Best performance, worst expectations

SK Hynix delivered its strongest quarterly report in history today (+557%), yet still fell 9.61%. The market says it's because they are 'too good at making HBM'—long-term contracts locked in prices, so they can't benefit from the surge in spot prices.
Samsung's share of general-purpose DRAM is even higher, and it also fell 5.23% today. The same logic doesn't explain both.
The real reason: The market is re-evaluating 'how long profits burned through cash can be sustained.'
Apple's market cap broke $5 trillion today, surpassing NVIDIA to reclaim the global #1 spot. On the same trading day, Apple rose while storage stocks crashed—the market voted with hundreds of billions of dollars: 'Restrained capex + high returns' is more valuable than 'Crazy capex + uncertain returns.'
Tomorrow, watch Samsung. The numbers don't matter (the forecast was released early); focus on three things:
① Is the HBM4 outlook consistent with SK Hynix? Only independent confirmation from both makes it credible.
② Have foundry + mobile segments improved? Cyclical resilience is the key to valuation.
③ Will the Fed be dovish or hawkish tomorrow morning? When the policy floor and interest rate ceiling appear simultaneously, short covering will be fierce.
SK Hynix hasn't gotten worse; the market has changed its mind. We've been talking about burning cash for two years; investors want returns, not more capex. Tomorrow's Samsung isn't just another earnings report; it's a battle for pricing power over 'who really makes money in the second half of the AI era.'