[00:00] [Rob Campbell] [RC] Here's some advice from the late Charlie Munger: go to bed every night a little wiser than when you woke up. On today's podcast, emerging markets portfolio manager Wen Quan Cheong on his passion for learning, and how both our team and the tools we're developing work together to help us be more productive, compress learning cycles, and reinvest that extra time into even more curiosity.

We end with the corners of the emerging world that have piqued Wen's interest the most: physical AI, low-earth-orbit satellites, battery technology and clean rooms, and something refreshingly low-tech — companies winning with customer obsession in markets dominated by incumbents who never had to try.

[00:40] [Disclaimer] This podcast is for informational purposes only. Information relating to investment approaches or individual investments should not be construed as advice or endorsement. Any views expressed in this podcast are based upon the information available at the time and are subject to change.

[00:57] [RC] Wen, welcome back to the podcast.

[00:59] [Wen Quan Cheong] [WC] Hey, Rob. Thanks for having me today.

[01:01] [RC] It's lovely seeing you. As listeners may know, Wen's based in Singapore, so we're doing this at odd time zones — at least this one, I think, a little bit more for Wen than for me. But Wen, today I wanted to walk through what investing is like in 2026. Of course, there's so much that's new and exciting, and I'm sure we'll talk about AI. 

But before we look externally, I wonder if we could start inside, with how our team works. You work most closely with our emerging markets team. Can you share what those dynamics look like in terms of making investment decisions, and how they've evolved with new technology?

[01:37] [WC] Yes, let me give you a flavour of the team dynamics. As a team, we're big believers in Ray Dalio’s concept of believability-weighted decision making.

At the heart of that is having the humility to say I don't have the answers all the time, and because of that, I'll need to learn to lean more on my teammates. But in order to do that well, we need to understand each person deeply: what are their genius areas? What are their blind spots? 

That includes understanding my own strengths and weaknesses too. When you know that clearly, you can be more intentional, you tend to lean on people where they're strongest, and you can be more measured in areas where they may not be as strong. That has always been our team's culture.

I really enjoyed a part of leadership which is modelled after our firm's culture, and more importantly, under Peter Lampert's leadership. This involves creating space for my teammates, whom I would call partners, to spend more time in their areas of strength, and to keep developing those strengths further. Investing is difficult, it's very competitive, lots of smart people out there. But I believe we have a very strong team with Peter, Josh, Shan, and myself.

I often draw the analogy of us being in the special forces. You have high-contributing individuals, each with different strengths and genius areas, and the only way we succeed is by leaning and depending on each other.
What makes it enjoyable is that I respect each of them as strong investors in their own right, and what I value is the diversity of thought they bring. There can be various instances where I don't agree with the investment idea, but we choose to go ahead with the recommendation, as I lean more on that particular partner's input.

[03:43] [RC] How do you choose to defer if you have great respect, as you said, for one of your colleagues in a particular knowledge area you think they're well versed in? How do you balance deferring to knowledge versus saying, no, I'm looking at this a little bit differently, I'm going to step in and challenge?

[03:59] [WC] My own personal process is to spend a lot of time voraciously learning about a subject first. I think the first step is to learn a lot about a subject and then build a case—be well prepared for the meeting, and that in itself will help start a very active and effective debate.

But ultimately, if I feel that, for example, if Josh or Shan have spent a lot more time than me on something and are very close to it, I think it's my role to try to find a blind spot through the initial research I've done. But at the end of the day, I might well choose to follow their recommendation, because for that particular instance I'd trust their research more than mine, given their closeness to the ground. But it's case-dependent at the end of the day.

[04:44] [RC] Can you give us a sense, just on the emerging markets team—you mentioned people have different strengths—one thing I've noticed about all of you is just the high degree of passion that you have.

[04:57] [WC] Yes, thanks, Rob, and that's a very valid point, because one common trait across our team is passion. All of us actually started investing at a young age.

Most of my partners started investing early in their twenties. For myself, I started in my late teens. And to keep growing in this industry, you have to have curiosity and a real hunger to keep learning.
I can speak a bit more about team dynamics. Maybe I'll start with Josh. Josh tends to be a bit more contrarian, and with that he brings very unique ideas. 

He pushes us constantly into areas we may not naturally look at: defence, capital-cycle themes, pharma, or other less obvious or even beaten-down areas. In my opinion, one of his strengths is just getting out there, meeting people, building relationships, and learning through good conversations.

One thing I've noticed about Josh is he's great at relationships because he's a very fun person to be around and have a conversation with, we always have a lot of fun chatting together. And through these relationships and conversations he has outside of work, he constantly brings back new insights and new ideas to the team.

As I mentioned, Josh is more contrarian, so even from a risk-management perspective, he tends to wear a black hat on the team and he's good at pointing out what could go wrong. And that perspective is very useful, because it keeps me in check. I tend to be a bit more on the yellow hat.

I can talk a bit about Shan as well. He joined last October, and he's a great addition to the team. Shan has this unique ability to quickly join the dots together and turn information into insight, which makes him very strong at idea generation, and also very good at drawing relationships and monitoring what's going on with our existing portfolio companies.

One thing I'd like to say about Shan is he has an incredible memory, almost like a walking encyclopedia. I often tell him, hey, do you have an embedded LLM in your brain? Because he goes through his internal database and comes up with very thoughtful insights. Perhaps because of this, he's an expert, in my opinion, in semiconductors, but he also has very strong breadth across all sectors and regions because he reads very widely. 

So I always see him as wearing the analytical hat, or what I'd call the grey hat—he's able to point out and see very clearly the positives and negatives and then provide a recommendation.

As for me, I enjoy making decisions. As I mentioned, I carry more of a yellow hat. I tend to look at where the opportunity is, where the insight is, and what the upside can be.

So I think as a team, we all balance each other very well.

[07:55] [RC] That's what stands out to me, just the balance, as you describe the various team members. Going back to the beginning, though, just to wrap up this topic, it's really about the dynamic, right? It's less about each individual. and how you come together.

[08:07] [WC] Yes, exactly. And above all, I think the most important thing is the culture, what we care about: being curiosity, openness, and candour, because good investing requires the consideration of different perspectives.

As decision makers, we have to see things through different lenses, genuinely consider the input of others, and remove ego from the process. By doing that, hopefully we can make better, more effective decisions, and in that process also reduce some of the blind spots we might have.

[08:41] [RC] When ChatGPT was released almost four years ago now—I guess about three and a half years ago—we've seen the technology evolve so rapidly. When we have other guests on the podcast—I recently recorded a conversation with Paul Moroz on the global side—he's constantly referencing you guys in Singapore as doing so much to push on how we can incorporate AI into our own process. 

I'm wondering if you can share in some detail what you're working on, what's working, and whether there are categories of AI use cases in our investment process that you're really excited about?

[09:15] [WC] Sure, I have tons to speak about that, so stop me if I talk too much. But, going back to first principles, I think what's really important, and I'm very proud of, is our firm culture because at Mawer we have a culture of constant improvement, and that really matters in the age of AI. Everything has risk, AI has risk, and we don't want to over-worry about some of these risks and completely stifle innovation. I've heard that happens at many other places.

So, I think the goal for us is to be thoughtful and measured, but keep experimenting, and that’s one of the things we do best here at Mawer. I think it's important, and this is my own story, to take a step back and think about what we want to achieve with AI.

After thinking for weeks and looking through my workflow, I came up with two broad categories. The first I'd term productivity gains. I call it learning compression. Let me start with productivity gains.

I spent time looking through my own workflow as an investor and then breaking it down into areas where I create more value versus areas where I create less value. For example, one area that's critical, but probably lower value from a human perspective, is forensic analysis. And the reason is because it is very important.

Forensic analysis protects us on the downside, but by itself it usually doesn't create the deepest insight or long-term alpha for us, but it's very much needed. What I've found is that with a well-structured checklist and a well-designed prompt, AI can actually run an even deeper forensic check across much longer time periods, and more efficiently than human input can.

Of course, I'm not saying human input is irrelevant: humans are the second layer of check and review. But AI can do a lot of the heavy lifting, and some of the forensics tools we've built can identify fraud as well.

[11:49] [RC] I can see the value of that. We have that step in our process. It's designed to make us slow down, consider what's in the numbers, and not fall in love with the narrative we might hear from management. For example, and this is a question I often get from clients, how important is it if you can have an AI agent do this forensic accounting analysis?

But to your point, you want to have the human review to make sure things weren't missed. How important is it to train the human, though. Meaning, if all you ever knew in your career was to hit a button and run the checklist, are you missing something by not having gone through and learned it yourself, by doing the checklist on your own?

[12:28] [WC] Yes. So what I'll say is that AI isn't a replacement for learning, it's a reinforcement for learning, it's learning on steroids. It doesn't mean you shortchange the learning process.
In fact, it should speed up and cram even more learning into a shorter amount of time. Effective learning, that's what I'd call it.

And I think that's really important. So to your point, I think the basics are still very important. If we're training new analysts, we shouldn't shortchange it, at times, maybe even do it manually, for the sake of learning.
But once you have a grasp of the basics, I think it's okay to automate, and then you become the higher-level check, doing more of the forensics analysis, having questions, and then investigating those questions. I think that's where the human layer comes in.

[13:21] [RC] I think that's the key on the productivity side. It's less about compiling the fact that the number of days receivable is getting longer. It's more about…that's strange, let me go investigate and push on that because I think this might be a sign of something about the business, something about the character of management, whatever it is. That's something to dig into and really spend your time on.

[13:40] [WC] Interestingly, the AI tool I built is able to even explain some of this: whether it's legitimate or whether something fishy is going on. It's pretty scary, but very interesting, and it's changing and evolving and getting better all the time.

What I'll say as well is that I think the most important thing, at the end of the day, in building all these AI productivity tools, is that it's not just about saving time for the sake of saving time. The real value is reinvesting that time into higher value-add areas. And what I consider higher value-add areas is reading out of curiosity. Just blind learning. Learning for its own sake.

Often when I do that, months or even years down the line, I join the dots and create a new insight. Going out, learning, spending more time on idea generation…that's a higher-value area. Turning over more stones, doing more scuttlebutt, meeting people, meeting experts, going to conferences, speaking to other credible, smart investors, and building those relationships. In my opinion, those all count as high value-add, and I'd spend more time on that.

So interestingly, when I look at my own calendar this year compared to last year, what I realize is that how I spend my time actually looks a lot different. I attend a lot more conferences, I join a lot more management calls, and every day I'm just aiming to be a learning machine. And Charlie Munger says as well, the aim is to go to bed smarter than you were the previous day.

[15:22] [RC] I like that, that's really good. So there’s aspects of the process that historically might have taken a lot of time. There wasn't value in the time itself; the value is more in the insights, so we're really pushing to be more efficient there. I like your idea that it's not just for time's sake, you've got to reinvest it into learning. It's the learning-compression part of it.

Can I ask about another part of our process that's long been a feature of both how we evaluate businesses and how we evaluate price, our Monte Carlo analysis that we embed into our discounted cash flow modelling. Is that still something useful?

Which category does that fall into? Is it one that's getting more efficient, or one we're doubling down on?

[16:04] [WC] Unfortunately, I haven't been able to fully automate the modelling of these yet. Over time, I think we'll get there. But for now, we still spend a considerable amount of time thinking through these scenarios. Let me take a step back and explain the purpose of the Monte Carlo process we use.

In short, it's an intellectual exercise in humility. We live in an uncertain world, and as investors we can have strong views, but we can never pretend we know exactly how the future will play out. 

So, the Monte Carlo process forces us to build scenarios around an investment case and ask: what are the different ways this thing can go very right or very wrong? It's not about creating a false sense of precision. It's not saying this is exactly what the stock is worth. It's about mapping out the range of possible outcomes.

Some people might call it the upside/downside scenario, plus the base case. We ask ourselves: what can happen in a base case, a bad case, a bull case, and more importantly, where does the skew lie today. Are we more skewed to the upside from the current share price, or the downside?

How does the risk/reward look? And that eventually feeds into portfolio sizing. If the range of outcomes is wide, but the upside is very compelling and the downside is manageable, that may justify a larger position, and vice versa.

Maybe I can talk about an example we have today with a very wide range of outcomes: the AI picks-and-shovels play, particularly the memory stocks. I think that's one of the most widely debated topics out there, especially given historical precedent.

The range of outcomes in memory can be very wide, because as I mentioned, it used to be a very cyclical industry. The key assumptions are things like ASPs, margins, and how long the cycle lasts. In the past, the range of outcomes on those factors was very wide, from substantial profitability to huge losses.

Given the environment we're in today where supply is quite constrained, the industry has consolidated, and we have a very strong demand environment charged by AI, we have to factor in that history often rhymes. I agree with that, but we also have to factor in the evolution of the industry.

This Monte Carlo approach helps us create scenario analysis to see what happens if the current demand and supply dynamics remain for longer. As you see with all the LTAs signed today, and the very constrained clean-room supply coming in, we might see that cycle dragged out past 2028 and that would mean the bull case could look very attractive.

We can also simulate a scenario where the cycle turns earlier. If you look at the consensus estimates out there, 2028 is the year everyone thinks everything would collapse and margins would normalize. We can simulate that scenario too.

But at the end of the day, we ask ourselves: does the reward outweigh the risk? And that feeds into how we manage the position and size it in the portfolio. I also believe a good Monte Carlo process requires a lot of curiosity, openness, and candour.

It's like an exercise in character, we also need people to challenge the assumptions. When working on these memory names, you need people to ask, what if we're wrong? You need teammates to look at the company through different lenses and provide input, and then you have to actively consider that input.

So it's not just a modelling exercise that happens individually; it's very much a team conversation.

[20:30] [RC] I can see, from our earlier discussion, how having the various people we have—not just on our emerging markets team, but across the firm—is important, since a number of people clearly have interest in this particular topic, though we're just using it as an example. I liked, where are we wrong, but also where might we be right? Or where might we be underemphasizing what things could happen and having a framework for that.

I always love asking you: what are some of the big themes you're following as an investor in emerging markets? I think last time we spoke I asked something similar, so I'd assume there's some overlap. We talked a lot about AI last time too. But sitting here at the end of June, what are you looking at?

Maybe some things that are more obvious, and some that are less obvious.

[21:12] [WC] Yes. A statement upfront: we're first-principles investors first. We don't just chase themes for the sake of chasing themes; it's a process of finding wealth-creating companies with good management teams, bought at attractive valuations. But at the same time, personally, I do like themes, because I like finding high-quality businesses supported by very long-term structural tailwinds, which allows them to redeploy capital at high rates of return.

I think that's the most important part I'd emphasize. With themes, the profit pools are also expanding, especially in a blue-ocean market where the upside is still hard to size. The opportunity set can be attractive, and the competitive landscape may not be fully formed.

So it can also be a lot less competitive. Fortunately for us, some of the themes we participate in are quite consolidated, with very high barriers to entry in most cases. And you rightly pointed out that we had this conversation a couple of months back.

I'd say nothing has really changed; the themes we listed then are still the big themes today. AI is obviously the big theme we've already spent time on.

So we're spending more time on other areas now. But what's changed since then is that valuations have run up quite a lot in most cases. Within that whole bucket of AI, though, you can't paint everything with one brush.

Some areas are obviously expensive, with a lot priced in; some areas are still reasonably priced. So, obviously, even great businesses and great investments can be poor investments at the wrong price.

We're still doing a lot of work here, continuing to monitor it closely and make sure our thesis still holds. But within AI, one interesting theme we've actually been spending more time on, out of curiosity, is physical AI. I think this could, one day—not now, but maybe in 5-10 years—be a very important structural theme.

It's useful to spend some time investigating this theme as well.

[23:30] [RC] Maybe to start…what is physical AI?

[23:32] [WC] Basically, humanoids, a lot of automation, drones, etc.

[23:39] [RC] Robotics.

[23:40] [WC] Yes, robotics, basically. You're right. Mostly used for industrial applications today.

There are limited use cases for them now, but who knows what they'll be used for. I think this will be big, especially as a lot of reshoring happens in the U.S., and in order to make the cost structure work, maybe that could force the industry to develop a lot quicker as well. So that's one area: physical AI.

Another area I'm interested in is low-earth-orbit satellites, or LEO for short.

[24:11] [RC] This is because of the SpaceX IPO?

[25:14] [WC] Yes, but I also think this connectivity and communications infrastructure (space-based networks) will be very important over time, especially in markets where traditional infrastructure is a lot harder to build.

There's a lot of interest and a lot of capital being deployed to invest in these areas, to build national champions, if I can call it that. In scanning these themes, we've identified a couple of interesting ideas with potentially free optionality within the physical AI and LEO segments that are still fairly priced today. Interesting opportunities.

We're constantly on the lookout, finding ideas in these areas as well.

[24:59] [RC] What I'm curious about, Wen, is with the situation in the Middle East, I imagine many companies and geographies are reconsidering energy security. Has that popped up as a theme for you?

[25:11] [WC] Yes. Energy storage, especially for data centers and renewables, has become even more critical in light of what you mentioned around energy security. Penetration rates are also being driven by improvements in battery cost, technology and energy density, which has created interesting opportunities.

In many areas they're not very attractively priced, but there will be spots of opportunity, even among large companies. We do own one of the industry-leading battery makers in our portfolio as well.

[25:50] [RC] You mentioned companies reshoring earlier, something we've heard about for some time, probably the last 5-6 years, with an acceleration in that rhetoric probably around a year ago. I'm curious how that's evolved. How far along the path of reshoring or reshaping supply chains are we?

If it's a baseball game, what inning are we in? We've heard a lot about "China plus one" as a strategy. How far along are companies on that front?

[26:17] [WC] Yes. Companies, to your point, are rethinking where they should manufacture, how they diversify supply chains, and how they manage this geopolitical risk. In my opinion, a lot of the thinking was done last year.

There were a lot of attractive bills in the U.S. aimed at bringing manufacturing back locally, either through punitive tariffs or dangling carrots like tax incentives, and moving production out of China. But many countries are also trying to build more local, resilient supply chains in strategic industries, whether in the U.S. or Europe.

You can see this in areas like semiconductors, batteries, even data centers, pharmaceuticals, and defence, because governments increasingly want these critical capabilities closer to home, or at least within trusted, friendly supply chains. That creates a lot of opportunity across various parts of the value chain—industrial parks, ports, and even specialized engineering services.

To give an example, we own a company called CTP, a vertically integrated developer and owner of industrial parks. This company is a Philippine-domiciled port terminal company. They acquire port concession rights around the world and operate them to world-class standards. These companies benefit from the shift in supply-chain and trade-flow infrastructure I mentioned earlier.

I've also mentioned before that we own a company called Acter Group. They have an oligopolistic market position in the clean-room engineering consulting space, which is very interesting because it benefits from the global clean-room shortage we're seeing now, and the broader build-up of clean-room capacity, especially in the U.S.

For a lot of these manufacturing industries, the clean room is the bread and butter of the capability required to manufacture high-end or sensitive products.

[28:28] [RC] Yeah, you've got to get it right. Beyond these bigger themes that get attention in the news or in papers, are there other themes you're looking at that might appear idiosyncratic to a casual viewer, but might be connected…a bit more bottom-up in nature?

[28:45] [WC] Yes. In my opinion, the key theme of EM, the bread and butter of EM, is finding companies that are customer-obsessed, with a shared-economies-of-scale principle. In emerging markets, many industries have historically been dominated by oligopolistic SOEs or politically connected private enterprises that didn't always have to be truly customer-obsessed to win, i.e., they could still get away with bad service and overcharge customers.

So it gets very interesting when some of these markets open up to competition. Great management with good execution can build a business that operates more efficiently, focuses on the customer, and shares the cost savings through lower prices. We like companies like that, and they tend to have a few common traits.

They have managers who act with integrity. They truly obsess over customers. And, as I mentioned, they combine all this with the ability to put profit second and the customer experience first, making things cheaper and more accessible.

The combination of these three traits is very powerful, and these companies tend to become noticeable, strong players in the market over time.

[30:19] [RC] Especially, as you say, if they're competing against legacy, entrenched, maybe less-aligned businesses. I can imagine that gives them a pretty strong advantage.

[30:30] [WC] Yes, and we do own several companies in that category, some in the finance industry, like NU Holdings, Bajaj Finance, and HDFC Bank. They're all examples of these models we look at.
There are a lot of other ideas we're looking at, and hopefully in time we'll own more of these types of companies as well.

[30:54] [RC] Trying to wrap up what we've talked about today, maybe let's go backwards this time. There are a lot of themes out there in the emerging markets world, some very exciting, some worth watching. But to your point, as first principles for us as investors, there may be themes out there we need to be aware of, and in some cases follow for idea generation, but it's really about the individual businesses.

AI is a bigger and bigger part of what enables us not only to find ideas, but to scale up the learning compression, getting more efficient in how we do things. But maybe going back to the start, underpinning all of that is the team we have. One that can take advantage of this, that knows how to work together, and understands how to challenge each other in a challenging world. Any last comments before we wrap up, Wen?

[31:40] [WC] I spent very little time talking about the very interesting tools we've built. Maybe I'll keep that for next time. There are tools like the idea-generation tool we have, and the knowledge base, which is kind of a pet project of mine. Those are really interesting projects.

I'm actually using AI to help me stay constantly monitoring and updated on the latest with the platform. This knowledge base has been a very useful tool to help me continually build my expertise on companies over time. There's just so much to know.

[32:26] [RC] I'm glad you've got Shan on your team, the walking LLM. Speaking of Shan, I think we're going to have him on in a little while to dive specifically into memory, given how important that's been, both in terms of what's been driving markets and some of the positions in our portfolio. 

So, Wen, thank you so much. Sounds like we'll have you on again. I appreciate your time. Thank you.

[32:46] [WC] I'd say you'll be in for a treat with Shan talking about memory, so tune in for that podcast, and take care, everyone.

[32:55] [RC] Hi, everyone. Rob here again. To subscribe to the Art of Boring podcast, go to mawer.com — that's M-A-W-E-R dot com, forward slash podcast — or wherever you download your podcasts. If you enjoyed this episode, please leave a review on iTunes, which will help more people discover the Be Boring, Make Money philosophy. Thanks for listening.

Companies Mentioned
CTP
Acter Group
NU Holdings
Bajaj Finance
HDFC Bank