Stéphane Nières-Tavernier

A look at... Echiquier Artificial Intelligence ǀ May 2025

Stéphane Nières Tavernier, Head of Technology Unit & Manager, Christophe Pouchoy, Manager of Echiquier Artificial Intelligence, La Financière de l’Échiquier (LFDE)

The aim of the Echiquier Artificial Intelligence management team is to identify the most attractive opportunities in artificial intelligence (AI), which over the last few months has been adopted even more rapidly. Apart from actors in cloud infrastructure (semiconductors, data centre equipment, hosting providers), which were the first to benefit, the monetisation of generative AI features has now spread into the software sector with the advent of copilots and AI agents. The AI market could amount to as much as USD 1.81 trillion by 2030 – 5.4 times is size in 2024[1].

TRANSACTIONS

The economy

At the very start of the year, the entire ecosystem was boosted by the announcement of major investment plans for AI infrastructure (including the Stargate project). This was reflected by surging share prices in the sector, which spurred us to take profits on certain overvalued stocks such as Cloudflare and Crowdstrike. Then Deepseek emerged, raising concerns of a “commoditisation” of generative AI models and causing a sharp fall in securities including Vertiv and ARM, in which we had just reduced our positions. We were thus able to take advantage of more attractive entry points to add to our holdings at the end of January.

The fall in the fund between February and mid-April can be attributed both to intrinsic factors, including doubts about spending on AI capacity and its returns, and to external factorssuch as the depreciation of the US dollar and worsening macroeconomic sentiment due to the potential impact of tariffs. Confronted with these uncertainties, in early March we scaled back our investment in companies with the highest valuations (Confluent, Snowflake, MongoDB, Eli Lilly, and Zscaler) in order to reduce the fund’s equity exposure and switch into stocks considered more defensive (Alphabet, Thermo Fisher). In mid-April, after quarterly figures and messages from management gave reassurance on the direct and indirect impact of tariffs, we added to our positions in a few of the less-affected software publishers such as Elastic and Dynatrace. We also brought Marvell (AI semiconductors) into the portfolio.

Fundamentals

Investments in the cloud infrastructure needed to operate AI models (or at least to train them) continue in 2025. The major host providers and main buyers/users of AI servers are still adding new capacity; this is now discounted in the future growth forecasts of Nvidia, Broadcom and other AI ecosystem providers, on which we took profits at the end of 2024 and early 2025. The second wave of AI monetisation is being driven by software publishers (apps, internet services, cybersecurity),with the integration of new AI products (data analysis, copilots, AI agents) making an increasing contribution to their earnings growth. We are therefore maintaining the fund’s high exposure to this sector. Lastly, we are taking a position on the boom in AI models in the real economy via Meta (digital advertising), ARM (Edge AI) and Thermo Fisher (medical robots); at the same time, we are holding on to more defensive stocks in case of a downturn in the macroeconomic environment.

INVESTMENT STRATEGY

In a universe currently consisting of over 900 stocks detected by our semantic algorithms, our strategy is to use our sector expertise to try and identify the best investment opportunities in the AI ecosystem through fundamental financial and non-financial analyses. Of the players that match our four target profiles – AI Sellers, Users, AI Infrastructure Providers and Facilitators – we favour firms that stand to benefit most from the amplifying effect of this universal technology. We have developed in-house tools to pinpoint new opportunities and compare their growth prospects and valuations. Our conviction-based portfolio focuses on some thirty international stocks that we believe are likely over time to tap into the value creation associated with this revolution.

The information, data, stocks and opinions of LFDE provided herein are for information purposes only and thus do not represent an offer to buy or sell securities, investment advice or financial research. Past performance is not a guide to future performance. The fund entails foreign exchange risk, the risk of capital loss, the risks associated with equities, markets and investing in small and mid-cap stocks, discretionary management risk, interest rate risk related to changes in rates, and sustainability risk. For more information on its characteristics, risks and fees, please read the regulatory documents available on our website www.lfde.com.
[1] GrandViewResearch study, April 2025