Episode Transcript
The Finance Ghost: Welcome to episode 264 of Magic Markets. It's great to have you here with us as usual. And as you've noticed in the past couple of weeks, we are doing something pretty cool this year, where we are introducing you to some local boutique fund managers who have some excellent thinking, some really cool ideas, and are here to share them with you as the Magic Markets audience.
And I'll introduce our guest this week shortly. Let me first say hello to Moe. Moe, welcome as always, and excited to do another one of these with you. Episode number 264, no less. We've done a few of these.
Mohammed Nalla: Indeed, and it's why I'm so excited to have some new fresh voices on Magic Markets, because I think that the specialist boutique fund manager niche in South Africa is definitely undertapped.
I mean, we've got some really smart people in that space, and for us to be able to glean their insights and share that with our listeners, that's really getting me quite excited.
So, Ghost, why don't you introduce our guest this week and let's jump straight into it?
The Finance Ghost: What Moe's trying to say is he's bored of me, basically, but he's not going to get bored so easily of David Eborall, who is a smart guy.
David, you are the founder, fund manager, chief capital allocator - take your pick - at SaltLight Capital Management, and you do some very cool stuff. You write some great letters, you manage some proper money, and you do it in a way that I think is quite unique actually, in the South African market.
So, welcome to this. Thank you for joining us here on Magic Markets. Thank you for valuing this audience and bringing them your insights and your thinking.
David Eborall: Yeah, guys, thank you so much for having me here. It's really an honour to come and chat to you. I must be just upfront with your listeners that I'm not the greatest salesman, I'm not the greatest podcaster. You guys sound brilliant on your show. I'm just a roll-up-my-sleeves kinda guy, trying to find the best opportunities.
The Finance Ghost: Not the greatest salesman, he says, and then go straight into flattery. I see what is happening here. Well, I'll tell you what, David, Speaking of flattery, 15.4% annualised return since inception five years ago. That is pretty good. Well done.
And of course, to keep you humble, you've just had your first negative quarterly return since 2022. Anyone who has managed any money in the market or invested will understand that that is a feature, not a bug, in this world. So we will obviously talk about that.
But I think before we get into any of those details, just give us the backstory to SaltLight Capital, to your background as the fund manager. And while you're at it, give us an explanation of the name as well - because “SaltLight Capital” is rather unusual.
David Eborall: I call myself an entrepreneur who happens to run a fund. And I've been - I founded my first business in high school. I was very fortunate, I'm showing my age here, but that was the internet days.
I started doing internet sites, which is very trivial now (you can just code it in some kind of AI agent). But at the time I was selling websites to corporates and that was my first entrepreneurial venture.
I've been very fortunate with three things.
Obviously, the coding background has helped. I started very young to pick up a few books. It was very hard back in the 80’s to get coding books as you can get today. That has helped to also add a lot of research ability for this fund.
And then the last is, I worked at RMB and Morgan Stanley. RMB was really interesting. They had this program called the Class Of Programme, which is started by the founders of RMB. And I got to see a whole bunch of different asset classes - all little ingredients and things I've taken to, to put in the fund.
So the fund started 10 years ago. The first five years, every entrepreneur makes a mistake in their business. I'm sure you've experienced it. You make every mistake in the world.
I was in the wrong country. It was the difficult years in South Africa, had the wrong fund structure and it was a difficult pool of capital to capital raise in. And it was also not purely our skill set - South Africa doesn't have a lot of technology opportunities beyond the typical one or two.
And so when we went to the global fund, that was really where everything took off. The right capital pool, right structure and the right opportunity sets. And obviously we'll talk a little bit about AI, which we started investing in way before ChatGPT came in.
Obviously after that, it's been this major opportunity set.
Mohammed Nalla: David, that's such a fascinating journey. I mean, I love these stories of someone that goes through the path less trod. That's really what you've indicated to us.
What I want to kind of get into - you've indicated a whole bunch of stuff here. But I recently read your latest newsletter. And something else that stuck out for me is that Saltlight deliberately avoids fitting into those neat little tick boxes. Like a growth manager or a value manager. We're going to talk about tech, but you don't explicitly say we're focusing on tech, we're a tech fund.
Now why this is interesting for me is because most institutional investors out there still build their architectures around those buckets. It's how they frame their thinking, it's how they frame their allocation.
So what exactly led you to abandon that structure, that approach and this move towards what you describe as a more “opportunity maximalist approach”?
And maybe even just a step back from that, because I don't think our listeners will be familiar with this. Maybe explain to us what an opportunity maximalist approach is, that defines SaltLight Capital?
David Eborall: The history of it was those difficult first five years. No matter how smart you were in domestic South Africa, it was very hard to generate above-average returns. I think being limited to one asset class, one market, and then obviously a particular fund type, it just limits you.
When we had the second fund, we completely inverted it. And so it's: go anywhere, different asset classes, any geography. We can go into - believe it or not, some funds can't go into all cash. We can go into all cash if we need to.
So it was really just to have that maximum flexibility to go where the opportunities are within of course, the regulatory framework. That is the maximum permissible we're allowed.
I think I agree with you in terms of a lot of big institutional funds. They want a product, they want a sliver of a niche - they want to allocate the assets to bonds or equities. And that's their decision level.
My retort always is that a lot of investors want their fund managers in the funds, their own capital in those funds, I think it's just very limiting. Some particular tech or asset or bonds or equities will work in different regimes and different markets and different periods. And five years can be a very long time in a fund manager's life.
And so having the ability to move around and have that flexibility and not be mandate restricted, although we've tended to stick with a lot of technology and eCommerce and new opportunities, we can go into those others.
The Finance Ghost: Yeah, it's like being a great footballer, and you're forced to just play for a bad club for your entire life, right? Sometimes you just need to go somewhere else. This team is never going to win. Don't be there. Go play somewhere else.
And that's actually how most individuals would treat their money. I've never in my life met an individual investor who says, “Oh, you know what, this year I'm absolutely only buying stocks in this market, even though it sucks.”
That's not how people do it. But for some reason that's how this entire fund management industry seems to work. And I understand from a capital allocation perspective, the argument is, well, the capital allocator tweaks the allocation. Your job is to stick to a mandate.
But I like your approach and I can understand as an entrepreneur why it resonates, because you need to go where you see the opportunity. That's how your brain is wired.
And that means that obviously you've got some interesting and different things in the portfolio. But I think what does set you apart from a lot of your peers and makes you quite interesting is that there's a really strong technical expertise underpinning this
I can say, as active as I am on X locally, and so are you, something that's always really stood out for me about your approach is I've seen you have these deep conversations where you are really, really skilled from an IT or a technical perspective.
For example, you really understand these platforms, the AI world, everything else. And obviously we'll get to this stuff still. But I think that your research process is something that is worth actually digging into, because you do the work, as people say.
And people use that term very vaguely and there are lots of different ways to do it. People who have great feel in the market, for example, can also do really well. There are many, many ways to make it work.
But my read on you is you do the work. Is that a fair assessment of your research process?
David Eborall: Yeah, thank you. I really appreciate that. I think, trying to know - for us as a boutique fund manager, with 1,300 other funds in this country across different asset classes and different funds, to be one out of that bunch, you have to obviously have a very differentiated and deep research process to at least find alpha.
My view is always that obviously there are US funds which we compete with, but to compete in that market you're obviously fighting in the premier league and you need to have a very good research process to understand.
AI has been one, very complex opportunity set, which is quite different to something like software or eCommerce or mobile, or social commerce, or social media. These were fairly easily understandable opportunity sets, but the nuance and the alpha is sometimes in the technicalities.
So we’re fortunate, having a bit of the coding experience, in two parts.
One, because we can understand, let's say other research or when companies talk about it, but the other is just rebuilding it for ourselves. We've used agents in the business probably a year and a half now in some of the research tasks. Obviously not making investment decisions but doing a lot of junior analyst work, going through transcripts, going through filings, just giving us a quick perspective.
In a couple of years that might be quite table-stakes, but for now it really helps us and we probably save one or two people. We've also got them involved in some of the operations. So we’re incredibly efficient in running our business, and I think the businesses of the future will have a lot of these capabilities.
And then by building you get to understand and learn the products. So we've tried every cloud out there: cloud products, every hyperscaler. And I can probably tell you what are the pros and cons of each, and where they'll fit.
And that just gives us a bit of a lead, compared to some others where you're relying on sell-side research or some kind of products to be announced, etc.
So, (1) the AI is complex, and (2) we’re actually using it in our firm - just really I think a good base to kind of navigate.
Now we may not be always in AI, I think AI is the big opportunity now, but of course there are lots of tangential opportunities within that as well.
Mohammed Nalla: Ya David, I mean I think that's fascinating in that it's very much a hands-on experience. I think the shortened growth curve for smaller companies to be able to leverage AI, build out their teams or their processes. That's really quite interesting.
But I want to maybe segue a little bit away from that - because that's around use of AI, your own experience, your coding journey - into the AI investment thesis. Everyone's talking about AI but I think you take it a step further - you've described it - I'm going to use your words here: “AI epoch”. Right?
And is that just not another way of describing a long megatrend? I always use that word, megatrend. Is this really the same thing or do you think there's something structurally different that's happening right now?
And I want to then extend that a little bit, if you'll permit me, is that if that framing is correct, if this is in fact something structural, it's different. We often talk about shovels in the gold rush here on Magic Markets. So, if it’s shovels in the gold rush and if that's the thinking around it, where in this AI ecosystem do you think investors are most likely to make money? And where might the market actually be getting the story wrong?
Because there's a lot of narrative out there, there's a lot of hype, and it takes a lot of work to sift through all of that noise. What's your view on where those opportunities lie, and is this structurally different?
David Eborall: I think a bit of background. Over a decade we've been working with machine learning. And obviously in 2021, when you first initially invested in it, the hard part was understanding how this technology was going to diffuse across society. Because it was in the purview of very highly paid research scientists.
And then of course, ChatGPT came out, which was distilling into something everybody could use. And that was the “aha” moment, so that this could diffuse across society.
Now this kind of subject of diffusion. A book that we read a couple years ago that really just changed our mind on this, was a book by Carlota Perez, who wrote this book called Technological Revolutions and Financial Capital.
She studied different epochs, so it’s called it an epoch - and why that's important is because some technologies are like the fridge or the microwave. It solves a small problem like making your food cook faster, or something like that.
Epochs are major technological changes that change all parts of society: regulatory, economic, social. And you can think of these as the railways, right, which made travel much cheaper. Horses went out the window. And you had the ability to start towns and cities away from the main urban areas.
Then, the last one she calls the information age, which she says started in the early 1970’s. So it wasn't the internet, it was this information age of digitisation and taking physical things into digital bits. And within that there are multiple different technologies, not just the internet. And she breaks this epoch into five to six decades. A 60-year opportunity set, broken up into two phases.
The first phase is an eruption phase where this early technology like the railway or the mainframe slowly comes in, and people are like, “Wow, this is incredible. We can digitise information”. And that lasts about two to three decades.
And then you get this turning point. In those first three decades, obviously society hasn't quite adapted to this new technology. It's very slow to work through. And then there's this turning point when suddenly, society reconfigures itself.
So, if you go to the information age, the internet was that - right around the dot com crash, when Amazon didn't work, Webvan.com didn't work. You needed this diffusion period where you had the last mile, you had a lot of the technology and of course then the mobile phone which could diffuse right across the world, where everyone's got 1.2 phones.
So that's the first phase.
The second phase is this phase where it gets deployed across society and suddenly then the new technology - now you have the mobile phone using digitised products and it's diffusing across society.
This book she wrote really widened the opportunity set for us. And so if you take this back, if we are an AI epoch, the six-decade opportunity, there's not just going to be a chatbot, it's not going to just be a few - solving some research. It's robots, it's self-driving cars, it's all these subsequent things. And you think of the information age was the fax machine, the PC, it was the internet, the modem.
And so all of these are investable opportunities in a greater epoch. And so we see it as not just AI, not just the chatbot, not just the hyperscalers, but there's going to be multiple opportunity sets and different things.
Now because there's so many and they're so long, we don't have to pick one thing - just the AI infrastructure trade. There are going to be so many in the next couple of decades, to actually make money as investors. And this is - at least from my age - how I think.
Before I get disintermediated by robots, how I hope to make my wealth, and hopefully my clients as well!
The Finance Ghost: Yeah David, as you say, no one's really sure about the future. This intermediation, I mean it's tongue-in-cheek - who knows what happens? I sometimes go down a road where I think about it, then I want to stop thinking about it. I just don't know.
I did laugh about your 1.2 phones because my spare phone is 0.2 of a phone. It's very broken. Perhaps my experience with AI will be a bit better.
But you're right, it's diffusion, it's technology that just becomes everywhere, essentially. And I think what's interesting, and that's maybe a point I want to just bring back. Earlier you mentioned you're using agents in your business. You're using them for research, you're using them for ops. So there's a real world use-case where you are experiencing it every day and you are essentially a boutique.
So obviously as a small business owner, which is what you are, you're looking for those opportunities to optimise. And in your view this is saving you money, it's making the people you have more efficient. That's clearly your take on it, right? Because I can tell, you believe in this technology. That's your lived experience, right?
David Eborall: Yes, I think definitely. Look, whether the economics are there, I think they're obviously subsidising what we - the tokens we're using and the value we’re getting heavily now.
In the same breath, I can also say that obviously the technology and cost of inference is going to get massively cheaper. And so I think that's the bet right now.
But I think the gap between the cost of a human resource and an agent is still very wide. So even if they doubled the costs, you would still do it.
I think the hard part now is the orchestration. We've had to self-write a lot of this infrastructure, incorporate it in our processes, and of course the data - I think that is the biggest bottleneck for most small, even medium or large businesses is that the data is not in a good format which agents can read. That's going to be the first thing they have to get right.
And then of course being able to orchestrate it all because these things do hallucinate, they do go off a tangent and so you don't want them operating your nukes yet. But I think for various low-value tasks, they can definitely take over those.
Mohammed Nalla: I think for anyone who's actually used these agents, used AI - and not only do they hallucinate, I showed my daughter this example where we asked ChatGPT how many “r” letters there are in strawberry and it got that wrong. I think we’re way off in terms of just being able to trust AI. But like you say, in these epochs, they run for a long period of time and we're really just starting what you can call the AI epoch.
I want to maybe just take a step back, because I can clearly see your passion for the subject coming through here and it's really quite fascinating.
One question, as you mentioned, diffusion, right? And what's changed for me, if we compare this to the information age that maybe started in the 1970s - versus the pace of technological change today - do you think that diffusion actually accelerates? Do you think we have 30 years worth of this epoch? Or do you think diffusion is happening so much faster that it might shorten the length of time within the epoch?
And the reason I ask this is because again, in your letter you discussed three contradictions that the market is trying to price. One is that the capex signals in the supply chain are accelerating (that maybe talks to the fact that we're early stage on the epoch). But then the market's also skeptical of Nvidia's forecasts. And then we had that big wobbly on SaaS stocks about two or three weeks ago, seeing that massive compression coming through because they think that that disruption is quite inevitable.
So how do we reconcile those contradictions and what do you think around the pace of that disruption and the epoch over the next couple of years?
David Eborall: I think the first point, back to Carlota Perez. The eruption phase - the first three decades, generally, there's never profits, right? It's in investment mode. The book is largely an interplay between the technology and financial markets which need to finance these build-outs. And so there's quite a good pattern of - everyone will bring up the railway examples where x percentage of the UK's GDP was invested in railways, etc.
I think it's all dependent on how quickly it diffuses. That diffusion question you mentioned there is a big one. Currently, if you think about how we consume AI, it's pretty easy on a little chatbot and it's very easily consumable, but the intelligence doesn't carry around.
Obviously we can't do it over clouds and, sitting in a server in the US where a robot's going to have to operate and make multiple decisions in a short space of time. We need to bring that compute closer and closer to where the work needs to be done. That is still a long way off.
And there's a number of bottlenecks - robots to be in your household is probably a decade away, just because they can't compute in your household to do all those tasks.
I think the market's trying to figure out how this all plays out. One, the ROI question, which is a favourite, I think, of a lot of investors. But in the same way they're saying that SaaS is going to sell-off and AI is going to disrupt it. I don't always know if those two things can be true - that it's going to be so successful in software and yet there's no ROI in the hyperscalers.
And the biggest probably known factors are that demand is very high and computers very short. If you're protecting your investment space in this, you want to build your moat within a very fluid platform space. And I think all the hyperscalers are investing a lot in that.
I think from our work, what we see, and I think probably people don't realise is, if you go back to the supply chain, the hyperscaler is probably step four of five or six. And if you start looking at the companies in step one of the supply chain, they start building for 2028, 2029. And if they suddenly, last quarter, started to signal that they’re going to invest a lot more, then you’re starting to look at additional capex in 2028, 2029 coming through.
And I don't think the market's quite reconciled that these numbers are going to get even bigger. I think a lot of people are forecasting a bit of a stabilisation period as these big numbers of $600 billion or $400 billion over the last few years.
But for someone who's followed this over the last couple of years, even mentioning $200 billion capex would have been a crazy idea in 2022. We’ll probably adjust to a trillion at some point in the next two or three years.
And I think this is just something everyone needs to get used to. It's not going to be profits. We haven't got to the Google and Facebook era when all those super profits were made, the greatest companies in that era. We still need to get there.
The Finance Ghost: Yup. That's a mistake that many of us have made. I made that mistake with Netflix. I thought, oh, the economics are going to suck forever. Bad error. I rectified it recently and thanks to all the distortions around the Warner Bros. Discovery deal, I actually got into Netflix at a nice price, which I'm happy about. So that's cool.
You've mentioned SaaS. I don't know what it is - like SaaS, Sasol - it's the most volatile syllable around! Basically, if it starts with “sass”, you can hurt yourself in the markets and you're going to have to be scared.
But really what it comes down to is a concept in your most recent letter that I think is worth digging into, which is called Expected Value Investing.
Now, I had this wonderful boss in my investment banking days. He always said to me that we know that our models and forecasts are not going to play out as expected, but the key is really to make sure that the process is right, and so is the maths.
I had another boss in those days who said to me, since money was invented, there's only one way to say thank you. So that gave you some idea of the culture that I spent a few years in, which was great. You'll be familiar with that world, definitely.
But the point is that the forecast cannot possibly be right. That's the point. There's no way of getting it 100% correct. The most you're ever aiming for is a directionally accurate range of outcomes that are more or less in the right place.
Is that essentially what you're talking about when you write about Expected Value Investing? Because it seems like this approach really underpins so much of your capital allocation process.
David Eborall: I think one of the big lessons over the years of being in the markets is that trying to make exact predictions of outcomes is a futile proposition. Simply because a financial market is a complex adaptive system where in the unit, one human being, their biases will act, maybe irrationally, irrationally, but as a whole it can act rationally as a collective. And trying to predict that is very difficult.
And so what I've learned over the years is - the share price is just a summary of all expectations of the future, weighted by probabilities. And within that time period, let's say we're having a five-year time period of 2031, as we look today, what could the world look like in 2031?
I don't know exactly what it looked like, but we can start to make some hypotheses about what certain factors would be there. I certainly believe that compute will continue. Some things will change and some things will massively not change.
And so what we're trying to rather do is use the price as a way to back out what each of these expected outcomes will be. There's no science to it, it's just an intuitive - you can't get the exact, that it should be 52% or 56%. But if we can build a portfolio of these ranges of outcomes and let the portfolio handle circumstances as they come, and have the ability for it to handle very bad scenarios, the five-standard-deviation event, or just good times.
Then, as an investor, we can mentally manage it. Because, okay, I don't need to be right all the time. I just need to make sure that I've got the right expected value in my portfolio, sized appropriately for the risks.
So obviously a very highly priced company, you need to have only the good stuff happen in a very good measure, or beyond that. If you've got a fairly priced business, then you've got a higher range of outcomes which could come and obviously a very positive event or middling events could happen, and you're still doing pretty well on a return-on-capital basis.
And so I think for me is freeing your mind and not trying to be right all the time has been, just (1) helps you sleep at night, but (2) just also gives you the tools: when to sell, when to say, look, maybe the odds are not great, the market's expecting too much here. Let me just downweight the position or upweight the position if there's something that's got it wrong.
It's an idea I think poker players have got very well. I don't go gamble at the casinos, but poker players work with positioning, work with money management and these expected range of outcomes depending on what they hold. And so your cards or your companies, and their moats and what they can do in their market. And so you just need to play those outcomes depending how that company could react.
Mohammed Nalla: I love that. I've seen certain concepts out there, like Probabilistic-Based Investing. And for me this Expected Value Investing is really along that kind of theme. I quite like it because throughout my career, certainly in the institutional space as well, we were quite fixated on scenario planning.
And I think that points to your whole thinking, is that no one's going to get everything right all the time. But if you actually build your portfolio so that it's robust throughout various scenarios, that's really what you're trying to achieve.
Unfortunately David, I think we've run out of time and I still have a long list of questions here. We can probably get into that in future shows. But I want to ask you one question. In terms of operationally, the business, does SaltLight Captial only target institutional clients? Are you targeting retail clients? What is your kind of client mix?
And how can our listeners, if they are interested in your ideas - because I've been fascinated by the depth of your thinking in this tech stack. You've certainly given me some insights that I don't have from a technical perspective.
If our listeners are interested in reaching out to you, how can they actually contact you? Where can they find SaltLight Capital? And what types of investors are you actually taking on at this point in time?
David Eborall: Most of our investors are ultra-high-net-worth investors who are ex-entrepreneurs themselves who get the opportunistic process and mindset.
We are available on EasyEquities, which is one of our portfolio companies. We've been there for many, many years and backed the team there, and so it is available for retail investors there.
We’re on Momentum or Allan Gray. Otherwise you can invest directly. If you want to find out more, I think our letters, as you’ve mentioned - you guys have read my letter, and compliance - I really appreciate it. The letters are all available on our website. It'll give a good flavour of how we think.
So saltlightcapital.com. Anytime, just give us an email if you want to find out more information. We're a 10-year investment horizon business so we're not going to be a short term play, but if you like the opportunistic type of investor, we can look after your capital.
The Finance Ghost: Yeah, absolutely. If I can maybe say, I've introduced David to a friend of mine who ticks that box of a really good entrepreneur, very successful, understands capital allocation. He was super impressed to the point where he's invested in the fund. So I think that says something.
David Eborall: Thank you for that.
The Finance Ghost: And David, I just enjoy the way you approach it. Expected Value Investing is how I run my entire life, really. It's why I love playing strategy games like Warhammer.
It's the same - it's like, what are the possible outcomes here? Okay, I don't know which one it's going to be, but if I'm just like more or less in the right direction, things are going to work out okay.
So very, very cool. If anyone's wondered where you are in South Africa, I'm not sure all the hadidah noises in the background are going to make it out of the edit. You are in Joburg. We'll leave the hadidahs in. Maybe that's what you should name your agent, actually!
David, we're looking forward to doing more of these with you this year. We've got loads to ask you - clearly - to understand what's in your portfolio, why you own it.
But I would encourage listeners, don't wait for that episode. If you like what you've heard, reach out to David. He wants to talk about what's in the portfolio (you'll probably have to stop him). And read his letter - because he's been very modest about how many people read it. It's actually very good.
So, David, thank you so much for doing this. First of several, or many, and we look forward to having you back.
David Eborall: Great, guys, thank you so much. Really, really enjoyed this.
Mohammed Nalla: David, again, thank you so much from the Magic Markets team. I really enjoyed this chat. And again my big gripe is that we just don't have enough time!
To our listeners, let us know what you thought of the show.
Hit us up on social media. It’s @MagicMarketsPod, @FinanceGhost and @MoehammedNalla, all on X. Or you can find us on LinkedIn. Pop us a note on there.
Mohammed Nalla: Or you can find David. We definitely have his X handle in the transcript to the show as well. So go and follow him. Certainly some fascinating insights coming down the line there and we hope you've enjoyed this.
Until next week, same time, same place. Thanks and cheers.
SaltLight Capital Management (Pty) Ltd is an Authorised Financial Services Provider, FSP number 48286. Past performance is not an indication of future performance. Please always speak to your financial advisor when making any investment decisions.
