WHEN ARTIFICIAL INTELLIGENCE BECOMES ABUNDANT IN INVESTING, WHAT WILL BECOME RARE?
For quite some time now, I have had a favorite little joke on the subject of artificial intelligence. I like to say that “the sum of human and artificial intelligence will remain constant.” The higher artificial intelligence climbs, the lower slips human intelligence. Of course, that’s probably just a bit exaggerated, although we see plenty of evidence around us lending support to this proposition. In recent months, I’ve been exploring this topic more seriously, and we even touched upon it at our annual meeting of shareholders. Some of you found it very interesting, and I promised that I would address it in more detail in a letter to shareholders—and specifically, how it plays out in investing.
A more precise and serious formulation of this thesis would therefore probably not be that “the sum of human and artificial intelligence will remain constant.” Rather, it could be said that, “As artificial intelligence grows, some part of human intelligence loses its economic scarcity while other human capabilities become even scarcer still.” I’d like to present five main points to you and then close with seven concluding arguments. This will be a somewhat philosophical look at the otherwise very practical and concrete impacts of A.I. on investing.
Main thesis
First, the main thesis. This is the idea that, in investing, value typically shifts to where something is in scarcity. A.I. and abundant machine intelligence will not strip investors of their competitive advantage, but they will shift it elsewhere. Activities such as gathering, summarizing, sorting, and transcribing information, together with basic modeling and even some aspects of pattern recognition will become ordinary commodities. What will gain in value will be capabilities for sound judgment, good taste, patience, an original perspective, strategic skepticism, and the ability to distinguish between what is and is not important. We view A.I. as an extraordinarily useful tool but a very dangerous substitute for one’s own judgment. Now, my five main points:
1. Intelligence is becoming more accessible, but sound judgment is not.
For many years, investors’ competitive advantage lay partly in their better access to information and the effort required to process it. Legendary investors such as Benjamin Graham and Warren Buffett had to search for data in libraries early in their careers, personally visit individual companies, and literally beg for even the most basic information. They were willing to spend long days poring over publications such as Moody’s and S&P manuals and the like. Their judgment was better than that of others, but they gained their edge even during the information-gathering phase.
That world, however, is long gone. Increasingly, everyone will have access to quick analyses, instant summaries, high-quality transcripts, comparable data sets, and decent models for initial analysis. Investing, however, has never been just about gathering raw information. It has always been about the ability to make sound decisions in an environment of uncertainty. This means distinguishing signal from noise, understanding motives, identifying what is unique, and resisting the alluring but false illusion of truth. While sufficient information may improve the average quality of analysis, it does not mean automatically that the average quality of judgment will also improve. On the contrary, if it creates a false sense that we understand the issues at hand, it can have exactly the opposite effect. When answers are at our fingertips, the value of a well-placed question increases.
2. The use of A.I. can reduce an investor’s information advantage but increase one’s interpretive advantage.
Information advantage depends upon who knows more and who knows it quicker. Interpretive advantage belongs to those who better understand what the information means. In a world where artificial intelligence is widely and abundantly available, the former advantage loses its significance. The latter, however, becomes more important. A successful investor will not necessarily be the one with the most data, but rather the one who can place facts into a broader context, recognize which variables truly influence intrinsic value, distinguish between lasting changes and fleeting fads, understand second-order consequences, and correctly define and assess the level of risk. Markets have always mistakenly equated a greater volume of information with deeper understanding. Artificial intelligence may further exacerbate this misperception. A.I. also is unlikely to replace experience acquired through many years of investment practice. Rather, it can complement that experience.
3. The greatest danger is not that artificial intelligence will surpass us in thinking, but that it will unify our thinking.
If professional investors use the same tools, and these tools are trained on similar data and designed to solve similar problems, then the likely result will be not brilliance or originality but a convergence of opinions. The danger lies not only in inaccuracy but also in the homogenization of thinking. Widespread use of A.I. can lead to greater consensus; greater consensus can lead to similarity of positions in the market; and similar positioning can lead to greater market instability.
Paradoxically, an abundance of A.I. and its use may create more opportunities for truly independent and original investors, because the average level of opinion will become more and more uniform. This is a very important consideration for active investors. Performance alpha may increasingly depend not on the use of better A.I. tools but on the extent to which investors refuse to be constrained by their outputs. Originality and independence thus gain in value. Once everyone has his or her own highly intelligent A.I. assistant, original thinking will likely become more rare, not more common.
4. The scarce commodities of the future are temperament and time horizon.
If artificial intelligence speeds up analysis, that will not make most market participants calmer or more patient. Quite the opposite, in fact. They may behave even more impulsively. A.I. will have a similar effect on investor behavior as do social media. The only difference will be in the form of interaction. While interaction with social media takes place on public platforms, interaction with A.I. takes place in private. This means that certain traditional advantages may become even more important: the willingness to wait, emotional stability in periods of market downturn, the ability to cope with uncertainty, the willingness to accept that one might be wrong for a certain period of time, or perhaps the ability to hold a different opinion without seeking its immediate validation.
An ability to think long term may offer even greater added value. Machine intelligence accelerates short-term processing of information in the market, so human investors stand to gain more by extending their time horizons rather than shortening them. The more the market focuses on speed, the greater will be the value added by taking a longer-term view. Market efficiency and competition in the very short term will increase, while market efficiency and competition over the long investment horizon are likely to decline even further. This creates growing opportunities for so-called time arbitrage.
5. Artificial intelligence is likely to improve the analysis of what is obvious, but not necessarily of what is abstract, obscure, and vague.
Artificial intelligence will likely be most effective when it has access to large amounts of data, language is standardized, and the subject matter is well-documented in accessible information. Examples include large companies with highly liquid shares; sectors where consensus rules the day; where there exist regular disclosure of information and mainstream narratives. Many of the best investments, however, are not found in these clear-cut and well-described areas. Often, we find them in places that are complex, underfollowed, under-covered, scattered around the world, culturally specific, historically conditioned, or difficult to translate into unambiguous data sets. In the spirit of my recent book, one could say that the era of abundant intelligence need not spell the end of hidden investment treasures. On the contrary, it may make them even better hidden and more valuable.
A deeper philosophical question might be: What is investing, really, if it is no longer primarily a competition in information processing? And our answer might be: Investing is the art of making informed probabilistic decisions regarding long-term cash flows within a social system influenced by incentives, narratives, reflexivity, and emotions.
Artificial intelligence helps with this to some extent, but not with everything, and especially not with the social and reflexive aspects. The legendary investor George Soros once presented his theory of reflexivity in his book The Alchemy of Finance. It posits that investors’ perceptions shape economic fundamentals, which in turn influence those perceptions as well as future market prices. By creating this feedback loop, this dynamic prevents markets from reaching an equilibrium state, and often times this, in turn, leads to exaggerated market cycles and volatility. Now A.I. models are entering this feedback loop between markets and investors. The MARKET – INVESTORS – MARKET feedback loop will become a MARKET – A.I. – INVESTORS – MARKET – A.I. – INVESTORS – MARKET loop. The more investors gradually come to rely on the outputs of A.I. models, the more the human factor will recede into the background and the feedback loop will approach a MARKET – A.I. – MARKET state.
Machines may learn to interpret the map very well, but the markets will continue to punish those who forget that the terrain is changing. Quite a few years ago, long before the term “A.I.” became a part of our everyday vocabulary, I had the opportunity to talk about A.I. with Garry Kasparov. He was in Brno at the time. I also had the chance to play a simultaneous game of chess against him, and afterward we talked about computers in chess. If my memory serves me correctly, Kasparov told me that computers play chess very well because chess is a game where the rules are precisely defined and the objective of the game is also precisely defined. It is a closed system with complete information and no uncertainty outside the chessboard. Investing, however, is a “game” where the rules are very complex to define and where the objective of the game is also very complex to define. It is an open system; players adapt, data are incomplete, and the system has its own reflexivity. Unlike in chess, there exists no objectively correct move in this game.
Seven summary arguments
1. A.I. is commoditizing basic analytical work.
2. The advantage is shifting from information to interpretation.
3. The value of judgment is increasing.
4. Also increasing is the risk of false certainty.
5. Consensus may strengthen.
6. Human strengths – temperament, patience, skepticism – are not losing their value.
7. The best opportunities may lie where data is incomplete, situations are chaotic, and context is difficult to standardize.
So, in a nutshell, this is our current view on A.I. and investing. Like any other opinion, this one is purely subjective. It is likely that each of you will agree with some points and disagree with others. I myself do not consider what I have written above to be unequivocal. We could debate, for example, whether I am underestimating A.I.’s ability to improve its judgment. The idea that what is unknown tends to be overlooked may be correct, but that is not always and not automatically the case. We also could debate, for instance, that while the risk of stereotypical thinking is real, the opposite risk is also important. That is the question of artificial originality.
The world is changing. Five years ago, for example, we would not have been having this conversation at all. In another 5 years, our views and experiences may be somewhere completely different. We therefore strive to remain as flexible, open, and adaptable as possible—and to keep learning. We use A.I. extensively on a daily basis, and it boosts our productivity incredibly. At the same time, we strive to leverage its capabilities while capitalizing on its shortcomings and making the most of increasingly scarce human values. Despite all the hype surrounding A.I., the basic principles of investing remain the same.
Changes to the portfolio
During the second quarter, we sold out four stock positions and added two new ones to the portfolio. We sold the troika Lam Research, Applied Materials, and KLA Corporation, all of which are key suppliers of manufacturing equipment and process control solutions for the semiconductor industry. These are undoubtedly excellent businesses, without which it would be impossible to produce increasingly advanced chips. Nevertheless, once the valuations of even the best companies begin to reach levels of 20× sales and 50× earnings, the balance between quality and price shifts significantly to the investor’s disadvantage. In such a situation, it is no longer enough simply to recognize that these are excellent companies. It is also necessary for future growth, margins, and return on capital to remain exceptionally high over the long term and for practically no significant risks to materialize. In our view, this is an overly demanding and speculative combination. Despite the ongoing semiconductor boom, it is still good to remember that this is a pretty cyclical industry. There remained no margin of safety between price and value to speak of, and the high prices were therefore our impetus to sell.
We also sold the Canadian Cenovus Energy. This was the only company in our portfolio focused on oil production. It is fair to say that we made more here than we ever expected, primarily thanks to the war in Iran. Investing in oil-producing companies combines long-term considerations of a company’s fundamental value with short-term fluctuations caused by oil price volatility. Moreover, markets tend to overreact to stock prices, in both upward and downward directions. A year ago, in the spring of 2025, when uncertainty surrounding U.S. trade tariffs was at its peak, the spot price for a barrel of WTI crude oil plummeted to nearly $60. At that time, Cenovus shares were trading at Can$16. In our view, this was significantly below the company’s intrinsic value and it was a signal for us to buy.
This year, in the second quarter of 2026, as concerns peaked about the impact of the war in Iran on the oil market, the price of WTI climbed to $114 and Cenovus Energy shares rose above $40. The company’s intrinsic value certainly did not increase by 150% in just 1 year. That value cannot be derived from current oil prices but from long-term expected oil prices. Although these, too, have risen somewhat over the past year in our view, we did not believe that they justified a share price exceeding $40. That is why we sold the shares. We believe that in this sector, it is wise to respond to its greater cyclicality. The market tends to extrapolate current trends far into the future, and this creates opportunities in oil company stock prices for both good buying and good selling opportunities. It is therefore quite possible that, given favorable conditions, we will return to this stock in the future. Among other reasons, this is because the stock acts as a form of hedge against adverse geopolitical events. In this regard, the stock performed excellently this year.
We bought shares in Visa and Kaspi.kz. Visa probably needs little introduction. Nevertheless, I’ll describe the company briefly, as people are often surprised by just how complex and sophisticated its business is. Visa is not a “card company” in the simple sense of the word. It is a global technology infrastructure for electronic payments that connects banks, merchants, consumers, payments processors, governments, and companies. Its main role is to ensure that payments between two parties are processed quickly, securely, reliably, and with minimal risk without regard to the country, currency, bank, card type, or sales channel involved.
The sophistication of Visa’s business lies in the fact that it is not just about forwarding the payment itself. Visa operates an extensive payment network full of rules, technologies, and security standards. It also handles transaction authorization, fraud prevention, settlement, tokenization, data analytics, and risk management. Every payment looks simple to the customer. A person taps with a card or clicks in an online store, but behind all of this lies a highly complex system coordinating among many participants in the financial ecosystem.
Visa’s strength lies in its vast reach, trustworthiness, and network effect. The more merchants that accept Visa cards, the more useful Visa is to cardholders. The more people who use Visa, the more important it is to merchants. This effect is very difficult to achieve, because any competitor would have to simultaneously secure a large number of banks, merchants, and customers, as well as regulatory approvals, technological reliability, and global acceptance.
Economically, Visa is an exceptionally attractive business because it does not expose its own balance sheet to the credit risk of ordinary cardholders. It does not lend money to customers like a bank. It collects fees for network operation, transaction processing, and related services. Its business therefore combines elements of critical financial infrastructure, a software platform, a global brand, and a regulated network oligopoly. It is precisely this combination of a simple user experience and an extremely complex system behind the scenes that makes Visa one of the most sophisticated companies in the world of financial services. As a result, its profit margins and returns on invested capital are very high.
In our view, Visa is one of the best businesses out there. We have known this for a long time, and, of course, other investors know it, too. Primarily for this reason, Visa shares have historically traded at valuations that were generally too rich for us. Now, their price has finally reached an acceptable level, which is why we were very pleased to add the shares to Vltava Fund’s portfolio.
Even from a global perspective, Kaspi.kz is an exceptional example of a digital ecosystem combining into a single application payments, e commerce, marketplaces, consumer finance, merchant services, travel, advertising, and even select government services. It is therefore neither just a bank, a payment app, nor an online store, but a deeply integrated infrastructure of everyday economic life in Kazakhstan. Its uniqueness lies in its penetration rate, frequency of use, and breadth of features. Kaspi reports more than 25 million consumers and 900,000 merchants across Kazakhstan and Turkey (thanks to its majority stake in the online retail shopping portal Hepsiburada). In Kazakhstan, it has achieved an extraordinary usage level of 77 transactions per month per active customer. This is an engagement level that even many global platforms might envy.
Kaspi’s dominant market position is founded not only on its size but also on the network effect among consumers, merchants, and financial products. The more customers use Kaspi to pay, shop, and manage their finances, the more important it becomes for merchants. The more merchants are in the system, the more valuable the app is for customers. This interdependence creates an ecosystem that – much like Visa’s – is very difficult to replicate.
The sophistication of the business lies in Kaspi’s ability to monetize a single customer relationship in multiple ways: through payments, commercial transactions, credit products, merchant services, advertising, logistics, and other digital services. From the user’s perspective, it is a simple application on one’s phone. From a business perspective, it is a comprehensive, data-driven platform combining elements of Visa, PayPal, Amazon, Shopify, a bank, a BNPL (buy now, pay later) provider, and digital public infrastructure. All of this is backed by an exceptionally strong local presence in a single market. Kaspi’s profitability is very high, and even the otherwise excellent Visa cannot match its return on capital.
Kaspi is a company that is very well known among investors. We ourselves have been closely monitoring it for about 6 years. During that time, we came close to buying its shares on several occasions, but this is the first time we have actually done so. The decisive factor was a business trip I took to Kazakhstan. I had been invited to lead a two-day workshop at the Narxoz University Business School. The audience consisted of professionals, including senior executives from the financial sector. This gave me the opportunity to discuss the economy, the financial sector, government economic policy, and Kaspi itself with financial professionals. It was also my first opportunity to take a detailed look at the Kaspi super-app in real life. I must admit that it literally took my breath away. It seemed to me to be a generation ahead of what we are used to in the Western world. I saw how people use the application on a daily basis, and I experienced firsthand that without it one is quite limited in Kazakhstan. Last but not least, Kazakhstan itself made a very positive impression on me. When you add it all up and combine this with the significant undervaluation of Kaspi’s shares, that was our reason for buying the stock. Following our very successful investment to date in the Mexican company Quálitas Controladora, Kaspi is our second investment in emerging markets.
And what can be said generally about current developments in the stock markets? Speculative frenzy in certain narrow segments of the market is very high. The main narrative flows from history’s largest ever boom in capital investment, all of which is connected to A.I. In the short term, this appears to be boosting the overall profitability of companies in the market, because capital expenditures are reflected in suppliers’ profits but not nearly as much in their customers’ costs, since the vast majority of these costs are capitalized. The stock market as a whole is trading at record-high earnings multiples driven by record-high margins. The sustainability of this situation is not high, and we are trying to stay as far away as possible from these segments of the market, seeking instead the calmer waters that are, paradoxically, full of high-quality and attractively valued companies. Holding these stocks does not entail high risk and does not carry so high a speculative element as do the shares we avoid. We remain cautious and highly selective. We are focused on companies whose value is based more on real and sustainable earnings, high free cash flow, disciplined capital allocation, and reasonable valuations than on optimistic market expectations. We believe that this approach best reflects our responsibility to you, our shareholders, at this time.
Daniel Gladiš, July 2026
For more information
Visit www.vltavafund.com
Write to us at investor@vltavafund.com
Follow us at www.facebook.com/vltavafund and https://twitter.com/danielgladis