Hynix is currently down 12.52%, while Samsung is down 8.42%.

Korea Investment & Securities predicts that SK Hynix’s operating profit in the second quarter may be 8% lower than market expectations.
The fundamentals are fine: SK Hynix’s earnings are still very good, with profit flexibility affected by long-term contracts for HBM. In the short term, there isn’t a “spot price increase” dividend as strong as that of ordinary memory.
Actually, the market’s views on HBM4 pricing are quite divergent right now. HBM4 comes with logical units and is highly customized.
In the end, it depends on many variables, including: the procurement contract price for Nvidia; HBM4’s performance, power consumption, and yields; who bears the advanced packaging costs; whether Nvidia is willing to pay a premium for stable supply; and how many order shares SK Hynix ultimately secures, and so on.
None of these things are fully transparent at the moment. That’s why KIS itself also admits that the price of HBM4 depends on the contract signed with Nvidia, and it’s indeed difficult to judge.
KIS may have adopted more conservative assumptions for HBM pricing and profit margins, so its second-quarter profit forecast lands toward the lower end of market expectations. But HBM4’s pricing is not a simple cost-plus model.
KIS’s so-called “after HBM4 goes into mass production, ASP reverts to the market average” can be understood more reasonably as not meaning that “HBM4’s own price returns to the average price of ordinary DRAM.”
That is impossible.
A more reasonable meaning is that as HBM4 ramps up in volume, SK Hynix’s overall DRAM blended ASP growth may gradually move closer to the market average.
In other words, the discussion is about the company’s overall DRAM average selling price growth rate—not that the HBM4 product itself becomes ordinary DRAM priced like ordinary DRAM.
So, once again: the memory fundamentals remain strong. The problem lies in the macro environment. $SKHY