Layer 3 · HY9H — SK Hynix Sponsored GDR
Thesis
The European wrapper for 000660 — same HBM monopoly, same record Q1, same 2027/2028 supply tightness — listed in Frankfurt for EUR investors. Tier B not Tier A purely to avoid double-counting Tier A exposure that's already in the book via 000660; this is a valid B-tier proxy for accounts that can't hold KRW directly.
What it does + financial print
SK Hynix is the dominant HBM3E supplier to NVIDIA (>50% share through Hopper, locked through Blackwell, primary on Rubin pending HBM4 qualification), the global #2 DRAM supplier, and a top-3 NAND vendor. The HBM monopoly is the entire bull case — HBM supply is sold out through 2027 and the TAM is expanding from $35B in 2025 to $100B+ by 2028.
Q1 2026 (April 23) printed KRW 52.6T revenue (+198% YoY, +60% QoQ) — a beat. OP KRW 37.6T at 72% operating margin. Net income KRW 40.3T at 77% NM. Guidance: HBM demand far exceeds capacity next 3 years; HBM4E samples 2H26. Stock at all-time-high KRW ~1.83-1.95M; HY9H GDR at EUR 594.
Bull case
HBM4 NVDA Rubin design-win confirmation, HBM4E sample-to-volume transition through 2027, and supply tightness extending to 2028 (per Samsung's tracker referenced in the data). EUR-based investors get the same fundamentals as 000660 holders with simpler settlement.
Gap / bear case
The whole memory complex is at cycle-high valuations — +35% in 30 days, +77% in 90. Any signal that Samsung's HBM4 qualifies at NVDA (potentially rebalancing share) would compress SK Hynix's premium. KRW FX risk is real for the underlying.
Trigger to upgrade / downgrade
Upgrade to Tier A on a pullback to EUR 520-560 (the breakout shelf), reclassifying as the primary HBM exposure for EUR books. Downgrade if Samsung wins majority HBM4 share at NVDA Rubin.
The trade
- Entry zone: EUR 520-560 on pullback
- Stop: EUR 480 close (50-day)
- Position size: 1.5% NLV (or 0% if already holding 000660)
- Catalyst date: Q2 print July 24 2026
- Conviction: 7/10