Citrini Semis
· Citrini Research
· July 20, 2026 at 02:32
· ⏱ 2 min read
| Read on Substack ↗
Summary
The AI selloff is not due to open-source competition or cost pressures but rather a leverage-fueled unwind of crowded trades. The author argues that aggregate token spend continues to grow, and opportunities exist in mis-categorized names such as a mobile-chip designer (likely Qualcomm) and beneficiaries of IBM's earnings miss.
•The recent market correction is attributed to the fragility of leveraged, crowded AI trades rather than fundamental shifts like open-source alternatives or Chinese competition.
•Aggregate token spend is growing rapidly even as per-token costs decline, indicating expanding overall demand.
•The author is specifically hunting for names that have been 'put in the wrong category,' including a mobile-chip designer that turns constraints into advantage and the biggest winners from IBM's earnings miss.
The article's subtitle 'Qualcomm’s Card' and the description 'a mobile-chip designer that turns constraints into advantage' point to Qualcomm as a potentially undervalued AI play, though no position i
The article's subtitle 'Qualcomm’s Card' and the description 'a mobile-chip designer that turns constraints into advantage' point to Qualcomm as a potentially undervalued AI play, though no position is disclosed.
Risk: Qualcomm's mobile-centric revenue could limit AI data center upside; execution on AI inference cards may lag competitors.
The article directly references 'the biggest winners of IBM’s earnings miss,' implying IBM's recent results disappointed and that the miss creates opportunities for others, suggesting a negative view
The article directly references 'the biggest winners of IBM’s earnings miss,' implying IBM's recent results disappointed and that the miss creates opportunities for others, suggesting a negative view on IBM itself.
Risk: IBM's legacy IT services and mainframe exposure may continue to underperform in the AI-driven market shift.
This newsletter, published July 20, 2026,
features Citrini Semis
discussing QCOM, IBM.
2 trade ideas extracted by AI with direction and confidence scoring.