Navigate Download

Market Commentary— Two Markets, One Year

Over the past few editions of The TANDEM Report, as well as within our other commentary pieces, we’ve spent a lot of time discussing the narrow focus of the S&P 500. This concern persisted throughout the fourth quarter as well. Ultimately, the S&P 500 ended the year up nearly 18% including dividends, but unlike the past couple of years, 2025 was a tale of two markets.

Over the first 100 days of 2025, the AI trade had fallen out of favor and the Magnificent Seven stocks stumbled out of the blocks. Trade policy out of Washington garnered much attention and blame for the index’s first quarter woes. The S&P 500 dropped more than 20% at its lows, the tech-laden Nasdaq fell more than 25%, while the Bloomberg Magnificent 7 index plunged roughly 30%. Part of this was trade/tariff related. However, one of the first dominos to fall had actually occurred in January in what was dubbed AI’s “DeepSeek” moment.

DeepSeek, a Chinese AI start-up, released an AI model which rivaled U.S. models. That would not necessarily be alarming in and of itself. However, the China-based company matched U.S. models from OpenAI (ChatGPT) or Anthropic (Claude), while spending a fraction of the cost. The implication was alarming. U.S. companies might have burned through hundreds of billions of dollars unnecessarily.

The DeepSeek moment, coupled with rising trade tensions that culminated with President’s Trump’s “Liberation Day”, rocked markets. But the Trump Administration quickly pivoted and following the rollback of some of the strongest threatened trade policy, markets staged a strong rally. Risk assets, specifically, rallied tremendously. As an example, the S&P 500 High Beta Index (which is designed to measure some of the most volatile names in the S&P 500) gained more than 50% in just 4 months. The aforementioned Magnificent 7 Index gained nearly 50%. Riskier pockets of the market were red-hot through the summer, while more cautious pockets, like the S&P 500 Low Volatility Index (which is comprised of much less volatile companies) gained only single digits.

From there, the divergence between risk-on and risk-off assets only accelerated and became more extreme. From August 22nd through October 29th, the market experienced a complete melt-up. During that time, the S&P 500 gained 6.5% while the Magnificent 7 rallied 15.5%. The S&P 500 Low Volatility Index fell 5.3% during that stretch, which is highly unusual behavior. Typically, the S&P 500 Low Volatility Index is directionally correlated with the broader market, just with less magnitude. That is to say, if the S&P 500 is up, it is typical to see the Low Vol Index also up, just perhaps less. If the S&P 500 is down, it is not surprising to see Low Vol also be down, but perhaps nearly not as much. Historically that has also been the case with Tandem.

The nearly 12% return differential between the S&P 500 and the Low Vol Index has historically occurred only in extreme environments: the final months of the dot-com bubble in late 1999 and early 2000, following the market downturns after the dot-com bubble burst and after the Financial Crisis, as well as following the Covid bottom. If history is any guide, such divergent behavior signals either a blow-off top or a generational buying opportunity. Given current valuations, persistent concentration, and the market’s singular thematic focus, I believe the divergence is more likely to be indicative of a blow-off top. It certainly does not appear to be a generational buying opportunity.



Source: FactSet

Note: This data is derived by subtracting the 2-month return for the S&P 500 from the 2-month return of the S&P 500 Low Volatility Index. For example, if the S&P 500 Low Volatility Index returned 10% and the S&P 500 returned 15%, then the 2-month return differential would be -5%. If the S&P Low Volatility Index returned 5% and the S&P 500 returned -5%, then the 2-month return differential would be positive 10%.

The tide turned in late October and the market favored the consistent, less-volatile companies Tandem seeks throughout much of November. The underperformance of the Low Volatility Index relative to the rest of the market reversed. By late November, the Magnificent 7 had fallen as much as 8% from its October 29th peak, while the Low Vol Index posted positive returns.

Helping drive the underperformance of AI was another development which reinforced doubts about AI spending requirements. Google released Gemini 3, which outperformed ChatGPT on key benchmarks, and it did so without using Nvidia chips. Google instead relied on its proprietary chips. Without any dependence on Nvidia, the hardware linchpin of the AI boom, Google bested OpenAI’s flagship model. While Google isn’t yet producing chips at scale, the strategic implication echoes DeepSeek – perhaps the massive capital expenditures aren’t necessary to achieve superior AI results.

Through year-end, the S&P 500 recovered its short-term losses, posting new all-time highs just after Christmas.

The market’s movements over the past two to three years have been overwhelmingly AI-driven. Recent months have brought intensifying debate over an AI bubble, focused on enormous spending coupled with lofty valuations. The DeepSeek and Google developments suggest that perhaps, just perhaps, this spending spree isn’t as essential as the market had assumed. Whether that thesis proves correct will likely define the market narrative for 2026.

Commentary— Investing in a Market of Speculation

We’ve never viewed it as our job to chase every rally. We’ve always sought to preserve and grow the capital that you’ve entrusted to us. Right now, that means doing something uncomfortable – staying disciplined while speculation runs rampant.

The S&P 500 has become unrecognizable to us. Ten companies (which make up just 2% of the index) now represent nearly 40% of its value – the highest concentration in at least 50 years. Since October 2022, ten companies account for nearly 60% of the S&P 500’s total gain. That’s not diversification. It’s become a concentrated bet on a single theme: Artificial Intelligence.



Source: FactSet

*Alphabet includes both Class A and Class C shares in the table above.

We believe this concentration represents a bubble. We’ve chosen to stick to our discipline rather than throw in the towel and chase this bubble.

History has taught us that speculating and investing yield vastly different outcomes. Speculators buy for the sole purpose of selling higher. Oftentimes the speculator is investing in the story, the hype, or the dream. The speculator is indifferent to the underlying business or its price. Instead, the speculator, whether consciously or not, operates under the maxim of the “Greater Fool Theory” – believing that they can purchase even an overvalued asset because another buyer will pay more for it later. In turn, the speculator’s success is dependent almost entirely upon another purchaser willing to pay a higher price.

Eventually, the music stops playing, the story dries up, and some speculator somewhere is left holding the bag.

Investors, by contrast, often look for the opportunity to purchase stakes in growing and/or undervalued businesses with the expectation that these businesses will continue to thrive. When the investor finds such an opportunity, she commits capital to that business with the expectation that the business will prosper and grow, which in turn would reasonably be expected to lead to a rise in the value of the business.

Today’s market rewards speculation. Among the best performing baskets of stocks this year have been things like “Meme Stocks,” “Unprofitable Tech” companies, and Quantum Computing. While these stocks have compelling stories, prices are seemingly not supported by any sort of fundamentals.

Consider these examples: There is a lithium-battery company that, at one time in the year, sported a market cap of $10 billion and was up 250%. This company has never sold a product. They have $0 in revenue in nearly five years as a publicly traded company—yet “the market” deemed it to be more valuable than Gap or American Airlines. There is another company making flying taxis with a market cap of nearly $12 billion, up more than 160% in 2025, with less than $100,000 in sales over its past 12 months. To this writer’s eye, both speak to exuberance, not rationality.

The market is also historically concentrated in one single trade – AI. If the trillions of dollars being spent on AI infrastructure does not bear significant income-producing fruit, then these names become hazardous. Morgan Stanley predicts $2.9 trillion in AI infrastructure spending through 2028, while Nvidia expects data center capex to reach $3 to $4 trillion by 2030. Yet an MIT study found that only 5% of businesses were generating “value” from implementing AI, while 95% saw either zero return or no measurable impact. Recently, the market has started to differentiate among the previously homogenous AI names. The share prices of those taking on debt to finance their massive AI build-out, like Oracle most recently, are being treated harshly when compared to their less indebted peers.

The common pushback on recent bubble talk has been that today’s tech giants are all highly profitable companies, unlike the dot-com companies of 1999. This is misleading. Yes, there were unprofitable names running rampant in ’99. However, the so-called “Four Horsemen” (Microsoft, Intel, Dell, and Cisco) in 1999 were all highly profitable large cap stocks. Comcast, Amgen, Qualcomm, Sun Microsystems, etc. were all credible businesses. That did not insulate these companies from sharp pullbacks. The dot-com bubble was inflated on the backs of profitable business on the way up, but as the bubble began to deflate those profitable businesses led the way back down. Why? Because the price one pays for a security matters.

Despite all of this, I still believe Artificial Intelligence seems likely to be game-changing technology. Similarly, the internet ended up being game-changing technology in the late 90s. However, the price you pay for a stock, or an index fund, matters. Today’s market, in my view, parallels the market of the Tech Bubble. Investors got the thesis correct. The internet changed the world. However, it turned out that price mattered. The S&P 500 produced an annualized return of less than zero for nearly a decade after the market peaked in early 2000. That’s a long time. Valuation and price ended up mattering more than hype. I believe we are in a similar place today.

At Tandem, we invest. We do this by scouring our investable universe of over 3,500 stocks for companies that can consistently grow their businesses through different economic cycles. Whether it be a financial crisis, a global pandemic, inflationary environments, or through global trade wars – we seek to identify companies whose business has done relatively well in good times and bad. Oftentimes that leads us to investing in companies that are remarkably consistent, even if unexciting. There is much to be said about a business model as resilient as Costco, the duopoly that is enjoyed by the likes of Mastercard and Visa, or the remarkable consistency of a waste management company like Republic Services or Waste Connections. However, these are not the stories that speculators seek, though they are the stories that consistent investors need.

Are we exposed to AI? Yes. Depending on the Tandem strategy you employ, there is some AI related exposure in the form of stocks like Microsoft or Amphenol. Furthermore, a number of our companies are implementing AI to help with cost savings measures. But we are not placing 40% of the capital you’ve entrusted with us on a single bet – no matter how compelling the narrative. That’s just not prudent risk management. As noted earlier, the S&P 500 has much more exposure to AI through its Magnificent 7 names, plus a few other highflyers like Broadcom and Palantir. Those stocks add up to nearly 40% of the S&P 500 today. We believe that we are not particularly representative of the S&P 500 today, so it is not our expectation that we will perform similarly– for better or worse.

So where does this leave us moving forward?

In a card game like poker or blackjack, we play the hand in front of us and do not concern ourselves with the last one or the next one. Yet in investing, investors often tend to look to the past for guidance about the future. There is an overwhelming tendency to select the best trailing performance. That would be brilliant if we knew that the next period was going to look just like the last one.

Today we find ourselves in a difficult market for our discipline. Does this mean we have made poor decisions? No, it does not. A bubble has inflated. Our quantitative discipline has by and large avoided the bubble. Making a good decision does not mean it leads to a good outcome every time. Sometimes we are unlucky. Similarly, making a poor decision does not guarantee a poor outcome. Sometimes speculators are lucky.

We believe we own good quality companies, and that our discipline has served us well for 35 years. For the last few years, the market has valued something other than what we do. Some would argue that chasing a bubble is a low-quality decision. However, even a low-quality decision can work out well sometimes. Over time, when repeated decisions are made, high-quality decision-making leads to success more often than does low-quality decision-making. Ask any good poker player. Eventually luck evens out and quality decision making prevails.

I hope that we have served clients well for 35 years not by chasing bubbles, but by avoiding them. Our discipline doesn’t always feel good in the moment. Sometimes the temptation to go with the crowd can be strong. But that’s not what we are here to do.

Owning companies that have historically displayed the ability to grow through different economic cycles is a good quality decision. Benjamin Graham, often seen as the father of value investing and to the mentor of Warren Buffett, once said that one should never buy a stock because it has gone up or sell one simply because it has gone down. Instead, investors should be comfortable owning a stock as if they had no way of knowing its daily share price.

We remain committed to this philosophy. We’re not trying to beat your neighbor, another manager, or even best the market. Investing is about preserving and growing capital over time – through discipline, patience, and focus on consistent businesses.

Tandem Update

Recently we have fielded a number of questions regarding the divergence in performance between our Tandem strategies and the broader market. One of the messages that we have tried to convey in our commentary pieces is that we are not the market. We are not invested like the market. Therefore, we do not expect to behave like the market. When looking at relative performance, there are times where that difference is to our client’s advantage and times when it may not feel so.

Looking forward, our game plan for 2026 remains unaltered. We are disciplined investors. Our discipline seeks stocks that can consistently grow revenue, earnings, cash flow, and dividends (if paid) throughout an economic cycle. That eliminates a lot of companies. That also eliminates a lot of companies that have been leading this market higher. Nvidia, for example, didn’t grow its dividend from 2019 to 2024, it then grew it in 2024 and hasn’t grown it in the subsequent 6 quarters. There is nothing consistent about that growth – we couldn’t own it if we wanted to. Meta basically lit its free cash flow on fire when it pursued the Metaverse and is now burning free cash at a quick pace once more in pursuit of AI – maybe it works out for Meta, maybe it doesn’t, but either way it doesn’t meet our quantitative criteria so we can’t own it. Tesla does not have the consistency we value. Microsoft meets our criteria, and we love Microsoft. We’ve owned it at Tandem for decades. However, it’s a full position in both our Large Cap Core and Equity strategies. No matter how much we love it, we are unlikely to ever match the S&P’s 6% weighting in Microsoft. So, no matter how well Microsoft does, we are always losing attribution relative to the market within that stock.

Across our three strategies, we own 45 separate companies. Of those, 31 are on pace to grow earnings by at least 10% in 2025 – a feat that we are quite pleased by. Of those 31 names, nearly ½ of them fell in 2025. That seems unsustainable. In the short-run, fundamentals do not always matter. Over the long-run, the market is driven by earnings growth.

Our approach differs markedly from the market’s current hyper focus on AI. We are not likely to change our stripes. We will continue to seek companies that can meet our stringent criteria of consistent growth. I suspect that stocks which grow double digits will eventually return to favor. We’ve been doing this for 35 years and we’ve seen markets before that have behaved in this manner. Ultimately, the price one pays for a stock matters, hype eventually subsides, and fundamentals typically pave the way higher.

Download
Back to Archives

Disclaimer: Tandem Investment Advisors, Inc. is an SEC registered investment advisor.

This audio/writing is for informational purposes only and shall not constitute or be considered financial, tax or investment advice, or an offer to sell, or a solicitation of an offer to buy any product, service, or security. Tandem Investment Advisors, Inc. does not represent that the securities, products, or services discussed in this writing are suitable for any particular investor. Indices are unmanaged and not available for direct investment. Please consult your financial advisor before making any investment decisions. Past performance is no guarantee of future results. All past portfolio purchases and sales are available upon request.

All performance figures, data points, charts and graphs contained in this report are derived from publicly available sources believed to be reliable. Tandem makes no representation as to the accuracy of these numbers, nor should they be construed as any representation of past or future performance.

This document was originally written/recorded in English. Tandem does not guarantee the accuracy, completeness, or reliability of any translated materials, and shall not be held responsible for any discrepancies, errors, or misinterpretations arising from the translation process.