FFM Fund Newsletter - Feb 2025

February 2025

Dear Friends, Dear Investors,

A few days ago, there came a sudden realisation that a Chinese company, DeepSeek had come out with a few good, efficient AI models at (supposedly) a fraction of the cost of the ones developed by their US competitors.

Since DeepSeek, being Chinese, has less to no access to the latest GPU chips because of US export controls, that came out as quite a shock to investors and American AI companies such as OpenAI. As the saying goes, necessity is the mother of invention and having to make do with less latest generation semiconductors pushed DeepSeek to be more creative in the creation of its AI model. If you want to go into the weeds on all the technical details, here is the indispensable Ben Thompson on all of it: https://stratechery.com/2025/deepseek-faq/.

All of this led to quite the selloff in everything AI-related, be they Nvidia, Taiwan Semi, Schneider Electric, Vertiv and many others, although names like Meta and Apple were up (the thinking being that they would have to spend less on AI than what was originally planned).

When such a shake-up in the markets occurs, one should always let the dust settle before concluding anything with certainty. Is all this an overreaction or the beginning of the end for the AI trade? We shall see, but we would argue that this was an overreaction. This is because of the new buzzword of the day: Jevons Paradox.

The Jevons paradox was first described by the English economist William Stanley Jevons in his 1865 book The Coal Question. Jevons observed that England's consumption of coal soared after James Watt introduced the Watt steam engine, which greatly improved the efficiency of the coal-fired steam engine. Watt's innovations made coal a more cost-effective power source, leading to the increased use of steam engines in a wide range of industries.

This in turn increased total coal consumption, even as the amount of coal required for any particular application fell. Jevons argued that improvements in fuel efficiency tend to increase (rather than decrease) fuel use, writing: "It is a confusion of ideas to suppose that the economical use of fuel is equivalent to diminished consumption. The very contrary is the truth."

This idea, applied to the first industrial revolution, and which turned out to be entirely true, is probably applicable to our current, AI, industrial revolution. As AI tools become less energy-intensive, demand for them will only increase, leading to steady demand for energy, semiconductors and all the other tools needed to power the AI revolution forward.

This is why we believe that the news out of China are probably a net positive for the whole AI trade. The only clear conclusion is that China, far from lagging, is just as good as the US on the AI front, whatever export restrictions were put in place. As a matter of fact, when looking at Nvidia financials, one notices that 20% of revenues come from Singapore. That Singapore would need that many semiconductors seems highly improbable, and one can conclude that some companies in Singapore act mainly as fronts for Chinese companies that then re-export those chips to mainland China.

In this context and despite the DeepSeek announcement in the AI sector, the US markets had a strong month, and Europe continued its excellent progress, which began at the start of this year. Meanwhile, tough negotiations regarding potential entry taxes to the United States are ongoing. It will be interesting to see how this develops over the next few weeks.

Best regards,

Your CaridaB Group Team