FFM Fund Newsletter - Jan 2026

January 2026

Dear Friends, Dear Investors,

“Life is just one damn thing after another.”

John D. Archbold, Standard Oil executive upon hearing of the news of the US Supreme Court’s order to dismantle the Standard Oil Trust, later president of Standard Oil of New Jersey (now known as Exxon).

Dear Friends, Dear Investors,

If there is one year for which the above quote seems particularly fitting, it is 2025. Events unfolded at such a pace that it is hard to remember them all. From DOGE attempting to dismantle the U.S. government apparatus, to “Liberation Day” (the day we realized Donald Trump wasn’t joking about tariffs), the ongoing war between Russia and Ukraine, the U.S. bombing of Iranian nuclear facilities, and the AI revolution, it has literally been one damn thing after another.

2026 begins at full speed geopolitically, with U.S. intervention in Caracas and increasingly large protests in Iran. However, our focus here is not geopolitics, but economics. There are two main points we wish to address in this letter: the current frenzy around AI (or is it a bubble?) and the question of whether the United States remains the best place to invest. As most AI-related investments are U.S.-based, these two topics are closely linked and likely the most important for our future returns.

Let’s start with artificial intelligence (AI). This is a theme we have written about extensively in our previous letters, and which we have also discussed in depth with you and during our webinars. In fact, most of our exposure (Taiwan Semiconductor, ASML, Amphenol, BWX Technologies, Alphabet, Meta, etc., except Nvidia, which we added later) was already in place before the AI frenzy began with the launch of ChatGPT at the end of 2022. This exposure was subsequently significantly amplified.

As we have mentioned before, it is impossible to know who the winners will be, which AI model will dominate, or even whether the current levels of investment make any financial sense. However, what does seem relatively clear from a valuation perspective is that we are in a bubble. This is not necessarily a bad thing; one just needs to be aware of it. The real question is: at what stage of the bubble are we?

As we wrote in 2024, the hyperscalers (Microsoft, Alphabet, Meta, Amazon) are engaged in a Red Queen race, where you must run just to stay in place. Since Q4 2022, these four giants have spent $1.3 trillion on capex and R&D, almost exclusively related to AI. These are almost unimaginable sums. To better understand this, one only needs to look at technology investments as a percentage of GDP for 2025, compared with other major U.S. projects of the past (chart courtesy of JP Morgan’s Eye on the Market).

If you look closely at this chart, the gray bar (the third from the right) represents fiber broadband investments in 2000. As many compare the current AI bubble to the internet and telecom bubble of 1995–2000, it can be alarming: the black bar (current AI investments) is already higher than at the peak of the previous bubble in 2000. This is somewhat concerning, but we must remember that the current AI bubble is extremely capital-intensive. You cannot simply create a large language model (LLM) in your garage with a friend, as one could create a website in the 1990s. So, again, nothing truly alarming… for now. This is confirmed by valuations. Although high, they remain far from the peak of the previous bubble, as can be seen here (still courtesy of JP Morgan’s Eye on the Market).

Our conclusion is therefore as follows: yes, we are in an AI bubble; yes, this bubble will burst, as all bubbles do; yes, it will end badly for the sector, but not immediately. At the start of this new year, despite geopolitical maneuvers, it is difficult to be too pessimistic. Both the U.S. and Europe are in a phase of fiscal stimulus and engaged in a cycle of falling interest rates. Certainly, there are long-term issues, such as public deficits and, particularly in the U.S., central bank independence, but these are unlikely to be problems in 2026.

That said, while we are aware that we are operating in a bubble environment, we must remain particularly vigilant. The goal is not to try to exit at the perfect moment—an almost impossible exercise—but rather to be ready to adjust our exposure when fundamental signals begin to deteriorate. In other words, we will likely need, at some stage, to reduce our exposure to the tech sector based on the evolution of company earnings and valuations. This will probably first apply to the hyperscalers (Microsoft, Alphabet, Meta, Amazon).

To understand why, we need to return to the original reasons for investing in these companies: they were all monopolies (or duopolies in the case of Alphabet and Meta in advertising), each staying in its lane without competing too aggressively with one another. They generated massive amounts of cash and did not need to spend much, allowing them to buy back shares and increase returns for investors. They continue to generate large amounts of cash, of course. But now they are engaged in a fierce war to develop the best LLMs. To do so, they must reinvest heavily in extremely costly capital expenditures. While earnings may still rise, free cash flow no longer follows. Our initial investment thesis is therefore no longer fully valid, as is clearly illustrated in the following chart:

This does not mean their stock prices cannot continue to rise, especially if investors continue to view LLM investments positively. It simply means that at some point, these companies will need to demonstrate a return on investment from these massive expenditures.

For now, we see no evidence of that. It is also important to note that these investments have mostly been funded from retained earnings, not debt or external financing, which distinguishes the current situation from the 2000 bubble, for example. We discuss it now precisely because this is starting to change, as can be seen in the following chart:

Debt inevitably appears at some point in a bubble, and it is usually this debt that triggers the inevitable collapse when it can no longer be refinanced. We are not yet close to that moment of truth, but we will monitor its evolution very closely.

Finally, we will likely reduce these companies first because, fundamentally, they are at war and are buying the weapons. We strongly prefer to liquidate the buyers of weapons (especially since the outcome of this war is unpredictable) and keep the suppliers, such as ASML (lithography), Nvidia (chip design), Amphenol (semiconductor connectors), or Taiwan Semiconductor (chip production). We feel we will hold ASML and Taiwan Semiconductor for quite some time, as the rest of the chain simply cannot function without them. But at some point, it will likely be necessary to reduce exposure across the entire value chain. For now, our approach is to gradually reduce AI exposure, like peeling an apple.

Now to our second topic: is it still relevant to remain invested in the U.S. medium to long term? With concerns about AI, this was the most frequently asked question by our clients in 2025. This stems from various Trump administration decisions, particularly on tariffs. The U.S. seems to have adopted a different approach to international relations and the post-WWII world order. It is also clear that the Federal Reserve will no longer be as independent as before, starting this year. Rate cuts are now the norm, regardless of inflation figures. In truth, no central bank is ever fully independent of political power, especially in the U.S., but this independence is clearly diminishing. Meanwhile, the U.S. fiscal deficit, above 6% of GDP, shows no signs of shrinking.

Looking at 2025, there is no indication that foreign investors have fled the U.S. market. How could they? The U.S. market represents nearly 70% of the MSCI World Index, and most AI-related stocks (except ASML) are listed in the U.S. There is also no sign of domestic capital flight. There was, however, a currency adjustment. In 2025, the U.S. dollar lost 12.65% against the Swiss franc and 39% against gold. This is normal: forex markets are the most liquid in the world.

For our part, we have always been underweight in the U.S., with most portfolios roughly evenly split between the U.S. and Europe (broadly defined, including the U.K., Nordics, and Switzerland). But, as with AI, it will be interesting to monitor real opportunities coming from the U.S., particularly for the type of companies we favor: sector leaders in growing markets (economic assets increasing 12–15% per year), with high margins (often over 50%), low debt, and strong reinvestment capacity, as reflected in high ROIC and ROCE, typically implying limited or no dividend policies.

Of course, some of these companies can practically only exist in the U.S., due to market size, venture capital depth, and the tech ecosystem. This is especially true for broad tech, large payment platforms like Visa or Mastercard, and certain very specific business models, such as Uber or large oilfield services companies like Schlumberger. For other activities meeting the same quality criteria (structural growth, high margins, reinvestment capacity) Europe could gradually offer increasingly attractive opportunities, as valuations normalize and local champions emerge or strengthen.

When you think about it, why have U.S. markets so consistently outperformed the rest of the world since 2009? Mainly because innovation and GDP growth were higher there. That hasn’t changed, but it could over time once the AI bubble bursts. Given the scale of AI-related investments, one could argue that once this phase ends, U.S. GDP growth could converge with Europe’s (below 1%).

We do not believe there should be zero exposure to the U.S., but it is reasonable to consider that, once the AI bubble bursts, and given the disproportionate weight of AI companies in the S&P 500, U.S. markets (measured by major indices) may stagnate for some time. If this seems extreme, remember it took fifteen years for the Nasdaq 100 to reach a new high after the 2000 peak. Like our AI exposure, this transition will be gradual, but it will happen.

Finally, a third point, as a bonus. It is perfectly acceptable to be invested in a bubble for part of a portfolio, provided one understands how it ends. That is where the money flows, and where investments are made. But it is also useful to ask the reverse question: which sector has seen almost no investment over the past ten years? One such sector is oil and natural gas. After a massive investment bubble between 2010 and 2015 (notably in U.S. shale oil and gas) and multiple bankruptcies, the survivors have since refrained from investing in new developments, reduced debt, and returned cash to shareholders via dividends and buybacks.

And so here we are today, with most states realizing that energy needs are increasing and that oil and natural gas remain essential parts of the energy mix. Everyone talks about the return of nuclear, but these are massive investments taking years to materialize. In the meantime, oil hovers around $60 per barrel, and no one is optimistic about the sector. It is in this context that we began buying certain oil-related stocks (Schlumberger, Exxon, TotalEnergies, Gaztransport & Technigaz).

We must remember that oil is cyclical, and if we are at the start of a new supercycle in oil and gas, this is a business that can generate extraordinary free cash flow. Adjusted for inflation (base year: 2008, U.S. CPI), $60 per barrel today equals roughly $40 in 2008. We believe the 2008 highs, around $140, could be reached and likely surpassed before the end of this cycle.

As always, and to paraphrase Mr. Archbold, we will take one damn thing after another.

We wish you an excellent 2026.

Best regards,

Your CaridaB Group Team