Article published Jul 6, 2026. Prices below use latest available snapshots.
Summary
SK Hynix — the world's #1 HBM maker, the memory inside NVIDIA's AI GPUs — is scheduled to list on Nasdaq as SKHY around July 10, 2026 in a ~$29.4B ADR offering, the largest ever. This report answers four things: what the listing means for an already-huge public company, what its Seoul tape looks like (and why it's "wonky"), why the HBM monopoly is US-listing at all, and how much money it actually makes (a lot). Verdict: structural confirmation of the memory supercycle, watch-the-debut not buy-the-open — the cleaner expression of the identical thesis is already MU.
Full working + Seoul price/RSI data + the four US-listing levers: investigation 2026-07-06-sk-hynix-skhy-nasdaq-adr-debut-why-us-list.
The listing (verified): SKHY, ~July 10, a record
Ticker SKHY on Nasdaq, trading expected ~July 10, 2026; ~$29.4B, 17.79M new ADS (per its SEC F-1 registration) — the largest-ever ADR listing, past Alibaba's 2014 debut (CNBC, BigGo). Underwriters BofA / Citi / Goldman / JPMorgan; proceeds fund Korean memory fabs + chipmaking equipment, explicitly ASML EUV scanners. 10 ADRs = 1 common share — one ADR prices ~$155–160 off a ₩2.3M Korean share.
What it means for "the ADR" — sponsored + listed is a different animal
Like most big Korean names, SK Hynix has been reachable in the US only through thin, unsponsored OTC ADRs (and a German proxy, HY9H.F) — no capital raised, no real liquidity, no index eligibility. SKHY is its first sponsored, exchange-listed, capital-raising ADR. That changes three things: it's primary (17.79M new shares — a ~$29B fundraise, not just access); it opens real liquidity + US index/passive eligibility; and it's the same equity in two venues, so SKHY tracks Seoul ± FX and the session gap. A US holder takes KRW + Korea-session risk, not escapes it.
The Seoul tape (000660.KS) — parabolic, and genuinely wonky
The parabola, annotated (live 000660.KS closes; the table has the full detail):
| Date | Close (₩) | Note |
|---|---|---|
| 2025-07-07 | 269,872 | ~a year ago |
| 2025-11-05 | 577,423 | HBM re-rate underway |
| 2026-05-04 | 1,446,758 | vertical phase |
| 2026-06-22 | 2,919,000 | near 52wk high ₩2,987,000 |
| 2026-06-23 | 2,555,000 | KOSPI −9.9% crash day (≈−12.5%) |
| 2026-07-02 | 2,187,000 | second air-pocket |
| 2026-07-06 | 2,343,000 | now — −21.56% from high |
Why it looks wonky: ~11x in a year (₩270K → ₩2.99M); it passed Samsung as Korea's most valuable company mid-run; 10–15% single-day swings for a trillion-won-plus company (the crash day, the July-2 air-pocket); a high nominal ₩2.3M (~$1,600) share price because Korean names rarely split (the 10:1 ADR ratio is partly a fix for exactly this); and now −21.6% off the high but still up enormously — a violent pullback inside a monster uptrend.
Why list in the US at all — four structural levers, not "more volume"
- Capital depth. $29.4B is too big to pull repeatedly from Korea's pool without moving your own stock; US markets are the only venue deep enough to absorb it cleanly.
- Dollar funding for dollar capex — a natural hedge. Proceeds buy ASML EUV and US-priced tools; raising USD to spend USD removes the FX mismatch.
- Escape the "Korea discount." Korean equities carry a chronic governance/chaebol discount; US AI-semis trade richer. A Nasdaq line is a bid for a valuation re-rate toward the US comp set.
- Investor base + passive flows + prestige. Direct access for US institutions that won't/can't hold KRX, an index-inclusion path, and a marquee listing planting the HBM monopoly in the US AI trade. Volume is a consequence of these, not the reason.
The money — monopoly-rent economics
Q1 2026, all-time records: revenue ₩52.58T (~$35.5B), +198% YoY / +60% QoQ; operating profit ₩37.61T at a ~72% operating margin; net ~₩40.3T at ~77% net margin (SK hynix IR, CNBC). HBM #1 at 56.4% share; HBM4E samples H2 2026, mass production 2027. A 72% operating margin on a hardware business is the entire thesis in one number — that is HBM scarcity rent, not normal semiconductor economics.
Fundamentals are company-reported: SK Hynix is a foreign filer with no US filings yet (pre-listing), so these ₩ figures are single-source.
Verdict — structural confirmation, watch the debut
The single most important memory company on earth is monetizing the cycle at a record dollar valuation, buying more ASML EUV, and re-rating toward US semis. That is the memory-supercycle thesis confirming itself in the primary market. For the desk's tradeable reads: MU is the US-listed comp and the cleaner instrument for the identical thesis (dollar-native, no Korea/session overhang); ASML is levered via the stated EUV proceeds.
For anyone eyeing SKHY directly: it's the cleanest pure way to own the HBM monopoly in USD — but it debuts into a parabola already −21.6% off its high with 10–15% daily swings, while printing $29B of brand-new supply the market must absorb in its first days (IPO-allocation selling + no trading history = a wide, jumpy debut). The business is extraordinary; the entry is a mega-cap IPO on top of a vertical move — not the same decision. Per the desk's trend-hold book: watch the debut, buy the base — not the open.
Watch triggers: the SKHY debut tape (allocation air-pocket vs. index-inclusion bid — enter on a base, not the first print); KRW/USD (the ADR carries currency); HBM4E timeline (samples H2 2026 → mass production 2027 — slippage is the first crack); the MU cohort (records that don't lift MU signal a Korea-idiosyncratic re-rate); and the lane's standing #1 risk, capex glut — a record raise funding record fab build-out is the supercycle's own seeds-of-reversal.
Sources
- Fundamentals figures: company-reported results (quarterly/annual filings) as available at the artifact date; predates the desk's EDGAR reconciliation gate — figures not re-verified after publication.