When Markets Narrow, Discipline Matters Most

24 nov. 2025

Dear Investors,

The year 2025 has unfolded in a way that tests the patience of any long-term investor. For us, it has been a year of two very different regimes. In the early months, when the United States imposed broad tariffs on much of the world, our portfolio behaved exactly as we hoped it would. Diversification, balance-sheet discipline and factor exposures that tend to be resilient in stress meant that we navigated the tariff shock reasonably well. But markets do not sit still, and as soon as that narrative died down and investor enthusiasm swung back to artificial intelligence with renewed force, the regime shifted sharply. In that environment, narrow mega-cap leadership returned with a vengeance, and our more diversified, fundamentals-driven approach has lagged. It is never pleasant to fall behind in markets that appear to reward the most concentrated and speculative exposures, but we have seen similar climates before — and they have never persisted indefinitely.

More recently, November has offered a welcome reversal, with our flagship fund ahead of the global equity index by several percentage points and our hedged strategy performing particularly well — though in a market this narrow, we are mindful that such windows tend to stay open only until the next AI earnings surprise by even a fraction of a percent!

The current AI enthusiasm has revived a dynamic reminiscent of the “vendor financing” era of the late 1990s, when telecom companies extended credit to customers who used that credit to buy even more telecom equipment. Revenues rose, orders grew, and investors applauded numbers that were flattered by circular flows of capital. Today’s version of that behavior is subtler, but the echoes are hard to ignore. We now see a small circle of companies making substantial, often interdependent investments in each other's AI capabilities — cloud capacity, generative-model development, chip supply, training infrastructure and data-center build-outs. Each company’s spending becomes another’s reported revenue; each commitment justifies the next. What emerges is an ouroboros-like system in which the same pool of capital recirculates through a handful of firms and appears as growth from every angle.

This dynamic has created a difficult environment for the factors that underpin quant investing. Value, conservative investment and even profitability — bedrock exposures that have delivered robust long-term premia — have struggled. Most of these factors are down meaningfully this year, leaving momentum as the only consistent outperformer in a market ruled by excitement rather than valuation discipline. It is no surprise, then, that we receive more questions about whether quant investing still works. We ask those questions ourselves from time to time; like Michael Burry famously admitted during his housing-market ordeal, there are moments when irrational markets tempt even the most rational among us to throw up our hands. But the discipline behind systematic investing does not disappear because sentiment temporarily sways elsewhere. On the contrary, these are precisely the moments when discipline matters most.

History offers helpful perspective. The late-1990s bubble was a challenging period for value-oriented factors. Value and conservative investment were relentlessly punished as investors became convinced that companies investing aggressively in the internet would inevitably become the long-term winners of a transformative technology. Profitability and capital discipline were ignored; and the belief that a select group of fast-growing firms could capture most of the economy’s future income became mainstream. That conviction proved fleeting. Years later, the companies that had been hailed as the “certain” winners were not always the ones that ultimately captured the economic benefits of the internet. Often, the adopters — not the infrastructure builders — earned the higher returns. And in the aftermath of that cycle, the very factors that had been left for dead experienced some of their strongest multiyear runs on record.

It would be comforting to say that today’s environment is entirely different, and in some ways it is. Many AI leaders today are genuinely profitable; the technology is real and economically meaningful; and the corporate balance sheets are stronger than those of the dot-com hopefuls. But in other ways the parallels are too meaningful to ignore. Market concentration has reached levels that should make any risk-conscious investor uneasy. NVIDIA has grown to roughly six percent of the developed equity market, reflecting an expectation that its technological edge will persist indefinitely and that AI-fueled demand will support extraordinary long-term growth. The top ten companies represent almost thirty percent of the entire world’s market capitalization, and nearly every one of them has committed aggressively to AI. The market is effectively making a collective bet that these few companies will not only commercialize AI successfully, but will capture a disproportionate share of all future economic benefits. That may turn out to be correct — but history suggests that such outcomes are rare, and that the benefits of foundational technologies tend to diffuse broadly across thousands of firms rather than consolidating in a handful.

This environment has left a clear imprint on factor behavior. The 2023–2025 window has produced yet another spell of unusually poor outcomes for the traditional quant factors: size, value and conservative investment have all performed poorly as capital has flowed almost mechanically into the largest, capex-devouring, AI-exposed companies. A chart of cumulative factor performance over this period reinforces what we feel day-to-day: these premia have been under sustained pressure. It is little consolation in the short run, but historically such periods of factor underperformance have occurred almost exclusively during speculative surges, not during periods in which fundamentals deteriorated — a point made clear by the broad set of attractively priced, fundamentally sound companies we identify each day and that are poised to benefit once the market’s attention returns to fundamentals.

The graph shows how the market (long-only) and several long–short factors behaved during speculative phases of the market: the dot-com boom, the AI boom and the dot-com bust. The “Small Caps” factor, for example, is constructed by shorting the largest 30% of global equities and using the proceeds to buy the smallest 30%. An underperformance of this factor therefore indicates that large caps are outperforming small caps. The same long–short logic applies to the other qualitative factors displayed.

The role of sentiment becomes even clearer when we split factor performance across the past decade into up- and down-market months. In rising markets — especially those driven by AI enthusiasm — factors such as value and conservative investment suffer the most, while in declining markets the pattern almost perfectly reverses.

The graph shows how several long–short factors have behaved in up- and down-market environments. Value and low-investment stocks tend to outperform growth and high-capex stocks during weaker markets, while the opposite is generally true in rising markets. Some factors, such as profitability, have delivered positive results across both regimes, whereas others — most notably size — have struggled consistently, with mega caps outperforming small caps regardless of overall market direction.

This leads us, inevitably, to the question of whether fundamentals will matter again. The honest answer is that we do not know when they will reassert themselves, only that they always have. If the massive AI investments paying for the current valuations truly do produce winner-takes-all economics, our more conservative, fundamentally anchored style will continue to lag. But for that scenario to play out, the world would need to converge into a market where a tiny number of companies — each already commanding enormous weight in global indices — become an oligopoly representing an outsized share of all global equity value. It would also require flawless execution, manageable energy constraints, a smooth and affordable expansion of data-center infrastructure, and an absence of meaningful competitive pressure. Even today, we already see that competition materializes quickly: China has replicated key capabilities at remarkable speed, and other new entrants are emerging with lower costs. The notion that one or two model developers will indefinitely dominate an entire technological field is possible, but far from guaranteed.

Experience suggests that enthusiasm eventually gives way to analysis. After the dot-com bubble burst in March 2000, the broad market declined over sixty percent on a risk-adjusted cumulative basis in the subsequent three years drawdown. Yet during that same period, the very factors that had lagged for years produced exceptional returns. These were not small rebounds. They were powerful reminders that the economic foundations underlying factor premia do not disappear simply because narrative momentum temporarily overwhelms them. Investors who held firm through those years, or who added exposure during the most difficult stretches, were rewarded handsomely as quant investing turned out to be the best hedge against the aggressive market downturn.

The graph above shows the cumulative performance of the market net of the risk-free rate (long-only), alongside the returns of the four long–short portfolios used in the Fama–French five-factor model. While it is clear that most quantitative factors struggled during the dot-com boom, it is equally evident that they benefited significantly from the subsequent market rotation once the bubble unwound. 

We do not claim that history will repeat itself. But we do recognize the patterns of sentiment, concentration and narrative certainty that have preceded major reversals in the past. AI is undoubtedly transformative, and we are not rooting against the companies building it. But we caution that the current market structure exposes passive investors to extraordinary concentration and thematic risk. By contrast, diversified factor exposure — though uncomfortable during periods of strong thematic rallies — remains one of the most reliable ways to counterbalance that risk. Our recent underperformance is real, and we are not indifferent to it. But we remain confident in the long-term power of fundamentals, and in the resilience of the factors that have delivered durable returns across market cycles for decades.

When the pendulum eventually swings back toward valuation discipline, as it has in every prior cycle, we expect our strategy to be well positioned to benefit from the normalization that follows. Until then, we thank you sincerely for your patience, your confidence and your partnership.

Best regards,

Werner, Thierry & Nils