2327 / 2327

2327 — Yageo Corporation

WIKI stockpassives-mlcc updated 2026-08-29

🔴 Live NT$562 (yf:2327.TW) · 2026-09-05 · research written 2026-05-20 Consensus NT$600–NT$1,465 (mean NT$1,082, 12mo, Yahoo consensus (14 analysts)) · 2026-09-05 🔔 TRIGGER HIT — your target >NT$540–NT$620 (own-DD, 12mo) — now in zone (+4%)

Thesis

Verdict: WATCH — high-quality business, wrong entry price. Do not buy at NT$572. Yageo is a real cyclical-plus-mix-shift winner: the world's #1 tantalum capacitor maker (designed in to NVIDIA H100/H200/B200/Blackwell server power rails, with >30% of tantalum revenue now AI-linked), #1 chip resistor maker, and #3 MLCC supplier behind Murata and Samsung Electro-Mechanics. It is riding a K-shaped recovery where high-end AI/auto/industrial pricing holds firm while commodity smartphone MLCC stays soft. The company executed a clean ~320-620bps operating-margin expansion (FY23 19.0% → Q1 26 25.2%), beat consensus EPS by 9.2% in Q1 2026, and is compounding on a base where AI is ~15% of group revenue and rising.

The problem is entirely price. NT$572 is the 52-week high, +207% off the NT$111.75 low, trading 35% above the mean analyst PT of NT$423 (50% above the NT$382.50 median) at 49.8x trailing / 26.8x forward P/E versus a 5-year average of ~22x trailing / ~17x forward. The deep-dive rates the business Medium-High and the people B+ with no red flags; the checklist scorecard reads 6 Yes / 4 No, where the four No's are valuation, technicals (52-week high + RSI extended), behavioral traps (FOMO + confirmation + recency bias all active), and a clear exit plan (target prices sit below current). Conviction is Medium-Low at NT$572, Medium-High at NT$420-460.

What has to be true for the bull case: continued AI capex into 2026-2027 plus disciplined capacity utilization that sustains pricing power — neither requires new capex (management explicitly guided no new MLCC capacity for 2026). The franchise compounds through a cycle; the question is not whether the thesis is right but whether the price has front-loaded too much of it. As the deep-dive puts it: the sector is inflecting; the stock is not. Cleaner same-theme alternative flagged: Sakai Chemical (4078.T), the upstream BaTiO₃ powder bottleneck at ~5% of Yageo's market cap, less consensus-crowded.

Snapshot

Yageo Corporation (國巨股份有限公司) · TWSE: 2327.TW · Currency: TWD. World's #1 chip resistor and tantalum capacitor maker, #3 MLCC supplier. GDRs listed on Luxembourg Stock Exchange. GICS: Information Technology — Electronic Components. HQ Xindian District, New Taipei City, Taiwan. Founded 1977 by Pierre T.M. Chen (still Chairman); IPO 1993 (TWSE). ~30,624 full-time employees (FY2024 year-end, down from 36,101 in FY2023, -15%; ~39,000 per the older profile pull — see note). Auditor: Deloitte & Touche (continuous; partners Meng-Chieh Chiu, Chun-Yu Wang).

Valuation snapshot (as of 2026-05-20, source: yfinance):

Metric Value
Share price NT$572
Market cap NT$1,178bn (~US$36.6bn)
Enterprise value NT$1,117bn
P/E (TTM) 49.8x
Forward P/E 26.8x
P/FCF (LTM) ~46x
EV/Revenue (LTM) ~8.0x
EV/EBITDA (LTM) ~25.5x
Price / book 7.0x
FCF yield ~2.2%
Dividend yield 0.96%
52-week range NT$111.75 – NT$572.00 (at 52-week high; +207% off the low)
Beta 1.40
Shares outstanding 2.06bn (2,058M post-split)
Insider held 23.5% (yfinance); true Chen-family control ~20-21%
Institutional held 25.4% (33.2% of float)

Analyst consensus: Buy (12 opinions). Mean PT NT$423.08 · Median PT NT$382.50 · Current NT$572 = 35% above mean, 50% above median. One Strong Sell rating has persisted four months (likely a Japanese broker bearish on the cycle).

Business

Pure-play passive-component manufacturer — the small ceramic and tantalum chips that filter and stabilize power on virtually every printed-circuit board: phones, EVs, AI servers, industrial gear, satellites, medical instruments. Yageo earns its money by being one of three or four firms in the world that can ship these at automotive- and AI-spec quality at volume. Revenue is ~100% one-time hardware sales — no SaaS, no licensing. Distribution is overwhelmingly through global distributors (Arrow, Avnet, WPG/Yosun) into a fragmented end-customer base; the company formally discloses no customer >10% of revenue and no supplier >10% of procurement.

Business lines and revenue mix

Yageo reports at consolidated level, not fixed segments; splits below are estimates from product-family disclosure and management commentary:

Product family Yageo position Est. % revenue
MLCC (multilayer ceramic capacitors) Global #3 (~13-15% share); mix shifting commodity smartphone → high-end AI/auto ~40%
Chip resistors Global #1 (~25-30% share); commodity-leaning but cash generative ~20%
Tantalum capacitors (legacy KEMET) Global #1 (company-disclosed >46% share); >30% of tantalum revenue now AI-linked ~20%
Inductors / magnetics (Chilisin + Pulse + Ferroxcube + Mag.Layers) AI/auto power conversion ~10%
Sensors (Nexensos platinum RTDs + Telemecanique industrial) + others (film/poly caps, AC line filters) High-end industrial specialty ~10%

Three economic engines

  1. Commodity scale base — chip resistors and standard MLCC where Yageo is #1 by volume; prices the floor; mid-single-digit operating margin; cyclical with smartphone/consumer demand.
  2. Specialty / high-end — auto-grade MLCC (AEC-Q200), AI-server tantalum, niche industrial; qualification-gated (18-36 months); >40% gross margin; sticky pricing through cycles.
  3. Sensors + inductors — Nexensos (platinum RTDs), Chilisin (power inductors), Telemecanique (limit/proximity sensors), Shibaura (NTC thermistors) — diversification away from pure capacitor cyclicality.

The technology — first principles

Every circuit needs capacitors (store charge, smooth voltage ripple, decouple noise, filter EMI), resistors (limit current, set bias), and inductors/magnetics (store energy in magnetic fields, filter, step voltages). An AI server uses 10,000+ MLCCs and several hundred tantalum/polymer caps; a BEV uses 8,000-10,000 MLCC vs ~3,000 for a legacy ICE car. MLCC physics: capacitance = dielectric constant × area / thickness; advanced barium-titanate ceramics hit dielectric constants of 1,000-15,000, modern layers are 0.5-1 micron thick, and a 1mm package can stack 500-1,000+ internal electrode layers. The case-size race runs 1005 → 0603 → 0402 → 0201 → 01005 (~0.25 × 0.125 mm); Yageo ships 01005 in volume with 0080 in R&D but still trails Murata on absolute smallest sizes and highest CV. MLCC process: Ceramic Powder → Powder Milling → Foil Casting → Screen Printing → Lamination → Cutting → Binder Burn Out → Sintering (~1,100°C, differential shrinkage is the worst yield killer) → Tumbling → Dripping → Curing → Plating → Testing → Taping. Tantalum physics: anodized tantalum forms a Ta₂O₅ dielectric — very high capacitance per volume, high temp/voltage tolerance — essential for CPU/GPU power decoupling where MLCC can't supply enough bulk capacitance. Constraint is the ore: concentrated in DRC/Rwanda/Brazil, conflict-mineral certified, price-volatile; KEMET has the most established certified tantalum supply chain. KEMET's T598 polymer line is class-leading for AI-server low-ESR transient response.

Key customers & end markets

No named customer >10%; end-customer presence inferred from design-in commentary. NVIDIA (via ODMs) — designed in for tantalum power filtering on H100/H200/B200/Blackwell GPU rails and high-end MLCC on HBM decoupling; AI now ~15% of group revenue. Apple + Samsung (via Foxconn/EMS) — commodity + spec MLCC in phones, tablets, wearables, Macs. Auto Tier-1s (Bosch, Denso, Continental, Aptiv, Magna) — auto-grade MLCC, inductors, polymer caps, multi-year design-ins. EV OEMs (Tesla, BYD, VW, Hyundai-Kia). Distribution: Arrow (ARW), Avnet (AVT), WPG/Yosun (3702.TW). Switching costs are high at the high end (auto AEC-Q200 + AI reference-design design-ins are sticky 2-5 years), low at commodity (drop-in at distributor level).

Moat and competitive position

Layered moat: (1) cost leadership in commodity — chip resistor #1, mid-grade MLCC #1, prices the floor; (2) tantalum monopoly — #1 globally, >46% share via KEMET, limited substitution in AI-server power delivery; (3) qualifications in auto-grade + high-reliability — multi-year sticky design-ins; (4) one-stop-shop breadth — only Yageo and Murata can offer chip R + MLCC + Ta + inductors + sensors at scale. What Yageo does NOT have: the top-end MLCC technology lead. Murata wins the smallest case sizes and highest CV grades first (R&D ~7% of revenue vs Yageo ~2.7-2.8%). Direct competitors: Murata (6981.T, MLCC #1, ~38-42% share), Samsung Electro-Mechanics (009150.KS, #2, ~18-22%), Taiyo Yuden (6976.T, ~11-13%), TDK (6762.T, #4, battery-dominant), Walsin Technology (2492.TW, lower-tier Taiwanese), KYOCERA-AVX (tantalum #2), Vishay (VSH, broad-line). Porter: rivalry moderate-to-high (5 firms control most high-end revenue), buyer power moderate (distribution dilutes), supplier power moderate (tantalum ore concentration real), new entrants low (~3-5 years and >$1bn to scale high-end MLCC), substitutes low (passives are physics-mandated; on-die decoupling chips at the low end only).

Manufacturing footprint

Kaohsiung, Taiwan (Dashe, Nanzih, Dafa — main MLCC + chip resistor capacity, high-end MLCC concentrated here, under-utilized as of Q1 2026 with management guiding to lift Q2); Suzhou + Dongguan, China; Malaysia + Thailand (2023-2024 buildouts for geographic diversification); KEMET legacy plants (US, Mexico, Italy, Portugal, Japan-TOKIN, Indonesia — tantalum/film/polymer); Nexensos (Heilbronn, Germany — platinum RTD sensors, since 31 Mar 2023); Telemecanique Sensors; Pulse/Mag.Layers/Bothhand/Magic Technology (inductors); Shibaura Electronics, Japan (6957.T NTC thermistors, acquired Oct 2025 via 87.3% tender). Asset-heavy: capex NT$6.6bn in 2024 (5.4% of revenue, below the 8-12% historic range). Geographic revenue mix 2024: Greater China 50% (rose from 47%), Europe 23% (doubled from 15% since 2020 on KEMET/Nexensos), Americas 14%, Asia-Pacific ex-China 13%.

M&A engine (strategic minority stakes also held)

Growth has been roll-up acquisitions, not JVs: Vishay-Vitramon (2007), Pulse Electronics (2014, US$166M), KEMET (2020, US$1.85bn cash — tantalum #1 + TOKIN Japan), Chilisin (2021, full acquisition — inductors, ~15% of group revenue), Nexensos (Mar 2023, €79.4m from Heraeus — platinum RTD #1), Telemecanique Sensors (2023, from Schneider Electric), Shibaura Electronics (Feb-Oct 2025, tender — NTC thermistor #1). Strategic minority stakes: XSemi (50%), Global Testing (28.85%), KAIMEI Electronic (11.97%), Advanced Power Electronics (29.95%), uPI Semiconductor (20.21%, increased via private placement May 2024 — vertical integration into power-management ICs downstream of passives).

Financials

Revenue and margin trajectory (NT$bn except EPS/%; source: yfinance financials + company):

Metric FY22 FY23 FY24 FY25 Q1 26 LTM (est.) FY26E
Revenue 121.1 107.6 121.7 132.9 ~140 ~155
YoY % — -11% +13% +9.3% — +17%
Gross profit 46.0 36.0 41.8 48.1 ~52 —
Gross margin 38.0% 33.5% 34.4% 36.2% ~37% ~38%
Operating income 29.0 20.4 23.4 29.8 ~33 ~38
Operating margin 23.9% 19.0% 19.2% 22.4% ~24% ~25%
Net income 22.7 17.5 19.5 23.6 ~26 ~31
EPS post-split (NT$) 9.05 8.50 (est) 9.55 (est) 11.51 ~12.7 ~14.5-15

Gross margin: 33.5% (FY23) → 34.4% (FY24) → 36.2% (FY25) → 38.1% (Q1 26) = +460bps over 8 quarters. Operating margin: 19.0% → 19.2% → 22.4% → 25.2% (Q1 26) = +620bps. Real and structural, not one-off. (Note: the older profile pull listed FY25 GM 36.2% / OM 22.4% and an FY25 EPS of 11.48 in one table vs 11.51 in the deep-dive — minor rounding discrepancy, the deep-dive 11.51 is the cleaner figure.)

Incremental margin analysis (Q1 26 vs prior-year quarter): Q1 26 incremental gross margin 49% vs reported 38.1% — new revenue is materially higher quality than the base. Incremental operating margin 45% vs reported 25.2% — new revenue drops through to OI at ~2× the base rate. This is the AI-server tantalum + high-end MLCC mix-shift working in real time. If the mix shift continues, steady-state operating margin should converge toward 28-30% within 2-3 years (current 25%, incrementing at 45%) — a +500bps margin-expansion runway.

Quarterly revenue and second-derivative (yfinance, NT$bn):

Q1 25 Q2 25 Q3 25 Q4 25 Q1 26
Revenue 31.1 32.8 33.1 35.9 38.2
QoQ % — +5.5% +0.9% +8.5% +6.4%

Sequential growth positive every quarter since the Q4 2024 trough; second derivative (acceleration) positive Q4 25 (+8.5%) into Q1 26 (+6.4%); Q1 26 YoY +22.7%. Order visibility commentary extends into H2 2026; Q4 2026 will be the test. (Cycle bottom was Q1 2025: the -23.9% EPS miss.)

Cash flow & balance sheet (NT$bn):

Metric FY22 FY23 FY24 FY25 LTM Q1 26
Operating cash flow 22.3 28.4 33.8 32.5 ~36
Capex -3.4 -9.4 -6.6 -8.0 ~-7
Free cash flow 19.1 19.0 27.2 24.6 ~29
FCF margin 15.8% 17.7% 22.3% 18.5% ~21%
Net debt 55.7 50.8 65.8 66.8 66.8
Net debt / EBITDA 1.4x 1.7x 1.9x 1.5x 1.4x
ROE 16.6% 12.9% 13.0% 15.9% ~16%
ROIC (est.) 12% 9% 9.5% 11% ~12%

Cash + ST investments NT$98.1bn (FY25); total debt NT$150.5bn; net debt NT$66.8bn. NT$38.5bn new syndicated facility (Hua Nan-led, 2024-2029) terms out maturity. The balance sheet is heavily levered relative to typical Taiwanese tech peers — residual structure of the KEMET deal — so dividend/buyback flexibility is constrained vs Murata or TDK (net cash). ROIC ~11-12% (post-tax, incl. KEMET goodwill) > WACC ~8-9% by ~300bps — value-creating but a modest spread; ex-goodwill ROIC is materially higher. No dilution story: share count actually declined from 2.52bn (FY22 diluted) to 2.06bn (FY25) on the 2022 buyback; the August 2024 capital-from-earnings stock dividend (83.9M shares, ~17%) is a Taiwanese stock-split convention, not dilution; real dilution from converts (Chilisin CB + overseas CB) + employee RSAs is ~6% over 18 months. No outstanding US-style convertibles or warrants.

Industry landscape

Passive-components / MLCC market is consolidated: top 5 in MLCC ≈ 78% of revenue (top 5 high-end ≈ 68%), top 3 in tantalum >80%, top 3 in chip resistors ~60%. Global TAM: MLCC ~$13-15bn (6-8% CAGR), high-end MLCC ~$8bn (8-10%), tantalum cap ~$2bn (AI-driven re-acceleration), chip resistor ~$3bn, inductors ~$5-6bn, industrial temp sensors ~$3bn. Total addressable ~$30-35bn; Yageo's share ~12-13%. Murata is the ceiling ($16bn+ revenue, $90bn+ market cap). The cycle is K-shaped (TrendForce framing): high-end (AI server, auto, industrial) in a Murata-led bull cycle with tight supply and firm pricing (JP/KR utilization >80%, TW/CN 60-70% with Yageo on the lower end and room to lift); commodity (smartphone, consumer) range-bound near multi-year lows. Q1 2026 saw Murata, Taiyo Yuden, and Yageo all raise high-end prices for the first time in 4+ years. Historical MLCC up-cycles run 6-10 quarters trough-to-peak; Yageo bottomed Q4 2024, ~5 quarters in. Emerging threats: Chinese MLCC scale-up (Sunlord 002138.SZ, Fenghua, Eyang — ~10% of global capacity, low-end only); on-die/on-package decoupling (multi-year low-grade headwind to very-low-end MLCC, not tantalum); the realistic top-of-mind threat is a Korean/Japanese (Samsung EM / Murata) price-war response, not a Chinese new entrant. Upstream alpha candidate flagged: Sakai Chemical (4078.T), Japan's specialist BaTiO₃ powder maker (~5% of Yageo's market cap, supply-tight on next-gen ultra-thin dielectric).

See sector page: passives-mlcc

Management

Management grade: B+ (mgmt-dd) / B to B+ (deep-dive). Trust verdict: Yes, with eyes open. Red flags: none. Two soft yellow flags worth monitoring (founder share-shuffling + audit-committee independence).

Leadership

Pierre T.M. Chen (陳泰銘 / Tie-Min Chen) — Founder & Chairman since 1977 (47 years; founded at age ~21-22; age band 61-70 per proxy, born ~1956-57). BS Engineering, National Cheng Kung University; Honorary Doctorate in Management, NSYSU. Built Yageo from a single-product chip-resistor startup into #1 chip resistor + #1 tantalum (post-KEMET) + #3 MLCC through organic growth plus the M&A program. One of Taiwan's wealthiest individuals (Forbes net worth ~US$6.1bn, May 2024); ranked among the world's top 10 art collectors by ARTnews (~200 works incl. Picasso, Bacon, Richter, Cai Guo-Qiang, administered through Yageo Foundation, founded 1999). Operationally active — personally chaired the 27 May 2025 AGM and chairs the Strategic Investment Management Committee (the body that makes M&A decisions; independent directors do NOT sit on it — the M&A engine is Chen's personal authority). Compensation NT$62,739k (~US$2M, 0.32% of net income) — his wealth is in equity, not comp. No disclosed regulatory/litigation history; personally clean.

CEO & GM: Deng-Rue Wang (王淡如) — CEO + GM + Director since 4 May 2020 (~5 years); career-long Yageo insider (ex-COO; prior SVP & CFO of Qsida); EMBA NTU + MS Computer Engineering, UMass; concurrently Chairman of Chilisin Electronics. Compensation NT$146,852k (~US$4.7M, 0.76% of NI) — the highest-paid individual, more than the Chairman; appropriate. Discrepancy flag: the older profile listed the CEO as "David D.R. Wang (王淡如)" — same Chinese name 王淡如, so this is the same person under a different romanization (David D.R. / Deng-Rue), not two people.

CFO: Eddie Chen — joined 15 July 2021; Wharton MBA; prior CFO at Fubon Financial Holding and Chimei Innolux; recruited to manage post-KEMET capital-structure complexity. Comp band NT$30-50M. No family relationship to the Chairman (proxy: "Managers who are Spouses or Within Two Degrees of Kinship: None").

Bench (11 senior managers, professionalized and externally sourced): EVP T.Y. Chang (ex-Philips MLCC), SVP Brian Liu, SVP C.T. Lee (long-tenured, 27,689 shares), CIO Chris Yang (ex-NXP/Philips, 38,766 shares), CHRO Alison Tung (ex-TSRC), CPO William Chen (ex-GlobalWafers CPO), CLO Benjamin Kao (ex-Delta EMEA legal/IP, ex-Giant GC), VP Nick Chen (ex-DHL/DACHSER logistics), Accounting/Governance Officer Kevin Yang (NTU MAcc; ex-Senior Manager Audit, Deloitte). Not a founder's-friends bench.

Ownership & alignment

yfinance reports 23.5% aggregate insider. The 2024 annual report (31 Mar 2025 snapshot) shows: Tie-Min Chen direct 29,589,231 (6.98%); via spouse/minor children 35,360,267 (6.82%); TMC Family Heritage (99.999% Chen) 29,220,353 (5.63%). True Chen-family aggregate ~20-21% (deduplicating the double-listed TMC Family Heritage / TIE MIN CHUAN CHENG block) — materially more concentrated than the 23.5% "all insiders" figure implies, because most of that 23.5% IS Chen. Chen's ~20% stake is worth ~US$1.7bn per the mgmt-dd holdings table but ~US$7.3bn per the deep-dive holdings table at NT$572 — figures disagree; the ~US$7.3bn figure is the arithmetically consistent one (NT$1,178bn cap × ~20% ≈ NT$236bn ≈ ~US$7.3bn), while ~US$1.7bn appears to net to a narrower direct+family+TMC-FH block. Roughly 50-60% of Chen's ~US$6.1bn net worth is the Yageo stake; the rest is the art collection (~US$1bn+, Yageo Foundation-administered) plus stakes in smaller related companies he also chairs (TONG HSING, Advanced Power Electronics, uPI Semiconductor, XSemi). Independent directors hold zero Yageo equity — standard Taiwanese practice (cash comp, not stock), a contrast to US/UK norms. CEO Wang's direct stake is small but his career is fully tied to the franchise (also runs Chilisin). Aggregate director comp 4.47% of NI; major-manager comp 1.50% — reasonable for Taiwan. Restricted stock awards introduced 2024 (YMIP short-term + YLIP long-term plans; ~0.2% of share count, not dilutive) — a positive governance trend.

Capital allocation track record (overall grade A- to B+)

KEMET (2020, $1.85bn cash) A — bought a struggling tantalum company in a smartphone-weakness cycle; now the linchpin of the AI-server tantalum thesis. Chilisin (2021) B+. Nexensos (2023, €79.4m) A-. Telemecanique (2023) B+. Shibaura (2025) C+ to B- — got the asset (87.3% acceptance, NTC thermistor #1) but paid final ¥6,635/share vs opening ¥4,300 (+54% over opener) after a MinebeaMitsumi bidding war; cash-funded via a ¥45bn loan to Yageo Electronics Japan LLC, so not dilutive to Yageo equity, but the price-discipline question is real. Buybacks: NT$6.2bn in 2022 at depressed levels (validated by subsequent action); none 2023-2025. Capex discipline: 5.4% of revenue in 2024 vs 8-12% historic, no new MLCC capacity for 2026. Dividend NT$6/share FY25 (~52% payout, ~1% yield). Timing test: clear cost-of-capital awareness at the KEMET (2020 trough) and 2022 buyback (trough); the 2025 Shibaura deal at elevated valuations is the one mark against, but cash-funded so not value-destructive to equity.

Governance flags

  • Board: 10 directors, 30% independent (3 of 10) — meets Taiwanese standard, below US/UK best practice (50%+). Single-class shares (one share = one vote), no poison pill, no staggered board. Controlling-family ~20% stake is the natural anti-takeover.
  • Joy Chen (Pierre Chen's eldest daughter) elected to the board at the 27 May 2025 AGM via TMC Family Heritage — widely read as the cleanest succession-planning signal. Chen-I Hsu added as independent director at the same AGM.
  • Audit + Compensation + Nomination committees all chaired by Cheng-Ling Lee — technically independent (ROC CPA, independent director since 2012, cooling-off long passed) but a former Vice Chairman of Yageo and ex-Deloitte audit manager (Deloitte is the current auditor). Procedurally compliant, substantively the soft-end of independence. Hsu Tun Son Lin (Whitesun Equity Partners Chair, PhD King's College London, Dominica nationality) and Lai-Fu Lin (UHY L&C partner) are more substantively independent.
  • Founder share-shuffling (Oct 2024 - Apr 2025): Chen moved blocks of personal shares into two short-lived shell entities (CHEN SHI JIA ZU Co. — 29,214,379 shares Oct 2024; TIE MIN CHUAN CHENG Co. — 6,139,914 shares Mar 2025) then merged both into TMC Family Heritage within 3-4 months. Net economic effect zero (same shares, same ultimate ownership). Disclosed transparently as "Director holds shares in name of others." Most likely Taiwanese estate/inheritance-tax planning (aligned with the Joy Chen appointment weeks later), but pattern-matches NongAap's "asset-shuffling-through-newly-created-entities" flag. Yellow flag — monitor; not insider selling.
  • Audit-quality soft yellow: Accounting/Governance Officer Kevin Yang was a Senior Manager at Deloitte (current auditor) before joining; the annual report formally denies a revolving door at the CPA-firm partner level but Yang's bio confirms ex-Deloitte audit. Non-audit fee 43% of audit fee (explained by ESG + M&A advisory).
  • Shell scan: 80+ subsidiaries across Bermuda, Cayman, Samoa, HK, China, Vietnam, Japan, US, EU — the product of 20+ years of M&A, not a tax/asset-shielding architecture. The 4 Samoa entities (Chilisin Holding, Magic Technology, Classic Magic Developments, Trendy Island Investment) are 100% Yageo-owned but their function is not described in the annual report — disclosed-but-unexplained (working hypothesis: Chilisin-legacy IP/treasury vehicles). ~14.4% of float held via opaque foreign custody wrappers (Dominant Investment Holdings 4.85% via CTBC custody, PRC Holding Limited 4.83%, Wholly Group Japan III 4.69% via Mega Bank custody) — beneficial owners not disclosed.
  • Litigation: none material. The KEMET-legacy US capacitor antitrust class action (2014-2019) settled substantially before the 2020 Yageo close; Yageo was not a primary defendant.

Credibility / follow-through (HIGH)

9 of 11 EPS beats over the last 11 quarters (the lone big miss was Q1 2025, the cycle bottom, -23.9%; a small -2.0% Q4 2024 miss preceded it; six consecutive beats since with widening margin: +0% → +4.5% → +5.7% → +9.2%). Wang's commentary runs measurably conservative ("moderate," "slight") and consistently undershoots delivery — a sandbagger posture that builds buy-side trust. Substantive operational claims (headcount cut 36,101 → 30,624, OM 19% → 25.2%, AI design-in, FY25 cash flow) all followed through; the one slip was the Shibaura timeline (guided Q3 2025, completed Oct 2025) due to the external MinebeaMitsumi bid war, not management failure. Key-person risk is real and material — Chen is the strategic engine; succession is in motion (Joy Chen + CEO Wang) but not closed; 5-15 years of continued Chen involvement is actuarially plausible.

Catalysts & risks

Catalysts (bull)

  • AI-server content uplift — content per rack rises ~2-3× per NVIDIA platform refresh; tantalum supply tight; AI ~15% of group revenue, targeted >15% by year-end 2026; durability 5-10y+.
  • Capacity-utilization recovery — Kaohsiung high-end MLCC under-utilized through the post-COVID smartphone downcycle; management guiding to lift utilization in Q2 2026 with no new greenfield — incremental volume drops to margin.
  • EV transition — 2-3× passive content per BEV vs ICE; auto-grade MLCC qualified at most Tier-1s; multi-year design-ins.
  • Tantalum scarcity — limited global supply, KEMET #1 (>46% share); structural pricing power.
  • Industrial-sensor consolidation — Nexensos + Telemecanique + Shibaura roll-up (~10-15% of group revenue at full integration); high-margin, less cyclical.
  • High-end pricing power — Q1 2026 industry-wide high-end price hikes (Murata/Taiyo Yuden/Yageo) for the first time in 4+ years.
  • Near-term events: 28 July 2026 Q2 earnings (consensus EPS NT$4.09); May/June 2026 monthly revenue prints; Shibaura integration/accretion disclosure; sell-side PT catch-up toward NT$500+. Medium-term: sustained AI density on NVIDIA Rubin/Vera Rubin platforms; EV cycle re-acceleration; Pierre Chen formal succession; possible MLCC capacity-expansion announcement when high-end utilization sustainably tightens.

Risks (bear)

  • Multiple compression as the cycle ages — High in the next 6-12 months and the dominant near-term risk; even if EPS grows ~25%, the stock can drop 15-25% on the multiple compressing from 27x fwd to 22x. Not manageable except by not buying at the high.
  • AI capex normalization — Medium-low 12-18m, medium 24m+; a hyperscaler digestion year H2 26 / H1 27 softens tantalum demand (~15% of revenue at risk). Mitigated by end-market diversification.
  • Commodity MLCC pricing — chronic/high; volatile through smartphone cycles; manageable via mix-shift (management pulling capacity off low-end grades).
  • China geopolitical / tariff exposure — Medium; ~50% of revenue is Greater China shipped; escalation could hit 5-10% of revenue. Mitigated by Malaysia/Thailand/KEMET footprint.
  • Tantalum raw-material concentration — Medium; African ore concentration + conflict-mineral/ESG exposure. Mitigated by KEMET certified supply chain + multi-source contracts.
  • R&D gap to Murata — persistent; Yageo R&D ~2.7-2.8% of revenue vs Murata ~7%, Taiyo Yuden ~5%; structural reason it trails on highest-end MLCC.
  • Korean/Japanese price-war response — low-medium; Samsung EM defending share or Murata accelerating capacity could slow Yageo's catch-up. The realistic top-of-mind competitive threat.
  • Family-control governance + succession — persistent (low operational risk, high tail risk); concentrated in the Chen family.

What would make the thesis wrong

Any quarter with sequential revenue declining + management guiding for sequential weakness; Murata or Samsung EM announcing aggressive price cuts to defend share; NVIDIA Rubin/Vera Rubin reference-design BOM showing reduced tantalum content; any auditor change or accounting restatement; any unexplained acceleration in the founder share-shuffling pattern.

Valuation / DCF

Yageo is the most expensive name in its comp set on every measure. Premium to its own history: trailing P/E 2.3x its 5-year average, forward P/E 1.6x, EV/EBITDA 2.1x. Premium to peers: trades 30-60% above Murata on forward P/E and EV/EBITDA despite Murata's technology lead, ~2.5x larger R&D as % of revenue, and net cash.

Multiple Yageo 2327 5-yr avg Murata 6981 Samsung EM 009150 Taiyo Yuden 6976
TTM P/E 49.8x ~22x ~38x ~22x ~28x
Forward P/E 26.8x ~17x ~28x ~18x ~22x
EV/EBITDA TTM 25.5x ~12x ~17x ~9x ~13x
EV/Revenue 8.0x ~2.5x ~3.5x ~1.2x ~2.5x
P/Book 7.0x ~3.0x ~3.5x ~1.5x ~2.5x

DCF framing (rough):

  • Base case: Revenue NT$155bn FY26 → NT$170bn FY27 (+10%) → NT$185bn FY28 (+9%); operating margin 25% → 27% → 28%; FCF conversion ~80% of OI net of capex; WACC 9%; terminal growth 3%. Equity value ~NT$900-1,000bn → ~NT$450-490/share (~15-20% below current).
  • Bull case: Revenue +15% FY26-27 on stronger AI tantalum + Shibaura accretion; operating margin 30% by FY28; equity value ~NT$1,400bn → ~NT$680/share (+19%).
  • Bear case: Cycle rolls over by H2 26, FY27 revenue flat, margin -200bps; equity value ~NT$700bn → ~NT$340/share (-40%).

Implied expectations at NT$572: the market is paying for ~12-13% revenue CAGR for 3-5 years, operating-margin expansion to ~28-30%, and continued multiple support above the 5-year average. Achievable but not safely priced.

12-month target (deep-dive base case): NT$540-620 — implies flat-to-+8% return as forward EPS grows ~25% but P/E compresses from 27x to 22-24x. Bull NT$700 (FY26 EPS NT$22+ at 32x), bear NT$380 (mean-revert to consensus PT). Expected return is asymmetric to the downside at this entry: upside +22% (bull), base flat-to-low-single-digit, downside -33% to -40% (bear).

Sell-side disconnect: mean analyst PT NT$423 sits 26% below current (35% below spot framing elsewhere) and median NT$382.50 sits 33% below — the sell-side has not followed the stock up the AI-tantalum re-rate. Either stale (lagging the mix-shift narrative) or skeptical the cycle holds; both could be true. One Strong Sell rating has persisted four months (likely a Japanese broker bearish on the cycle). Would I still buy at 10-15% higher? No — NT$640-660 is in the bull-case fair-value zone with no margin of safety. Valuation verdict: NOT reasonable at current.

Decision log

2026-05-20 — Profile, deep-dive, mgmt-dd, and pre-buy checklist all written (4-stage swarm pass). Deep-dive: conviction Medium-Low at price / Medium-High on business; 12m base PT NT$540-620, bull NT$700, bear NT$340-380. Mgmt-dd grade B+, trust verdict "Yes, with eyes open," no red flags, two yellow flags (founder share-shuffling + audit-committee independence).

2026-05-20 — Checklist verdict: WATCH (do not buy at NT$572). Scorecard 6 Yes / 4 No — the four No's are valuation, technicals (52-week high + RSI likely overbought), behavioral traps (FOMO + confirmation + recency bias active), and a clear exit plan (targets below current). All 5 FundamentEdge hard-rule gates pass (revenue-growth primacy, positive 2nd derivative, valuation-is-not-a-thesis, quality-as-oligopoly-moat, upward estimate-revision direction) — the thesis is structurally sound; the discipline question is entirely price.

Conditional buy plan (entry discipline mechanism — 0% allocation at NT$572):

  • Tranche 1 (starter, 30% of target = 1.0-1.5%): NT$480-520 — first 10-16% pullback / sector consolidation.
  • Tranche 2 (base, 40% = 1.5-2.0%): NT$420-460 — at/below mean analyst PT NT$423.
  • Tranche 3 (conviction, 30% = 1.0-1.5%): NT$340-380 — bear-case zone, only if cycle headlines confirm digestion.
  • Target portfolio weight at full conviction: 3.5-4.5%. Hard stop -25% from each tranche entry. Time stop H2 2026 / re-evaluate Q3 2026 if no tranche triggers.
  • Exit triggers (full liquidation): auditor change (Deloitte departure); financial restatement; any founder share movement that turns out NOT to be estate planning; Murata/Samsung EM aggressive price-war; AI tantalum content reduction on NVIDIA reference design.

Peer-swarm context (2026-05-20): the MLCC swarm (Yageo + Taiyo Yuden + TDK + Walsin + PDC/6173 + Holy Stone/3026) is uniformly at or above analyst PT mean — the cycle re-rate is distributed across the whole sector. Yageo is the highest-quality core MLCC name but also the most expensive; Walsin is the higher-quality Taiwanese-tier pullback name; Sakai Chemical (4078.T) is the upstream BaTiO₃ bottleneck alpha candidate (~5% of Yageo mcap, less crowded) flagged for a separate /profile run. A parallel swarm node reportedly ranked PDC > Yageo > Murata > Taiyo Yuden > Walsin > Samsung E-M > Fenghua at then-current prices.

2026-05-23 — Briefing: "MLCC Sector — Top 2 With Crowding in Mind" (handoff). Yageo scored Pass (basket 17/30, Crowding 2/5, Revision Velocity 2/5) on a NT$629 limit-up vs GS PT NT$346 = 82% above target; reframed as a tantalum story, not an MLCC story.

Filings cross-check (per checklist): no restatement, no auditor change, no SEC actions, no material weakness, no insider open-market selling. Shibaura tender (Feb-Oct 2025) was the major M&A event; Joy Chen director appointment (May 2025) the major governance event. No red flags that change the verdict.

Current stance: WATCH — do not buy at NT$572; scale-in plan is the discipline. Same-theme cleaner alternative on the watchlist: Sakai Chemical (4078.T).

Sources

Consolidated from four vault fragments (2026-05-20 swarm pass): 2327.md (profile, originally written 2026-05-20, updated 2026-05-23), 2327-deep-dive.md, 2327-mgmt-dd.md, 2327-checklist.md. (A 2327-filings.md stub in the folder was an empty placeholder and was not merged.)

External sources cited across fragments:

Related vault coverage: serenity-method.md (MLCC roll-call), santec-vs-jem-vs-anritsu-and-more-versus.md (Taiyo Yuden as Japanese MLCC peer); briefing briefings/2026-05-23-mlcc-sector. SemiAnalysis mirror: no dedicated Yageo or passive-components primer (only indirect hits — a 2022 Roundup bearish on Murata MLCC, wrong over 4 years); no contradiction to flag.

Consolidation queue (merged 2026-05-30)

These four fragment files were folded into this canonical page on 2026-05-30 and remain live pending Pink's archive confirmation.

  • [ ] 2327-mgmt-dd.md
  • [ ] 2327-deep-dive.md
  • [ ] 2327-checklist.md
  • [ ] 2327.md

Source updates (auto-maintained)

Drop/MLCC (May 24, 26) - 东莞证券_MLCC行业深度报告：供需矛盾加剧，高阶MLCC价格有望上扬_260325 (1)

Dongguan Securities (Mar 2026) reports Yageo's MLCC delivery times and prices are already lengthening in Q1, and charts Yageo's MLCC revenue rising through Q4 2025, consistent with the supply-demand tightening thesis driven by Murata and SEM running at full capacity with 15–35% price hikes starting April 2026.

Relevant to your thesis: Directly supports the bull case — high-end MLCC pricing power is inflecting now, which underpins the mix-shift margin expansion already observed in the wiki's Q1 2026 operating margin data.

Source: dropfile://MLCC/东莞证券_MLCC行业深度报告：供需矛盾加剧，高阶MLCC价格有望上扬_260325 (1).pdf

Drop/MLCC (May 21, 26) - global_passives_basket_comparison

The basket comparison scores Yageo 17/30 (Pass) at TWD 520, with Crowding, Valuation, and Revision Velocity each at 2/5; sell-side EPS revisions of 3–7% have not kept pace with the ~178% 12-month price move, leaving shares ~50% above the most recent major-bank target.

Relevant to your thesis: Independently corroborates the core bear case: a separate framework flags the same revision-vs-price disconnect the wiki thesis names as the central risk.

Source: dropfile://MLCC/global_passives_basket_comparison.pdf

Drop/MLCC (May 20, 26) - Yageo Corp. (2327.Tw) Profitability driven by solid AI deman...

Goldman Sachs (Apr 16, 2026) reiterates Buy on Yageo with a revised NT$346 TP, citing 1Q26 earnings 4-7% above consensus, AI revenue at 14-15% of sales (guided to exceed 15% in 2026), tantalum holding the highest BB ratio across all products, and management guiding 2Q26 UTR to rise to 75%/85% for standard/premium lines.

Relevant to your thesis: Confirms the AI-driven tantalum and mix-shift thesis — but GS's NT$346 TP (now ~37% below the NT$551 live price) is the strongest analyst anchor for the bear case on valuation.

Source: dropfile://MLCC/Yageo Corp. (2327.Tw) Profitability driven by solid AI demand will be a new norm; Buy, with new TP of NT$346.pdf

Drop/MLCC (May 21, 26) - Murata vs Yageo (5.21.2026) (1)

Murata vs. Yageo comparison confirms Yageo's Q1 2026 record: NT$38.2B revenue (+22.7% YoY), 25.2% operating margin, and AI tantalum revenue >30% of tantalum sales; the piece also argues Yageo's AI MLCC narrative is overstated — the real driver is inventory-correction recovery plus tantalum, not GPU-rack MLCC BOM exposure, and Yageo is excluded from the 4-supplier ultra-high-CV MLCC tier.

Relevant to your thesis: Reinforces the wiki's existing read that Yageo's AI story is primarily a tantalum story, and corroborates the WATCH/valuation concern — 26x forward P/E with a PEG of 0.58 is cheaper than Murata, but the cyclical (not structural-moat) nature of the margin recovery supports caution at ATH pricing.

Source: dropfile://MLCC/Murata vs Yageo (5.21.2026) (1).pdf

Drop/MLCC (May 24, 26) - J.P. Morgan-MLCC Industry：Growing likelihood of tight supply...

J.P. Morgan (Apr 2026) upgrades Murata and Taiyo Yuden to Overweight on rising AI server MLCC demand, forecasting capacity utilization reaching 90% mid-2026 with AI server MLCC units growing 50-60% annually through 2027 and GB300 content up 50-60% vs GB200; Yageo is mentioned as a market participant but not rated.

Relevant to your thesis: Corroborates the K-shaped recovery and AI-content-per-server expansion thesis, but JPM's preferred plays are Japanese names, consistent with the wiki flagging Yageo as crowded relative to less-consensus alternatives.

Source: dropfile://MLCC/J.P. Morgan-MLCC Industry：Growing likelihood of tight supply demand； Murata Manufacturing and Taiyo Yuden up to Overweight-260403.pdf

Drop/2. Frontend (May 24, 26) - MS Rubin Rack_20May26

Morgan Stanley's bottom-up Rubin rack BOM analysis estimates MLCC content per rack at ~US$4,320 for VR200 vs. ~US$1,530 for GB300, a +182% increase, driven by higher content on compute and switch boards plus new BlueField and ConnectX modules.

Relevant to your thesis: Directly corroborates the AI-server MLCC demand thesis — the +182% MLCC content step-up in Rubin validates the structural unit-content tailwind underpinning Yageo's AI revenue mix-shift narrative.

Source: dropfile://2. Frontend/NVDA/MS Rubin Rack_20May26.pdf

Drop/Bottleneck (Jun 17, 26) - 0 - JPM 150626 - MLCC substrates

JPM (June 2026) confirms Yageo benefits from an AI server tailwind in low-end MLCC — high-end capacity consumption by AI servers tightens low-end supply/demand balance, lifting Yageo's low-end MLCC pricing and utilization rates; tantalum and resistor pricing also forecast to firm.

Relevant to your thesis: Supports the K-shaped recovery and AI-as-demand-catalyst narrative, with JPM's UTR and pricing forecasts reinforcing the bull case on margin expansion from current under-utilization.

Source: dropfile://Bottleneck/MLCC/0 - JPM 150626 - MLCC substrates.pdf

Intake (May 23, 26) - mops-diligence-2026-05-23

The MOPS diligence sweep confirms Yageo's ISS QualityScore of 4 — materially better than peer Walsin (ISS score 10, worst decile) — and validates that Yageo's full-ownership acquisition model (KEMET, Chilisin, Nexensos) carries less related-party opacity than the PSA group's partial-stake cross-entity structure.

Relevant to your thesis: Reinforces the governance quality gap that supports Yageo's premium over Walsin and PDC, consistent with the wiki's B+ people rating and absence of red flags.

Source: intakefile://mops-diligence-2026-05-23.md

Intake (May 21, 26) - 2327.TW-filings

Yageo's 2024 annual report and Q1 2026 results show six consecutive beats since the Q4 2024 cycle bottom, with Q1 2026 EPS of NT$3.90 (+9.2% vs. consensus), operating margin expanding to 25.2%, and the Shibaura Electronics tender completing Oct 2025 at 87.3% acceptance after a bidding war raised the price 54% above the opening ¥4,300 bid.

Relevant to your thesis: The unbroken beat streak and margin trajectory confirm the operating-leverage thesis, while Shibaura's overpay risk and China revenue rising to 50% reinforce the existing bear points on valuation and geographic concentration.

Source: intakefile://2327.TW-filings.md

Drop/Bottleneck (May 24, 26) - 东莞证券_MLCC行业深度报告：供需矛盾加剧，高阶MLCC价格有望上扬_260325 (1)

Dongguan Securities (March 2026) shows Yageo's MLCC delivery times and prices already trending up in Q1 2026 alongside Murata and Samsung EM, with Murata raising AI-server and auto-grade prices 15–35% from April and Samsung EM following with double-digit increases.

Relevant to your thesis: Confirms the high-end MLCC pricing inflection underpinning Yageo's margin expansion narrative, reinforcing the bull case on mix-shift and AI-driven pricing power.

Source: dropfile://Bottleneck/MLCC/东莞证券_MLCC行业深度报告：供需矛盾加剧，高阶MLCC价格有望上扬_260325 (1).pdf

Drop/Bottleneck (May 21, 26) - global_passives_basket_comparison

The May 2026 basket comparison scores Yageo 17/30 (Pass) at TWD 520, flagging Crowding (2), Valuation (2), and Revision Velocity (2) as the drag dimensions, with the stock trading ~50% above Goldman's just-raised PT of NT$346 and sell-side EPS revisions of only 3–7% not keeping pace with a ~178% 12-month price move.

Relevant to your thesis: Directly reinforces the wiki's core bear point — price has run far ahead of fundamentals — and adds a peer-relative frame: Yageo ranks below even Kingboard (Watch) in the basket, with Murata rated Buy as the cleaner risk-reward in the same theme.

Source: dropfile://Bottleneck/MLCC/global_passives_basket_comparison.pdf

Drop/Bottleneck (May 20, 26) - Yageo Corp. (2327.Tw) Profitability driven by solid AI deman...

Goldman Sachs (Apr 16 2026) reiterates Buy on 2327 with a revised 12-month TP of NT$346 (up from NT$302), citing 1Q26 results 4-7% above consensus, AI revenue at 14-15% of group and guided to exceed 15% in 2026, tantalum as the highest BB-ratio product with solid 2H26 visibility, and UTR expected to rise to 75%/85% for standard/premium MLCC in 2Q26.

Relevant to your thesis: Confirms the AI mix-shift and tantalum pricing-power bull points, but GS's NT$346 TP (vs. current NT$572) reinforces the wiki's core bear: the stock has already traded well through sell-side fair value.

Source: dropfile://Bottleneck/MLCC/Yageo Corp. (2327.Tw) Profitability driven by solid AI demand will be a new norm; Buy, with new TP of NT$346.pdf

Drop/Bottleneck (May 21, 26) - Murata vs Yageo (5.21.2026) (1)

Murata vs. Yageo comparison (May 21, 2026) confirms Yageo's Q1 2026 results (revenue +22.7%, operating margin 25.2% record, EPS NT$3.90, AI revenue ~14-15%), notes Yageo's moat is breadth not depth, and argues the AI MLCC thesis is overstated — the real AI driver is tantalum (>30% from AI servers), not MLCC rack BOM exposure, which represents only ~3-6% of annual revenue.

Relevant to your thesis: Reinforces the wiki's own framing ("Yageo's AI story is more accurately a tantalum story") and the valuation skepticism, while adding the specific data point that AI rack MLCC is ~3-6% of revenue — a bear point against AI-MLCC hype, bull point for tantalum.

Source: dropfile://Bottleneck/MLCC/Murata vs Yageo (5.21.2026) (1).pdf

Drop/Bottleneck (May 24, 26) - J.P. Morgan-MLCC Industry：Growing likelihood of tight supply...

J.P. Morgan forecasts MLCC capacity utilization reaching 90% by mid-2026 (now 87–88%), AI server MLCC demand growing 50–60% annually, and GB300 carrying 50–60% more MLCC content than GB200; JPM upgrades Murata and Taiyo Yuden to Overweight but does not cover Yageo.

Relevant to your thesis: The UTR inflection supports the wiki's "sector is inflecting" call; JPM's explicit preference for Japanese names over Yageo is a soft contra on consensus crowding at current prices.

Source: dropfile://Bottleneck/MLCC/J.P. Morgan-MLCC Industry：Growing likelihood of tight supply demand； Murata Manufacturing and Taiyo Yuden up to Overweight-260403.pdf

from the vault · open in obsidian ↗