A critical re-read of the consensus 'Neosem = Core / OpenEdge = long-dated option / Exicon = afterthought' framing. Once you separate the CXL purity prism from the 2026 earnings-momentum prism, Exicon (092870 KQ) emerges as the strongest near-term J-curve — KRW 79B in cumulative orders against 2024 revenue of KRW 31.6B, single-source CLT 11,520-parallel tester, Samsung qualified. Neosem is a mid-cycle bet that needs 1Q26 confirmation. OpenEdge stays a 3-5 year IP option with royalty mix at 0.4%.
BOE's Apple line utilization collapsed from 82% to 48%. Korean panel-makers are getting headlines. But the real alpha is one layer down — the equipment and materials suppliers exposed to both the China 8.6G OLED CapEx cycle AND the HBM/DDR5 semiconductor cycle. AP System (265520), Soulbrain (357780), Dongjin Semichem (005290).
KOSPI +49% YTD leads every major market by 30pp+. Three structural drivers fire at once — HBM supercycle, legislated governance reform, and Sell America rotation. Here's the 3-sleeve framework foreign allocators need to own Korea properly.
Crimson Desert's 29GB 1.04.00 patch landed April 23 with difficulty modes, 1,000-slot housing, and pet systems. New review positivity hit 94.9% (+11.2pp). The next Patch 1.05 watch point is boss-rematch and endgame-loop delivery. At P/E 15× this pushes fair value to ~83,000-86,000 KRW.
Full 3-year US tenbagger census: 40 names from 1,690 tickers, 33% are AI-infra full stack, median 15.7× beats Korea's 13.8×. Where the US leads, Korea follows on a 3-6 month lag — here's the pair-trade map.
Only 52 KOSDAQ names get real sell-side coverage (2+ brokers, 3M). Six archetype theses explain the small-cap cluster — and Amosense + Coats Technologies sit on choke points the market hasn't priced in yet.
Full 3-year census of KOSPI + KOSDAQ: 2,613 stocks, 27 passed 10×, a 1.03% hit rate. Three-quarters are profitable, power/grid is the hidden cluster (8 of 27), and ROE≥40% with P/E≤15 would have caught several ex-ante.