4971 — MEC Company Ltd.

Thesis

Stance: PASS / WATCH at the current price; re-engage ¥7,000-9,000. MEC Company Ltd. (4971.T, TSE Prime) is a best-in-class Japanese specialty-chemicals franchise selling the copper microetching chemistry (CZ-series) that has been spec-locked into the world's top PCB and IC-substrate fabs for two decades. The business is qualitatively excellent — recurring spec-locked consumable, ~100% global share in copper surface treatment for PC CPU substrates, 61.6% FY2025 gross margin / 27.4% operating margin, net cash, a 24-year operator-CEO with a $49M personal stake, and A-grade capital allocation. The problem is the entry price, not the business.

What has to be true for the thesis to work: CZ-8101 holds its qualified spec at Ibiden / Unimicron / Shinko through the Blackwell → Rubin → next-gen substrate transition. If CZ stays in the recipe, the whole thesis works mechanically — AI server PCB layer-count expansion and ABF substrate growth push CZ consumption up structurally through 2028 as a volume-and-mix story (more inner layers and tighter loss budgets = more CZ-treated surface per board).

Why PASS now: the stock ran +377% off its 52-week low into a 40.5-41.1x TTM P/E — roughly double its 5-year normal multiple range of 14-22x — while the mean analyst PT of ¥8,900 sits ~19.5% below spot. The buy-side has already priced what the sell-side will revise to; new buyers are paying for execution that hasn't happened yet. The asymmetry is poor: bear case ¥4,400 (-60-61%) vs bull case ¥13,200 (+18-19.5%) over 12 months. Three independent signals — valuation, behavioral traps (FOMO + narrative seduction + recency bias), and technicals (overbought, at 52w high, parabolic) — all point the same way: do not chase. STF Research called this at ~20x P/E on 2026-03-01; that thesis has played out cleanly and is now consensus, so there is no longer an out-of-consensus edge.

The single biggest risk is not a thesis crack but multiple compression: even with flat fundamental delivery, 41x → 22x is -47% on the stock. The business deserves a position in the portfolio at the right price. This is not the right price. Wait.

Snapshot

One-liner: CZ-series copper microetchant supplier with ~100% PC CPU substrate share; the chemistry layer of the AI server PCB + ABF substrate + 5G AiP build-out.

• Ticker / exchange: 4971.T — Tokyo Stock Exchange (TSE) Prime. GICS Materials / Specialty Chemicals. HQ Amagasaki, Hyogo, Japan. Founded 1969. 508 employees (Dec-2025). Fiscal year = calendar year (Dec). Website https://www.mec-co.com
• Price (2026-05-15 EOD): ¥11,050 — down 1.3% from the ¥11,190 reference in the 2026-05-12 entry; -9.6% from the new 52-week high of ¥12,230 set 2026-05-14. 52-week low ¥2,385 (the 2026-05-12 entry cited ¥2,345; minor discrepancy between fragments). Stock is +363% off the 52-week low (+377% per the 2026-05-12 checklist).
• Market cap: ¥201.8B (~$1.34B USD at ¥150). Enterprise value ¥217.8B. Beta 1.06.
• Valuation snapshot (2026-05-15): P/E TTM 40.5x (was 41.1x on 5/12); forward P/E (FY+1) 56.9x (was 57.7x); EV/EBITDA TTM 75.7x depressed by Q1-windowed EBITDA — use FY2025 EBITDA ¥7.3B → ~29.8-31x normalised; P/B 14.5x (was 14.7x); dividend yield 0.95% (was 0.76% on 5/12); P/FCF FY25E ~48x; EV/Revenue ~9.9x.
• Key stats: FY2025 revenue ¥20.9B (+14.9% YoY), gross margin 61.6%, operating margin 27.4%, net margin 24.0%, diluted EPS ¥272, ROE 12.8%, ROA 7.9%. Net cash position (total debt ¥1.25B). Q1 FY2026 diluted EPS ¥25.44 on 18.73M weighted shares.
• Analyst coverage: 2 analysts (thinned from 3 on 5/12); mean PT ¥8,900 (high ¥10,000, low ¥7,800) — ~19.5% below spot.

Business

What MEC sells. MEC is a Japanese specialty-chemicals company selling microetching and surface-treatment chemistries to the printed circuit board and IC substrate supply chain. The franchise rests on one product family — the CZ-series copper microetchant — which has effectively monopolised inner-layer copper roughening for high-end multilayer PCBs and ABF substrate cores for two decades. Per sponsored research, MEC claims ~100% global share in copper surface treatment for PC CPU substrates, and CZ accounts for roughly half of corporate revenue (CZ-series 66.9% of chemical-segment sales per the FY25 breakdown in the deep-dive).

The economic pitch in one line: MEC sells a few hundred yen of chemistry per substrate panel, but it is the chemistry that lets PCB fabs hit the adhesion, dielectric-loss and laser-drill specs that AI servers, ABF substrates and 5G AiP modules now demand. Layer-count expansion, finer line/space and tighter loss budgets all raise consumption intensity per board.

Product lines (no formal segment disclosure beyond geography):

• CZ-series microetchants / adhesion enhancement — copper roughening for multilayer PCB inner layers, ABF/BT IC substrate cores, high-frequency substrates. ~50% of revenue.
• AMALPHA — proprietary metal-to-resin direct bonding chemistry. Smaller revenue today; positioned as the future structural-bonding play (5G AiP, power modules, EV battery cell packs, HBM RDL adjacency). MEC owns it, so direct-bonding substitution risk is internalised.
• V-Bond series — environment-friendly alternative to black oxide; legacy multilayer PCB application.
• AP-series — next-gen no-roughening platform; medium-term substitution risk MEC has internalised, extends runway into 2030+. Forecast was cut for timeline maturity, not product failure (IR-confirmed to STF Research).
• Other surface-treatment chemistries — pre-lamination treatment, DFR pre-treatment, laser-direct-drilling copper prep, degreasing/anti-tarnish, residue removers.
• Process equipment — small business; spray/dip processing lines sold alongside chemistries.

Business model. Recurring consumable chemistry sold by the litre/kg through long-qualified specifications at PCB and substrate fabs. Customers cannot swap suppliers without re-qualifying the board recipe with their end-customer (Intel, AMD, Nvidia, Apple). High switching cost, low individual ASP. Switching cost > chemistry cost.

The moat:

• Spec lock-in. Each board recipe is qualified at the OEM (Intel, Nvidia, AMD) level. Re-qualifying a different microetch is a 6-18 month cycle that PCB fabs and OEMs avoid unless the incumbent fails.
• Process know-how. CZ behaviour at sub-0.1µm roughness is held tacitly; competitors can match the bulk chemistry but struggle on consistency across millions of panels.
• Margin signal. 61.6% gross / 27.4% operating in a chemicals business is the moat — if a credible substitute existed, pricing would have collapsed.
• Scale at the high end. AI server PCB and ABF substrate volumes are concentrated in ~10 fabs globally; MEC is qualified at all of them. A blank-check disruptor would need 5-7 years and ~$200M R&D plus a multi-year requalification cycle.

Customers (inferred, not disclosed). MEC publishes no named customers or top-customer %. Inferred top customers: Ibiden (4062.T, ABF IC substrate), Unimicron (3037.TW, ABF substrate + HDI PCB), Compeq / Tripod / Nan Ya PCB (2313.TW / 3044.TW / 8046.TW, AI server HDI PCB), Shinko Electric (6967.T, IC substrate), Kinsus / Simmtech (3189.TW / KQ:222800, IC / memory substrate). If Ibiden + Unimicron together >40%, that is structural concentration risk in the substrate cluster — Yuho follow-up outstanding. Dependency flag: if Korean (Samsung, LG Innotek) or Chinese (Shennan, Victory Giant) entrants displace incumbents and specify different chemistry, MEC loses with the incumbents even though CZ itself is not the loser; the Suzhou expansion is the de-risking move.

Competitive position. Top competitors: Atotech / MKS Instruments (MKSI, lost high-end substrate share to MEC over the last decade); Uyemura International (private Japan, plating-focused, limited in microetching); Shipley / Rohm and Haas (within Dow, DOW, legacy, less focused on CZ-equivalent); Resonac (4004.T, adjacent on substrate films, not a direct microetchant competitor). Porter snapshot: low rivalry at the high end (CZ ~100% PC CPU share), moderate at legacy; commodity input suppliers (sulfuric acid, hydrogen peroxide); asymmetric buyer power (re-qual cost > chemistry cost); high entry barriers. STF Research frames MEC additionally as "an unrecognized HVLP4 player" — the chemistry side of the HVLP4 (very-low-profile copper foil) transition, tying it to the Mitsui Mining & Smelting / Furukawa Electric / Fukuda Metal HVLP4 thesis tracked elsewhere in the vault.

Operations footprint (six reporting bases): Japan, Taiwan, Suzhou (China), Zhuhai (China), Thailand, Europe — Asia ex-Japan is the majority. Amagasaki HQ + Japan plant (R&D + core production); Taiwan plant (Tripod, Compeq, Nan Ya PCB, Unimicron cluster); Suzhou (+~30% capacity by Dec-2026, AI server PCB/substrate); Zhuhai (south China cluster); Thailand (SE Asia); Europe (small, HDI); Kitakyushu new Japan plant under construction, scheduled to start operations Dec 2026 (AI substrate chemistry hub; depreciation drag factored into 2027 OM guidance). No JVs — wholly-owned subsidiaries in each geography; PCB-fab relationships are pure supplier-customer, not equity-linked.

Why the technology matters (first principles). Every advanced multilayer PCB and IC substrate needs a roughened copper surface so the next dielectric layer adheres without delaminating. As line/space shrinks (HDI → mSAP → ABF) and dielectric-loss budgets tighten (M9-grade CCL, Megtron 8/M9), conventional brown-oxide roughening leaves copper too rough — signal loss and yield collapse (the "pink-ring" failure mode). CZ gives a controlled ~0.1µm roughness profile that holds adhesion without killing high-frequency loss, which is why CZ has held share through 20 years of PCB tech transitions while competitors rotated through alternatives.

Financials

All figures JPY unless noted; FY = calendar year (Dec). Live data 2026-05-15 EOD (yfinance) plus MEC IR.

Revenue and margins.

Metric	FY2022	FY2023	FY2024	FY2025	FY2026E
Revenue	¥16.3B	¥13.5B	¥18.2B	¥20.9B	¥23.8B
Revenue growth YoY	—	-17%	+35%	+14.9%	+13.8%
Gross margin	60.0%	n/a	n/a	61.6%	n/a
Operating margin	24.5%	n/a	n/a	27.4%	n/a
Net margin	18.8%	n/a	n/a	24.0%	n/a
EPS (diluted)	¥161	n/a	n/a	¥272	¥275 (n=2 consensus)

FY2023/FY2024 P&L middle rows are gaps that the Japanese-language Yuho would fill. 3-year revenue CAGR FY22-FY25 ≈ 24% (strong).

Margin discrepancy across fragments (flagged): the 2026-05-12 buy-checklist cites 9M FY25 gross margin 62.4% and operating margin 28.3% ("expanding"); the 2026-05-15 profile and deep-dive cite full-year FY2025 61.6% GM / 27.4% OM. Both are kept — the 62%/28% figures are 9-month interim; the 61.6%/27.4% are the full-year prints. The earlier checklist also references "62% gross / 28% operating" as the round-number moat signal.

Q1 FY2026 (filed 2026-05-12 post-market). Diluted EPS ¥25.44 (basic ¥25.44) on 18.73M weighted shares. Annualised run-rate ¥101.76 — well below FY2026E consensus of ¥275 — but Q1 is seasonally the weakest quarter (PCB-substrate fabs ramp utilisation through the year; AI server PCB load is back-half). The seasonality argument carries the bridge from ¥102 to ¥275; the H1 print (Aug 2026) is the binary catalyst. Q1 revenue was ¥5.96B (+11% beat per the 2026-05-12 checklist; the deep-dive price table shows the 5/12 close at ¥11,190 was the Q1-print day). yfinance reports TTM revenueGrowth +87.7% — the TTM-vs-prior-TTM step from the AI server PCB inflection (vs +35% FY24 vs -14% FY23). The 2026-05-12 checklist separately cited Q3 FY25 chemical revenue +18.8% YoY / +15.6% QoQ, and CZ-8101 +20.5% YoY over 9M FY25 (+37.5% in Q3 alone).

Cash flow and balance sheet (FY2025). Operating cash flow ¥3.98B; capex ¥2.77B; free cash flow ¥1.21B; FCF margin 5.8%. FCF discrepancy (flagged): Nippon-IBR FY25E FCF was ¥4.27B while yfinance reports ¥1.2B (net of buyback impact) — both kept. FCF yield on FY25E ~2.1%. Cash & equivalents ¥10.35B (yfinance Q1 shows ¥4.63B — reconciliation needed; likely a reclass). Total debt ¥1.25B. Net cash position (-¥9.1B / -0.4x net debt/EBITDA per the buy-checklist; "¥9B+ net cash" per the delta checklist). Interest coverage effectively infinite (debt ¥1.25B vs OP ¥5.5B+). ROE 12.8% (the 2026-05-12 checklist cited 15.3% ROE on 9M FY25 annualised — interim vs full-year discrepancy, both kept). ROA 7.9%. ROIC ~13-14% vs WACC ~5-6% → clear value creation (+700-900bps).

Red flags: none. No aggressive accounting, no equity dilution, no leverage, no convertibles / warrant overhang / ATM / shelf. Share count flat-to-declining. One of the cleanest balance sheets in Japanese mid-cap chemicals. Net cash + FCF self-funds the Kitakyushu programme.

R&D ~5-6% of revenue (per Yuho): next-gen CZ for sub-0.1µm roughness, AMALPHA bonding extensions, photoresist-adjacent ancillary chemistry. M&A: none disclosed; organic growth historically.

Medium-term plan (2030 Vision Phase 2, Feb 13, 2026): FY2027 revenue target ¥25B (held flat in the revision; +13% revenue CAGR FY25-FY27 from ¥18.2B → ¥25B per the buy-checklist, which frames the base as FY24); FY2027 operating margin guidance raised to 26-30% (mix shift to semiconductor-package-substrate chemistry, partially offset by Kitakyushu start-up costs); FY2027 ROE target ≥10%. FY25 9M actuals were running ahead of plan: ~¥20.3B annualised revenue, 28.3% OPM, 15.3% ROE vs the ≥20% OPM / ≥10% ROE targets.

Industry landscape

MEC sits in the wet-process chemistry layer of the PCB / IC-substrate supply chain. TAM is not cleanly broken out by IDC/Prismark; the closest proxy is the wet-process chemical line inside the ~$80B PCB market and ~$15B substrate market, giving an order-of-magnitude $2-3B annual addressable spend across copper adhesion, microetching and cleaning. MEC's ~$140M revenue on ¥21B implies high-single-digit share of total etch/adhesion chemistry, dominated by the high-value CZ niche. Secular tailwinds: AI server PCB layer-count expansion (24L → 32L+); ABF substrate area expansion through Hopper → Blackwell → Rubin (STF Research: substrate area 3,025mm² → 5,625mm² → 8,100mm² — a direct CZ consumption multiplier); HBM4 base-die RDL (AMALPHA optionality); low-loss copper specs (PCIe 6/7, 224G SerDes) tightening roughness tolerance in CZ's favour. Wafflebun's ai-server-pcb-primer.md flags MEC as an "indirect play" with ~$2B market cap and ~15% supply-chain share — consistent with this page.

(Note: this consolidation is filed under the passives-mlcc sector slug per the reorg task; MEC's actual industry is PCB/substrate surface-treatment chemistry, not MLCC passives — the sector tag is an organizational bucket, not a business descriptor.)

See sector pages: advanced-packaging · passives-mlcc

Management

Governance verdict: PASS / A-grade. Long-tenured operator-CEO with meaningful skin in the game, clean board structure, committed and executed capital-return policy, improving IR responsiveness, no red flags. Governance is not the question on this stock — valuation is.

Leadership.

• Kazuo Maeda — Representative Director, President & CEO. Age 63. President since 2002 (~24yr); CEO since 2015 (~11yr). Career MEC operator; built the Asia international footprint (Zhuhai, Taiwan, HK, Suzhou) before the CEO seat. Owns ~727K shares = 3.98% of outstanding, ~$49M USD at current price — materially above the non-founder professional Japanese-CEO norm and a meaningful alignment signal.
• Sadamitsu Sumitomo — Director, EOO Global Operations. ~5.2yr board tenure. ~20K shares / 0.11% (~$1.4M).
• Tetsuya Taniguchi — Director, EOO. ~1.3yr board tenure (joined early 2025). Recent internal promotion; the visible succession candidate being seasoned.
• Katsuaki Kitauji — Operating Officer, GM Accounting & Finance. Below-board financial lead since May 2025. No board-level CFO — Japan small-cap pattern.
• Hiroyuki Maruoka — Executive Operating Officer, Head of Administrative HQ. Career insider.

Ownership. Insider ownership 18.45% (per yfinance major holders); aggregate insider stake ~3.54M shares (~17-18%). Outside-director stakes: Mitsutoshi Takao (Outside Director, Audit Committee Chair) ~3K / 0.016% / ~$202K; Kaoru Hashimoto (Outside, AC) ~600 / 0.0033% / ~$40K; Eiji Miyashita (Outside, AC) ~500 / 0.0027% / ~$34K. No insider buys or sells in the last 6 months. Institutions 51.96% (63.7% of float); 68 institutions. Top institutional holders (long-tail, no dominant active anchor — top 9 ≈ 4.5% of shares): Vanguard Total Intl Stock Index (211,970 / 1.08%), Vanguard Developed Markets (139,594 / 0.71%), Fidelity Japan Fund (128,600 / 0.66%), Hood River International Opportunity (119,511 / 0.61%), Vanguard International Explorer (67,018 / 0.34%), Dimensional Intl Small Cap (60,000 / 0.31%), DFA Japanese Small Company (57,800 / 0.30%), Vanguard FTSE All-World ex-US (54,400 / 0.28%), Schwab Intl Small-Cap (30,188 / 0.15%). Hood River and Fidelity Japan are the only thesis-driven active holders of size. No 13D-equivalent activist filings; no identified PE/strategic.

Board & governance flags. TSE Prime independent-outside-director requirement (≥1/3) met since the listing-tier reclassification. Audit Committee chaired by independent outside director Mitsutoshi Takao. No dual-class shares, no poison pill, no staggered board. Short interest: yfinance returns no short-ratio for 4971.T (Japanese short interest runs through JSDA/JPX — manual lookup if material).

Capital allocation (A-grade). Policy: 35%+ payout ratio + 4%+ DoE (Dividend on Equity) floor + opportunistic buybacks (per 2030 Vision Phase 2). Execution: dividend raised ¥45 → ¥85; 500K shares cancelled Aug 2025; FY2025 ¥935M dividend + ¥1.29B buyback = ¥2.23B total capital return (well within FCF capacity). Treasury stock 1.31M shares (~7.2% of issued) — buyback fuel and acquisition currency. DoE introduced mid-MTP; Nomination & Comp Committee majority-independent.

Incentive-fundamentals alignment: Aligned. Maeda's $49M stake material across multiple cycles. Comp plausibly tied to FY27 ¥25B sales / ≥20% OPM / ≥10% ROE MTP targets, but the company is running materially ahead of all three at current run-rate — a "low bar but aligned" structure where management gets paid for execution already substantially delivered. SBC hurdle detail is not disclosed in English IR (Yuho follow-up needed) — the one verifiability gap.

Succession + key-person risk. Tetsuya Taniguchi is the visible candidate; no explicit succession plan disclosed in English IR (the single meaningful yellow flag). CZ chemistry tacit know-how lives in the R&D team, not Maeda personally — which lowers true key-person risk.

What changed 2026-05-12 → 2026-05-15 (governance lens): (a) Q1 FY2026 Tanshin filed with simultaneous earnings and dividend forecast revisions — the dividend revision is the higher-signal disclosure, implying management has line-of-sight on a beat (they rarely revise dividend guides without it). (b) STF Research disclosed direct MEC IR confirmation on the AP-series forecast cut ("the broader market development simply is not fast enough yet to necessitate the AP product… the core CZ-series business remains robust and completely unaffected") — modestly positive evidence of improving English IR responsiveness against MEC's historically thin posture. (c) Analyst coverage thinned 3 → 2 as the stock punched 52w highs — a small negative if a covering analyst exited rather than upgraded. Items to track next pass: whether the 3→2 coverage drop reverses by Q3 2026, whether English IR catches up to the Japanese Tanshin pace, whether Taniguchi gets an expanded role/title at the next AGM, and whether Kitakyushu lands on time and on budget.

Catalysts & risks

Catalysts (bull path).

• Analyst PT upgrade cycle (late May – mid June 2026). No analyst revisions in the 3 days since the Q1 print; sell-side typically refreshes 2-3 weeks after a Tanshin. If a single analyst lifts PT to ¥12-13k post-Q1, consensus PT could close 30-40% of the gap in one cycle — historically the lift that triggers algorithmic re-rating across passive momentum funds. This is the path to closing the PT-vs-spot gap from below rather than via price drifting down.
• H1 FY2026 print (Aug 2026) — the binary catalyst. H1 run-rate vs FY consensus ¥275 confirms or breaks the seasonality bridge.
• AI server PCB layer-count expansion + ABF substrate area growth through Blackwell → Rubin — direct CZ consumption multiplier.
• Suzhou +30% capacity online (Dec 2026) and Kitakyushu plant start (Dec 2026) — volume capacity, though Kitakyushu also brings margin drag.
• Capital-return execution — dividend forecast revision filed with Q1 signals confidence.
• HBM4 base-die RDL adoption — early-stage AMALPHA optionality.

Risks (bear path).

1. Multiple compression (highest-likelihood downside). Even with flat fundamental delivery, 41x → 22x = -47% on the stock. Valuation re-rate risk moved from "High" to "Active": the PT-vs-spot gap (-19.5%) has widened in % terms as price rose; two ways to close — analysts upgrade (constructive) or price drifts to PT (destructive).
2. AI capex pause in 2027. Substrate volume disappoints; FY27 OPM compresses 28% → 22-23%. Bear case ¥4,400.
3. CZ substitution at a top-3 customer. Low probability, not zero — Atotech or a Chinese entrant qualifies a substitute at Ibiden or Unimicron; would invalidate the thesis in 2-3 quarters.
4. Customer concentration in the ABF substrate cluster (Ibiden / Unimicron / Shinko) — undisclosed; partially closes if MEC is qualified into new entrants (Suzhou) before share shifts.
5. Kitakyushu start-up cost drag on FY2026-FY2027 EPS — near-certain; band 26-30% already factors depreciation; closes once Kitakyushu hits budgeted utilisation (late FY2027 / FY2028).
6. AP-series fails technical milestones / AMALPHA substitution long tail (low near-term, medium 5-10yr — internalised since MEC owns AMALPHA).
7. FX / JPY translation — non-Japan customers price in USD/TWD/CNY; no active FX hedge programme disclosed; structural.
8. NEW (deep-dive): hyperscaler capex commentary cut by NVDA, AMD, MSFT, GOOG, META, AMZN at the next earnings cycle would mark the top of the AI-PCB cycle and invalidate the volume-multiplier thesis.

What would make the thesis wrong: CZ-8101 share loss in advanced packaging; AP-series technical-milestone failure; AI-substrate capex reset. Downside scenario: -47% to -61% over 12-18 months depending on the multiple-compression × earnings-miss combination.

Behavioral traps (active, flagged honestly). FOMO (primary risk — stock +377% YoY, at 52w high, parabolic; grew on the new 52w high); narrative seduction (the AI-PCB / CoWoS / chiplet-packaging story is genuinely good — the risk is paying any price for it; STF's "Last Cheap AI Stock" piece is exactly the narrative that drove the rally — STF was right at ¥5k, doesn't mean right at ¥11k); recency bias (Q1 beat + Q3 record print pull toward "this is the new normal"); anchoring on a possibly-stale ¥8,900 PT; loss aversion ("I should have bought at ¥5k" is the wrong frame). Confirmation bias and authority bias actively counter-checked (Nippon-IBR is sponsored research, paid by MEC, discounted accordingly).

Portfolio correlation flag. Adds to the Japan-substrate cluster already in the vault — 4062 (Ibiden), 6278 (Union Tool, drill), 8027 (E&R Engineering, TGV), 5706 (Mitsui MS, HVLP4 foil). High intra-portfolio correlation already exists in this theme; concentration risk if the substrate cluster reverses together. Dilution risk: effectively zero.

Valuation / DCF

Stance: valuation NOT reasonable at ¥11,050. Margin of safety: negative. The stock trades at roughly double its 5-year normal multiple range. Pre-rally (Dec-2025) Nippon-IBR had FWD P/E at 22.3x; today it is 36-41x — nearly doubled in 5 months while FY25E earnings were revised only +20-30%.

Multiples vs history vs peers.

Multiple	Current	5yr range (pre-rally)	Peer comp
P/E TTM	40.5-41.1x	14-22x	High end
FWD P/E	36-41x (56.9x on FY+1 per yfinance)	14-22x	High end
EV/EBITDA FY25E	~30-31x (75.7x TTM, Q1-windowed)	10-14x	High end
P/FCF FY25E	~48x	15-25x	High end
EV/Revenue	~9.9x	3-5x	High end
P/B	14.5x	—	High
Dividend yield	0.76-0.95%	1-3%	Low end
FCF yield	~2.1%	4-6%	Low end

Implied expectations at ¥11,190: the market is pricing ~12-15% perpetual revenue growth at stable 28%+ OPM for 8-10 years — possible if AI-PCB plus AP-series both deliver, but with no room for execution slippage.

Simple DCF read (from the deep-dive / buy-checklist):

• FY27E base case (per Phase 2): ¥25B revenue × 25% OPM × 0.78 tax × 18,000K shares × 22x FWD P/E ≈ ¥6,500/share.
• Bull-case FY27 ¥28B × 28% OPM ≈ ¥9,000/share fair value at a normalised 22x.
• At ¥11,190 the stock is 25-70% above the DCF range and ~25% above the analyst mean PT of ¥8,900.

Multi-method scenario set (12-month, per deep-dive):

• Bear ¥4,400 (-60 to -61%) — AI capex pause + multiple compression.
• Base ¥9,000 (-18.5 to -19.5%) — analyst PT zone, consensus realises.
• Bull ¥13,200 (+18 to +19.5%) — FY26 EPS ¥330 + 40x multiple holds (the delta-checklist frames bull on a 40x hold; ¥330 EPS exceeds the ¥275 consensus).

The asymmetry is poor: bull +18-19.5% vs bear -60% does not justify a starter position. "Would I buy 10-15% higher (¥12,500-13,000)? No — not even close."

Entry / sizing plan (if/when conditions met): ¥7,000-8,000 — starter 1.0%; ¥6,000-7,000 — scale to 1.5-2.0%; <¥6,000 (deep correction, no thesis break) — opportunistic add to 2.5%. Maximum position 2.5% (medium conviction, single-product platform risk). Maximum loss accepted: 35% drawdown from cost basis (a thesis-break scenario where CZ qualification is lost triggers full exit). Exit / take-profit: ¥18,000-20,000 if entered <¥9,000 (>100% gain, stretches fair value); trim zone >¥13,000. NEW stop/re-evaluate triggers from the deep-dive: ¥12,230 break-up with volume → revisit (market may be discounting an H1 beat ex-ante); ¥10,000 break-down (5-day support) → first sign of distribution, wait for stabilisation.

Technical read (2026-05-12 → 2026-05-15). 50-day MA ¥7,758 (price +44% above — extended); 200-day MA ¥5,185 (price +116% above — very extended); RSI(14) 73.2 (overbought); Golden Cross intact; parabolic move from Dec-2025 (late-stage momentum, not early-trend). By 5/15: 5-day range ¥10,350-¥12,230; 20-day avg volume 337k; 5/13 volume 865k (4x — institutional re-marking on the Q1 print); 5/15 volume 147k (0.4x — light, retail profit-taking not distribution). Spike-to-¥12,230-then-fade pattern = "expensive consolidation," not a buy signal and not a panic signal. No clean support below ¥10,000 until the ¥7,000-9,000 consensus-PT zone. Technical verdict: do NOT buy now — wait for a pullback to ¥8,000-9,000 (50dMA + analyst-PT zone) or 3-6 months of base-building above ¥9,500 that turns the 50dMA into rising support.

Decision log

• 2026-05-12 — Initial research suite (profile / deep-dive / mgmt-dd / buy-checklist). Verdict: PASS at ¥11,190; WATCH / re-engage ¥7,000-9,000. Pre-buy scorecard 6 Yes / 3 No / 1 Partial — the three Nos (valuation, behavior, technicals) all point to not entering at the 52w high. FundamentEdge gate: 3 pass, 2 amber (quality-as-risk; estimate-revision direction). Business quality, financial health, and incentive alignment all PASS; valuation NO; behavioral traps YES (FOMO + narrative + recency); technicals NO (overbought, parabolic). Action: set price alerts ¥9,000 and ¥7,000; re-run on trigger; re-evaluate full thesis after H1 FY26 print (Aug 2026).
• 2026-05-15 — Delta refresh (all four documents). Verdict unchanged: PASS at ¥11,050; re-engage ¥7,000-9,000. In three days: stock pushed to a new 52w high ¥12,230 (5/14) then faded -9.6% to ¥11,050 — within ~1% of the 5/12 entry; the 5/12 verdict held through the test. Q1 FY2026 print landed 5/12 PM (EPS ¥25.44). Analyst coverage thinned 3 → 2; mean PT unchanged ¥8,900 (-19.5% below spot); no upgrade cycle yet. STF Research's Mar-1-2026 bull case (at ~20x P/E) surfaced via Substack and has played out cleanly — stock roughly doubled to 40.5x TTM; no out-of-consensus edge remains. Scorecard again 6 Yes / 3 No / 1 Partial. Governance refresh: PASS / A-grade unchanged; dividend forecast revision read as a positive confidence signal; STF IR-confirmation channel modestly positive on IR responsiveness.
• Standing exit / re-evaluation triggers: take profits ¥18,000-20,000 (if entered <¥9,000); stop/re-evaluate if CZ-8101 quarterly volume turns negative or Ibiden/Unimicron cut substrate volume materially; hard stop if a substitute microetch is qualified at a top-3 customer; Kitakyushu delay beyond Q1 FY27 → review. Next genuine decision point: H1 FY2026 print, Aug 2026.
• Cross-checks: SemiAnalysis mirror searched 2026-05-12 and 2026-05-15 — no SA coverage of MEC, no contradiction. STF Research cross-check 2026-05-15 — thesis played out, no edge remains, no post-rally STF follow-up. Filings: JP issuer, no SEC filings; Q1 FY2026 Tanshin (2026-05-12), Feb-2026 MTP revision, Aug-2025 buyback completion/cancellation, Dec-2025 Nippon-IBR sponsored update; no restatements, auditor changes, material-weakness or going-concern flags.
• Open follow-ups: customer-concentration % (Yuho / channel checks); SBC hurdle detail (not in English IR); FY2023/FY2024 P&L middle rows (Japanese Yuho); Q1 segment/geography line items (Tanshin PDF); cash reconciliation ¥10.35B vs yfinance Q1 ¥4.63B.

Sources

Fragments folded into this canonical page (2026-05-30 consolidation):

• 4971-buy-checklist.md (pre-buy checklist, 2026-05-12)
• 4971-checklist.md (pre-buy delta refresh, 2026-05-15)
• 4971-mgmt-dd.md (management DD delta refresh, 2026-05-15)
• 4971-deep-dive.md (deep-dive delta refresh, 2026-05-15)
• 4971-profile.md (company profile, 2026-05-15)
• 4971.md (prior canonical / suite index, updated 2026-05-15)

External and primary sources cited across fragments:

• Live yfinance data (price, technicals, holders, estimates, financials) — pulls dated 2026-05-12 and 2026-05-15.
• MEC IR: 2030 Vision Phase 2 medium-term plan (Feb 13, 2026), https://www.mec-co.com/en/ir/mtp/pdf/medium-term_business_plan2025-2.pdf ; March 2025 Shareholder Briefing; corporate-governance and directors pages.
• Q1 FY2026 Kessan Tanshin (Consolidated Financial Results, three months ended March 31, 2026; filed 2026-05-12) — earnings + dividend forecast revisions; EPS ¥25.44.
• Nippon Investment Bespoke Research UK — MEC Co., Ltd Research Update, 8 Dec 2025, https://www.mec-co.com/en/ir/library/pdf/report/5703.pdf (sponsored research, paid by MEC — discounted accordingly).
• STF Research, "Finding the Last Cheap AI Stock in the Japanese Rally" (2026-03-01, paid Substack, stfbutnou.substack.com), https://stfbutnou.substack.com/p/finding-the-last-cheap-ai-stock-in — pulled via opencli browser session 2026-05-15; includes direct MEC IR confirmation on AP-series timing.
• Internal vault cross-references: ai-server-pcb-primer.md, packaging-glass-substrate-primer.md, 4971-vs-4062-vs-5706-showdown.md.
• Briefing: 2026-05-15 MEC (4971), https://pink.sakdiarpa.com/reports/4971.html and vault briefings/2026-05-15-4971.
• SemiAnalysis mirror (~/Dropbox/Wafflebun/KB/wiki/semianalysis/) searched 2026-05-12 and 2026-05-15 — no MEC coverage.

---

Consolidation queue (merged 2026-05-30)

The following research fragments were folded into this canonical page and stay live pending Pink's archive confirm:

• [ ] 4971-buy-checklist.md
• [ ] 4971-checklist.md
• [ ] 4971-mgmt-dd.md
• [ ] 4971-deep-dive.md
• [ ] 4971-profile.md
• [ ] 4971.md

Source updates (auto-maintained)

Intake (May 12, 26) - cu-wiring-resin-primer

The primer positions MEC's CZ-series as the spec-locked copper-to-resin adhesion chemistry at step 4 of every AI substrate build-up cycle, consumed 20-plus times per substrate, with CZ-8101 now qualified into CoWoS chiplet packaging — and explicitly frames the AP-series no-roughening platform as MEC's own internalised successor.

Relevant to your thesis: Directly reinforces the spec-lock moat and layer-count volume driver; the CZ→AP succession being MEC-owned is the key internalised-substitution-risk point the wiki flags but this article elaborates with process-step granularity.

Source: intakefile://cu-wiring-resin-primer.md