SK hynix: Cheap Despite A Historic Memory Boom

SK hynix reported 257% revenue growth and 557% operating profit growth, lifting operating margin to 76%. HBM4 entered mass production in Q2, with HBM4E targeted for 2027, supporting its technology lead. About 10 customers signed long-term agreements (typically ~5 years) with deposits and volume commitments, improving demand visibility and underpinning a likely positive read-through to semiconductor/AI-memory sentiment.

Analysis

This reads less like a one-quarter beat and more like evidence that AI memory has become a scarcity market with quasi-contractual cash flows. The economic value is shifting from generic DRAM exposure to qualified HBM capacity, which should keep price realization above cost curves and support a premium multiple versus traditional memory names. The spillover beneficiaries are not just the direct supplier but also the semiconductor tool stack and, indirectly, GPU vendors that need secure memory supply to ship systems on time.

Near term, the stock can keep re-rating as investors chase the combination of growth, margin, and demand visibility. The bigger catalyst over the next 1-3 months is whether management signals incremental capacity adds without sacrificing yield; that would validate that the business can keep compounding rather than simply harvesting a temporary bottleneck. The main risk over 6-18 months is that HBM remains a capacity race, not a moat: Samsung and Micron can close the gap faster than the market expects, and once supply normalizes, today’s extraordinary margin profile can compress quickly.

The contrarian point is that long-dated customer commitments may be masking demand prepayment rather than proving durable end-market growth. If AI capex pauses, contracts may preserve revenue visibility but not protect pricing power, and the market could discover that the 2027 pipeline was discounted too aggressively. What would falsify the bullish view is any sign of sequential margin rollback, slower HBM mix expansion, or a visible step-up in competitive yields from Samsung/Micron on the next two quarters of data.

Key Decisions for Investors

• Long SKHYV on any pullback over the next 1-2 weeks; thesis is continued multiple expansion as the market prices in multi-quarter scarcity and contract visibility. Risk/reward is favorable until there is evidence of sequential margin compression.
• Pair trade: long SKHYV / short MU over 3-6 months. The bet is that the current leader keeps pricing power while Micron spends into catch-up mode before seeing comparable operating leverage.
• Use SSNLF as the cleaner competitive hedge if accessible; short SSNLF against SKHYV when next HBM supply-chain updates or earnings commentary approach. The pair should work if the market starts differentiating leader vs follower on yield and qualification timing.
• Watch AMAT and LRCX as second-order beneficiaries for a basket trade. If HBM capacity additions accelerate, tool vendors can monetize the buildout regardless of which memory vendor wins share.