Why we do not invest in semiconductors

Here are a few examples:

Looking at this list, I feel like I am traveling back in time, particularly to the late 1990s!

We did not invest in these stocks back then and we have not done so this time either. Although this decision may make us look wrong in the short term (as was the case in the late 1990s), we do not believe we will change our approach. Here are the main reasons behind this decision:

• Circle of competence. The semiconductor sector is particularly complex. At COTE 100, we have a simple rule: if we cannot properly understand a company’s business model, we prefer to stay away.

• Structurally cyclical. The semiconductor sector is notoriously cyclical, characterized by boom periods where margins are very high and others where most players struggle to generate profits. It is a sector that requires substantial capital expenditures, resulting in high fixed costs, all combined with prices that fluctuate significantly depending on supply and demand (commodity-like). For example, since 2010, Advanced Micro Devices, whose earnings per share are expected to reach a record high in 2026 (after a record in 2025), has reported losses on four occasions.

• Unpredictable industry subject to rapid technological change. Strong demand for chips and high prices encourages many companies to develop their own chips to reduce dependence on certain manufacturers. This is notably the case for Google, Amazon, Microsoft, Meta, and Apple.

• Margins well above historical norms. Taking AMD again as an example: its net profit margin is expected to exceed 26% in 2026 – compared to an average of around 14% over the past 10 years. Micron’s margin is expected to reach 50% in 2026, compared to an average of 17.3% over the last decade. It is difficult to believe that recent margins are sustainable over the long term.

• Dependence on a single theme: AI. Demand from hyperscalers – particularly Google, Amazon, and Microsoft – reflects massive investments in data centers that will increase their capacity to meet AI demand. But what would happen if these companies reduced their investments?

Furthermore, it seems to me that there is currently significant enthusiasm for everything related to AI, which is greatly benefiting semiconductor stocks.

The necessity of an industry does not always translate into a good investment, nor does a revolutionary technology. One only needs to think of the airline, automotive, or telecommunications industries to be convinced of this.

Philippe Le Blanc, CFA, MBA 
Chief Investment Officer at COTE 100