SK Hynix reports second quarter results on Wednesday, July 29 (Asia morning). With memory at the center of the AI infrastructure trade, this is arguably the most important semiconductor earnings release of the week.
July’s selloff priced a demand scare into this print, so the first headlines will land on a nervous market. Keep this email open during the call. These are the four headlines most likely to shape the initial reaction, and the condition under which the bearish interpretation would actually be correct.
1. “HBM fell as a share of revenue.”
That may simply be denominator arithmetic. Conventional DRAM pricing rose sharply this quarter, while HBM pricing is largely fixed through annual contracts. HBM’s share can therefore fall even as HBM revenue grows.
Bearish if: HBM revenue itself declines, or management attributes the lower share to lost business or weaker customer demand.
2. “SK Hynix failed to capture the full market price increase.”
It never does. Market benchmarks do not flow immediately into reported results. Blended DRAM pricing near +40%, with conventional DRAM around +45%, would still be consistent with contract timing against a +58% to +63% market benchmark.
Bearish if: Blended DRAM pricing lands closer to +30%. That would validate the bearish realization case.
3. “Net income missed.”
Q1 net income included roughly ₩11.5 trillion of one-time gains, making the sequential comparison misleading.
Bearish if: Operating profit also misses, particularly because realized pricing or conversion was weaker than expected.
4. “Q3 bit guidance disappointed.”
Lower conventional bit growth is not automatically bearish if pricing remains firm. HBM4 consumes more wafer capacity through larger dies, base dies, and stacking yield, so weaker conventional volume can coexist with a tightening supply constraint.
Bearish if: Softer shipment guidance arrives alongside weaker pricing or softer demand commentary.
Today’s email is about interpreting the headlines. Tomorrow, before results, we’ll publish the call itself: the three thresholds we registered on July 24, the probabilities assigned to each, and the one number that decides between our view and the bear case.
The complete House View is already available to Tessara Pro members since last week, including the segment-by-segment bridge from Q1, the physical evidence, and the probability engine behind our forecast:
Every registered call will be graded publicly, hit or miss, after earnings:
If someone you know is watching this print, feel free to forward them this email.
Cheers,
Teng
Tessara publishes research, not investment advice. Nothing here is a recommendation to buy or sell any security.