SMIC shares climb on strong Q2 earnings, broader chipmakers up

SMIC rallied after posting sharp Q2 results: revenue rose 36.1% YoY to $3.01B and shareholder profit surged 261.7% to $479.2M, with gross margin expanding to 25.3% from 20.4%. Shares jumped as much as 6.4% to HK$71.9, and the company guided Q3 revenue up 2%–4% sequentially with gross margin of 26%–28%, citing continued AI-related industrial momentum and improved capacity utilization to 93.7% from 92.5%. The strength lifted broader Asian semiconductors, with SK Hynix up ~3%, Micron up 4.2%, TSMC up 1%, and Advantest up ~4%.

Analysis

This reads less like a one-off China beneficiary and more like confirmation that the semi cycle is broadening beneath the AI headline. The important mechanism is utilization discipline: when a foundry can keep fabs hot while lifting mix, pricing power usually migrates first into memory and then into the more levered parts of the supply chain. That makes MU and SKHYV the cleaner expression than TSM, whose quality is already well owned and whose upside depends more on sustained valuation support than incremental demand surprise.

The second-order risk is that the strength is partly a mix/ASP story rather than a real volume inflection. If so, the rally can fade once channel inventory normalizes and sell-side models catch up; that tends to matter most for names with the most operating leverage, not the most scale. Over 6-18 months, rising Chinese domestic capacity can also substitute away some mature-node demand, which is a quieter headwind for non-AI logic/industrial suppliers than for leading-edge AI winners.

Consensus is probably overweighting the idea that "AI demand" lifts everything equally. The better framing is that AI tightness is still most visible in memory and advanced packaging economics, while foundry names are more of a confirmation trade. If TSM monthly revenue or MU pricing commentary fails to validate this over the next 1-2 quarters, the sector move is likely to compress back toward a narrower AI-only trade.

Key Decisions for Investors

• Long MU on 2-3% pullbacks over the next 1-2 weeks; use the next earnings/revision window as the catalyst. Falsify if DRAM spot pricing flattens or MU signals inventory build.
• Pair trade: long MU / short TSM for 1-3 months to express higher operating leverage and lower valuation in memory versus the more fully owned foundry winner. Thesis weakens if TSM monthly revenue reaccelerates materially.
• Add SKHYV as a smaller high-beta confirmation position for HBM/DRAM pricing strength, but size below MU due to FX and liquidity risk. Exit if HBM lead times or pricing commentary soften.
• Avoid chasing direct SMIC OTC exposure; use liquid U.S./ADR proxies instead. If the sector rally persists without follow-through in MU guidance, treat it as a beta trade and reduce.