Stock Movers: SK Hynix, MUFG, Baidu (Podcast)

SK Hynix extended a two-day slump in Seoul to more than 20% as a selloff in global AI hardware stocks reignited concerns that the chip rally ran too far, too fast. MUFG shares surged to the highest market value in Japan, reclaiming top spot among the three megabank groups. Baidu fell in Hong Kong on mounting worries about upcoming earnings after an analyst preview.

Analysis

The AI hardware tape is shifting from fundamentals to positioning. SK Hynix’s drawdown matters less as a read-through on end-demand than as evidence that the most crowded part of the AI complex is being de-grossed; that usually spills first into memory peers and semiconductor equipment before it shows up in customer capex. Near term this is a technical air pocket, but over 1-3 months the market will test whether hyperscaler spending and HBM pricing still justify 2026 margin expectations.

MUFG’s rise is a cleaner macro signal: investors are paying for balance-sheet duration, domestic rate normalization, and capital-return optionality in Japan. If that regime persists, the next beneficiaries are other lenders and insurers, while JGB duration proxies and dividend substitutes can underperform as capital rotates toward earnings leverage. The move is most vulnerable if BOJ tightening expectations fade or if deposit beta/credit costs rise faster than loan yields.

Baidu’s weakness looks like skepticism around AI monetization rather than simple revenue drift, which is usually a multiple problem before it becomes an earnings problem. The contrarian angle is that profitable China internet names with real ad cash flow can absorb a slower AI payoff, while BIDU’s premium can compress further into the print. Falsifier: a clean guidance beat with improving cloud/search mix; if that happens, the stock can mean-revert quickly because positioning is already fragile.

Key Decisions for Investors

• Long MUFG on a 3-6 month horizon; treat pullbacks as entry points for a reflation/rates beneficiary. Thesis breaks if BOJ turns explicitly dovish or bank NIM expansion stalls.
• Buy BIDU downside into earnings via a put spread, or run BIDU vs KWEB as a relative-value short for 1-2 months. Target 10-15% downside on a soft guide; cover if AI/cloud commentary re-accelerates.
• Do not chase a fresh short in SKHYV here; wait 1-2 sessions for stabilization and the next hyperscaler capex updates. If HBM pricing rolls over and the break persists, express the view with SMH/MU rather than the single name.
• If you want a cleaner Japan-relative trade, overweight MUFG versus weaker Japanese low-rate proxies rather than adding broad beta. That captures curve-steepening and governance flow with less market risk.