1 August 2023
Close your eyes and allow your mind to gently recount the melody of your favourite song. Visualise the graceful dance of each note, the emotional rise and fall of the chorus, and the commanding beats that set the rhythm. When this ensemble works together in magical ways, time can be lost to great music. This intricate creation transcends the mere arrangement of composition; it’s a display of the power of musical alchemy.
Envision now the artist, the architect sculpting this musical masterpiece. Every lyric, beat, tempo, and sound is meticulously interwoven, forming an emotional tapestry that is as perplexing as it is beautiful. Each note and word create an unseen bridge between the artist’s talents and your mind. Tugging at your heartstrings, igniting your thoughts, and awakening your emotions. THIS is one of the most remarkable products of humanity.
Every musical artist aspires to those moments of connection with a listener. Just one moment of the glimpse that you just had. It’s why they get up in the morning.
Curating Opportunities, Weighing Odds
At SaltLight, we are motivated by the same creative passion. We pour our hearts into our work, viewing it not as a mere vocation, but as an act of creation. Instead of melodies and rhythms, our composition unfolds on a different stage. We see ourselves as treasure hunters, sifting through a vast terrain, discovering hidden opportunities, and appraising their potential for reward. Over time, as we collect these coarse rocks, we observe, adjust, and refine our approach.
Our singular goal is laser-focused on one event: the grand opening exhibit set for the year 2028. This is when we present our Hope, Cullinan and Kohinoor diamonds. The economic ensemble that we have assembled: our anticipated opportunities, the innovative products and services that have changed customers’ lives; the problems that were once impossible but now solved.
Most artists work in hidden, poorly lit studios until their creations are revealed to the world. We share the journey of this curation through our letters to you – our co-investors. These letters are our backstory. This is the Netflix documentary about what goes on behind the scenes (with assurances that we will avoid the dramatic tearful moments and emotional music).
This is our journey, our mindset, our process, and our responses to the unexpected twists and turns that inevitably influence our portfolio.
In this letter, because of the size of the opportunity, we continue to talk about AI. The hype, the drama in the rapidly evolving landscape and we share our perspective on where we are focused at the moment. Our discussion then turns towards our exposures in South Africa, which have unfortunately been detracting from our performance. Time will tell how this twist in the story will unfold. We take this opportunity to provide clarity on our approach and share our perspectives on the opportunity set that this region presents.
Artificial Intelligence with Calculated Caution
In our last quarterly letter, we extensively discussed our broad thoughts on Artificial Intelligence (AI), hypothesising that AI heralds a multi-decade investable opportunity. To borrow the terminology of Carlota Perez, we see the diffusion of AI as a techno-economic epoch of sweeping transformation that will reshape our economic, political, and social landscape.
But despite the intentional grandeur of this opening statement, our investment approach to this promising future is navigated with calculated caution. History has demonstrated a few recurring themes that we would be remiss to ignore.
Firstly, emerging technologies tend to exhibit significant instability during their immature stages.
Secondly, despite the inherent uncertainties and wide-ranging non-linear outcomes, the Zeitgeist tends to nurture visions of unlimited opportunity, often disregarding the inevitable “valley of despair” present in every historical arc.
Lastly, when the eager embrace of Financial Capital latches onto a narrative, the gap between anticipated growth and reality can become meaningfully far apart.
Beyond our cautious tone, it is exhilarating to observe how the innovative spirit of both individual tinkerers and corporate pioneers has been unleashed. It must be remembered that in previous technological epochs (e.g., the Age of Steel or the steam engine), only the very wealthy or connected could tinker with the technology. Today, because of the vast tentacles of the Internet, **everyone** can dabble. However, there is a democratic ceiling, especially when it comes to creating robust neural network models on domain-specific data. The computational capacity required is both costly and scarce. Most enterprises resort to using existing, closed models because they lack the resources and expertise to build their own.
In time, we expect these constraints to lessen. At this moment, constraints are investable opportunities and we’re leaning into them. Over the long term, we really want to seek the entrepreneurs and businesses that democratise the technology. We don’t know how that story exactly unfolds yet, but we remain dedicated to identifying the signs and positioning our portfolio to benefit from this unfolding narrative.
AI Infrastructure Problem
The crux of our discussion last quarter was that SaltLight’s approach is derived from the lessons of previous technological epochs. We have deliberately concentrated our portfolio investments on the infrastructure layer of the AI stack because every preceding epoch underwent a substantial ‘installation period’ before the technology fully diffused. Today, moving atoms rather than electrons is the primary challenge.
This is our working hypothesis at present. However, the entire AI landscape is rapidly evolving. Therefore, as we continue to learn more, and as events progress, our assertions will undoubtedly be refined and updated.
With that in mind, here are a few insights from our portfolio companies. We hope that these snippets will give you a glimpse of the intricate canvas that we are gradually painting:
NVIDIA has become the de-facto AI name in the market (rightly so) and has delivered a remarkable performance for us. We have been gradually trimming our investment to, what we call, an ‘optionality’ size position as the stock price has dramatically gone up. Conditions are chalk and cheese to when we first purchased our investment.
Recent news makes us think twice: PyTorch 2.0, the machine learning framework recently announced support for AMD GPUs which is a negative for NVIDIA.
For many readers, this may be getting into the weeds but in our May 2021 letter, we discussed one aspect of NVIDIA’s competitive edge – its proprietary CUDA software. Operating as the interface between their GPUs and machine learning (ML) libraries, CUDA has become the standard layer in the AI stack to enable parallelised processing, a critical function for complex neural network models. Competing, cheaper, GPUs with inferior working libraries struggled to gain meaningful traction because the early versions of machine learning libraries (PyTorch and TensorFlow) did not directly interface with them. This created friction for developers and effectively blocked cheaper competitor GPUs from handling machine learning workloads.
Wide adoption of PyTorch means that PyTorch has become the de facto layer that most AI developers interact with. ML libraries are now the gatekeeper. In March, PyTorch 2.0 was released with out-the-box support for AMD’s GPUs.
It’s early days, but we believe this development could ‘chip’ away at the decade-long competitive advantage NVIDIA has enjoyed. This development warrants our attention and careful observation as it unfolds.
LAM Research, a leading toolmaker for memory and logic semiconductor fabs, has demonstrated its resilience and competitive positioning in the semiconductor ecosystem. Chip densities are now approaching the limits of physics and Moore’s law is stretched to its atomic limitations. Future innovation is around 3D stacks of chips and advanced packaging. LAM has heavily invested in tools that push innovation at the atomic level. Despite this investment, it requires little shareholder capital to grow and therefore it returns capital through healthy dividends and share repurchases. LAM has been a wonderful performer for us over the last two years.
ASML has no competitor. We’d characterise it as a monetary ‘apex’ predator. This privileged position allows the company to capture value within the semiconductor ecosystem. You can see it in their margins and operating leverage. ASML’s customers are so eager for their products that they are willing to pay upfront, two years before a machine is even delivered! This is a testament to the company’s unrivalled capabilities and the high demand for its unique offerings. It will keep this privileged position as long as it continues to push the frontiers of chip density.
ASML announced, this quarter, that their first High-NA machines (costing $200m per machine) are going to start shipping in 2024. This is a new source of its technological moat and earnings for the next decade. For the curious reader, we recommend looking up Extreme Ultraviolet Lithography tools to truly grasp how this technology could be considered ‘magic’ for what it accomplishes.
Beyond infrastructure, we do own some portfolio investments that we call ‘AI companies in sheep’s clothing‘ (we’re considering making a T-shirt with this on it). These traditional ‘ legacy’ businesses (we can see you rolling your eyes already) have created proprietary AI tools that will drive their future value creation.
One example is Applovin, which connects game publishers and advertisers through its AI model. Applovin earns a ‘tax’ on each successful advertisement and therefore it depends on as much data going through its network as possible. This also widens their moat because more data means better pricing and better targeting for the ecosystem. It’s worth mentioning that having a ‘legacy’ label has been somewhat beneficial for us: last year during the height of the Apple App Tracking fears, we bought this business at a single-digit multiple.
Over the last two quarters, our US portfolio has had us quickstepping down the road like John Travolta in Saturday Night Fever with Stayin Alive playing in the background. Thanks, in part, to a moderating pace of interest rate hikes and easing inflation data.
Conversely, our South African exposure has been a bumpy ride, akin to navigating a road riddled with potholes while Coldplay’s “Fix You” plays on repeat.
Presently, valuations of South African public companies are wallowing at unprecedented lows. Private businesses are being sold at much higher valuations than public companies, and as a result, the number of listed companies over the last decade has halved. Changes in recent pension fund rules, allowing an increase in foreign exposure, have exacerbated the exodus. This past year, market players have been grappling with a litany of concerns: frequent power outages, growing geopolitical tensions, and political dysfunction. It’s probably not superfluous to say that voluntary capital has left the country.
Let’s momentarily put aside the stark imagery of the Titanic’s final moments. This challenging backdrop indeed invigorates our contrarian instincts, reminding us of the wisdom conveyed in Edward Chancellor’s seminal book, “Capital Returns”. The central premise is that such moments of extreme capital scarcity often present ripe opportunities.
The general principle of investing is that good things need to happen. Distressed investing, conversely, is usually about widely anticipated negative events that do not materialise.
Our collection includes a distressed portfolio of businesses that are a unique breed of “survivors” and “thrivers” amid entropy. These are five highly researched, highly asymmetric opportunities.
Make no mistake about it, some of these portfolio companies are priced as ‘distressed’ assets but scratch the surface, and there are robust businesses underneath. Our wager is that the market has overweighted the likelihood of negative outcomes.
What we are primarily seeking is for the perceived risks to be mitigated over the next two years. If just two or three of these opportunities pan out as anticipated, the return on this portfolio should adequately compensate for the risks taken.
A few comments about the portfolio:
Our bet is that the sum of the various parts (SOTP) exceeds what the share price suggests. We think there are multiple ways to slice and dice value creation.
The cold-hearted, short-term, game-theory optimal action would be to simply let the SA Taxi business fail and let the ‘cross-default firewalls’ prevent contagion to the rest of the business.
Yet, despite the recent missteps (own goals and uncontrollable ones), the market has forgotten that SA Taxi has a unique position in SA. The minibus taxi industry is not going to go away. It is vital for millions and the assets have inherent economic value for at least 10-15 years. SA Taxi needs to find a sustainable model that balances value capture between riders, owners, and financiers.
The disparity in price vs SOTP, we believe, stems from the market discounting a high likelihood of a punitive rights issue to bolster capital adequacy levels in SA Taxi. This might still happen, but there are many levers that they can pull.
Meanwhile, We Buy Cars is the buried treasure of the whole company. We recently sold a vehicle on We Buy Cars and it was a frictionless, positive experience. We concur with management’s strategy of growing bays and market share.
Looking at the trends among international peers, adding finance and insurance products substantially enhances margins. Notably, peers do not use their own balance sheets to fund these financial products. Likewise, We Buy Cars recently announced that they have partnered with a local bank (still unnamed but we have our guesses) to fund this book which they believe can get to R20bn. The market, at present, does not attribute any value to this potential opportunity.
SA Taxi has a six-month timeframe to roll over its intermediate bullet payments. We will soon gain insight into the lenders’ course of action. If a successful rollover is achieved (albeit at a higher cost), this would represent the mitigation of the risk we’re anticipating.
This quarter, we invested additional capital in Purple Group, a name that many will associate with its largest business, Easy Equities (EE). The capital was primarily raised to expand their operations into the Philippines in a partnership with GCash (a platform with a staggering 66 million active users).
EE’s significant South Africa success can be attributed to its focus on what most competitors consider the least-profitable customer segment (young investors with small asset bases). EE aggregated over a million customers by leveraging complementary online distribution channels (Capitec’s 9m app users, Discovery Bank and Telkom Mobile).
An analysis of the data reveals that the attrition rate amongst active users is negligible once they deposit a modest sum of capital. Considering the relatively young age of the average EE user, retaining these customers throughout their careers could significantly compound unit economics. This is achieved through the growth of investable earnings and product upsells, making EE’s lifetime value and returns on CAC highly attractive.
As EE expands into the Philippines, it intends on following a similar partnership playbook. EE’s expansion is not without its challenges. For context, the retail investing landscape in the Philippines echoes South Africa’s market conditions of 15-20 years ago. EE is having to lay the groundwork with regulators by introducing them to fractional shares and allowing domestic retail investors to invest offshore.
Despite these hills to climb, if 1) EE can successfully overcome the regulatory humps and 2) convert even a small portion of GCash’s customer base, the CAC would remain extremely attractive. As a result, even a modest volume of trading activity could yield satisfactory returns on capital. With EE investing about $3.3 million of the $8.5 million (approx. R150m) capital raised to tap into this 66 million user base, they are taking a relatively minimal risk for the potential of substantial returns.
Blue Label Telecoms
Over the past year, our hypothesis is that the risk-reward dynamic at Blue Label Telecoms (BLU) has fundamentally altered. Last month marked a significant turning point for Cell C, as they powered down their last owned towers. This shift means Cell C has transitioned from a capital-intensive cellular operator to a more capital-light cellular wholesaler (a super-MVNO of sorts), signifying a substantial reduction in risk.
You’d think management would be on the business press circuit peddling the changing environment. Not quite. In our recent meeting with BLU’s management, they displayed minimal enthusiasm for selling the evolving narrative. Typically, managers of publicly listed companies excel at promoting their equity story, making this deviation all the more remarkable.
We believe there are several key developments to monitor:
- Updated information on the effectiveness of the partnership with Capitec.
- Reduced accounting complexity of the combined business if (or when) BLU obtains control of Cell C
- Insight into Cell C’s free cash flow (FCF) generation following the decommissioning of its towers.
If one or more of these factors turn out to be positive, they’ll serve as ‘risk killers’ to the substantial uncertainty implied in the stock price (the core BLU business is currently valued in single digits).
We’ll know more after their year-end results.
Being Paid for Taking Risk
Let’s address a question that many co-investors may justifiably ask: With the substantial risks currently present in South Africa, why would anyone take the plunge? From our experience, a lesson that we’ve learned over and over again is an aphorism: the money is made in the uncertainty. As long as we’re being paid for the risk, it is worth taking. This is our art – seizing on good risks.
Consider our experience in October last year. Investing in U.S. technology seemed like a gruelling proposition. Inflation was on an uphill climb, interest rates were soaring, and the tech sector was nursing a post-COVID hangover. The magnitude of uncertainty was palpable, a feeling that resonated deeply within us.
While we tread these uncertain waters, we find solace in one aspect: the companies in our South African portfolio are ‘survivors’ and ‘thrivers’, demonstrating resilience regardless of the macroeconomic environment. Of course, their performance would soar should the economic landscape improve. But it’s not crucial because of the margin of safety that we have. Idiosyncratic ‘risk killers’ are what we need to play out.
As always, we take this moment to remind our co-investors that the lion’s share of our liquid wealth is invested right alongside yours in the same fund. We are with you in the ups and inevitable downs. And that’s how it ought to be, we would not have it any other way.
Collective investment schemes are generally medium to long-term investments. The value of participatory interest (units) or the investment may go down as well as up. Past performance is not necessarily a guide to future performance. Collective investment schemes are traded at ruling prices and can engage in borrowing and scrip lending. A s chedule of fees, charges, minimum fees and maximum commissions, as well as a detailed description of how performance fees are calculated and applied, is available on request from Sanne Management Company (RF) (Pty) Ltd (“Manager”). The Manager does not provide any guarantee with respect to the capital or the return of the portfolio. The Manager may close the portfolio to new investors to manage it efficiently according to its mandate. The Manager ensures fair treatment of investors by not offering preferential fees or liquidity terms to any investor within the same strategy. The Manager is registered and approved by the Financial Sector Conduct Authority under CISCA. The Manager retains full legal responsibility for the portfolio. FirstRand Bank Limited is the appointed trustee. SaltLight Capital Management (Pty) Ltd, FSP No. 48286, is authorised under the Financial Advisory and Intermediary Services Act 37 of 2002 to render investment management services.
 Source: https://pytorch.org/get-started/pytorch-2.0/
 A wonderful article on EUV: https://spectrum.ieee.org/high-na-euv
 Cell C no longer maintain their own fixed cost tower network but ‘roam’ off the other two major networks (MTN for prepaid and Vodacom for postpaid)
 In April, Capitec disclosed that 500,000 SIM cards (55% conversion rate) had been sold and they’re selling 5,000 per day.
 Blue Label Telecoms has undertaken an elaborate process to avoid full ownership of Cell C until it can be approved by local regulators.