Part 1 traced the four-layer recognition matrix (revenue / margin / channel / market perception) of Pamicell's transformation from a bio company to an AI-CCL materials company. Five new data points landed within a week. KRX sector classification officially changes May 4 from 'basic pharmaceutical' to 'electronic components' (the accounting closure of the market-perception layer); Citi and Goldman call CCL undersupply through 2026–2027 a 'new normal' (industry-cycle verification); DS Investment & Securities and Meritz Securities converge on 1Q26 estimates (revenue ₩36.2B, OP ₩12.1B, OPM 33%); the value-chain matrix expands into a 4-tier structure; and the precise OPM match of Doosan Electro BG (1Q26 30.1%) and Pamicell (2025 30.1%). Five new pieces all point in the same direction, with a fifth analytical layer — industry-cycle duration — being added to the framework.
Kiwoom Securities (039490) is no longer a 'cheap brokerage' at PBR 0.6–0.8×. PBR moved from 0.55× (2024) to 1.39× (2026E) — the market has already reclassified Kiwoom as an ROE-20% capital-efficiency brokerage. The April 30 1Q26 strong-print sell-off (-6.02%) is not a model rejection. It is the natural self-stabilization signal of a price that already reflects the model. If Meritz Financial is the static 'ROE × payout' compounder, Kiwoom is the dynamic 'ROE × trading-volume beta × capital turnover' variant on the same matrix — and inside the post-recognition market, what gets verified next is no longer discovery but model durability.
Once Meritz (capital cancellation), Kiwoom (capital turnover), and KB (foreign access) settled as three distinct peaks of Korean financials, the market began watching something else — the speed at which followers move toward those peaks. Shinhan Financial (055550) is the company moving fastest toward the 'foreign access' coordinate KB reached first. ROE 11.9%, CET1 13.19%, ₩700B treasury share buyback-and-cancel, Value-up 2.0's ROE × payout-ratio formula. And an internal NVR (Net Volume Ratio) read of +18% — the accounting fact that, in a window where price barely moved, up-day volume exceeded down-day volume by ~44%. The signal that volume recognized the 'transit between peaks' before price did. The first case of a new chapter in the Korean financials re-rating story.
If Meritz Financial is the static peak of 'capital-buyback compounding' and Kiwoom Securities is the dynamic peak of 'trading-volume beta', KB Financial Group (105560) is a peak on a completely different dimension — the 'foreign access proxy' peak. Foreign ownership at 75.72% is not 'capital still to come' — it is 'verified flow infrastructure already priced in'. MSCI Korea 25/50 weight 2.00%, EWY at $20.9B AUM, CET1 13.63%, 2026E payout ratio 60.6% (83.0% including legacy treasury cancellation), PBR 0.88×. When global allocators look at Korean financials, they look at KB first — what that means inside a recognition-completed market.
The repricing of Korean financial stocks has already happened. The 'low-PBR discounted asset' era is over; the market now evaluates Korean financials through ROE × payout ratio × EPS accretion. Meritz Financial Holdings (138040) is the company that built this new standard and embedded it most deeply: ROE ~22%, 50–60% net-income deployed to share-buyback-and-cancel, and EPS growing +13% even as revenue contracts. Per-share value compounds while top line shrinks. The cleaner question now is the *speed* at which KB Financial (105560), Hana Financial (086790), Shinhan Financial (055550), DB Insurance (005830), and Korea Investment Holdings (071050) adopt the same standard — and the time gaps among them define where alpha still exists inside Korean financials.
Samsung Foundry customers in 2026 — confirmed via earnings calls, customer disclosures, and supply-chain reporting — include Tesla (HW5/Dojo successor at SF2), Tenstorrent (Wormhole/Blackhole at SF4X), Qualcomm (selected modem and Snapdragon nodes), Google (TPU successors and Pixel SoC at SF4LPP/SF3), Ambarella (CV3-AD ADAS), and Samsung's own System LSI (Exynos). The honest answer is that Samsung Foundry remains a credible #2 to TSMC at advanced nodes, with the customer roster heavily weighted toward AI accelerators, automotive/ADAS SoCs, and customers willing to take a calculated yield-risk premium for capacity availability or sovereign-supply reasons.
Korea's May strategy isn't about selling—it's rotating from overheated AI infra winners into under-the-radar second-order proxies. Here's the playbook.
Pamicell topped my April foreign-plus-institutional accumulation screen by market-cap intensity. The more important discovery is that 97% of 2025 revenue came from biochemicals, with low-Dk electronic materials tied to Doosan Electronic BG's AI CCL cycle now the core re-rating question.