This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.
SK Hynix reports Q2 on July 29 — and the market is fighting over the wrong number. Not the profit. The bar. This is not financial advice — a view from Seoul, not a recommendation to buy or sell anything. (Won-to-dollar at ~₩1,480/$; verify at the time of reading.)
SK Hynix (KRX: 000660 / Nasdaq: SKHY) reports second-quarter results on Wednesday, July 29 in Seoul — the evening of Tuesday, July 28 US Eastern time. The Street consensus for operating profit was about ₩65.3 trillion (~$44.1B). Last week, a Korea Investment & Securities analyst cut her estimate to ₩60.4 trillion (~$40.8B) — about 8% below consensus (with a revenue estimate of ₩80.9tn in the same note). And yet she kept her price target at ₩3.8 million (~$2,570). Trim the earnings, keep the target: that combination is the whole story. Seoul sold the stock down ~4% on the cut — but the analyst who made the cut is treating it as a timing adjustment, not a hit to the franchise. That’s the checkpoint this stock is standing at.
A quick sanity check, because that number looks like a typo. ₩60 trillion of operating profit in a single quarter is larger than what most US megacaps earn — so a US reader’s first instinct is “that can’t be right.” It is. SK Hynix just posted Q1 2026 operating profit of ₩37.6 trillion (~$25.4B) on ₩52.6tn (~$35.5B) revenue — a 72% operating margin, its best quarter ever, up 405% year over year. This is what a memory maker with an HBM near-monopoly prints at the top of an AI super-cycle. The ₩60tn Q2 estimate isn’t a decimal error; it’s the run-rate. (That, by the way, is the “no US equivalent” part — no American chipmaker runs a 70%-plus margin at this scale.)
This piece takes the downgrade apart — what came down, what got pushed out, what’s structurally capped — and why a lowered bar can, paradoxically, make July 29 easier to clear. And where that paradox breaks.
The timeline to the checkpoint — the market already sat one exam
| Date | Event |
|---|---|
| Thu 7/16 | Bank of Korea’s first rate hike in 3½ years (2.50→2.75%) · KOSPI −6.37% |
| Fri 7/17 | Seoul closed (holiday) — the same day Nvidia, TSMC and AMD fell together in New York; oil spiked past $90 on Middle East tension |
| Mon 7/20 | Deferred selling landed at once: KOSPI 6,516.27 (−4.46%), KOSDAQ −5.33%, sell-side sidecars on both boards. SK Hynix closed ₩1,764,000 (−4.23%, ~$1,192); Samsung ₩244,000 (~$165, −4.31%). The net seller that day was institutions (−₩922bn / ~$623M), not foreigners |
| Wed 7/29 | SK Hynix Q2 results (verify time — per IR) |
Look at the back of the value chain on July 20. Chip-equipment names fell three to four times harder than the stock itself — Wonik IPS −18%, PSK −13%, Jusung −11%. When the tools sell off worse than the maker, the market isn’t repricing one quarter; it’s re-examining the whole capex cycle.
🎩 Under the Gat — The stock down 4%, the equipment down 18%. The question the market is really asking isn’t “how much did Hynix earn,” it’s “will it keep spending that earning.” That’s why, on July 29, the guidance will matter more than the number.
Dissecting the downgrade — three screws, not a demand collapse
The Korea Investment note (analyst Chae Min-sook) got to ₩60.4tn from a ₩65.3tn consensus by loosening three screws — none of which is “demand died.”
1. Prices rise, but Hynix captures less of it. The analyst estimates SK Hynix’s Q2 DRAM average selling price (ASP) rose ~30% year over year — but says its heavier HBM mix means it benefits less from the commodity-DRAM price rebound than peers. In a business where price is profit, Hynix gets a smaller slice of the very rally lifting the sector. Note the sentence: this is “prices rise, our mix just doesn’t carry as much of it,” not “prices broke.” July 20 sold as if they were the same thing. And the same HBM concentration flips to a defense when the commodity cycle rolls over — it’s a double-edge.
2. HBM4 ramps in Q3–Q4, not Q2. Full HBM4 volume was pushed to the second half, so an upside piece analysts had penciled into Q2 came out. The same note expects Hynix’s ASP growth to normalize toward the market average from Q3–Q4 as HBM4 ships — i.e., the “captures less” drag from screw #1 eases exactly when HBM4 ramps. This is a calendar problem, not a demand one: deferred profit, says the bull; “delays are rarely one-and-done,” says the bear.
3. LTAs cap the upside — and the downside. Long-term supply agreements lock the price on much of the volume, so a spot-price spike doesn’t flow straight into the quarter. That re-realization — that LTAs cap earnings upside — was the actual trigger of Seoul’s worst session on July 13. The structure sits in the Nasdaq prospectus too; we broke down the filing in the SKHY prospectus autopsy. What gets left out: the same contracts put a floor under profit when prices fall.
| Item | Consensus | KIS note (Chae) | Nature of the gap |
|---|---|---|---|
| Q2 operating profit | ~₩65.3tn (~$44.1B) | ₩60.4tn (~$40.8B) | −8% — a re-rate, not a collapse |
| Q2 revenue | — | ₩80.9tn (~$54.7B) | still a sharp jump QoQ |
| DRAM ASP (YoY) | — | ~+30% est. | rising, but HBM mix limits the benefit |
| HBM4 ramp | penciled into Q2 | Q3–Q4 | timing shift, not lost demand |
| Price target | — | ₩3.8M (~$2,570), unchanged | earnings cut, valuation held = timing call |
Why a “cut” can be healthy — the mechanics of a reset bar
US investors know this grammar. Nvidia has beaten consensus and fallen anyway — the textbook of “priced to perfection.” When expectations run ahead of results, even a perfect card fails. The reverse holds too: when the bar comes down before the print, the same result becomes a pass. A downgrade nine trading days out — and a −4% stock — is also the cut-off score sliding lower before the exam. The unchanged ₩3.8M target is the tell: the analyst lowered this quarter’s number without lowering the company.
Korea’s checkpoint bolts on two structures the US version doesn’t have. First, unlike Samsung, SK Hynix issues no preliminary flash — the final number lands in one shot on July 29. The process and channels are laid out in the Korean earnings-season guide. Second, the time gap between the Seoul release (morning) and SKHY trading (that night): Seoul prices first, then the ADR recomputes its premium on top — the same premium that ran 38% and compressed to ~22% in the debut week (the ADR premium autopsy). July 29 is that premium’s second exam, too.
🎩 Under the Gat — Setting your bags down before the checkpoint and having them confiscated at the checkpoint are different events. The downgrade may be the first kind. But that only says the odds of passing went up — not that the stock goes up. That second sentence is one this blog doesn’t write.
Leave a seat for the other side — where the paradox breaks
Fairly, here is the bear’s read, undiluted. (1) The ASP-benefit cut may be the first data point of a trend, not a one-off — if next quarter’s note comes down again, that’s not a checkpoint, it’s a decline. (2) The HBM4 deferral is also a competitive question — if Q3–Q4 volume slips again, the “calendar problem” defense dies. (3) As July 20’s ₩922bn of institutional selling shows, pre-earnings position-clearing is still ongoing; margin debt is down more than 10% from its peak, but nothing says the de-leveraging is finished. (4) Above all — if the bar comes down and the result still prints below it, a lowered cut-off is no comfort.
The July 29 checklist — four things to watch besides the number
- Realized ASP — the gap between the estimate and the actual. That gap, more than beat/miss, sets the direction of the next note.
- HBM4 timing — whether “Q3–Q4 ramp” holds, and whether volume guidance is attached.
- Second-half capex commentary — this is where the −18% equipment move is judged over- or under-done.
- LTA renewal hints — any mention of repricing existing contracts rewrites the “capped upside” story itself.
This is not financial advice. Won-to-dollar uses ~₩1,480/$ (provisional; verify the publication-day rate). Figures are as of July 20, 2026 closes and public materials; intraday numbers are labeled “intraday.” The July 29 timing and format are per company IR, not a consensus-calendar guess. Estimates cited are one brokerage’s; re-verify against primary sources before acting.
The SK Hynix four-parter: the prospectus, broken down · how to read Korea’s earnings season · why the ADR trades at a premium.
Frequently Asked Questions
When does SK Hynix report Q2 earnings?
Wednesday, July 29, 2026, in Seoul — which is the evening of Tuesday, July 28 US Eastern time. Verify the exact time against company IR. Unlike Samsung, SK Hynix does not issue a preliminary flash, so the final numbers arrive in one release. This is not financial advice.
Why did an analyst cut the estimate?
Three reasons: (1) DRAM prices are rising, but SK Hynix’s heavier HBM mix means it captures less of the commodity-price rebound; (2) full HBM4 volume ramps in Q3–Q4, not Q2; (3) long-term supply agreements (LTAs) cap how fast contract prices reset. It was not a demand-collapse call — and the same note kept a ₩3.8 million (~$2,570) price target.
Is an estimate cut a sell signal?
Not necessarily. A pre-earnings cut can lower the bar and make a ‘pass’ easier (expectations reset), or it can be the first step of a downtrend. The deciding data is the realized ASP and the second-half guidance on July 29. Not financial advice.
Which moves first — the Seoul shares or Nasdaq’s SKHY?
The Seoul ordinary shares react first, on the morning of the report; the SKHY ADR reprices that night in New York, layering its premium back on top. That premium compressed from ~38% to ~22% in the debut week.
This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.