Wall Street tech analyst explains why investors should look past Samsung selloff

Semiconductor stocks fell as investors digested Samsung’s preliminary Q2 results, but Mizuho’s Jordan Klein says the sell-off looks like a momentum unwind rather than fundamentals weakening. He argues Samsung’s operating profit was a strong beat (excluding one-time bonus expenses) and highlights memory operating margins implied at over 80%, with DRAM/NAND pricing in Q3 tracking very strong vs Q2 levels and NAND forecasts up 35–40% QoQ. Klein adds that upcoming TSMC/ASML results should matter more than Samsung prelims for semi price action, while citing attractive valuations and a bullish server CPU outlook supporting Intel and DRAM suppliers.

Analysis

This looks like a classic momentum de-risking rather than a clean fundamental reset. The market is probably over-penalizing anything with “semiconductor” in the name, but the real signal is that AI demand is increasingly being separated from legacy handset/TV/display exposure; that favors foundry/compute leaders over diversified hardware names. In the next 1-2 weeks, the key catalyst is not Samsung’s print itself but whether TSM and the equipment supply chain confirm that advanced-node demand and capex are still intact.

The second-order effect is that a strong memory upcycle can actually be bullish for certain CPU and server silicon names: rising DRAM/NAND pricing tends to improve OEM willingness to spend on higher-ASP systems and can support server refresh cycles, which is why INTC deserves attention despite its own execution issues. For AMD, the setup is more event-driven: into the July 23 AI event, the market can re-rate if management narrows the gap between product ambition and monetization, especially if sentiment has been washed out by the broader semi tape.

Contrarian view: the consensus is likely missing that this is a breadth issue, not an AI demand issue. If Samsung’s weakness is concentrated in non-memory businesses, then the stock reaction may be a bad read-through for TSM and the premium AI complex; however, if July memory commentary disappoints or TSM softens on orders, the current dip-buying narrative breaks quickly. Over 6-18 months, the risk is that AI capex remains strong but margins get competed away, so the winners are those with pricing power and clean exposure to advanced process demand, not the broadest exposure to semis.

Key Decisions for Investors

• Buy TSM into any 1-3 day post-Samsung weakness; target a move higher into the next earnings/guide print, with the thesis invalidated if TSM comments suggest advanced-node order moderation or AI capex digestion.
• Add AMD on weakness ahead of the July 23 AI event, preferably via calls or a call spread to limit premium bleed; this is a 2-4 week catalyst trade with upside if management upgrades AI revenue visibility, and invalidated if the event is heavy on TAM but light on bookings/margins.
• Initiate a tactical long INTC position as a 1-3 month CPU-cycle expression, since rising server CPU demand can improve mix and narrative more than the market expects; keep sizing modest because execution risk remains the dominant brake on multiple expansion.
• Use any rebound in the crowded AI semiconductor basket to trim exposure rather than chase it; if TSM/ASML commentary fails to confirm the tape, the next leg is likely a de-momentum unwind rather than a fundamental drawdown.
• Watch SSNLF only as a sentiment canary, not a high-conviction short; if the stock stabilizes after management commentary, that would confirm the market has over-discounted legacy weakness and is misreading the memory cycle.