RV Capital: Exploring AI’s Impact on Jobs and the Future of Work

RV Capital’s commentary for the fourth quarter ended December 31, 2025.

Read more hedge fund letters here

Dear Co-Investors,

The NAV of the Business Owner Fund was €1’353.70 on 30 December 2025. It increased by 12.8% since the start of the year and by 1’264.3% since its inception on 30 September 2008. The compound annual growth rate since inception is 16.4%.

Three Sell Decisions

There were no new investments to report on in the second-half of 2025, but there were three sales. I typically do not discuss selling decisions in these letters, mainly because I would rather not discuss any misgivings about companies I otherwise admire. I am doing so now partly because all three decisions have very different reasons and give me the opportunity to discuss the different motivations for a sale; and partly because, well, there are no buy decisions to discuss.

A Framework for Selling Investments

Before getting to the discussion of the companies I sold, it perhaps first makes sense to lay out a framework for how I think about selling.

There is only ever one reason to sell an investment: based on a prudent evaluation of a company’s future cash flow, an alternative investment appears more attractive. It involves forming a view on the long-term cash flow-generating potential of an investment, comparing it to the best alternative, and selling it if the alternative seems more attractive.

Framed this way, the decision to sell sounds purely quantitative – weighing the expected return of one investment versus another. It is, but this fails to capture some of the nuance. Sometimes, the expected cash flows of an investment are explicit, i.e. they can be known with near certainty. Other times they are implicit, i.e. they are downstream from many assumptions and are far less certain. Most of the time, they are a mixture of the two.

An example of the former – where the cash flow expectations are explicit – is a high-grade corporate bond. The amount and the timing of the cash flow can be forecast with near certainty. A decision to sell is based on explicit quantitative assumptions, such as one bond yielding less than another. Even here, though, some degree of judgment is required, for example regarding the likelihood and severity of a potential default.

An example of the latter – where cash flow expectations are implicit – is an equity. The amount and the timing of future cash flow are less certain, especially the further out into the future the forecast extends. It requires judgment on qualitative aspects of the investment case, such as the quality of the management, the duration of the moat, and the likelihood of low-probability but potentially catastrophic risks. All of these qualitative elements will manifest themselves quantitatively over time – through the cash flow the company produces – but they defy easy quantification a priori.

How do I weigh a change in my view about something I can easily quantify vs something I cannot? One possibility would be to simply ignore factors that cannot be easily modelled in an Excel spreadsheet. Obviously, that would be insanity. In practice, if I have lost confidence in an important element of an investment case but it is difficult to quantify, I will sell the investment.

Examples where this might be the case include losing confidence in a CEO’s ability, having doubts about the durability of a moat, or seeing a potentially fatal regulatory threat. In theory, I could adjust my cash flow forecasts downwards to reflect the change in view. In practice, I do not as it would constitute false precision. How can you quantify the impact of a CEO who has lost motivation, for example? I would prefer to invest in a company where I have greater confidence in the core elements determining its future earnings power.

Why We Sold Prosus

The first sell decision was Prosus, the Netherlands-based Internet holding company, which we have owned since early 2020. I wrote about why we bought it in my first-half letter for 2020.

The rationale for owning Prosus was always to own Tencent, and we used the proceeds to buy Tencent directly. I suppose it was also a buy decision (of Tencent), but I view it primarily as a decision to sell Prosus, as I made up my mind long ago that I want to own Tencent, and nothing has changed in that respect. The only change is that we are now invested directly.

Why Sell?

Prosus’ value consists primarily of its stake in Tencent, which is over 100% of its Net Asset Value, but it also has several other investments and a large organisation dedicated to managing those investments and finding new ones.

I was never particularly enthusiastic about its other businesses or its ability to allocate capital. I have nothing against the people working there, who seem well-intentioned – it just happens to be my belief that large organisations with multiple investment committees and diffuse responsibility are not set up to succeed at allocating capital. I am more of a Berkshire Hathaway fan with its lean headquarters and single decision-maker, as you probably guessed.

Despite this, I decided to buy Prosus rather than Tencent in 2020 primarily because it traded at a significant discount to its sum-of-the-parts value. A few years later, the investment case strengthened further when it announced an aggressive share buyback program, repurchasing the maximum permitted amount of its own stock per day, using the proceeds from the sale of Tencent shares to fund the purchase.

The share buyback was incredibly value-creative. At times, Prosus could effectively buy Dollars for 50 cents. I was not super enthusiastic about selling Tencent to finance the repurchase. Prosus could have funded the repurchase through the sale of peripheral holdings such as Meituan or Trip.com, with cash, or with the regular dividend it receives from Tencent. I personally would have first exhausted these sources of cash before selling Tencent, but on balance, the program still made a ton of sense, and I was supportive of it. It meant that the intrinsic value per Prosus share would grow faster than Tencent’s, even if there were some sub-optimal acquisitions along the way. Importantly, thanks to the discount not just to the sum of the parts but also the value of its Tencent stake, ownership of Tencent per Prosus share would go up, even as its overall stake went down.

What has changed today is that the company has significantly dialled back its share repurchases. Part of the motivation is to sell down its Tencent stake less rapidly, which, on balance, is a positive and I approve of. The other motivation is to allocate capital more aggressively into new investments. This is something that I am not a fan of. As I already stated, I do not believe that it is something Prosus is good at. As a result, it fundamentally changes the calculation of owning Prosus vs Tencent directly.

In terms of my sell framework, how does it fit in? It is primarily a quantitative decision. When Prosus was aggressively repurchasing its own shares, I thought the opportunity cost of holding Prosus was lower than holding Tencent directly. Since Prosus changed its capital allocation policy, this calculation has flipped.

There is a qualitative element to the decision as well. When I bought our stake in Prosus in 2020, the company’s M&A track record was pretty good. I felt some disquiet about its corporate structure, but on balance I thought the way it allocated capital would not harm the business and may even be a source of positive surprise. Since then, I have cooled on its ability to allocate capital. A lot of the decisions have not worked out, and in hindsight, the stellar track record at the time I invested was probably due to the secular tailwinds enjoyed by Internet stocks, especially at the onset of the pandemic.

No doubt, the company would argue today that it has a new team, led by a successful entrepreneur, Fabricio Bloisi. I, of course, wish him well. My concern about anyone’s ability to allocate capital well in a corporate structure such as Prosus’ remains. I also have mixed feelings about Fabricio. He seems to be a talented entrepreneur and a decent guy to boot. However, in my experience, great operators are rarely great capital allocators, as they require two fundamentally different skill sets.

Furthermore, I am not a fan of his incentive structure. He has a “moonshot” package that envisages a huge payout if certain stretch goals are met. This makes no sense, as Prosus’ share price development will be mostly determined by Tencent, where he has zero agency. It also encourages taking big, high-risk bets rather than patiently building the company for the long term. The rapid pace of acquisitions since he took over suggests the incentive package is working as the Board intended. This is not what I am looking for, though.

The Idiosyncratic Reason for Selling PDD Holdings

The next sell decision was PDD, the Chinese e-commerce company known in the West for its Temu business. We first invested in PDD in 2023, and I described the investment in my 2023 year-end letter.

I am not one to shy away from a controversial investment, but even by my own lofty or, depending on your perspective, lowly standards, I may have outdone myself with PDD. Literally everyone hated it. It offended environmentalists who shuddered at the idea of millions of packages being flown through the air every day, financial journalists who bemoaned the company’s lack of transparency, and investors who thought the company’s competitive advantage was transient, based on the de minimis exemption, a trade rule that allows low-value goods to be imported without paying duties or taxes.

I disagree with these controversies, and they were not the reason I sold the stock. I see considerable value in the service PDD provides. It saves consumers money by cutting out middlemen and delivering directly from the factory to the consumer, and it also eliminates a lot of the waste in more traditional retail models, which involve storing goods in multiple warehouses, much of which ends up in landfills.

Its financials were always opaque, but I do not mind this, provided the goal is to hide how good the business is, rather than how bad. As an aside, when I wrote about Tencent in 2020, I pointed out there was quite a lot of intransparency, in particular around its segment reporting, but that was not necessarily a bad thing.

Finally, I never feared the consequences of removing the de minimis threshold in Western markets. I always felt that its competitive advantage was far more than a tax arbitrage. The business model reduces costs across the value chain by eliminating middlemen. And in fact, the de minimis exemption has been removed in most countries, and the business continues to thrive.

Given that I appear to still love the business, why did I sell?

I sold for an idiosyncratic reason. Shortly before Christmas, at a time when I was certainly not planning to sell a company I had a lot of conviction in, Chinese investor friends drew my attention to a Bloomberg article with troubling news:

At least two fistfights broke out last week between PDD Holdings Inc. employees and Chinese regulators who were performing checks at the e-commerce company’s Shanghai premises, according to people familiar with the matter.

I found the article deeply concerning. It is never a good idea to get on the wrong side of regulators anywhere, but you do not need to be a China guru to know that it is a particularly bad idea in China. I cannot think of a single instance where a company has fallen foul of Chinese regulators, and shareholders have not subsequently spent many years in the doldrums. Alibaba is probably the most famous example, but others include Didi, after-school tutoring company New Oriental, and, more recently, subprime lender Qifu. Furthermore, what did PDD employees want to hide so much that they would rather engage in a fistfight with regulators than disclose it?

For both these reasons, I thought the prudent course of action was to sell, especially as the share price had changed little on the news. Within my sell framework, it is probably best described as a qualitative decision. It is not hard to imagine the downstream quantitative ramifications, though.

Fast forward to today, and it is not clear to me whether the decision to sell was justified. The story seems to have died without any observable consequences for PDD. Perhaps it continues to bubble below the surface. Perhaps it was a nothing burger. Time will tell.

Aside from this, I remain enthusiastic about PDD. My conviction in the strength of its business model and the opportunity ahead is unchanged or perhaps even greater. As a result of last year’s trade war, it was forced to evolve, in particular, in how it fulfils customer orders. As is often the case when a great company undergoes a period of stress, it has emerged stronger. I can imagine investing again when I have confidence that its regulatory problems are firmly behind it.

Why We Sold IPCO

The final sell decision, and by far the most difficult one, was International Petroleum (“IPCO”), the Canadian oil and gas company. I wrote about our original investment in my first-half letter for 2024.

The decision to sell IPCO was the most purely quantitative of the three sell decisions. When I originally bought our position towards the end of 2023, it was based on a set of assumptions about IPCO’s annual production and the prevailing oil price. On that basis, I thought it was a great opportunity with a lot of upside ahead.

Then, in 2025, a strange and, from my perspective, unexpected thing happened. The oil price fell significantly, and, all things being equal, you would expect the share price of an oil company to fall as well. This would have been fine by me. As I wrote at the time of the investment, I expected the oil price to have periods of strength and of weakness, and I was perfectly happy to sit out the periods of weakness. But instead, IPCO’s share price went up – nearly doubling in USD Dollar terms vs our initial investment.

I do not have any belief in my ability to forecast the oil price – clearly, as if I did, I would not have sold IPCO a few months before a spike in the oil price of epic proportions. When I recalculated my fair value based on the same production assumptions and the new oil price, the opportunity seemed less attractive. In fact, it looked like IPCO’s operating cash flow would turn negative. As a result, I decided to sell.

It is unclear to me why the share price rose so much despite a falling oil price. I suspect there are two reasons. Firstly, when we initially invested, energy companies were deeply out of favour. ESG investing was very popular. Obviously, ESG funds could not invest, but I suspect many non-ESG funds could not either for fear of offending some portion of their investor base. I, too, wondered how my investors would take the investment, but, as always, they were very supportive. Second, at the time we invested, IPCO had a major investment phase ahead to build Blackrod, the first new “SAGD” (Steam-Assisted Gravity Drainage) oil project in Canada in many years. There were fears that the project could suffer from delays or cost overruns, which would negatively impact the share price.

Fast forward to the end of 2025, and there was disillusionment about ESG investing and perhaps also a realisation that the world cannot do without oil, a realisation that will surely grow in the coming weeks, given what is happening in the Strait of Hormuz. Also, the IPCO team has done an incredible job on Blackrod. At the time of writing, the project is virtually complete on budget and a little ahead of schedule.

I remain a fan of IPCO

I want to be clear that I did not sell our stake in IPCO for any qualitative reasons, such as a loss of confidence in the management. To the contrary, I have been incredibly impressed by Will. It truly amazes me how someone from such a privileged background can be filled with so much drive and passion. He has also surrounded himself with some first-class executives, many of whom I enjoyed meeting on a site visit to Blackrod in June last year.

Will reminds me of one of my favourite memories from attending Berkshire Hathaway’s shareholder meetings. At the 2011 meeting, an Indian gentleman observed how difficult it is to get your kids to work hard if they live in an affluent society, as they do not need to. He then asked Warren and Charlie how they would incentivise his children to compete with the hungry, highly motivated kids from emerging markets like China and Brazil. The 30,000 attendees collectively held their breath, wondering what possible advice Warren and Charlie could offer on such an intractable problem. In his inimitable, deadpan style, Charlie then responded:

My advice to you is: Lose your fight as gracefully as you can.

The audience howled with laughter. The full clip is here.

Fortunately, Will’s parents were not listening to Charlie. Nor, for that matter, were Ernie Garcia’s.

Given my admiration for Will and his team and the rebound in the oil price, should I repurchase our holding? So far, I have not. The share price is up this year, so it already reflects expectations of higher near-term earnings to some extent. More importantly, as I have come to know the oil market better, I have become progressively more pessimistic about the long-term trajectory of the oil price. Prior to investing in IPCO, I was new to the sector, and the story I repeatedly heard was that the world was running out of oil due to years of underinvestment, driven by pressure from the ESG crowd.

The more I have studied the sector, the less certain I am of this. I have visited and studied many oil companies, and it is rare to find one that does not have plans to increase production. Furthermore, there is ongoing innovation in techniques for extracting oil that was previously considered uneconomic to produce. I met a private Canadian company, called Spur Petroleum, which has had considerable success using an enhanced oil recovery method called waterflooding in the Clearwater heavy oil play of Western Canada. I suspect it has applicability far beyond Canada.

At the same time, I am more pessimistic on the demand side, or, rather, optimistic from a CO2 emissions perspective. The progress being made in electric cars and batteries, especially in China, is astonishing. Electric cars have overtaken combustion-engine cars in terms of quality and desirability. Battery technology is already good enough but continues to improve along multiple dimensions, including cost, charging time, energy density, and safety.

The shift to electric vehicles is likely to be one of the second-order effects of the events in the Middle East. You may not see this in the West, which seems to have decided that protecting its incumbent car manufacturers is more important than reducing CO2 emissions. It is observable elsewhere and not just in China. Ethiopia, a country with no hydrocarbon production but abundant hydropower, has banned the import of combustion-engine cars.

Overall, I suspect supply will be higher and demand lower than generally believed. Of course, as everyone is aware, there are some very troubling events happening in the Middle East at the moment. The closure of the Strait of Hormuz used to be considered a worst-case scenario, but is now the status quo. The new worst case is the destruction of all the energy infrastructure in the region. If that happens, there will be an energy shortage for years to come. I would not make such a scenario a central plank of an investment thesis, though. My approach to catastrophic events is to survive them, not to depend on them.

Why We Kept Aker BP

Given my pessimism about the oil price at the end of 2025, why did I not sell Aker BP? Aker BP is a different company from IPCO. It has a far lower cost per barrel of oil due to the different geology of the Norwegian Continental Shelf. As a result, even in the fourth quarter of 2025, it remained profitable and could internally finance a dividend of around 10% of its share price. In the same quarter, IPCO was loss-making.

Furthermore, with its “alliance system”, Aker BP has one of the most impressive operating models I have ever come across in a company, irrespective of sector. It enables Aker BP to develop projects faster and more cheaply than competitors, which in itself should provide it with sustainably higher returns on capital. It also makes it the partner of choice for other oil companies looking for an operating partner for their acreage, which should enable Aker BP to grow faster. IPCO, the first company to successfully complete a SAGD project in many years, is well on track to build a comparable level of expertise in Canada, but it will take time to develop the domain expertise and reputation that Aker BP enjoys.

In summary, I am happy to own Aker BP even at the oil prices prevailing at the end of 2025. In the case of IPCO, I would want more confidence around the sustainability of a higher oil price or a lower valuation to invest again. I hope one of these two scenarios materialises, as I would love to be invested with Will and his team again.

Understanding RV’s Elevated Cash Position

As a result of these sales, we ended the year with a higher-than-normal cash quota of 14% of the fund’s assets. Several of you have asked me why. The simple answer is that I had more ideas about what to sell than what to buy. It was not an attempt to time the market and did not reflect a view that a stock market crash was imminent. I cannot predict stock market crashes, and I do not believe anyone else can either.

On balance, I would far rather be fully invested than hold cash. I explained why in my 2014 letter. In short, I believe that the power of compounding is so powerful that stock market crashes are unlikely to be frequent enough or severe enough to offset the lower returns from holding cash in the intervening period. Over time, I expect an investor who remains fully invested to outperform one who attempts to time stock market cycles.

That being said, it is perhaps inevitable that when ideas about what to sell are plentiful, ideas about what to buy are not, and vice versa. In the run-up to 2022. I held off on selling several investments as I lacked compelling new ideas. This was partly because, with the benefit of hindsight, valuations were toppy, and partly because my new-idea funnel was empty due to Covid travel restrictions (I like to meet the managers I invest in before buying). When the share prices of the companies I had wanted to sell crashed a few months later, I rued that decision. It is a mistake I am determined not to repeat.

If there is a delay in deploying the cash, so be it.

The Crans-Montana Tragedy and Regulatory Implications

On the 1st of January, I, like millions of other people, woke up to terrible news. A fire had torn through a cellar bar in the Swiss mountain village of Crans-Montana, killing dozens of young people and leaving dozens more with life-changing injuries.

I did not know anyone directly involved in the tragedy, and yet it affected me deeply. I struggled to think of anything else for several days and took little pleasure in small everyday indulgences that otherwise make life enjoyable.

I wondered why that was. Comparable or even worse tragedies happen somewhere in the world every day. Yet, they do not throw me off kilter. I suspect it was because I identified closely with the fire victims. I have daughters the same age who were out celebrating the new year that evening. It could have been them. I remember frequenting similar bars in my younger years – my Cambridge college had a cellar bar called the LNB that remains the stuff of legends at the university. On a busy night, it was so packed that a full-body massage came free with your drink if you made it to the bar. It could have been me. This led me to question whether I am really the type of person who only feels empathy for people from a similar background, which darkened my mood further.

Stepping back, the questions I ask myself, and other people, including regulators, will be asking themselves are: Was this tragedy avoidable? Was it due to mistakes by individuals, first and foremost, the bar’s owners, but also fire inspectors and regulators? Is fire regulation inadequate or inadequately enforced in Switzerland? Or, was it a terrible misfortune caused by an improbable sequence of events?

We will not know until the investigation is completed, and maybe not even then. My best guess is that the answer will be “Yes” to all of the above. Elements of the tragedy seem avoidable. Escape routes were inadequate; children were in the bar who should not have been; the soundproofing foam that lined the ceilings was flammable; and indoor fireworks had no place in a crowded bar. Some of the responsibility lies, no doubt, with the owners, but by no means all. It is a mystery to me why flammable soundproofing foam is a thing, as by definition, it is only used in confined spaces. If indoor fireworks are permitted in bars, they should not be. Both may be failures of regulation.

Above all, though, it strikes me that it was a terrible misfortune. The ceiling was ignited by a young lady in a cellar bar, sitting on someone’s shoulders, wearing a motorcycle helmet, holding a Champagne bottle with a firework in it. What are the chances? Because of the helmet, she likely could not see what was happening above her head.

The answers to these questions are important because they inform what the correct response from regulators and, potentially, prosecutors should be. If, on the one hand, the tragedy was caused by a mixture of poor decision making by the owners, missed opportunities by fire inspectors, inadequate regulation, and above all terrible misfortune, then the correct response should be mild punishment of those held responsible, reflecting that there were other contributory factors involved, and targetted regulation addressing the shortcomings in existing rules.

If, on the other hand, the answers are that the owners were solely to blame and fire regulation in Switzerland is completely inadequate, then draconian punishment of the owners and sweeping regulation should be the correct response. That said, I struggle to see how both can be true at once. Either regulation was adequate, and the owners ignored it, or it was inadequate, in which case they did not break any rules.

With memories of this terrible tragedy still fresh, I imagine many people will lean towards the latter, arguing that no punishment for the owners is too severe and no regulation is too stringent if it can prevent a similar tragedy from happening again.

I feel the weight of this event, too – I described my state of mind after learning about it to make this clear. I nevertheless think that such a response would be wrong. When business owners mess up, they should expect a punishment commensurate with their wrongdoing, but no more than that. Otherwise, the long-term impact will simply be to discourage entrepreneurship. Similarly, where fire regulation is inadequate, it should be tightened, but the temptation should be resisted to introduce new rules or tighten existing ones where there would be no obvious safety benefit, as the effect here too will be to stifle economic activity.

Zooming out, this tragedy illustrates one of the central tensions we face as a society today. Nobody wants pointless regulations that stifle economic growth. But when something goes wrong, nobody wants to be the person who opposes the regulatory backlash. Nobody wants to be the person who pushes back against regulations that would supposedly have prevented the deaths of dozens of young people.

When you look for it, you see this tension everywhere. The lack of domestic opposition in Germany when it shut down its nuclear power stations in the aftermath of Fukushima and the overregulation of financial institutions post the financial crisis are two examples that spring to mind. There are countless others.

It is easy to be against a pointless regulation long after the event that gave rise to it. But that is too late. The time for voices of reason to speak up is when it is most uncomfortable to.

Exploring AI’s Impact on Jobs and the Future of Work

From an investing perspective, the dominant topic in recent months has been the accelerating pace of improvement in AI. It is only three years since ChatGPT was launched, and yet autonomous AI agents are already overtaking humans at many tasks. It’s anybody’s guess what they will be able to do in a few years’ time, especially as the pace of development is, if anything, accelerating. This has led to doomsday predictions about employment, with many fearing that most white-collar jobs will be obsolete soon, followed shortly by blue-collar jobs, given the progress in robotics (see video).

I am 95% certain the doomers are wrong. The history of humanity is that when one job has been automated, it has freed up capacity for another, leading to an increase in both the quantity and variety of products and services that consumers enjoy, and ultimately an improvement in living standards.

I have heard counterarguments that this time it is different, given the sheer pace of technological change. I would counter that if the pace at which old jobs are being replaced is accelerating, so too will the pace at which people can find new ones. Rideshare was not a thing until Uber came along, but within a few years of its founding, millions of people were driving strangers around in their private cars. Such a swift reallocation of human resources would have been impossible one hundred years ago.

I can imagine a pushback to this might be that it is far easier for workers to transition between unskilled than skilled jobs. What about white-collar workers? I would counter that the question was whether there will be enough jobs to go round, not whether the winners in the old paradigm will be the winners in the new one. That aside, I expect there will be plenty of opportunities for people with drive and intelligence to attain well-paying jobs, equivalent to white-collar jobs today. They may just be different ones.

The only scenario in which there will be no jobs is one where everyone has everything they could possibly wish for. That day seems far distant. There is still unmet demand for existing products and services. Furthermore, AI will create the conditions for new products and services that do not exist today and are impossible to envisage. Who would have thought 20 years ago that being an influencer is a thing?

The influencer example is pertinent, as it illustrates that those new jobs may seem trivial or even pointless from today’s perspective. This was often the case in the past. After the agricultural revolution freed people from working in the fields, one of the professions that grew the fastest was law. Freed from having to work in the fields, we used our newfound freedom to sue each other. It will likely be the case in the future too. With millions of young men with time on our hands, perhaps the equivalent today will be a renaissance in invading other countries. That seems to be the path many of the world’s leaders are leaning towards. Oh dear, what a bleak thought! I think Crans-Mantana may still be weighing on my mood.

What about the 5% probability that there will not be enough for humans to do? My residual fear is not an acceleration in the pace of change or a technological discontinuity that changes everything overnight. My fear is regulation, or more precisely, that the market is prevented from doing what it does best – helping supply meet demand, enabling people with unmet needs to purchase services from people who can fulfil them.

For example, if AI fulfils its promise and leaves people with more free time, we will need more venues for concerts and bars to socialise in. It is easy to imagine though how regulation can get in the way. After Crans-Montana, how many entrepreneurs will be inclinded to open a new bar? How will concert venues host more spectators if they are forced to halve their capacity to comply with new fire regulations? And how will new bars or concert venues be built if it becomes trivially easy for NIMBYs to spin up a detailed planning objection using ChatGPT? After careful consideration of fire risks, of course!

When I call for less regulation and more entrepreneurship, I imagine my readers nodding their heads so vigorously that they risk straining their neck muscles. After all, an investor letter selects for a business-friendly readership, and in any case, who is in favour of pointless rules that stifle innovation?

Now imagine yourself sitting in a courtroom across the devastated parents of the victims of the fire in Crans-Mantana and arguing that the guilt that will accompany the bar’s owners for the rest of their lives is probably punishment enough and, beyond tweaking some regulations around soundproof foam and the use of fireworks indoors, no major changes to regulation are necessary. Still nodding as vigorously? I thought not.

Regulation needs to be thoughtful and to take account of trade-offs. In theory, this sounds obvious; in practice, it is less so as Crans-Montana demonstrates.

Analysing the AI Scare Trade and Software Moats

The sector considered most at risk from advances in AI is the software sector. At times over the last few months, it felt like every product announcement by Anthropic or blog post about the potential impact of AI catalysed steep declines in software companies’ share prices. The terms “AI Scare Trade” and “SaaSpocalypse” were coined to describe this phenomenon. Several of you reached out to me to ask what I think of it.

There is a lot of uncertainty about the long-term impact of AI. The only thing everyone can agree upon is that the impact will be profound, and nobody knows what it will be. In this respect, it reminds me of the Internet in the 1990s. It was clear that the impact of the Internet would be profound too, but nobody foresaw or could have foreseen the advent of social networks, the sharing economy, or video streaming.

What can be said with certainty today is that investors are panicking. When share prices fall 20% or more in response to an innocuous news article or a minor earnings miss, it is clear that investors are being driven by base emotions rather than thoughtful analysis.

As a rule, it is generally good to be a buyer when there is panic in the air. Of course, sometimes panic is justified. However, over an investing lifetime, you will generally come out ahead if you buy when people are selling.

Is the panic justified? My sense is the pessimism may be overdone. The value of any business lies not so much in the products or services it makes or provides per se, but in the moats around them that prevent competitors from entering the market and competing away excess profits. If a robot could make tomato ketchup, it would not necessarily ruin Heinz’s business. The value lies in its brand, not its manufacturing sites.

Similarly, the value in a software company does not lie in its code alone. It lies in the competitive advantages it has built around its business. These include distribution and customer relationships, network effects, regulation, domain know-how, proprietary data, and switching costs. Nearly all software companies have at least one of these moats, and many have several. The following chart does a great job of making this point using the iceberg analogy (see image).

What seems to be getting less attention is the opportunity that AI represents for software companies. Whilst it is reasonably well understood that productivity gains from using AI will help software companies reduce costs, what seems less well understood is that it will also help improve the quality of their products. Software companies will be able to incorporate customer feature requests more quickly and update the software more frequently. This should lead to greater customer satisfaction and, as a result, make customers less likely to switch than they already were. Smaller, resource-constrained software companies should benefit from this trend disproportionately.

The lens through which investors seem to be evaluating the risk of AI disruption is switching costs, i.e., which companies are most entrenched in their customers’ workflows and therefore difficult to displace. Highly regulated companies are viewed as being particularly safe bets in this regard, as regulation will protect them from AI-first competitors. The problem with switching costs is that they reduce the urgency to innovate and adapt to customers’ needs. I personally prefer the companies that seize the opportunity presented by AI to improve their customer value proposition. Delighting customers is always the best antidote to competition.

Mark Miller, the CEO of Constellation Software, summed up this idea in the Q4 earnings call:

Building products and features faster will not be what differentiates us long term. That capability will become widely available. It’s going to be table stakes. What will matter is what our businesses have spent many years developing, deep vertical knowledge, a genuine understanding of customer workflows and processes, the data inside their solutions and the trusted relationships they’ve built. I believe AI will help us do all of this better.

What about the risk from seat-based pricing if companies need fewer employees in future? I do not see the potential transition from seat-based to usage-based pricing as a distinct risk, but part of the broader question of whether software companies continue to have a moat.

Imagine a scenario where a company previously had 10 white-collar workers earning €10’000 per month using software that costs €100 per seat, and, thanks to improvements to that software, the company can do the same work with half as many workers, saving €50’500 per month. Clearly, in such a scenario, the company will be more than willing and able to pay far more than €1000 per month on a usage basis. The question is whether there is any kind of lock-in with the software vendor. If there is, the vendor will capture a good part of the value created. If there is not, the value is more likely to accrue to the customer. Either way, the question is whether the software company has a competitive advantage, not whether it can transition to a different billing model.

Software companies are out of favour at the moment. I have heard people opine that this is unlikely to ever reverse, as AI creates uncertainty about terminal values, and there is no scenario in which this uncertainty dissipates, since you cannot disprove a negative. Maybe. But in my experience, this is not how capital markets work. Attention spans are short, and investors’ focus tends to move on when the next big thing comes along. The risk does not go away, but the investors attention to it does.

As I write this, investors’ attention is already moving on to the next big thing – the war in Iran, the spike in the oil price and the potential impact on the global economy. That is a topic for another day. However, in this scenario, investors may suddenly warm to a sector with recurring revenues, low cyclicality, and depressed valuations.

RV Capital’s 2027 Annual Gathering

It was great to see so many of you in Engelberg in early January. The date for next year’s event is fixed for the 16th and 17th of January 2027. It is one week later than this year, which I hope makes it easier for my friends in the Southern Hemisphere to join us.

Registration will open at 3:00 p.m. on 4 September here. I pick a time that works for most time zones, but if you are a Pacific Islander, feel free to email me to plead for an exemption.

The tickets were snapped up in a few seconds this year, so set an alarm to avoid disappointment. Tickets may be even scarcer next year, if the capacity of the room is reduced for obvious reasons.

As a reminder, investors in the fund have a guaranteed spot, so if you are based in the Americas, feel free to sleep in, and if you are based in Asia, don’t let me disturb your Friday evening cocktails.

Rooms at the Terrace are only bookable through its website. The promo code is RVCapital27. The code only works from the 15th to the 17th of January. If you want to arrive earlier or stay longer, you need to make a separate booking. Rooms can be cancelled up until the 13th of January free of charge. Further details are available at rvcapital.ch.

In the meantime, stay healthy and happy investing!

Rob