The S&P 500 Index returned -4.33% and the S&P Global Broad Market Index returned -2.68% in the quarter ended March 31, 2026.1
Watching speculative AI bets outperform the quality businesses we own has not been easy.
But as capital rotates away from profitable, high-quality businesses and increasingly into concentrated AI bets, we believe this divergence has created one of the most compelling environments for our strategy that we have ever seen.
In our last letter, we argued that the market environment has become increasingly speculative and risk-seeking, as evidenced by the fact that:
- Unprofitable companies had significantly outperformed profitable ones;2
- The most-shorted stocks had outperformed the overall stock market;3
- Momentum, a key feature of most speculative markets,4 was the best-performing factor in both the United States and the European Union during the two-year period of 2024-25;5
- High-quality stocks, which have a history of outperforming global indices over the long-term but also of lagging in speculative markets, have recently experienced historic underperformance;6
- Artificial Intelligence (AI) stocks have significantly outperformed non-AI stocks since the release of ChatGPT, resulting in historic market concentration in a single technological theme.7
Since the beginning of the year, these trends have largely continued. Most notably, many AI stocks, especially the most historically cyclical ones such as semiconductor capital equipment and memory storage, continue to dramatically outperform most quality stocks, including ours. Moreover, recent product releases by Anthropic, one of the biggest AI large language model companies, have heightened investor concern about AI disruption of a wide range of businesses. Given the disruptive technological and geopolitical environment in which we find ourselves, we believe it’s important to help you better understand your portfolio’s exposure to these potential AI disruptions. Hopefully, this exercise will leave you feeling as optimistic about your portfolio’s future risk-adjusted returns as we are.
Before analyzing the businesses we do own, we want to explain again why we don’t own many of the current AI winners. As we extensively discussed in our Q4 2024 and Q3 2025 letters, it’s not because we are AI skeptics. In fact, we believe AI is quite likely to be transformative to both the economy and the long-term economics of many businesses. Rather, our concerns about AI stocks and the market’s concentration in them are based on our study of past technological transformations, such as the internet and railroads. These technologies changed our world in profound ways, but most of the stocks driving these changes performed poorly over the long term, and for a simple reason: too many investors and companies chased the riches that exciting new technologies promised, pushing down the returns that these investors and companies ultimately earned.
This simple truth is why most companies perform poorly over time. Companies, such as some of the current AI winners, generate products and services that earn high returns on their capital. Investors chase the high growth and high returns of these businesses, bidding their stock prices up. Then rival companies compete and innovate, driving down growth and returns on capital and causing the stock prices to collapse.
Our job is to find and purchase, at reasonable prices, businesses that are resistant to these forces of competition and innovation. In other words, we want to own businesses where it’s 1) very hard for new competitors to enter and 2) highly likely that demand for the businesses’ products and services will continue to grow over time. If these two conditions are met, then, as long as we purchase them at reasonable prices and regulators don’t cap their returns, we believe these businesses are highly likely to enjoy enduring pricing power, increasing returns on invested capital, and attractive risk-adjusted returns over time. Furthermore, as we’ve said before, these characteristics essentially make each of these businesses toll takers on large segments of economic activity.
With this backdrop in mind, let’s examine whether the toll takers we own are built to withstand – and in some cases benefit from – an AI-enabled future.
The first category we’d highlight is businesses that own dominant physical networks with key assets that are protected by not-in-my-backyard dynamics and/or regulatory and customer risk aversion. The railroads and industrial gas companies we own in Global Industrial Production, the waste management companies in North American Consumption, the aggregates companies in Global Construction, Copart in Global Miles Driven, and Transdigm in the Global Air Travel all fit in this segment and have, in our view, near-zero disruption risk from artificial intelligence. No matter how good artificial intelligence gets, the physics of hauling rock or waste and the opposition of local residents to a new rock pit or landfill being built nearby will not change. If anything, these businesses may benefit from artificial intelligence because they can use advances in robotics and AI agent capabilities to increase reliability and labor productivity.
The second category we’d highlight is businesses that have fewer tangible assets but that are similarly protected. Artificial intelligence can’t duplicate the heritage and exclusivity that power the status signaling belief networks of Hermès and Ferrari in Global Wealth, and it can’t replicate the capital efficiency and risk reduction benefits that accrue to CME Group and its customers in Capital Markets from owning the platform/network where a large percentage of financial markets derivatives trades are cleared. In fact, rising numbers of AI agents and AI-driven increases in global wealth will most likely increase luxury demand and trading volumes. Procter & Gamble and Colgate also probably belong in this category.8 These businesses benefit from psychological switching costs and social proof belief networks that could be impacted by AI agents recommending cheaper and/or better private label products. However, habit, risk aversion, and small purchase prices probably make any market share erosion quite slow.
The third category is businesses that own networks that have become globally accepted common languages and that are used to efficiently communicate positioning, manage risk, and benchmark performance. As a result, these proprietary languages are deeply entrenched in workflows, regulatory documents, and contracts throughout the world. In many cases, these businesses also possess deep proprietary databases that are not publicly available and, therefore, cannot be analyzed or manipulated by non-company-owned AI agents. The businesses in this category are FICO, S&P Global, MSCI, and Moody’s in Global Capital Markets and Verisk in Insurance. While these businesses have some ancillary divisions that could be negatively impacted by AI, we believe most of their revenues are likely to become more valuable over time as AI increases economic growth and, thus, insurance and capital markets activity as well. Moreover, these businesses could simultaneously improve their cost structures and profit margins if AI agents meaningfully enhance workforce productivity.
The fourth category is businesses with multi-sided network effects. Aon and Marsh in Global Insurance and Visa and Mastercard in Global Payments are the businesses in this category. Aon and Marsh own unparalleled networks that efficiently match buyers and sellers of global commercial insurance. Given their global presence, the unique data they generate from their brokerage network, and the complexity and ever-changing nature of risk, they give their often-global customers the information, judgment, and confidence required to navigate the bespoke challenges of what risks to insure in each country and how to most effectively purchase the protection. Visa and Mastercard connect billions of consumers with millions of merchants and thousands of banks, leading to a deeply entrenched payment infrastructure. Most consumers and banks are quite satisfied with the efficiency, reliability, reward benefits, and costs of this payment paradigm. Moreover, Mastercard and Visa continue to aggressively invest in new payment technologies and to partner rather than compete with potential disrupters such as PayPal, Apple, Google, most of whom quickly realize that they can’t solve the coordination problem of getting consumers and merchants to switch en masse. Lastly, Mastercard and Visa provide dispute resolution, chargeback infrastructure, fraud liability allocation, and other legal and regulatory protections that significantly reduce the theoretical cost advantages of alternative stablecoin payment infrastructures.
The fifth and final category is mega-cap tech companies. Microsoft and Amazon (AWS) in Global Enterprise IT and Amazon (E-Commerce), Alphabet, and Meta in Global Advertising & E-Commerce are our holdings in this category. Artificial intelligence is both a huge opportunity and a huge potential threat for these businesses. These businesses all own dominant multibillion-person networks that are, in many ways, everything stores for their customers. Their one-stop shop nature, combined with deep distribution channels, has historically made these networks uniquely durable. As technologies and customer-desired product offerings change, some of these companies’ business lines have shrunk or evaporated. However, as technology continues to grow in importance to the economy, these companies have been able to replace dying product lines with new and bigger revenue streams. Interestingly, because innovation comes from unpredictable places, most of these products and services weren’t even invented and popularized by these companies. However, their networks’ unmatched scale and scope advantages have allowed them to quickly become number two or three players in these large and growing markets. Artificial intelligence is different from these other advances in that it is far more capital intensive with a more uncertain payoff and potentially more competitors. However, we believe the unparalleled cash flow thrown off by these businesses combined with the adaptability provided by their networks’ scale and scope makes their downside more limited and their upside more exciting than many investors currently fear. In other words, we think they will generate attractive risk-adjusted returns from these prices. Furthermore, when we consider their place in your portfolio, we believe they provide valuable diversification benefits through their direct exposure to the AI boom. And we don’t think we are paying a price for these diversification benefits. As we’ve discussed, while the mega-cap tech companies’ future returns on capital are more uncertain than our other categories, we believe we are being more than adequately compensated for increased uncertainty. The mega-cap tech stocks we own trade at large discounts to our more predictable and durable companies despite much faster growth, leading to, in our opinion, similar risk-adjusted returns to the rest of our portfolio.
Now, if you’re particularly eagle-eyed, you may have noticed that the chart above includes one new business, Waste Connections. Where appropriate, given tax considerations, we’ve added this business to our waste management bucket. The Waste Connections investment thesis is very similar to the other waste companies, except that the company deliberately focuses on secondary and rural markets where it is frequently the sole service provider – creating economics that are, if anything, more attractive than those of its larger peers. Its smaller size also leaves more room for meaningful acquisition opportunities. However, we didn’t invest initially because we believed that these advantages were more than offset by the valuation premium that the company commanded. Partially due to containable environmental problems at one of its landfills, this valuation premium has shrunk significantly, giving us the opportunity to initiate a position.
You may have also noticed that a few businesses are missing from last quarter’s chart. The reason is that the increased volatility due to AI, the Iran war, and stock-specific factors pressure tested and further clarified our conviction in the portfolio positioning. Without exception, we found we were much more excited to add to our largest, highest conviction names than to our lower conviction, smaller positions. While we continue to believe these smaller positions are all great businesses, we also believe they are each exposed to significant and growing longer-term risk factors. For instance, while Intuit, Adobe, CoStar Group, and CBRE have strong current pricing power, AI progress will probably increase competitive supply and reduce switching costs over a multiyear period, putting these businesses’ enduring pricing power into question. Similarly, while Apple has overlapping network effects and powerful switching costs that create strong current pricing power, we have grown concerned both that Apple’s high margin services business may be undermined by price-shopping AI agents that evade Apple’s large app store toll and that the company’s supply chain in China would cost more in time and money to replicate in the event of a war than we previously understood. Additionally, while LVMH’s core heritage brands remain powerful status signals, we have grown increasingly concerned that their scale and ubiquity are eroding the scarcity that sustains their enduring pricing power. Lastly, while Progressive is an incredible business, we have never felt comfortable owning a large position because insurance underwriting is both opaque and requires significant leverage to generate attractive returns on capital. This combination can be very dangerous during financial crises. Moreover, as Progressive’s market share expands, its risk pool will inevitably converge with the industry average, making it increasingly difficult to sustain its historical outperformance in underwriting. And that’s not to mention the risk that AI agent proliferation shrinks their underwriting and cost advantages over time.
As a result, we used the stock market’s recent volatility to shift these positions into higher conviction ones that had experienced similar valuation compression but for reasons that we believe to be temporary in nature. In other words, we believe we sold great businesses facing legitimate threats to long-term pricing power to buy equally cheap great businesses facing threats to near-term earnings but not to long-term pricing power, which we have found to be a more reliable path to attractive risk-adjusted returns over time.
Concluding Thoughts
We understand that seeing your portfolio down in a period where certain parts of the market are surging can be uncomfortable. This is doubly true when the surging parts of the market are new and exciting AI technologies. We feel that discomfort right alongside you. Our personal capital is invested in the same strategy as yours. We hope this letter has clarified where we stand, but we’ll conclude by reemphasizing the main points we hoped to convey.
We believe AI will be transformative. We also believe it will be disruptive to many business models and industries.
While the broader market may be taking on increased risk to chase performance, we are maintaining a disciplined approach that is intentionally designed to be resilient to these disruptions.
Roughly half of your portfolio is invested in businesses that have little to no exposure to AI disruption — waste management, railroads, construction aggregates, industrial gases, auto salvage, aerospace parts, and luxury goods. These are physical, real-world toll collectors on global economic activity. They possess enduring pricing power that no chatbot can replicate. Garbage still needs collecting. Trains still need to run. Hermès bags don’t get disrupted by software.
The other portion of your portfolio is not a bet against AI — it is a bet on who will ultimately benefit. Alphabet, Amazon, Microsoft, and Meta own the cloud infrastructure layer on which AI runs. S&P Global, MSCI, FICO and Visa own industry protocols that are deeply embedded in the regulatory frameworks and contractual relationships of their respective markets. AI does not displace these moats, it reinforces them. These positions should benefit from AI adoption regardless of which AI applications ultimately win.
Some of the best entry points into high-quality businesses come during periods of fear and indiscriminate selling. As our portfolio changes this quarter demonstrate, we are actively looking for chances to put your capital to work at prices we find attractive. This approach supports our strategy of solid fundamentals and our goal of long-term growth.
We are here for you. If you have questions, concerns, or just want to talk through what’s happening, please don’t hesitate to call or email us directly. That’s what we’re here for.
Sincerely,
The YCG Team
Disclaimer: The specific securities identified and discussed should not be considered a recommendation to purchase or sell any particular security, nor were they selected based on profitability. Nothing said in this piece may be considered to be an offer to buy or sell any security. Rather, this commentary is presented solely for the purpose of illustrating YCG’s investment approach. These commentaries contain our views and opinions at the time such commentaries were written and are subject to change thereafter. The securities discussed do not represent an account’s entire portfolio and in the aggregate, may represent only a small percentage of an account’s portfolio holdings. These commentaries may include “forward looking statements” which may or may not be accurate in the long-term. It should not be assumed that any of the securities transactions or holdings discussed were or will prove to be profitable. Data presented was obtained from sources deemed to be reliable, but no guarantee is made as to its accuracy. S&P stands for Standard & Poor’s. All S&P data is provided “as is”. In no event, shall S&P, its affiliates or any S&P data provider have any liability of any kind in connection with the S&P data. No further distribution or dissemination of the S&P data is permitted without S&P’s prior express written consent. All MSCI data is provided “as is.” In no event, shall MSCI, its affiliates or any MSCI data provider have any liability of any kind in connection with the MSCI data. Past performance is no guarantee of future results.
1 For information on the performance of our separate account composite strategies, please visit www.ycginvestments.com/performance. For information about your specific account performance, please contact us at (512) 505-2347 or email [email protected]. All returns are in USD unless otherwise stated.
2 See https://fortune.com/2025/10/21/russell-2000-companies-unprofitable-stock-outperforms-tech-bubble/.
3 See https://finance.yahoo.com/news/goldman-basket-shows-painful-month-093000123.html and https://www.hedgeweek.com/european-short-sellers-hit-hard-as-most-shorted-deliver-surprise-results/.
4 See https://www.investopedia.com/terms/s/speculativebubble.asp.
5 See https://x.com/WTCM3/status/2009707666827030918 and https://x.com/WTCM3/status/2009707819256459645.
6 See https://privatebank.barclays.com/insights/market-perspectives-september-09-2025/qualitys-quiet-strength-why-it-may-be-due-a-rebound/, https://x.com/WTCM3/status/2009707666827030918, https://x.com/WTCM3/status/2009707819256459645, and https://www.trustnet.com/news/13473153/quality-stocks-could-rebound-as-early-as-the-end-of-2026-says-global-manager#:~:text=Trustnet%20/%20News%20&%20research%20/%20Quality,pulled%20capital%20away%20from%20quality.
7 See https://www.investopedia.com/ai-stocks-have-fueled-the-bull-market-for-3-years-will-the-momentum-continue-11820797.
8 We have excluded these holdings from the above graphic because they don’t fit quite as neatly into our toll taker framework. However, we continue to believe they are great businesses with powerful and enduring competitive advantages.
