don’t lose money,” it’s about avoiding permanent impairment of capital — avoiding investing in a business at, say, 200x earnings, or in a business that has no profits and never will — like what happened in the Nasdaq — or as seems to be happening now in unlisted businesses, which is perhaps the best parallel to the late 2000s today it’s not in the public market—it’s in the private market. When you have a company valued at 500 billion dollars, you have OpenAI—which truly has already built a valuable consumer base—and then everyone looks by analogy: “OpenAI is worth X, so I’ll take something worth half X, a third of X.” But in practice there will be one, two, maybe three winners; the rest are worth zero. That’s where the great capital destructions are, when you pay for a business that doesn’t exist, or when the business did exist but was completely displaced by innovation, by competition, by a regulatory failure—by any number of factors that render your profit estimate totally wrong. So, making a mistake and losing money is more about getting the business wrong than about what the market is paying. What the market is paying is normally much more about the prospective return. Because businesses change very little day-to-day. The problem with going to the office every day is that every day the prices fluctuate, but the businesses don't change every day. Businesses change every 5 or 10 years. So, when the price of the business changes, what has basically changed is the prospective return of that business.
{GABRIEL S} [00:34:33]: And the true impairments—when you look at history, the real impairments, the effective capital losses—they come exactly from that. Sure, there are extreme cases, like paying 70× earnings for a business that just generates cash with no growth, and naturally you end up with a substantial loss when that happens. But the bulk of impairments—if you take Cisco in the 2000s, Intel in the 2000s (I won’t even mention outright frauds like WorldCom)—but if you look at those kinds of businesses, the big impairments come from total earnings mismatches. Again, whether it’s through reflexivity—Intel in ’99 was reporting big profits riding on an infinite flood of VC money in the industry, same for Cisco, same for Sun Microsystems. What happened there were earnings impairments.
So, it's very important to have this understanding that there's something—for an investor, this is somewhat self-evident—but it's important to have this mindset that the mark-to-market can happen in two ways. It can be merely a change in the multiple while everything else stays the same, and this does happen; markets are not as efficient as people tend to believe, and we can comment more on this during the Q&A if it comes up. But there are also, in fact, wrong forecasts about businesses. There are businesses where it's imagined the future will be brilliant, and that future never arrives. So, it's more important to have this orientation towards the fundamentals than a mere fear of the normal fluctuations of the market.
{DAVID K} [00:36:08]: And once you set out to look at new geographies — or remove the geographic restrictions — and expose yourself to all the listed businesses out there, a key part of the investment job is deciding which businesses you’re going to study, having the ability to differentiate businesses. We see this as maybe the main gap in the training or investment process of a Brazilian investor—an investor who’s locked into a purely geographic game. Because you really need to understand and differentiate businesses to decide which types of companies fit your personal style, what you set out to do.
In practice, I think it’s safe to say—there’s little doubt—that the best businesses in the world aren’t here, of course there are exceptions. In Brazil there are two companies with the same qualities, the same characteristics as the world’s best businesses: Nubank and Mercado Libre. Curiously, neither was founded by Brazilians. One day it’ll probably take a sociological treatise to explain why no Brazilian has managed to build a company at that level of sophistication and understanding. I think that’s relevant—because there are other companies that are almost as good, but the level of comprehension among the entrepreneurs, the business builders, is very different from those truly running a playbook of increasing marginal returns, it’s very different.
So, clearly, we couldn’t accept being trapped by the randomness of having been born here. It’s part of our lives—we’re Cariocas, happy to have been born here—but we didn’t need to spend the rest of our lives only looking at Brazilian businesses, investing only in Brazil. The only real track that made sense was to look at allocations abroad.
And here we show a chart that complements the earlier one. In the first one, we showed the pain, the discomfort that a drawdown brings, and how time corrects those returns. Now we bring another chart, covering a different window. That one was 2004–2024; this one is 2014–2024, so a shorter window. But it’s the window where the foundation was laid for the exceptional performance of very high-quality U.S. businesses.
We used the simplest cut—the “Mag 7.”, we don’t even like it that much, but in practice it’s the aggregate of the largest U.S. businesses. And here it becomes very clear what we said earlier.
There is some American exceptionalism. Okay, that’s a joke—it obviously exists. The U.S. is the country that attracts the best people in the world; everyone wants to move there. It’s the country that most incentivizes competition, where states specialize in specific industries, the wisdom of the Founding Fathers in that design is truly impressive.But for us as investors, what matters isn’t national exceptionalism, it’s business exceptionalism. What matters for us, what makes the difference in our long-term compounding, is the exceptionalism of businesses. And these companies truly did grow earnings in an exceptional way over the last 10 years.
{GABRIEL S} [00:39:50]: And I think another piece of evidence here — again, somewhat counterintuitive, especially given Brazil’s current moment, where people tend to weigh things in overly geographic terms — is this: earnings growth in the Ibovespa, excluding Petro and Vale, denominated in reais (which is the right way, because you can’t look at Brazil in dollars; you start from the wrong FX rate and everything is wrong), was better than the S&P 493, better than France’s CAC index. So the evidence here is the following, again: what matters are the truly differentiated businesses. What created this differentiation in the American case is the earnings growth of a subset of companies that create the future. The top quartile is what truly makes the difference, is the very fact that enables a top quartile to exist. And in the United States — and this is why it’s a joke — the U.S. will always enable top quartiles, you can assume that going forward, that will still be the case. But if you’re an investor stuck in the leftovers—the S&P 493—there’s practically no differentiation. Of course this is a generalization, sure, investors can nitpick good names within each subset, but the broad carry — the baseline compounding — shows no exceptionalism versus other conventional countries, and in Brazil’s case, actually, as theory would suggest, it was better.
{DAVID K} [00:41:23]: In the next step, when you look abroad, you must define which companies you’ll be able to put on your shelf. My partner must be sick of me—I spent the whole week searching for the right word to define what we’ve accomplished in the last three years and what makes us happy. As Gabriel said, it’s not performance — performance obviously makes us happy, but it’s not what makes us most proud, what we achieved over the last three years as investors—and what, in a way, Gabriel will show in the upcoming slides, what changed us as investors — was our ability in those three years to expand our repertoire of companies, to expand the repertoire of industries we know and can draw on over time as capital allocators.
And when you talk about repertoire, it doesn’t necessarily mean things we’ve invested in, along the journey, you look at new businesses—many you discard, some are interesting but don’t fit us, others looked much better at first glance, but when you dig deeper, they aren’t that good, or have risks we’re not willing to take. And then, when you find a business that truly fits what you’re looking for and your way of investing, you put it on the shelf and because of that our repertoire has expanded over these last three years.
Everything we’ve said so far only makes sense if the alignments are right. Because we’re looking at the long term, the way to measure us is by the long term, so there’s no mathematical evidence of what we do that will show up quickly, it takes time to build consistency. At best, you get some point-in-time evidence of consistency between what we say and what we execute. So, believing in our alignments is extremely important—understanding them. Again, as Gabriel said, our goal here is to be as frank as possible about our intentions as investors, and showing our alignments is a big part of that. There’s a second trait we think is very relevant: showing our temperament as investors.
Again, in that original interview (we had already put this in the presentation, and then when I rewatched that original Berkshire interview), Buffett talked about this in the 1980s—that an investor with the right temperament is one who has the patience to capture compounding, who understands that price volatility means nothing compared to the volatility of business quality, and who has a certain comfort in his work, not serving the market but using the market to serve him. So a person without the right temperament suffers greatly with day-to-day volatility. When prices correct, they think businesses have deteriorated fundamentally, when in fact what happened was just price oscillation.
That temperament—the ability to be patient and endure the long term—is crucial for us, but also for our investors. It’s important to gather a group of people who share those same values. Maybe that’s the main goal of today.
And when you combine that temperament with the right alignments, you should have the ability to find the opportunity set—the substrate, the substance—that is the companies, the repertoire, this group of assets we want to keep on our shelf over time as potential investments, when they remunerate us, when the prospective return looks attractive enough. And when you have strong conviction in the businesses you can choose from, you should also be comfortable building a reasonably concentrated portfolio — so that you actually deliver interesting performance over time.
{GABRIEL S} [00:45:47]: A quick, brief addition here: there are two variables that are undoubtedly the most important—being a misaligned manager or lacking the right temperament. Alignment, I think, can be constructed formulaically, so it’s easier to assess. Temperament—we must hope we have the right temperament, and we’ll see that over the years. Obviously, this binomial is crucial, but it’s not sufficient. It’s important to keep in mind that to cross this Rubicon — recognizing there’s substance to be pursued outside the particular “box” one invests in—you also need a certain bet that this is doable. I think that comes from a specific trait: believing it is possible to do. So a certain amount of growth mindset—believing you can cross certain borders. If Swiss investors invest abroad, if Australian investors invest not only in the Commonwealth but anywhere in the world—why couldn’t a Brazilian investor do the same? If anything, we’re closer to the rest of the world than Australia is. Just a small point.
{DAVID K} [00:46:56]: And here we bring a data point we already disclose monthly in our factsheets, which after three years shows some consistency: the concentration of the top five positions and the top ten positions in the portfolio. It’s a portfolio that truly runs with a reasonable level of concentration. For us, this is a heuristic — a kind of rule of thumb to help us navigate over time, to signal when the market is offering fewer opportunities or when we’re finding fewer opportunities. We use it as a sort of beacon that guides some of our discussions, it’s not set in stone; it’s not something we target per se, we don’t aim for this aggregate concentration, but we do like to have individually meaningful position sizes, and when you hold meaningful positions, the practical outcome is this picture.
And, together with the time slide and the alignment slide—which for us is essential—that’s where Gabriel and I spent a lot of time when founding the firm: thinking about the best ways to give the right, clear, objective signals internally (to ourselves and the investment team) and externally about what we do.
So here are the two that, in our view, are the most important and most straightforward: first, the amount of our own capital—keeping a consistently high level of our own capital invested by the day-to-day stewards of the vehicle—and using that as a guideline to manage firm AUM growth. You know the rules; if you don’t and are curious, just ask—we always make them explicit: the percentage of partner capital and how we intend to grow over time. It’s a fund of very liquid assets; when we do the math on capacity limits, the capacity looks almost infinite. It doesn’t seem right to grow indiscriminately. So we’ll use partner capital percentage as a rule of thumb—a heuristic—to guide the firm’s growth. That’s one key alignment in our view. The other is something we did that we see as quite different locally: since we set out to invest with a long horizon, it made little sense to charge on short horizons, for consistency, as you know, the fund crystallizes performance every three years. Three years felt to us like the shortest feasible period that still balances keeping an investment firm with other partners alive while charging on a minimally reasonable horizon—so that investors can also evaluate us on a minimally reasonable horizon. We accept being judged on performance, but we ask that it not be six or eighteen months; at a minimum, judge us on the horizon we use ourselves — which isn’t even three years; in practice it’s longer—but three years works as a heuristic, a rule of thumb.
Temperament, as Gabriel mentioned, is personal — there’s an innate side, a trait you’re born with, but we also believe using alignments to help shape an investor’s temperament is very important. This combination creates a feedback loop that points in the right direction and, in practice, keeps us from freezing up — which is hard, it’s hard to get off zero. It was hard fifteen years ago when Atmos started; it’s hard today at TB, at the very beginning, to build a portfolio on day one and be fully invested in an equity vehicle, it’s not easy. So keeping our research pointed in the right direction for an equity fund is not something easy on a day-to-day basis, but it also guides you in the direction of not trying to be a genius, so alignment plus temperament is what makes you seek the light at the end of the day, it is what makes you seek rationality as an investor and what we shouldn't do is any big mistake, the return from compounding capital over time, it comes from avoiding big errors in practice while seeking to aim for the long term..
{GABRIEL S} [00:51:58]: Just a brief addendum as well, this largely comes down to sustaining our set of ideas. If any criticism from our investors were to be fair, given everything we’re saying, it would be if we started investing in a way that has nothing to do with what we’re laying out here, it happens, so we’re laying out a series of pillars partly to give you a guideline of what to expect from us
{DAVID K} [00:52:20]: It shouldn’t happen—but it does.
The last slide on the investment pillar — who does this on a day-to-day basis, it's not just me and Gabriel; on the contrary, we are a team of 5 people, 5 partners, who have been here since the first day of the company, on the management team. We function on a collegiate model, which is not the perfect definition, but it is approximately correct. I think 'collegiate' ends up being interpreted a lot like a participatory democracy; that's not what we do day-to-day. The collegiate model has much more to do with the way we build knowledge. It is a model in which a group of investors, a group of people who identify with each other and have mutual respect, gets to understand where the other is intellectually—what they’re reading, what they’re thinking—sharing that journey of building knowledge, and, in practice, because we’re investors, understanding the relative opportunities and risks of all the businesses we invest in. This ability, within a team, aligned in personal traits—with integrity and openness in sharing opinions—I like to say that anyone who knows me and Gabriel knows we’re quite different in many ways, but on everything that matters—especially in investing— we almost always agree with each other. So having the ability to discuss and agree on what matters in order to build a portfolio is very important.
{GABRIEL S} [00:54:18]: In that vein, adding a few more points, the collegial model really is crucial to our build-process. There are more isolated processes out there, and less isolated ones; in our case, collegiality is very important because it recenters the discussions. In investing, it’s normal for discussions to drift into totally esoteric topics, the analyst has more information than the counterparty, more than colleagues—it’s very natural in investing for day-to-day topics to stray from the basic environment of questioning premises, and if there’s one thing we can help everyone with, it’s this: all errors stem from mistaken premises.
There’s a quote we’ll put at the end that’s very good: “the better the logic, the worse the conclusions that flow from a bad set of premises”.
So wrong premises are the biggest problem and for a group of investors, what you want most is an environment that enables so-called “dumb questions,” so that you can always debate the basics—what originally shaped that idea—and step away from excess, from the overflow of information, which is very much a function of the world we live in. That would already be true elsewhere, but in today’s world there’s an enormous capacity for people to “over-learn” things. It’s normal: you have a huge volume of information—you should be unlearning at the same rate you learn new things. That’s what the collegial model does. It’s almost like a set of antibodies that clears out the noise—the noise really is parasitic—and guides discussions in an analytically correct direction over time.
{DAVID K} [00:56:09]: There’s a problem, which isn’t so easy to explain to people who don’t work with investing day to day—but when you explain misalignment it becomes clearer. In an investment team, where you have the figure of a portfolio manager or more senior partners, and a team that always has some seniority gap, and you have someone deciding everyone else’s compensation—it’s very easy for the junior part of the team to chase information. Because information is linear, it scales linearly, meaning that if I work X hours a week, I can generate a certain amount of information, therefore, the more hours I work, the more information I generate, the more information I generate, the more I bring to the portfolio manager. This is a very profitable industry when things are working—so the more information you generate, the better you look at the critical moment of compensation, the bonus time. So, if someone is linearly generating lots of information, they position themselves well at the snapshot to be well compensated. What’s the problem?
For investing, information isn’t that relevant; what matters is the ability to understand businesses and exercise judgment; judgment involves risk. I always give the example of the car rental industry in Brazil: a few years ago, there was a shift in the substrate—you were moving from combustion cars to electric cars, with Chinese car competition. It’s useless to ask the company what depreciation will be going forward, they know as little as you do—they know nothing. So, in investing you have to be very careful, pay attention to which questions you can and cannot answer, and accept that some questions you simply cannot answer.
So, when you give yourself more degrees of freedom—and that's what the choice between being a generalist and a specialist is about—when you say you're a generalist, you are giving yourself more degrees of freedom as an investor. You give yourself the freedom to accept that you're not going to hit every ball. I think this is very relevant: accepting that you're not going to hit every ball, because along the way, there are many questions that have no answer. When you find a question that has no answer, you must accept it, move on, and go to the next opportunity.
{GABRIEL S} [00:58:31]: Sometimes the question isn’t the most relevant one and when it isn’t, you can invest anyway, and sometimes it is the most relevant one. There are a series of investments in which the unanswered question is the key question, then the best option is usually to put it in the proverbial “too hard box” and move on.
And bringing it back here—we’re already near the end of the presentation—this slide is also from our foundational presentation, where we put back then how we thought, roughly speaking, or how we had idealized at the start of the company, what our circles of competence would be. SAM is an acronym for serviceable attainable market, so that's the more captive market, let's say, where we would invest. We would break down the investments into two quadrants, one quadrant for Brazil and one global, and the same for the broader market.
What happened — and I think it’s important as evidence for everyone—is that practically at no moment in TB’s history, from T0 onward, we actually thought this way in practice. So, this was just a way of illustrating our thinking at the time—the most basic output we had to try to convey our worldview. But in reality, we never actually thought that way. The collapse of those two quadrants into one single quadrant happened around month four of TB—exactly in line with how we always thought about investing. Especially because one of our biggest strengths is the ability to cross-pollinate ideas.
I think that’s the great strength of a generalist—though don’t get us wrong, generalists have many weaknesses too — but one of the big strengths is being able to look at a Latin American e-commerce company and a U.S. e-commerce company and then draw potential insights about how the e-commerce play-out might look in South Korea. For example, one company we hold today is Coupang, a huge market-share gainer in South Korean e-commerce.
This ability to cross-pollinate has many other examples too. For instance, in merchant acquiring businesses: you should be able to look at Fiserv at the same time as you look at Stone; in digital acquiring, you should be able to cross-pollinate between an Adyen and a Dlocal.
This has always been TB’s great strength. So, from the start, we effectively executed our day-to-day worldview with these building blocks. Curiously, the output of that ended up being a fairly balanced portfolio between Brazil and global—maybe because of our starting apperceptive mass. Going forward, it’s unlikely that will remain the case, but anyway, that’s the breakdown.
And some comments here: Brazil didn’t exactly help itself along the way. I think everyone here who invests in Brazil knows Brazil hasn’t been kind to investors. So several sectors we thought would generate ideas ended up leaving our circle of competence, crowded out by sectors that made more sense globally.
So that light gray area represents the sectors we cover much less nowadays—because obviously there are limits, and we specialize where it makes sense.
This is the last slide, and it’s also a recap of our foundational presentation. Back then, we ended with that idea of how analysis would be applied, with a final breakdown of how the portfolio would balance. And as I said, curiously, the outcome ended up pretty much in line with that, along the way.
We ended up with a genuinely balanced portfolio between global and Brazilian equities. But again, that was just output versus our actual knowledge production, especially at the start. A few comments here—maybe more relevant than ever. We remain investors across different tranches of corporate capital structures. We were able to take advantage of that for a time—particularly in early 2023, but thinking this way means those opportunities will appear opportunistically, we won’t proactively seek bond investments; we’ll only invest in bonds when they have equity-like returns for us; that has already happened at certain times, and we think now, more than ever, that big judgment matters more than big data.
We started TB in a pre-AI era. In the AI era, it’s become very clear that the idea that additional data helps the investment process is rather naive. And in the end — where do we go from here? Where does the portfolio balance out? What’s the wrap-up of this idea? Where are we today?
I think this picture is the graphic representation that best illustrates how TB thinks about investing today. It’s a geographic breakdown; today, we’re very clear that the geographic straitjacket we had at the firm’s beginning has already been broken. Again, partly because of Brazil’s shortcomings and partly because of the merits of the companies we cover. There’s been clear crowding out of certain idea sets. Structural investments—almost physically—are moving more toward structural opportunities which, by definition, are global.
Brazil is still very present in today’s portfolio—particularly in the opportunistic part. And we think Brazil will always have an intertemporal role, mainly as a balancing factor in the portfolio, very relevant in that opportunistic slice. And curiously, this even allows us to have large positions in the few Brazilian companies we do hold. Our bar is high, so when a Brazilian investment makes sense, it’s one that has broken through so many barriers that it very clearly deserves to be one of the fund’s largest positions. We’ll probably never have small Brazilian positions—and at times we may have none.
So the portfolio migrates from a strict geographic orientation, a priori, to what is in fact a much more open orientation—one that balances structural investments with opportunistic investments. I think this sets the general tone of what we wanted to convey. And I don’t know if you’d like to add anything.
{DAVID K} [01:05:36]: Thank you. The idea now is to have a long Q&A session—this is the time to ask questions. I also forgot to thank Bia and Jana, who had a lot of patience with us over the last two or three weeks organizing the event.
It didn’t turn out well because of the five of us—it turned out well because of the two of them. Thank you for the patience—we complain a lot in our day to day, and they put up with us.
{GABRIEL S} [01:05:59]: If the presentation was bad, at least the breakfast was good.
{DAVID K} [01:06:03]: There will be two assistants with microphones for anyone who wants to ask questions.
As I said, at eleven we’ll pause for a fifteen-minute break, and then come back for more Q&A.
{GABRIEL S} [01:06:20]: We went a bit long, but I think there’s plenty of room.
{DAVID K} [01:06:24]: Yes, plenty of room.
Q&A: First Question
{Question} [01:06:30]: Within your investment style and philosophy, which you just talked about at length, how do you think about risk management?
{DAVID K} [01:06:43]: I think when we show the drawdown, it’s clear in our minds that it’s important to build a vehicle that’s balanced, that helps the investor stay invested and feel less discomfort during some drawdowns. The vehicle invests in highly uncorrelated businesses, this means that while we have significant positions in Brazilian financials and utilities, we also have significant positions in technology businesses headquartered in the U.S. — but in reality, it’s a global, very stable technology businesses, so the typical characteristic of the businesses is uncorrelated, more so than day-to-day price.
This lack of correlation is what lets us reallocate capital between buckets, and that’s what we’ve done over these three years—suffering little in individual drawdowns, that’s very relevant. When I say “individual,” I mean either a single asset or a specific asset class. While one class is going through a drawdown, the other is genuinely uncorrelated, and that’s what allows you to reallocate capital. In practice, you also must understand there will be moments when correlation goes to 1. This happened in 2008. Those moments are exceptions rather than the rule. And when Gabriel mentioned the risk of “overlearning”, of trying to minimize drawdowns, that’s not what we do. And there are investors who do minimize drawdowns.
{GABRIEL S} [01:08:32]: But it’s a portfolio that, by design, has room for relevant drawdowns at certain times. By construction, it should be able to mitigate those fully synchronized drawdowns — we should always do somewhat better than the worst index at the time — but it is still a portfolio that allows for drawdowns. At the level of business risk understanding — which I think is more relevant —in portfolio composition through time, we also understand we won’t hold just one business type.
That’s the core of how we think about risk: TB’s risk thinking is grounded in the types of businesses we invest in. For example, we’ll rarely have a portfolio fully allocated to businesses with lumpier returns, whose margins fluctuate a lot, whose return on capital also fluctuates, even if the evidence shows that at a given moment those kinds of businesses can deliver very high returns, prospectively, we’ll rarely be fully allocated only to those.
We will always prioritize having some counterbalance on the other side, that is, from more stable businesses with stable characteristics. And this also implies over time that a good part of our research, on the margin, as we migrate and produce more knowledge in the world, a good part of our research also goes in the direction of getting to know these more stable businesses globally. Because we used to have a part of our input that was very oriented towards stable Brazilian businesses: the distribution businesses, the utilities businesses, the infrastructure businesses. And our job over time is also to come to understand what these global 'utilities-like' businesses are, which have nothing to do with regulated utilities. Meaning, what are the businesses that in fact have a propensity for extremely stable cash flow generation—from software-as-a-service to networks—that can complement the other side of the portfolio? Which is a side of the portfolio with businesses of tremendous quality, but that are eventually more susceptible to going through swings over time. So, this is our view of risk in terms of the business.
{DAVID K} [01:10:58]: And because we don’t minimize drawdowns by vehicle design, there’s no active effort to seek individual or tactical hedges. That’s not what we do day to day. During the moments in the past three years when there were external drawdowns, the fund held up very well. In fact, these three years produced exceptional returns—exceptional in the sense that they were an exception. The fund barely dropped at any point because when there were drawdowns abroad, Brazil happened to do well. The decorrelation worked in the best way possible; you could say nature was very generous with us during that in these last three years.
{GABRIEL S} [01:11:44]: But again, that’s by design — that’s the right risk answer: by construction. There’s no active risk management per se. I know that’s not exactly the question, but there’s no active effort to treat risk as anything other than an output of how we construct the portfolio. Our risk view is really tied to business selection—building a balanced portfolio. Because at the end of the day, you have to arrive, you have to survive to the end. If you die midway, you don’t get there. That’s our base orientation.
{DAVID K} [01:12:20]: May I add one more? Because this is a question that historically has generated a lot of discussion for us. An important part of what we do daily is having conviction in the businesses we hold. And I think that’s a meaningful shift in who we are today as investors versus ten years ago. When you’re in a market where businesses aren’t of such high quality, where you don’t have as much conviction, in practice you spend every day hoping the stock goes up, maybe it’s best start praying for it to rise—because when it does, you pocket the money and move on. What we do today is very different. In practice, the best thing that can happen to us when we invest in a business is for the stock to go down, because probably nothing happened to the business itself — it’s just a price fluctuation and if you have conviction in the business …
{GABRIEL S} [01:13:18]: As long as our original analysis was correct.
{DAVID K} [01:13:20]: Exactly, if your analysis is right, you can build much larger positions in the same asset. That might be one of the most important characteristics of the fund. Those who spoke with us in the two to three weeks leading up to the event know we wanted to show this visually in the presentation. It turned out to be harder to depict than we expected, but the ability to re-underwrite investments we already hold, that may be the most important day-to-day execution trait we have, and that’s what has allowed us to build the returns of the past three years. The ability to hold positions, resize them, without changing names too much—the portfolio picture shifts somewhat, the names change slightly, but the ability to carry them intertemporally, rebalancing along the way—that’s critical.
{GABRIEL S} [01:14:14]: At the end of the day, this is nothing more than generating a higher IRR on an investment than the first underwriting, which is only possible when there’s genuine belief that the business is perennial, stable, and that market swings are just that—market swings. We have many examples of executing this idea over the past three years, but we’ll keep bringing them up as we go along in this Q&A.
Q&A: Second Question
{Question} [01:14:50]: Good morning, given this collegial model, how do you decide team compensation?
{GABRIEL S} [01:15:08]: So, everyone is a partner. At TB, the idea is to build a business where—if everything goes right—over time we end up with a set of partners whose stakes converge to something quite similar.
That’s the orientation of a small firm with few people. From T0, our plan was to have a small partner group precisely because of that endgame, and for it to converge over time. So, compensation is a function of each person’s partnership stake. Those stakes were set at the start—there’s no magic, it was a heuristic—but a number that seemed appropriate at that initial moment, based on people’s backgrounds, and so on. Over time, there’s a catch-up between my stake and David’s and the more junior partners’ stakes.
{DAVID K} [01:16:05]: The two of us are equal partners—equal stakes. Our stakes and those of the “second generation” are equal within each cohort—and in practice there won’t be a third generation, at least for the next 20 years. The idea isn’t to create a third generation. Everyone earns according to the evolution of the partnership—where the firm is headed.
We don’t measure people by short-term results; the short term is too contaminated by noise. We didn’t invent anything here — we’re repeating what we already did before, and what Atmos did before. It worked very well to attract and retain people who buy into the project and contribute to long-term knowledge building.
So, trying to measure results is impossible in the short term, in what we do. If we are saying that the result is long-term, what one could do was to pool the money in the firm for 20 years and at the end say, "Folks, let's distribute it here." Maybe that would be the ideal, I'm not even sure it would be, because you have the randomness of who looks at what, and it's unfair to compensate a person poorly because the industry she landed in is a little bit worse than the other. What matters is the collective contribution to the entire process. The contribution happens from an individual company that the person covers, to what she brings to the day-to-day discussion, to the externalities, disposition, ability to represent the firm externally, how she brings relevant discussions from the outside in, how when she talks to a good investor she can articulate what we are doing. All of this is part of what we do. It's no use for us to do everything we do, to have a spectacular return for 10 years if we don't have the slightest ability to communicate and no one from the outside understands what we are doing. So, it is in fact something a bit more holistic and that has a certain stability by design. So, as everyone is compensated by the partnership and the stakes are equal by vintage, Gabriel and I earn the exact same, down to the decimal point; the second generation earns the exact same, down to the decimal point, and there is no intention of changing this in the medium term.
{GABRIEL S} [01:18:30]: There’s no outsize performance bonus. We included a phrase in the presentation: “ship, shipmate, self.” The ship comes first. Since people have boarded the project, regardless of anything else, this is the project—and over time we need balance, everyone contributing to the same direction, and that being clear to all. So far that’s worked very well. But given our view of capital allocation, big differentials are off the table—they end up destabilizing relationships.
{DAVID K} [01:19:16]: We accept a certain inefficiency, understanding there’s a large collective benefit to aiming long term. Markets have a problem: everyone gets Cartesian and thinks they’re great at math—when they’re not. We know we’re not great at math; at best we’re okay at algebra. If you try to pay one person 20% more than another, you generate so much long-run discomfort that it kills the collaborative spirit. The value lies in collaboration, not the individual. I wouldn’t do this alone; Gabriel wouldn’t; nobody would. The value is a small team that collaborates. We prefer long-term efficiency over trying to be precise in the short term — especially since the short term is more noise than information.
Q&A: Third Question
{Question} [01:20:00]: If you could tell me a bit more about your model for decision-making; a curiosity, given that you have a collegiate model—even if it's not a participatory democracy as you've put it—[I'd be interested in] any eventual challenges that you may have encountered in these three years or that may still arise in the future. And through two perspectives: the first, given that you have a collegiate model in which you have a broad discussion, how did you construct your decision-making process to, first, avoid eventually having an inertia in the decision-making on the theses that should exit the portfolio? And on the other side, how did you also seek to build it so that the good ideas, in order to be accommodated in the collegiate body, didn't have to be diluted to find the common denominator among everyone.
{DAVID K} [01:20:52]: I’ll start with inertia, then you can cover collegiality. It’s a good question—we reflected a lot on it over the last two years, maybe after year one. Inertia has a benefit when you’re invested in an approximately correct substrate. As I joked earlier: in Brazil you’re praying for the stock to go up; abroad, the best thing that can happen is for it to go sideways—or even down—so it helps you earn return and to have patience, it is important.
Our understanding today is that, barring any major changes in the world—and the world does change, it's not as if it's static, and our job is to consistently re-evaluate it—but looking at the snapshot today, the portfolio has changed very little because there's a benefit to inertia. You rebalance a position when it goes from a very attractive multiple, with very high prospective return, to a somewhat less attractive multiple, with somewhat lower prospective return, but you shouldn’t exit it —especially abroad, in very high-quality names. Closing out a position is a very strong statement; if you believe your analysis is right—especially as time passes and you get signals you’re on the right track—you should be patient and let it play out.
So, I think an important change is this: our effort must be to avoid over-activity, to avoid trying to be too active. So, I believe that regarding inertia, this patience is about letting time pass, but it's also about doing nothing for a long time, and you can go for months without doing anything. We go for months—we just went, I think, a month and a half, two months, without doing anything, without discussing anything in the portfolio, without discussing anything active, real, nothing practical to be done in the portfolio. The moment comes when you have a discussion and something has to be done. But, let time pass, let nature play out, and then you adapt to that afterward.
{GABRIEL S} [01:23:17]: On the collegial side: idea origination has a dose of serendipity. It’s hard to make it formulaic. Ideas arise mainly because we’re oriented toward what we’re looking for—the base is what matters.
Internally we’re clear on which sectors and company types we evaluate. So the chance we’ll suddenly jump into a direction that makes no sense is practically nil—and when it does happen, which it does occasionally, the “antibodies” show up and the group collectively says: look, this isn’t our direction; this has nothing to do with our style of investing, that’s the first point on origin. However, more than the origin of ideas, there is a process of vetting the ideas, which is more relevant, which happens on a day-to-day basis.
So, the day-to-day is an ongoing process, but our collegiate process is a process that demands a bit of what David mentioned, which is to give things some time. So, the timing is less a matter of weekly committees or what have you — which is the way analysis is conventionally conducted — and is more a process of trying to judge baskets of ideas in a horizontal way every three to six months.
So, we do this with a recurrence, with this recurrence that is a bit more elongated, where — and if this isn't clear, please ask me — we end up scrutinizing the ideas within a few hypotheses and a few topics that are predefined, like competition, like the business's potential for disruption, like valuation. So we discuss all these hypotheses in a more horizontal set, in such a way that we can weigh the ideas in parallel. Both ideas that are part of the portfolio, so old ideas, as well as new ideas.
So, when a new idea is already clearly very fleshed out on an analytical level —we've already done a presentation, we do internal presentations, so this is relevant, or what have you—this culminates in one of those collegiate meetings where we dissect the hypothesis testing of the ideas in more detail, and this is what ends up leading an idea to enter the portfolio or not. And then the sizing, naturally, is defined as a function of that.
{DAVID K} [01:25:50]: Sizing is an important discussion, because no position goes from 1% to 15% overnight. Before deciding sizing, you have to understand the collective comfort level with that company. To Rodrigo’s question, the best possible answer is Buffett’s: risk is not knowing what you’re doing, first, you need to know whether you understand the business.
There’s a kind of visceral comprehension you need—practically, a maturity in looking at the N possible futures, at N universes that exist. At any given moment you look ahead and see a million possible universes. In each of them, what is my level of comfort or discomfort with this company? What is my ability to re-underwrite this investment? Is this a business I—we as a firm—will be comfortable holding as a 10–15% position?
No it’s not or yes, it is; and from there, sizing is mainly a function of that. So, our sizing and portfolio construction approach has changed over time. In the beginning, we didn’t have a clear notion of “optimal sizing” or “optimal portfolio construction.” That only became clearer as we went along, as we understood businesses better, as we learned how we did or didn’t suffer with price adjustments in the short term; you get to know yourself as an investor, and you figure out what sizing each position should have over time. From that, sometimes you discover you sized “wrong”— when a stock drops and you don’t have the ability to increase.
Usually that’s because something happened you didn’t foresee—either an analytical mistake or a lack of creativity on our part. Maybe that’s an important discussion—the importance of creativity in the investment process; or it might be that the sizing was wrong because you moved faster than you should have—scaling too quickly relative to the opportunity or the business’s moment. That fine-tuning is very much tied to our personality as investors.
{GABRIEL S} [01:28:27]: Just to add—because this theme is very important—by way of example, it means that for a thesis to be underwritten at inception, it can be anything if the sizing is very small. But for a thesis to be underwritten with a meaningful initial size—one that allows future re-underwritings—it needs to have gone through a practically visceral refinement process, that has to answer several of those questions. For example: at T1, suppose a series of cataclysms occurs, maybe beyond what we can predict today.
Is the substance of the business, or of parts of the business, so strong that we’ll still want to re-underwrite it? That might sound trivial, but if you ask that question across various investments, in many cases the answer will be no—you won’t be able to re-underwrite in certain states of the world. So the ability to re-subscribe to the investment across multiple possible futures—that’s crucial. It ties to a question we once heard in a podcast with Todd Combs. He said something very basic—but the best insights are basic: “What’s the probability this business will be better in five years than it is today?” That’s one of the key primitives that assists in judging whether an investment can be truly meaningful.
That combination is what leads us to decide to size bigger at the start—and what gives us comfort to re-underwrite ideas over time.
Q&A: Fourth Question
{Question} [01:30:17]: You mentioned judgment versus Big Data. The big debate today is AI. I imagine you’re focused on it. What’s your big judgment on AI? Every day there’s something new— like a negative Oracle from a software standpoint. How do you see it? And second, from the perspective of the firm’s development—you’ve had great performance, a very good stretch, with some beta too. What have been the key learnings? Type I or Type II errors—you must have had more of one.
{GABRIEL S} [01:31:01]: Good question. I’ll start with AI. It’s hard to talk about now, because we’re in a moment with very obvious pockets of excess—for anyone who reads a newspaper, really. They’re clear. A stock jumping 40% in a day on the back of an announcement that basically says “I’ll be a junior participant in an unlisted company that needs funding”—that should set off red flags.
Our view today is this: several things in AI are material—almost beyond doubt. I’ll split between generative AI and core AI. We’re certain that several businesses we own in the portfolio have outright improved over the last three years thanks to the evolution of the transformer paradigm and what it enables in terms of predictability and accuracy of predictive models.
A basic example: Meta. Today it has internal systems for modeling consumer profiles—user profiles on the network—that are orders of magnitude better than three years ago. That’s simply the result of the huge number of GPUs the company has acquired on one side, and on the other, all the modeling built on top of that—models like Andromeda, Lattice, and so on, which are nothing more than very heavy machine-learning models that lean on this new generative AI paradigm. You can already see output. And to be clear—real output. In Meta’s case, this type of argument is better when exemplified.
When Meta manages to grow impressions—that’s the number of ad slots—and simultaneously grow pricing, that’s very unusual. It’s never happened in the company’s history except for brief windows. Normally, when impressions go up, prices fall. When Stories launched, impressions exploded—price fell.When Reels launched, impressions exploded—price fell (cost per mille, ad CPM).
But that’s not what we’re seeing now. Which shows conversion tracking capacity is exploding, even as new ad slots are being created. So yes—AI is a real thing. It’s important that this is stated. And we also start to see this inside the clouds, which is another angle to interpret this—specifically inference workloads, that matters most, more than model training— actual inference use cases are happening across many businesses.
That’s why Microsoft—a much more conservative company than Oracle—is seeing Azure grow high-30s, aiming at 40%. That’s a reacceleration of cloud growth. It’s mind-blowing. And GCP—Google’s cloud—which is better adapted to the AI era—when you strip out Workspace, which is a drag on growth, is also growing around 40%. So AI is the real deal. We have little doubt. But—and this is important—there’s also clear pockets of excess, and we start to question what part of the cycle we’re in.
That’s the key point. The Oracle example you gave—that’s exactly a case that raises an orange flag about where we might be in the cycle. The cycle could last three more years, five more—we can’t predict that. But it does seem—though I won’t say with total certainty, because this time it might be different, although this has never been true before—that there are pockets of excess. That’s evident. Especially when some startups raise their cash burn from $50 billion to $120 billion and investors applaud.
We’ve seen this movie in other eras. It’s a bit scary. Our view is—most importantly—our portfolio doesn’t rely on this. At all. That’s what I want to make clear.
We don’t operate thematically. AI is an addition for some companies that we already considered exceptional compounders, regardless of AI’s arrival. Microsoft, for instance—part of our portfolio since the beginning—two quarters ago was showing cloud growth in non-AI use cases higher than AI-related growth, briefly. It’s a company that grows come rain or shine.
So AI is an add-on. It’s an important one for us to think about in terms of where the world is going. But all our investments are businesses that stand on their own, regardless of this paradigm. And the main question we ask—more from a downside perspective—is: could any of our holdings be too reliant on reflexivity of the moment? Reflexivity is what made Intel extraordinarily profitable in the 2000s.
We don’t see that in almost any of our cases. Google’s core business—search, and its cloud—was already moving toward positive margins before AI. That’s the natural trajectory of a cloud business with huge infrastructure and sophisticated platform capabilities on top. So, that’s broadly our answer—how we think about it, and where we see ourselves in the cycle.
That’s how we’ve been approaching AI.
{DAVID K} [01:37:27]: From the fund’s start, as Gabriel said, we’ve never invested by theme. In the presentation we even used the Mag7 as a company cut—today it’s Mag7, it used to be FAANG, soon it’ll be some AI-only label. Themes are irrelevant to us.
Practically speaking, we haven’t made any investment that you could call a pure-play AI. We never sought that at any time. We never owned NVIDIA in the fund, for example.
{GABRIEL S} [01:37:56]: And it’s not for any anti-NVIDIA reason. We just don’t have the apperceptive mass to invest in NVIDIA. We can invest in a fab company like TSMC, where the mental-model reading is much clearer—you’re basically interpreting a scaled-out monopoly business.
Plain and simple. It’s much easier for us to invest in that kind of business, even if the substrate looks like rocket science. The substrate doesn’t matter that much. You don’t need to understand how an electron travels inside a transmission line to invest in a utility.
What matters is understanding the business’s microeconomic characteristics. Businesses with clear microeconomics are easy for us. NVIDIA might end up the largest company in the world for the next 50 years—we don’t have the apperceptive mass to invest in it.
Hyperscalers are something entirely different. And why not NVIDIA? Because NVIDIA created its own reality over the last three years. It was broadly prepared to do what it did—but those economic profits appeared out of nowhere.
Before that, it was a GPU provider for gaming. It created this modern paradigm. We don’t know where that ends. We don’t know what the equilibrium looks like at the end. Will ASICs—the hyperscalers’ proprietary chips—be competitive enough over time? Today it seems not, but maybe yes.
No one can answer those questions. Honestly, that’s one of those unanswerable buckets. Along the way there was the DeepSeek moment. What did the DeepSeek moment imply for everything?
It implied zero for the hyperscalers. But what could it imply for a company that, besides selling GPUs, also sells networking and scaled compute—basically new mega-clouds. It could imply that the efficiency required ends up such that what NVIDIA offers isn’t what will be needed over time.
As happened with Sun Microsystems in the 2000s, by the way. The world shifts in this kind of business. It’s an incredible company, but we lack answers to a series of questions. That’s what makes us limit certain investments.
Errors and learnings from the first three years. Just to circle back to that question—and then whoever wants can keep asking—we have a few points. We have a couple of points to comment on. First is evidence about the types of businesses we’re willing to buy.
We touched on this during the presentation, but a major learning—definitely more learning by doing than anything else—was our need for strong support in the ability to re-subscribe theses in order to actually hold an investment. And our difficulty in investing—even in businesses that are speculatively very interesting—in almost any size in that type of name. Again, this is personality. Some people do it well.
I don’t think that’s a trait we have much in our DNA. In terms of calibrating what we set out to do, I’d say that’s where we learned the most. We have a problem when it comes to talking about “errors” in the first three years because—for better or worse (and I think more “worse” than “better”)—few portfolio theses went through truly problematic situations in these first three years.
Real ones, I mean. There was a lot of mark-to-market. In some cases, more than 50% drawdowns in some holdings. But at no point did we feel deeply uncomfortable about business perennity, and there wasn’t much fluctuation in the businesses themselves.
When I mention those price hits: we started the fund in August 2022. In October 2022, one of our relevant theses—Facebook—reported and dropped 30% right off the bat. Then it became one of our biggest compounders. For us, that moment was very self-evident in terms of what was happening: a mix of existential fear via TikTok—where we see growth curves going to diminishing returns.
On the other side, Apple’s ATT (App Tracking Transparency) coming in at full force, hitting results hard. Apple framed a privacy move that was basically a mechanism to kneecap Facebook. These are two companies with a long, intense rivalry, even if they’re allies at the business level.
And on top of that, Meta was upping CapEx—blowing out CapEx and R&D with Reality Labs. A perfect storm. From a pure business-fundamentals view, we looked at that and asked, “What on earth is going on?”
In terms of the reality of it: there was a point—just to give a concrete example—when Meta traded at only a small premium to Salesforce. Salesforce is a very successful CRM company. We joked that Meta, internally (they had a product called “Customer,” later sold), probably had a Salesforce more valuable than Salesforce itself.
So it turned into existential fear and massive hate-selling. We even identified it at the time by looking at the company’s bonds. Curiously, we saw Meta bonds trade at a significant premium versus other Tier-A bonds in the american market—Microsoft or Google bonds.
It made no sense. Meta was net cash and had tons of PPE in servers. That actually helped us re-subscribe the thesis with more conviction over time. Why am I saying this?
Because it’s a textbook example of something that swung wildly in price and, in fact, helped us—by reinforcing the re-underwriting discipline. Of course, when you’re inside the storm, it’s never easy. But with discipline, we re-subscribed. The company may face other issues over time, but those specific issues were correctly identified as non-issues.
By now they’ve been largely transcended. So, at the fundamentals level, we didn’t experience that much pain—like owning a company whose profits got cut in half and thinking, “I’m facing a permanent earnings impairment.” Going forward, if we keep selecting reasonably well, we shouldn’t go through many such cases.
But when we do, perhaps we’ll have more to say on the “errors” angle. That’s about it.
Good question.
And again, within our apperceptive mass, we joke there were two moments to invest in NVIDIA—two. One was during the inventory write-off—crazy times—NVIDIA traded at a $250–300 billion market cap, also around 2022, writing off GPU inventory because of crypto. Back then NVIDIA was a different company; results were boosted by GPU sales for crypto—Bitcoin mining.
That first moment: it was clear the accounting impairment had zero predictive value for the business—but we knew that business far less than names like Meta, which looked equally discounted. So we didn’t buy then. The second moment to have bought NVIDIA was when it hit a $1 trillion market cap—after it had already tripled.
Why? Here’s the joke—but it’s also true. By that second moment, NVIDIA had just made its first profit step-change, which—looking in hindsight—was relatively evident (and honestly, it already felt that way at the time).
The questions were exactly these. “We won’t invest—but it’s obvious this is the case.” The company was on a ramp-up, at that point effectively the only game in town for high-end AI GPUs, and was going to keep delivering absurd growth with hefty margin. In effect, you were buying it on a two-years-forward basis at something like 8–10× earnings—if things played out.
And that’s what happened—the business exploded further. We still couldn’t pull the trigger. At that same time, we even had TSMC to buy—a business we understood far better. So, is that an error? Honestly, psychologically, we don’t think so. You can’t swing at every pitch, and trying to swing at pitches in things you don’t know catalyzes a series of real errors.
So—it’s ambiguous.
{DAVID K} [01:47:33]: In this case, there’s a stylistic choice—almost aesthetic—about what you’re willing to do and what you don’t want to do. You say: this pitch, I don’t need to swing at. We had the alternative of TSMC, which was already a position that, in a way, reinforced our comfort—and eventually we increased it.
So, looking at NVIDIA actually gave us more apperception (a concept we talked about in our second letter, at the start of the year). It made us better investors in TSMC, via NVIDIA. And I think the other part of this case—why I also don’t consider it a mistake—connects to the previous question: it’s about not putting yourself into an investment where, in several possible future universes, you’d find yourself in total discomfort. You thought you were swinging at one pitch, but it was something totally different, and then you’re left unsure what to do.That comprehension—what I call a stylistic choice—is understanding how that business will play out over time and saying: “There are scenarios where I don’t want to be in that position. I don’t want to find myself there.”
That made us better investors in TSMC. If we had invested and some of those universes had played out, it could have made us worse investors overall. Missing badly on a pitch, or making a big mistake, can hinder you for a long time, because it makes you start questioning your premises. So accepting that you’re not optimizing performance—just as we don’t try to minimize drawdowns, we also don’t try to maximize performance. It’s a balance. That balance point isn’t exact—it depends on your personality, our personality as investors.
{GABRIEL S} [01:49:40]: Equally important here is knowing what you don’t know. Of course, we knew something about NVIDIA—we could have invested. But we knew much less than we did about other theses. And generally we’ll prioritize theses we know much more about, where we have greater degrees of comprehension, rather than theses that look like they have incredible optionality but that we understand less. That doesn’t mean you should never invest a percentage of the portfolio in some of those —but it’s more about choice.
Next. Actually, let’s just mention this briefly. This first phrase is very relevant—up on the screen—it’s a Bezos quote we like a lot. He said: in business, people often ask “what will change in the next 10 years?” but they should ask more “what won’t change in the next 10 years?” We use that a lot at TB, both in terms of selecting businesses, and also in terms of what inside TB won’t change in the next 10 years—which is very important for us.
That ties back to what we said at the beginning: above all, we’re very comfortable that analysis will change. Today we use ChatGPT5 for a bunch of things, we’ll use it more and more. But the philosophy guiding what we do—that’s probably our basic pillar. If we change that, then we should be broadly questioned. But just to make the point: we think that quote really symbolizes what we said earlier.
Q&A: Fifth Question
{Question}[01:51:27]: Building on the AI theme, and in the lens of what you just said—about what doesn’t change—have you thought about, or do you think about, having an AI agent actively debating with you in the future of intelligence?
{GABRIEL S} [01:51:52]: We think that’s very interesting. We already do it, in a more subtle way. We believe these systems will improve a lot over time. The most important thing is—there’s a phrase from economist Tyler Cowen: people should already be writing for the AI of the future, not just for humans, but writing for AI. It’s an interesting mental model, because it forces you to think how you’ll feed information into a much better system. Because clearly we’re using the worst AI we’ll ever see and that system will help us as decision-makers, as an additional tool. Just like Excel is an exceptional and important tool.
Today what we do: we record all our management meetings, transcribe them using Whisper, and systematize them into a structured set of questions. All our theses pile up into those questions, and every meeting we re-evaluate them again through that framework—based on AI, basically ChatGPT 5 today, which is in our view the best available. Over time, we accumulate more and more data, and with system memory, in the future we’ll be able to query and recall what we said at any time. So today we’re mostly building a database for the future, more than anything else.
As a decision-support system today, it’s positive, but mainly as a blind-spot identifier. Sometimes we have long debates, structured by these questions, and we use AI to point out topics that were under-discussed or under-reflected. That’s helpful—it lets you revisit certain premises.
At the analytical level, its effectiveness is still ambiguous. But we think it will keep improving, especially in filling blind spots. So right now, we’re building an informational database for the future. That’s our main approach.
Q&A: Sixth Question
{Question} [01:54:30]: Given the different types of businesses you cover, how do you build your cost of equity, your ke? What do you consider as the risk-free rate, and how do you adjust the equity premium?
{DAVID K} [01:54:50]: In practice, we have two main reflections on cost of capital. The vehicle—our fund, which has existed for three years—is denominated in reais, so its cost of capital is also in reais. So, every investment we make abroad—for example, in Microsoft—we short the equivalent dollars, capture the interest-rate differential, and bring that investment back to our functional currency, the real.
Our cost of capital in reais is the real interest rate, plus something. What’s that “something”? In our view, it’s not science. There’s no magical number.
The second reflection is: once you set yourself up as a generalist investor, you give yourself more degrees of freedom and start observing many more businesses. The corollary, which we’ve lived in practice, is that our definition of a “good business” changed, it has become stricter. The bar is higher. Correspondingly, the required return we seek on any investment has gone up too.
So, our required return is higher—both in Brazil and abroad. It’s interesting. It becomes clear that, to justify an investment at 10–11% real in reais—let’s say a very stable business, and this is the return calculated in the spreadsheet—but with no capital redeployment optionality—that’s not enough for us. We seek returns above that, to capture compounding.
Eventually, optionality of cost-of-capital compression may happen, but it shouldn’t be in your base case. If the cost of capital compress, you should benefit. Cost of capital opens and closes—especially in an emerging market—at the mercy of the wind. Brazil is an interesting country.
There’s a great line by Nelson Rodrigues—it’s the best description of Brazil, I think. He was talking about Brazil watching the World Cup in the 1950s or ’60s: “Brazil oscillates between the most delirious optimism and the most obtuse pessimism”. That’s exactly it. Right now we’re in obtuse pessimism—you look six months out and see nothing, can’t predict anything, the cost of capital explodes.
And no one doubts that if we get a positive government, with a positive agenda, given the amount of capital in the country and the historical bias of big savers to keep it here, by 2027 we’ll be in delirious optimism again—real rates at 4–5% perhaps. So trying to actively navigate these swings is not what we do. We don’t trade those openings and closings. But we do want a reasonable premium over Brazil’s real interest rate.
{GABRIEL S} [01:58:01]: And just technically—because, in fact, since we hedge the FX, the functional currency is the real.
So we level everything. I don’t like giving numbers, but I’d say our prospective real return in dollars—since you earn the real-rate differential on top when investing abroad—oscillates anywhere between 11% real to 18–20% real. Today, many of the companies have appreciated a lot, so at times the prospective return was much higher.
It’s compressed since. That real-adjusted return is a great premium over the NTN-B. That’s how we think—always along the same lines, as an interest rate differential.
{DAVID K} [01:58:50]: The fund’s exceptional return these past three years is because you had that compression abroad—and we captured it. Just a quick comment.
{GABRIEL S} [01:58:59]: Which, ironically, is the worst thing that can happen to the investor. We’d have preferred it hadn’t happened, so we could have stayed longer in those compounders.
And this is mathematical. Anyone who invests knows this—bond people get it even better, credit people—the reinvestment rate is the most important rate. So, it’s a bit tragic, but it happened.
{DAVID K} [01:59:22]: Important thing is—there’s still return. There will be moments when it will be hard to find returns at all.
Q&A: Seventh Question
{Question} [01:59:34]: Could you explore the concept of “vision,” more from the perspective of business owners—what makes the vision of a Brazilian entrepreneur with a great business diverge from that of someone like the founders of Nubank or Mercado Libre, who manage to create businesses with increasing marginal returns over time? Where and why does that difference happen?
{GABRIEL S} [02:00:08]: We’re quite clear on this: founder-led businesses, last year people made a lot of jokes about Brian Chesky, Airbnb’s founder, over that ‘Founder Mode’ thing—poor guy— it’s obvious what he’s saying is true; it’s plain to see it’s true. Companies run by hired management typically can’t make a series of tough, forward-looking decisions. So the cutoff line in this answer—at least from our angle, the way we read things—is that we prefer to have a portfolio that, in most cases, holds owner-run companies.
Founders have the ability to “make hard choices” at virtually any point. Examples: Zuckerberg in 2022, blowing out CapEx buying GPUs without even knowing we were entering a generative AI era—simply to be prepared to fight TikTok’s move into video. No hired manager would do that. And frankly, when the market first questioned CapEx growth—with the stock tanking 70% from peak—you’d have had a problem staying the course.That’s why founders are so relevant—they can endure the challenging moments of the cycle, the times they’re questioned. The good founders, at least. And they can make the tough calls.
And tough calls really are tough—you’re under market fire. Mercado Libre was questioned for giving free shipping for years, but that was obviously the decision of someone thinking about demand aggregation in the most perfect microeconomic way. So, founders make a huge difference in our investment selection.
{DAVID K} [02:02:07]: I understood your doubt is also about how founders realize whether their business is better than another?
{Question} [02:02:17]: What happens in Brazil? Why do we have good businesses here, but the founders don’t seem to have the same vision—don’t manage to build businesses with increasing marginal returns? We have good businesses, but sometimes it feels like that comprehension is missing.
{GABRIEL S} [02:02:33]: The first part has an issue, that there simply aren’t that many businesses in Brazil with the characteristic of increasing returns. So that’s a prior constraint—not just a founder problem, but a typology problem. And this is systemic. It’s the historical lack of human capital. And it’s the fact that global demand-aggregation and platform businesses are global.
You have one Google, one Meta, one TSMC. You don’t replicate those at scale except in very particular markets—like South Korea, which has something like Naver. But generally, those businesses are inherently global. The Brazilian entrepreneur—we even think in traditional businesses they’re quite competent.
But traditional businesses, by definition, are limited in their ability to scale beyond one jurisdiction. So then you ask: why wasn’t Nubank founded in Brazil? I think exploring that is very deep, very sociological—hard to answer. Why wasn’t Mercado Libre founded in Brazil?
Again, sociological, hard to answer. Why one was founded by a Colombian, and why Argentina’s startup and VC ecosystem is more sophisticated than Brazil’s—that’s just a fact. Again, sociological, deep, and more a matter of history than anything else.
{DAVID K} [02:04:03]: I have some hints of the reason. I think Brazil 80 or 100 years ago was actually more connected to the world than it is today—education, families here linked with families abroad. So there was more global comprehension. For example, when media shifted from print to radio and TV, the players were local, and they managed to execute on a plan to create local media groups.
Chateaubriand bought a TV station, brought it to Brazil, made theater actors perform in front of cameras to create the TV market. He was seen as a madman—and he was a bit crazy. There was also the sense that business was somewhat local—companies were local.
{GABRIEL S} [02:05:03]: The global constraint didn’t exist; it really broke a global barrier. You could replicate the American media entrepreneur in Brazil—that was the reality. Now Facebook eats your cake. That’s kind of what happens. It’s much more challenging.
{DAVID K} [02:05:24]: And the world has truly become more global. Technology, the internet—moving into a connected world of bits—broke barriers. It allowed Shopee, for instance, to operate here, even using bizarre local logistics players. But since it connects directly with the consumer, knows the consumer, it can push product into the market anyway. It finds the crack.
When the ship is cracked, water finds a way in. They find the breach and compete. So there’s some local incapacity—really, a lack of vision.
There’s also the capital constraint. In practice, Nubank is a case that, coming from its background, managed to attract capital to place a bet on a..
{GABRIEL S} [02:06:12]: Why wasn’t it Brazilian? That’s the big question. And that’s a question for everyone here as well.
Q&A: Eighth Question
{Question} [02:06:27]: I’d like to go back to that part, in essence, in the initial presentation, where you had those three buckets: fixed income and bonds, the global and local slice. I remember I questioned that a lot, along with David.. I used to ask: if you were starting again, back in the past, would you still set that local vs. global floor? And invariably, there was always that view of a “bucket of opportunities” versus a “structural portfolio.” Along that evolution, how do you keep this discipline? If, in practice, the portfolio drifts toward global opportunities, for example, and stays concentrated there, what makes that a conjunctural move and not a portfolio that, structurally, drifted and then, due to bias or whatever, ended up stuck in a world we had questioned back then? Because, for better or worse, back then there was that discipline of: no, I’ll at least force myself to question and have some regional diversification.
{DAVID K} [02:07:35]: I’ll start, then you add. I think there’s some personal path dependence—since we were born here, we’ve invested in Brazil for 20 years, we know the companies. And Brazil has a disadvantage that works in our favor.
Since so few businesses emerge, the “substrate” is already given. We already know the businesses. The maintenance cost is relatively low. So there’s some inertia in continuing to follow these companies.
There shouldn’t be inertia in keeping allocations in the portfolio, but there is some personal inertia in continuing to track those names, in the ones that make sense, the ones we decide to keep following. What we wanted to remove now—ripping off another band-aid, really—is the obligation to be allocated here. We know we’ll likely remain allocated here for a significant part of the time, but we don’t want to obligate ourselves to it. There’s no reason to be structurally invested here.
It can happen opportunistically, but there’s no reason to be stuck here. Again, our repertoire of companies and business types is big enough now to let us look at multiple possible futures and say: in this kind of universe we may end up with nothing in Brazil—and still build a portfolio we’re comfortable with. European investors, there’s an important comparison here: I think the European investment industry is more similar to the Brazilian one than the American.
We all look to the U.S.—because that’s the dream for Brazilians, from the simplest to the wealthiest, it’s America. That’s who Brazilians look up to. But in terms of how the industry developed, we may have much more in common with Europe. Europe’s investment industry looked a lot like Brazil’s. It was European family money handed to European managers to buy European stocks.
That’s what Brazil was like 15 years ago. At some point, some people flipped the switch and became global investors. Today, if you look for a European-only investor, they don’t exist. That creature is extinct. Gone.
{GABRIEL S} [02:09:50]: Maybe you’ll still find a European small-cap manager, something like that—very specific cases.
{DAVID K} [02:09:54]: Exactly, very specific things—or mutual funds. But the kind of investor we are, that doesn’t exist in Europe anymore.
Why? Probably because those who didn’t adapt died out, and those who survived became global. And if you look at their portfolios, many have nothing in Europe. In practice, they’re just buying businesses.
And then, when there’s a deal—he’s going to buy a physical business, an infrastructure asset—he thinks he can find it in Europe, because he understands the regulation better. Perhaps the only analytical barrier that’s hard to overcome is regulation. Because regulation is, by design, local. The real skill is understanding regulation dynamically rather than statically. Reading regulation as a snapshot—looking at it one day and understanding it—isn’t that complex, especially now with ChatGPT.
You can ask it for help and in X months you’ll understand. But understanding how regulation evolves over time requires knowing who the regulator is, who the players are, the industry associations. I think a lot about this in Brazilian power utilities—that’s our work. Who are the players, who’s the regulator, who’s the granting authority, the Ministry of Mines and Energy, who are the consultancies.
That has huge value—because they tie everything together. We don’t realize it, but there are only a few consultancies that act as real think tanks, that at the end of the day set the direction of what’s “right” in regulation and create consensus. So, regulation may be a barrier. Beyond that—we invest in businesses. Just one last comment: for example, a business we looked at recently—we didn’t invest—but we put it in our ‘possible-for-the-future’ repertoire box, is the aircraft-turbine business.
We weren’t familiar with it, GE went through a transformational phase—splitting energy and healthcare into two new companies. What was left is the turbine business—narrowbody aircraft engines, like those used on shuttle flights. That market is split between two companies: GE, in the U.S., and Safran, in Europe. It’s the same global market. Where the company is located is practically irrelevant—hardly relevant at all.
You might need to pay a bit more attention to the European one—governance issues, government shareholder—but that’s about it. It’s a global business.
{GABRIEL S} [02:12:20]: In this case, Safran, right?
{DAVID K} [02:12:21]: Yes, Safran. The business changes very little—it’s global. So we select businesses. Geography is clearly being abstracted over time.
{GABRIEL S} [02:12:32]: And just to complement briefly—because in this sense, we do work on understanding “utility businesses,” which we mean in the basic sense of the word utility—not just regulated ones, obviously. That’s exactly the point David made. Investing in regulated companies anywhere in the world is difficult.
But investing in businesses with utility-like characteristics—from railroads to credit card networks… off the top of my head: rating agencies, clouds—that’s all utility business.
Again, they typically have absolutely global exposure. There’s no advantage at all in being where the headquarters is, much less the tax domicile. The advantage is analytical, not informational. And these businesses have very stable characteristics.
These are the ones we should, over time, migrate toward—if opportunities of this kind diminish in Brazil. Especially if we lose the ability to get exposure to that kind of stable carry in Brazil. But, to your question: today, we feel Brazil still has a relevant role in the combined portfolio—particularly a few names that still provide structural carry at reasonable valuations. And also some businesses with a slightly more opportunistic connotation.
By “opportunistic,” it doesn’t mean we don’t think they’re good businesses—it just means that, unlike some of the structural ones, which have a duration (not in the bond math sense, but a perennity where you don’t know when it ends), these have shorter thesis timeframes. Sometimes they’re trading, given current cash generation, at spreads far wider than what they should in a steady state. So today Brazil has that characteristic. And if a big repricing happens, we want the freedom not to be trapped. That’s the idea.
{DAVID K} [02:14:53]: The past three years were a poor window for deployment. In practice, investing in Brazilian equities is hard for investors. All the more striking, then, that a meaningful share of the fund’s returns in that period came from local assets.
It also raises a point about the about the work itself: you do the same thing every day so that you can notice the day that’s different. We go to the office every day; it feels like nothing’s different from the day before, until suddenly something screams at you and it seems obvious. There’s value in giving yourself degrees of freedom—and then realizing at some point that Brazilian businesses aren’t as good as foreign ones. But at prospective return levels, like at the turn of last year to this year—we’d never seen that before.
In 20 years, I’d never seen a company buy back over 10% of its float in 10 months—like Stone did—while generating 15% of capital and handing back part of that capital. That doesn’t happen historically in Brazil. And trading at 5× earnings. Then reality slaps you in the face—because yes, it’s great to buy very high-quality businesses, provided you’re being paid for it. Sometimes the business isn’t quite as good, but it’s at such a steep discount…
And it still has inherent growth. In practice, the company has inherent growth—here, it grows almost naturally, with a tailwind. If I’m being paid enough to take the risk of being in Brazil and buying a company whose cost of capital is inherently Brazilian—local—and the return is globally competitive—then that’s fine. So again, this “doing the same every day to be ready for the day that’s different” is very important. That consistency of craft, of practice, is what lets you capture those special moments when nature presents them.
Q&A: Ninth Question
{Question} [02:17:10]: Picking up on the point you made—that you’d like us to call you out if you ever change your investment process or philosophy. The “opportunity bucket” or opportunistic allocation has two sides: it allows shorter-horizon allocations, but it also lowers the quality bar for the business itself. How do you ensure that doesn’t become a broader permissiveness in analytical quality? And what’s the point of having this book, given it’s also a noise generator—and you’re a lean team, only able to cover everything because your playing field is so focused?
{DAVID K} [02:17:57]: I think it’s more a change in communication than in what we actually do. In practice, I’d say—like I mentioned earlier—our bar, our definition of what’s a “good business,” has actually become stricter, not looser, over the last three years. Our demand for comprehension, our comfort with the business, has gone up. Our demand for quality in the business itself has gone up. So those two points partly answer your question. What we’re trying to do here is just make explicit something that already happens. For those who’ve seen how we present the portfolio— we even considered showing a breakdown here. We dropped that idea quickly.
There are local assets in the core of the portfolio—alongside global names. Namely, Nubank is there with Apollo, Meta, Google—the best companies in the world. Equatorial is also a core holding for us—it’s in Brazil, but it’s structural, not opportunistic. It sits alongside Microsoft, Visa, Mastercard—very stable carry assets. And then, further down, you have the companies where, in fact, we have less ability to predict what they’ll be in five or ten years—less ability to answer Gabriel’s question: “Will this business be better in five years?” That’s harder to answer. But it’s paying us well enough for it to make sense and to be part of the portfolio.
Our intent is just to remove the “guardrails” we had imposed on ourselves—stop looking geographically and instead by type of carry we’re buying. Whether it’s a more cutting-edge business or a more stable one, the point is: understand what kind of carry we’re buying. Even as an internal heuristic to monitor how the portfolio evolves. This gives us room for businesses of less-than-exceptional quality to have some space in the portfolio and that helps maintain sanity.
We think a lot in terms of the search for returns—it’s kind of natural: where does the return come from, how is it going to be generated? A big part of the answer is about trying to avoid mistakes and bringing sanity to the process. Understanding that when there’s a business abroad that might be really good, but it’s trading at 30 times earnings, while you have a local business trading at 5… across possible future scenarios, what are the chances I don’t make a lot of money on this one—or, at the very least, avoid taking a big mark-to-market hit with such an outsized position? We did have outsized positions abroad six months, a year and a half ago—Meta, TSMC were disproportionately bigger than the rest. Not anymore.
Today, the portfolio is much more balanced—position sizes are more even. Because there’s nothing abroad with prospective returns that justify outsized positions. And where we do find those returns today is in local assets—both the high-quality, that give us comfort, and in those we’d call a bit more opportunistic. So this heuristic helps us keep control. Removing the guardrail is just preparing for the future, not a present necessity. Today we’re at a high in Brazil exposure compared to the past two years. That might actually be why now is the right time to rip off that band-aid—to prepare for the future, both on the downside and the upside. If next year’s election is favorable, things re-rate very quickly, and opportunities fade fast. On the negative side: we don’t minimize drawdowns, but we also won’t stay invested here if there’s no opportunity. We might want to leave. So, while we’re currently at peak Brazil exposure, giving ourselves that freedom seems the right move.
{GABRIEL S} [02:22:03]: And just a very brief comment—because I think David covered it all—there are a few pillars that are important at the analytical level for the businesses that, to your point, allow us to look at them with emphasis and step away from “pure opportunism.” Because we don’t believe in pure opportunism. So it’s very important that the businesses we hold as “opportunities” also have analytical links with global businesses. We’d rarely invest in a sector that has zero correlation with things we follow globally. Take Stone, for example, which David mentioned. We study the acquiring sector in depth globally—we look closely at Fiserv, we also look very deeply at the networks.
So if Stone were some random company in a random sector, it would be much harder to form an opinion—we’d let that pitch go by. But – and this is important - it’s within financial services. Another thing: in the opportunistic bucket, the bar gets raised to the point where it’s necessary for these companies to generate a lot of cash and be returning cash to you. That’s a prerequisite. In businesses where we don’t have such strong conviction, or where the future is harder to see, we set a requirement that they return a significant amount of cash over the next five years. Because if we’re wrong on everything else, at least we got part of the cash back. So those are some of the drivers.
{DAVID K} [02:23:41]: It’s also a major drawdown limiter—a protection against permanent impairment. Let me add one last comment—we keep stretching, but that’s the point of the meeting: it’s not to be linear, but more random.
Another important part of how we invest is drawing the no-go zones. And here there’s some path dependence from the industries Gabriel and I have covered over our careers. I looked at industrials in Brazil, which basically means auto parts and WEG, which is a good company. I also covered education. Gabriel covered real estate. These are businesses we never want to look at again. After so long in bad industries, these are now no-go zones—we won’t enter them opportunistically.
{GABRIEL S} [02:24:42]: And again, it’s not arbitrary—it’s about the microeconomic characteristics of those businesses. There’s a clear microeconomic bar, and a bar for cash generation capacity. Real estate, for instance, could be one of the few investable businesses, no matter how bad it is, because sometimes in the cycle it returns cash. But I’m just making a small addendum here—the probability of that is 0.1%, but in other states of nature it could happen. It has physical backing.
But other businesses, with low asset backing and low cash-generation potential, are completely off the radar. We won’t buy something just because it trades at a low P/E.
{DAVID K} [02:25:17]: That’s maybe the biggest limitation: not even knowing what’s happening in education and auto parts companies—and I am glad I don’t.
Q&A: Tenth Question
{Question} [02:25:29]: My question follows on what you said earlier about capital impairment. One reason you gave was business displacement. How, at TB, do you analyze the possibility of displacement in your portfolio companies?
{DAVID K} [02:25:45]: Again, in this work—you use multiple mental models to analyze any company. Munger described this best. If there’s one book you should read for investing , that you should reread every year—and I took many years before rereading it myself—and it’s important to do so, because each time you read it, as you gain maturity, you read it in a different way it’s Poor Charlie’s Almanack. The new Stripe edition is spectacular. It talks a lot about the mental models you should use: alignment, adaptability (borrowing from biology), lots of microeconomics.
I studied economics—wasn’t a great student—but at least I saved some of the bibliography. When I started working, I realized it was important to revisit those concepts I hadn’t had the maturity to grasp back in college: understanding what competition really is, what marginal return is. These are the concepts that protect you over time. And you end up with two main paths. One is direct to the business: what is competition, what is technological disruption, what businesses might be displaced when something new emerges. From a software business perspective, what is AI that could eventually displace the business — what will be the first software that people will abandon along the way?
And what’s the software bundle—like Microsoft—so entrenched, with such high switching costs, that companies are afraid to change it, the CTO knows if he changes and fails, he loses his job. It’s something that will remain there for many, many years — it’s a very stable business. The other path is, again, the re-underwrite: it’s your personal ability, in possible future scenarios, to not be able to reinvest — when reality presents itself in some way, whether through market pricing, or because at some point the controlling shareholder or the company made a capital allocation decision you’re not comfortable with. A lot of people didn’t invest in Meta in 2022 and 2023 because they weren’t comfortable with the move Zuckerberg was making.
It’s really important—I recommend to everyone I talk to—to listen to Meta’s earnings calls from early 2022 onward. To see the consistency and clarity of what he was doing. Along the way, even he broke into a bit of a cold sweat with the decisions that were made, because they start to impact the company. Abroad, the company compensates through stock, and when the stock drops 70–80%, probably some of the executives become uncomfortable. But the consistency of his decisions was clear. For us, that became evident over time. And in the end, he reaffirms everything he had said. As an investor, you have to keep reassessing. It’s not static work. You don’t just say “this is set for 10 years.” At every opportunity, you should revisit. It’s ongoing work—you can’t avoid that.
{GABRIEL S} [02:28:52]: Some questions open, others close. But like David said, the key is that some businesses rarely open existential questions. Those are the truly exceptional ones. Usually, they trade at very high multiples—so nature doesn’t help.
{DAVID K} [02:29:13]: Today, if there’s another pocket of excess beyond AI-related private companies, it’s in the most stable businesses in the world, which weren’t seen as unregulated monopolies but have come to be viewed that way over the last 10 years, they traded at 15–20× earnings. Now they trade at 40–70×. We’re talking about Costco, Walmart. Their competitive positioning is very stable. They grow little—but they trade at 40–50× earnings. Walmart even has some e-commerce, Costco maybe even more relevant.
We’re also talking about the rating agencies. Even GE, which we mentioned as an interesting and that could provide opportunities over time, already trades at 50×. The very stable businesses have reached very high valuation levels.
{GABRIEL S} [02:30:08]: And it’s not as if these businesses are trading at those valuations the way Oracle trades—where the expectation is of massive future growth. No, no. Expectations are minimal—just that current growth continues. That’s it. So nature balances things out very starkly at certain times.
{DAVID K} [02:30:36]: There’s a reason for this. In this case, you’ve had a movement for over 10 years now of capital migrating from active management to passive management. These are the quintessential businesses for investors wanting a stable portfolio to pair with passive strategies. It’s almost global capital flow. Look at the portfolios of most global investors nobody holds that many positions, they become marginal positions. But the very stable buckets—where capital floods in via ETFs—they’ve become extremely relevant.
So part of it is also avoiding these “stability errors.” Overpaying there can lead to painful corrections.
{GABRIEL S} [02:31:11]: That’s a permanent loss all the same. If you buy a business with a 2% earnings yield, at some point that 2%—unless you compound capital, unless the business redeploys all earnings at very high rates—will converge to something more reasonable, and you’ll have taken a capital loss. A 2% end yield is far too low to be compensated purely by earnings-growth carry. So that’s a very relevant variable.
{DAVID K} [02:31:43]: I’ll take Luiz and one more to keep us on schedule.
Q&A: Eleventh Question
{Question} [02:31:48]: First, very cool initiative doing this event. We see this a lot abroad—less so in Brazil. Let’s hope more funds do it, and that it’s annual, not every three years. {No, it will be annual. It will be annual.} My question ties to the last comment—the growth-investor dilemma, especially abroad. How do you think about the framework for investing in growth when starting valuations are high? You have some portfolio names at ~30×. How much time do you spend on valuation within your framework, and how do you balance it in the portfolio?
{DAVID K} [02:32:37]: On the margin, more and more of our positions reached those valuation levels—but the amount of active management we did over the last year and a half—really from mid-last year through mid-this year—was significant. It’s hard to measure the active effort because our portfolio has two “sides” that sometimes move very differently: one side falling, the other rising.
Not every portfolio is like that, some move more uniformly. Our way to measure it—especially reductions and active moves— is the number of company shares we hold for each share of the fund, thinking of Gabriel and I as investors ourselves: how many Meta shares did we own last year and how many we have today? Looking at the main names that moved from ~20× earnings to something closer to 30×…
{GABRIEL S} [02:33:41]: In some cases, from 10×. Meta, for instance—that was extreme. Some names truly flew relative to the starting point. We still hold them, but that end-state multiple is largely a function of the company’s success.
{DAVID K} [02:34:01]: Our active effort cut share counts by 55–60% to as much as 80%. Even positions that are 4–6% of the fund today—if we had done nothing, they’d be 20%. We trimmed a lot.
There’s some inertia— we have, maybe as a characteristic, the willingness to hold on to these businesses. At 60×, it’s nearly impossible to justify. At 30×, given the real growth capacity these businesses still have, especially when you understand how that growth will play out over the next 10 years—cash generation, maybe even margin expansion/operating leverage like we saw these past 2 years—we’re comfortable carrying a reduced position. For example, we once had over 10% in Microsoft; today it’s under 5%.
The best thing that could happen to us now would be Microsoft dropping 30%—out of the blue, as Trump says something. A 5% position becomes ~4 or 3; we’d take it back to 6–7 or10. Things move, and it’s hard to go from 0 to 10—given our personality.
We’ve never done that, it’s not our style. We move gradually. It doesn’t have to be zero
When you fully exit a position, there’s the inertia of getting back in, and we want to avoid that inertia as well. And there’s the inertia of gradually increasing a position. So there’s also an element of respecting our own tendencies, we’re not optimizing short term we’re exposing ourselves, potentially, to some mark-to-market hits from market corrections, but believing that, over the holding period—and given the drawdowns that will occur—we can achieve a better IRR by carrying these assets, on the premise that we’re getting the qualitative thesis right: the qualitative analysis of how the business evolves over time.
{GABRIEL S} [02:35:55]: Technically—we run a kind of “impermanent impairment” exercise (not permanent impairment, since it isn’t earnings loss). For these names, we project earnings growth and then exit at a much lower multiple to see if the compounding still works. A basic example today - there are still businesses of this kind - is Amazon: EPS growth over the next five years — even with revenue growth being much lower — can still be in the range of 18% to 19% per year, plus some yield, in a very stable business. It trades ~30× earnings on ~4.5% margins. If you compound for 5 years and exit at, say, 20×—an even lower multiple—you’re still far from any problematic capital loss.
It will likely return the Brazilian real rate, selling the interest rate differential. So this is an important heuristic—actually, almost an art—for a sanity check. If a name surpasses some absolute valuation base, it becomes uninvestable. We don’t have much of that today in the portfolio. Out of epistemic humility and respect for business quality, we prefer to keep a capped position.
{DAVID K} [02:37:18]: The closest we got was Microsoft at one point last year—relative to peers it was above 35× while others were ~25×—and it became our smallest relative Microsoft position, under 5% again, as it is now. That’s it.
{GABRIEL S} [02:37:42]: That’s tactical. Structurally, once you find one of these businesses, you should have a big doubt about selling—bigger doubt than about holding. So we revisit constantly with sanity checks—but that’s the gist.
{David K} [02:38:04]: As Munger says, nobody makes money buying and selling—everyone makes money holding. So accept it and carry—because you’re usually paid for patience.
{GABRIEL S} [02:38:14]: And accept that some drawdown is inevitable, leaving room to rebalance. If correlations don’t go to 1; there will be room to take advantage. Last one.
Q&A: Twelfth Question
{Question} [02:38:30]: Which funds or investors do you consider your peers in the industry—both in Brazil and abroad?
{David K} [02:38:42]: In Brazil, finding a peer is hard—what we do is quite different. What we do have are people we admire. The obvious answer is Atmos—that’s where we came from; we learned a lot of what we know there.
The rest in Brazil, as investors, are quite different from us—because of the substrate, the types of businesses they look at, and the lack of self-imposed guardrails. As an investor, for being willing to look at everything, I think the one we respect is Atmos Abroad, Americans are less similar to us than it seems.
It’s everyone’s impulse—again, Brazilians dream of being American without understanding what that actually means. We find we’re much closer to Europeans as investors—by history, by where they came from, by how capital is allocated, by the relationship with the capital’s owner. When you look at how these businesses were built, many European managers were built on relationships with the capital owners—often families in Europe—whom they respect and feel accountable to.
Personally (not sure Gabriel agrees), the event most similar to what we’re trying to do here is by a Swiss investor, Robert Vinall. He inspired me to do this—he hosts a social gathering for his investors in Switzerland. We don’t live there; maybe someday we’ll do it in Itaipava or Campos do Jordão—meet in the middle, spend the weekend together.
Everyone gets to know each other—dinners, events—to create that sense of identification and to be able to build a group of investors who share the same values. We say this often, to everyone who asks: TB hasn’t actively sought new relationships for some time. We believe the relationships we already have—the people investing with us—are more than enough for any ambition we might have. What matters is building a liability base that understands what we do and wants to allocate capital.
And Europeans are much more capital allocators than Americans—they keep some distance from the day-to-day Investment Banking noise in New York. That detachment makes them real investors.
{GABRIEL S} [02:41:22]: The U.S. industry obviously has exceptional examples—but it’s also over-specialized. That’s why it’s so different from us. You have the TMT manager—the guy in Silicon Valley doing TMT.
And then he has to sell—he’s out selling the future, trying to convince you that the things he’s looking at are the best things that exist. It’s a market of over-specialization. Even East Coast folks—rare exceptions aside—tend to be sector specialists.The European manager—building on David’s point—and here we can go from someone like Robert Vinall, who truly is a generalist at the highest level and runs a one-man show, to TCI or Egerton—are people who take much more of an approach of looking at everything, analyzing, filtering, understanding where to specialize, but always with a much more genuinely generalist mindset, and a bias toward allocating capital regardless of jurisdiction or sector.
That’s much closer to us.
{David K} [02:42:30]: Many families in Brazil are only now beginning the journey of becoming global investors with their own capital. It’s a relatively new journey, with a lot to be learned along the way. Only someone who has invested in a fund and experienced a 50% drawdown truly knows what a drawdown is. And only those who have gone through it understand the level of pain and discomfort that a 50% drawdown brings.
And abroad, if you invest in a TMT fund, you’ll discover those drawdowns happen over time. We don’t minimize drawdowns, but we also don’t want 50% drawdowns. There’s a balance. Could it happen? Sure. But we try to build a vehicle that helps the investor earn the carry.
That’s it, right? We respected 12:30—not noon… 12:30 was what the email said, so we took poetic license to go until now.
Folks, thank you so much. This was great. I’m very happy we pulled it off. Thanks, Gabriel. Thanks, everyone. That’s it.
See you next year.