DEEP RESEARCH · SEMICONDUCTORS
[Semiconductors] Thoughts on DeepSeek and the Semiconductor Sector
A risk-management note on increased semiconductor uncertainty after DeepSeek and the decision to raise software exposure
0. Bottom line first
After DeepSeek, I saw greater uncertainty in the semiconductor sector, reduced semiconductor exposure, and increased software exposure. Because I do not feel I understand this field deeply enough, I judged that reducing weighting was the right response when conviction was low.
1. Purpose of the note
I am writing this simply to record my own thoughts. Since I feel my understanding of this field is limited, I wanted to reduce exposure when uncertainty increased.
Interpretation: The core issue is risk management, not confidence in a forecast. Rather than focusing on whether I am right or wrong, I prioritized reducing portfolio exposure to risks I do not sufficiently understand.
2. Portfolio adjustment after DeepSeek
Official fact: The author wrote that, because of DeepSeek, he reduced exposure to semiconductors, which he viewed as being in a highly uncertain environment, and increased exposure to software.
- Semiconductors: reduced weighting
- Software: increased weighting
- Glass substrates: retained
Interpretation: Glass substrates were retained because the author viewed them as not yet a mass-production technology and not something that appears easy to sanction at this point.
3. Risk judgment
A rebound from excessive declines may happen. Still, I viewed the risk as larger because the United States will likely try to prevent China from succeeding in some way, which could keep pressuring companies in this field.
Interpretation: On that basis, I judged that watching from the sidelines to some extent and increasing software exposure, where this risk is lower, could offer a better risk-reward balance.
4. Conditions for increasing exposure again
This view could of course be wrong. But the important point is not whether it is right or wrong; it is that risk should always be considered first.
- When I judge that there is no further negative news left to come
- When even new negative news no longer drives further declines
- If that moment comes, I plan to increase semiconductor exposure again