4062 — Ibiden Co., Ltd.

Thesis

Verdict: WATCH, not BUY (as of 2026-05-15 refresh). Ibiden is the global #1 maker of high-end ABF / FCBGA IC substrate at the precise moment AI accelerator package areas crossed organic substrate's prior physical comfort zone — exactly the regime where Ibiden's process moat is structurally widest. The business quality is real and the trajectory is consensus-backed; what is contested is the multiple. The stock trades at 74.6x TTM / 52.0x forward P/E, and even after a 3% pullback the base-case DCF target (JPY 14,350) sits ~11% below spot. This is a quality franchise priced for perfection, not an asymmetric entry.

What has to be true for the bull case. (1) The substrate franchise keeps compounding through the Vera Rubin / Rubin Ultra cycle, with Ibiden defending flagship share against Unimicron; (2) the glass-core transition (Intel GlassCore, 2027-2030 commercial) becomes a bridge Ibiden builds rather than one built around it; (3) sell-side targets re-rate up to the real JPY 14-18k consensus range, removing the structural overhang of stale screener marks.

Why WATCH and not BUY at JPY 16,065. The FundamentEdge hard rules split: revenue-growth primacy PASSES (FY3/26 +34%, FY3/27E +18%), second-derivative PASSES (accelerating), estimate-revision direction PASSES (upward, with the caveat sell-side has been slow), but valuation-as-thesis FAILS (rich multiple, no margin of safety) and quality-as-risk is MIXED (oligopolistic, but the Unimicron share-loss dynamic is genuine). The asymmetry favors waiting. STF Research's coverage reinforces this independently: they treat Ibiden as the consensus center of the AI substrate value chain but route their actual stock picks (Uyemura, MEC, Taiyo, Kitagawa Seiki, Maruwa) around Ibiden — the implicit message that Ibiden's quality is real but the price has caught up and the alpha now sits in the under-covered chemistry / equipment names that feed it.

Conviction: Medium — high quality of business plus high quality of trajectory, but valuation has not reset enough to make this an entry-priced compounder. Flip to BUY on a pullback to JPY 13-14k (mid-30s forward P/E), a sell-side re-rate above JPY 17k, a Rubin / Rubin Ultra sourcing confirmation, or an FY3/27 Q1 beat with margin expansion.

Snapshot

Ibiden Co., Ltd. — ticker 4062.T, Tokyo Stock Exchange Prime. Global #1 ABF / FCBGA IC substrate maker; Intel server CPU + Nvidia data-centre GPU (Hopper / Blackwell / Vera Rubin) + AMD + hyperscaler substrate franchise. Oono (Fukui) plant phase 2 commissioning; glass-core transition leader. Founded 1912 as Ibigawa Electric, renamed Ibiden 1982. HQ Ogaki, Gifu Prefecture; ~14,000 group employees.

Valuation snapshot (2026-05-15 close)

Metric	Value	Δ vs 2026-05-12
Spot	JPY 16,065	–2.9% (from JPY 16,550)
Market cap	JPY 4.49 tn (USD ~29.9 bn at JPY 150/USD)	–2.8% (from JPY 4.62 tn)
Enterprise value	JPY 4.64 tn	+8.4% vs JPY 4.28 tn (net cash drawdown on capex)
P/E (TTM)	74.6x	down from ~77x
P/E (Fwd, FY3/27E)	52.0x	down from ~54x
EV/EBITDA (TTM)	~26-28x	–1 turn
EV/Revenue (TTM)	~9-10x	unchanged
52-week range	JPY 2,477 – JPY 18,365	new ATH JPY 18,365 printed May 12-14
Dividend yield	0.24%	unchanged (token, capex-prioritised)
Beta (5Y, JPY)	1.56	unchanged
Shares out	279.2M	unchanged
Yfinance consensus PT	JPY 10,441	STALE — do not anchor

Stock printed a new 52-week high of JPY 18,365 in the May 12-14 window before fading to JPY 16,065 — a ~12.5% drawdown in two sessions on no specific news, read as mechanical / positioning rather than fundamental.

Business

Three segments. Electronics ~70-75% (the IC-substrate franchise), Ceramics ~20-25% (diesel particulate filters / DPF), Others ~3-5% (includes a specialty graphite sub-line). Latest IR materials are the FY3/26 Q4 / full-year deck released late April 2026 (https://www.ibiden.com/ir/library/).

The moat — substrate process leadership

Ibiden's edge is in high-end ABF / FCBGA substrate process: warpage management at large panels, the multi-build-up dielectric ladder, embedded passives, and panel-level yield compounding. The thesis turns on package-area physics: AI accelerator package area crossed 5,625mm² (B200) → 8,100mm² (Vera Rubin), and B300 NVL16 pushes beyond the JEDEC 120×120mm package size envelope — the regime where Ibiden's process moat (low warpage at large panels, build-up cycle yield) is structurally widest. Glass-core substrate is the next bridge, and Ibiden is positioned as Intel's lead glass-core development partner.

Customers

Top customers: Nvidia (~15-25%; Hopper sole-supplier / Blackwell majority / Vera Rubin qualified), Intel (~20-25%; server CPU multi-decade relationship), AMD (~10-15%), hyperscalers AWS / GOOGL / MSFT / Meta combined ~10-15%, with Apple / Marvell / Broadcom / MediaTek as a ~5-10% long tail. Diesel DPF automotive ~20% (Cummins, Daimler Truck, Stellantis). Concentration flag: Nvidia + Intel combined approach ~50% of electronics revenue and ~35-40% of group revenue — real single-customer risk, partly offset by the Unimicron dual-sourcing dynamic (which protects against Nvidia squeezing prices via take-or-pay leverage, since Ibiden can substitute Nvidia volume into Rubin-tier supply at richer pricing).

Competitive position

Top-5 organic ABF substrate makers globally: Ibiden, Unimicron, Shinko Electric (renamed Resonatec post-Toshiba spin), Nan Ya PCB, Kinsus — with AT&S the European challenger, mainly Intel-aligned. Ibiden remains #1 by AI / data-center mix and process moat. Unimicron is the credible #2 and has been the share-taker at Blackwell GPU substrate (per SemiAnalysis 2024). Industry structure is a consolidated tier-1 oligopoly with 4-5 players capable of building flagship-tier substrate.

Value chain (upstream nodes)

Key upstream suppliers feeding the substrate: Ajinomoto (ABF resin), Mitsui Kinzoku (MicroThin / HVLP Cu foil), MEC (CZ adhesion chemistry), plus Atotech and Uyemura on plating / surface treatment. Per STF Research, Uyemura's ENEPIG final-pad surface-treatment share is ~50-55% in Japan (Ibiden, Shinko Electric) vs ~70% in Taiwan — implying the Japan substrate ecosystem is more chemistry-source-diversified than Taiwan, marginally reducing single-source upstream risk.

Financials

Income statement & margins (consensus from 17-broker average per QUICK / Bloomberg compile, late April; spread is wide):

Metric	FY3/24 (FY-2)	FY3/25 (FY-1)	FY3/26 (FY0, latest)	FY3/27E (FY+1E)
Revenue (JPY B)	369	369	~495 (+34%)	~585 (+18%)
Operating margin %	~14	~14	~17-18	~19-21
Net income (JPY B)	~35	33.7	~58	~80
EPS (JPY)	~125	121	~209	~287
FCF (JPY B)	negative	–79.6	turns positive 2H FY3/26	clearly positive

EPS roughly doubles FY3/25 → FY3/27E (JPY 121 → ~JPY 287). Latest reported quarter is FY3/26 Q4, disclosed late April 2026. No new earnings between May 12 and May 15.

Balance sheet & cash flow

Net cash; capex self-funded (cash on hand + operating cash flow + bank facility). FCF was negative in FY3/24 and FY3/25 (–79.6B) on the Oono build, turns positive in 2H FY3/26, and is clearly positive in FY3/27E. FCF yield is currently negative — Oono phase 2 capex still flowing; P/FCF is n/a on negative LTM FCF. The Oono plant (Fukui) carries a JPY 250-300B capex commitment, with phase 1 operating since early 2024 and phase 2 commissioning across 2026-2027.

Note on EV. The enterprise value move from JPY 4.28tn → JPY 4.64tn at a lower market cap is consistent with a net cash drawdown of ~JPY 350-400B over the 3-day window — possibly a capex tranche payment to Oono phase 2 contractors or a bank-facility drawdown for working capital. Flagged for the next earnings reconciliation, not actionable today.

Dilution

Dilution risk: low. No ATM, no shelf, no convertible overhang. The capex is self-funded. Incremental-margin analysis over the prior 8 quarters shows operating leverage building at the Electronics-segment level.

Industry landscape

High-end ABF / FCBGA IC substrate is a consolidated tier-1 oligopoly — Ibiden, Unimicron, Resonatec (ex-Shinko Electric), Nan Ya PCB, Kinsus, with AT&S as the European / Intel-aligned challenger. Demand is inflecting on AI accelerator package-area growth (5,625mm² → 8,100mm² → JEDEC-exceeding); supply is tight; Unimicron is taking structural share at Blackwell even as absolute volumes grow for everyone. The next architectural shift is glass-core substrate (2026-2030 depending on source). Upstream bottleneck nodes — Ajinomoto (ABF resin), Mitsui Kinzoku (HVLP Cu foil), MEC, Uyemura — are themselves under-covered beneficiaries.

See sector page: advanced-packaging (ABF/FC-BGA IC-substrate; also a major MLCC peer — see passives-mlcc)

Management

Leadership. President / CEO Koji Kawashima (since June 2017; prior CTO / COO; substrate-engineering lifer with a Toyohashi University of Technology PhD background). CFO is listed as Takeshi Aoyama in the profile fragment and as Yoshitaka Hidaka in the mgmt-dd fragment — DISCREPANCY between source fragments; both names retained pending verification against the FY3/26 securities report. Chairman Takeshi Ozawa (transitioned from the CEO role). Other named executives: Kazuhiro Sakai (EVP) and Hiroshi Nakamura (Electronics segment lead). Not founder-led — professional, internally-promoted management with multi-decade Ibiden tenure (typical Japanese corporate model).

Ownership & alignment

Insider ownership low — combined directors + ESOP <2% of outstanding, typical for a Japanese listed. Top holders (Japan structure): Master Trust Bank of Japan ~12-15%, Custody Bank of Japan ~5-8%, Toyota / Toyota-affiliated entities (small), foreign institutionals (BlackRock, Vanguard, JPM, Fidelity via passive + active sleeves) combined ~20-25%, self-held / ESOP ~2-3%. No 13D-style activist position disclosed; no recent material insider transactions in EDINET.

Compensation

Compensation is heavily cash (base + cash bonus) per the FY3/25 securities report; stock-based comp is modest (~JPY 100-300M per NEO range); no golden-parachute structures. METI / TSE governance reform is pushing TSE Prime listeds toward equity-linked compensation, but this is not yet substantive at Ibiden. Verdict on alignment: aligned-by-default given career concentration, weakly aligned-by-design given the cash-comp emphasis.

Capital allocation & governance

Capital allocation graded A−: the Oono capex is a clear, bet-the-house capacity expansion at the right point in the cycle; no M&A history of concern; modest annual buybacks at TSE-typical rates; no major equity issuance in the last decade. Guidance tendency is conservative-to-straight-shooter, with multiple FY3/25 → FY3/26 in-line beats and no weasel-language pattern. Board: ~11 directors, ~4 outside (independent) — below US best-practice on independence ratio but compliant with the revised Japan Corporate Governance Code; board stable, no activist involvement. Overall mgmt grade: A− / Green-with-yellow-on-disclosure-granularity — the yellow is structural (Japanese listed disclosure regime, no Form-4-equivalent same-day insider filings, no 10b5-1-equivalent plans), not Ibiden-specific. A team to trust with capital; the open question is the price.

Next governance reads

FY3/26 Annual Securities Report (有価証券報告書) expected June 2026 — the next material governance read (updated NEO comp tables, related-party summary, beneficial-owner list, board attendance, audit-firm tenure). FY3/26 AGM late June 2026 — watch for any activist / minority push on capital allocation, equity-linked comp, or board independence.

Catalysts & risks

Catalysts (bull)

Near-term (0-90 days): FY3/27 Q1 earnings (late July / early Aug 2026) — first print under Oono phase 2 partial commissioning; watch Electronics-segment operating-margin trajectory, any Vera Rubin substrate ASP / volume guidance, and Unimicron co-supply commentary. This is the next material datapoint. Nvidia GTC Fall 2026 (late October) — Rubin Ultra detail; SA's 2025 GTC piece flagged "beyond JEDEC 120mm" package sizes, and confirmation here is bullish on Ibiden's per-package dollar content. Intel earnings late July — Granite Rapids volume + GlassCore qualification timeline.

Medium-term (3-18 months): Vera Rubin / Rubin Ultra substrate sourcing confirmation (Q3/4 2026 watch); Oono phase 2 commissioning milestones (multiple announcements through 2H 2026 / 1H 2027); hyperscaler custom-silicon ramp (Trainium 2, TPU v6/v7, MAIA 200, Axion 2); METI advanced-packaging subsidy updates; glass-core Intel qualification gating. No new contract awards or capacity commitments disclosed in the 72-hour window.

Risks (bear)

1. Valuation / consensus-ahead risk — THE NUMBER ONE RISK. Stock is 12.5% off ATH on no news; the marginal buyer is becoming sensitive to multiple. Until sell-side actually re-rates targets to the real JPY 14-18k range, the stock is structurally exposed to multiple compression on any minor disappointment. Likelihood Medium-High in 3-6 months. Not closeable except through sell-side update cycles and continued execution.
2. Dual-source structural acceleration to Unimicron. The single most important fundamental risk. Mitsui Kinzoku (5706) primer flagged Unimicron and AT&S pulling MicroThin Cu foil volume that previously went exclusively to Ibiden, weakening Ibiden's de facto exclusivity at flagship Cu-foil specs. Likelihood Medium-High over 12-24 months. Partially closeable — Ibiden can defend at glass-core / next-gen but cannot reverse organic ABF share loss.
3. Capex overrun / Oono phase 2 commissioning execution slip.
4. JPY / USD FX — Ibiden books in JPY, sells in USD-tied pricing; a strong yen compresses reported margins.
5. Glass-core substrate transition execution — leadership on organic ABF does not automatically translate. Discrepancy / open risk: the base deep-dive assumed Intel GlassCore commercial 2027-2030 with Ibiden as the bridge; STF Research's January-2026 datapoint (Absolics — the SKC / Applied Materials JV — shipping volume glass-core samples to AMD for MI400, commercial plausibly 2026-2027) suggests glass-core may commercialise 12-24 months earlier and via AMD-through-Absolics rather than Intel-through-Ibiden. If it proves out, Ibiden's "bridge to glass core" optionality weakens materially. Watch AMD MI400 substrate sourcing announcements in 2H 2026.

Dilution risk: low (no ATM/shelf/convertible). Key-person risk: unchanged — Kawashima tenure secure, no succession event flagged.

Behavioral traps: watch for FOMO (most run-up Japanese semi name, easy to chase) and narrative seduction (the AI substrate story is genuinely good but the price has caught up).

Valuation / DCF

Three-scenario forward P/E framework on FY3/27E EPS of JPY 287:

Scenario	Forward P/E	Target	vs spot JPY 16,065
Bear	35x	JPY 10,050	~–37%
Base	50x	JPY 14,350	~–11%
Bull	60x	JPY 17,200	~+7%

At spot JPY 16,065 the stock sits between base and bull but closer to bull — risk-reward is asymmetric to the downside on multiple-compression risk. Base-case expected return over a 12-18 month horizon is ~–11%; the 3-day pullback closed only ~2pts of the gap (from –13% at the 2026-05-12 write). The DCF does not work at the current multiple — this is not a margin-of-safety play.

Deeper bear scenario. If the multiple compresses to the peer mid-cycle high-20s / low-30s forward P/E, the stock prints JPY 8,000-10,000 (–40% to –50%). Trigger: Unimicron displaces Ibiden as primary Rubin Ultra source, OR Oono phase 2 slips ≥6 months, OR JPY strengthens to JPY 140/USD.

On the sell-side consensus. The yfinance mean PT of JPY 10,441 is STALE / un-rerated sell-side data, not a real consensus mark — it reflects analysts who have not caught up with the AI substrate re-rating after the ~6.7x TTM run. Bloomberg / QUICK house-view consensus circulates JPY 14-18k. Do not anchor on the JPY 10.4k figure for any sizing or signal work; verify against the real consensus number on the next earnings cycle. A re-rate of the sell-side mean above JPY 12k would signal the catch-up is underway; above JPY 17k would remove a structural overhang and is a BUY trigger.

Position sizing. Conviction Medium; suggested action watchlist, not buy. If forced to enter today, 0.5-1.0% portfolio weight, scaling strictly only on dips to the JPY 13-14k range (mid-30s forward P/E).

Decision log

• 2026-05-12 — Base research cycle: full profile, deep-dive (~58k words across 17 sections), mgmt-dd, and pre-buy checklist. Checklist verdict: WATCH (quality business, consensus-ahead valuation, wait for better entry). Spot JPY 16,550; base-case target JPY 14,350 implied ~–13%. Mgmt grade A−.
• 2026-05-15 — 3-day delta refresh across all four research files (part of the Japanese-semi GAA-ramp swarm). Spot JPY 16,065 (–2.9%); 52-week high reset to JPY 18,365 (printed May 12-14), stock ~12.5% off ATH on no news. TTM P/E recompressed 77x → 74.6x; fwd P/E ~54x → 52.0x. Base-case target gap improved to ~–11%. Verdict unchanged: WATCH (still not BUY). No new IR disclosure, no insider EDINET filing, no SA mirror post, no governance change in the 72-hour window. Mgmt grade A− carries.
• 2026-05-15 — flags raised for next cycle: (a) yfinance mean PT JPY 10,441 is stale, do not anchor; (b) STF Research Absolics→AMD MI400 glass-core datapoint suggests faster glass-core timeline (2026-2027) via a non-Ibiden path — track AMD MI400 substrate sourcing; (c) EV jumped JPY 4.28→4.64tn on a ~JPY 350-400B net-cash drawdown, reconcile at next earnings; (d) re-fetch STF "Substrate Wars" and "Kitagawa Seiki" posts (fetch failed this session).

Triggers to flip to BUY

1. Pullback to JPY 13-14k (mid-30s fwd P/E) — resets multiple to defendable level.
2. Sell-side consensus target re-rates above JPY 17k — catch-up done, overhang removed.
3. Rubin / Rubin Ultra sourcing confirmation (likely late Q3 2026) — gates dual-source risk.
4. FY3/27 Q1 earnings beat with operating-margin expansion (late July / early Aug 2026).

Triggers to flip to PASS (permanently uninteresting)

1. Unimicron confirmed as primary Rubin Ultra source.
2. Oono phase 2 commissioning slips ≥6 months.
3. Multiple stays above 50x fwd P/E through FY3/27 Q2 — the consensus-ahead problem becomes chronic.

Next material datapoint: FY3/27 Q1 earnings, late July / early August 2026 (Japanese fiscal calendar). Calendar reminder worth setting; also FY3/26 Annual Securities Report (June 2026) for the next governance forensic read.

Sources

Consolidated from the following vault research fragments (all 2026-05-15 three-day delta refreshes pointing back to a fuller 2026-05-12 base cycle):

• 4062-profile (2026-05-15 profile)
• 4062-buy-checklist (2026-05-15 pre-buy checklist, verdict WATCH)
• 4062-mgmt-dd (2026-05-15 management DD, grade A−)
• 4062-deep-dive (2026-05-15 deep dive, Section A delta refresh)

External / cross-referenced sources cited within the fragments:

• Yfinance (2026-05-15 quote, valuation multiples, beta, mean PT)
• Ibiden IR page (https://www.ibiden.com/ir/library/); TDnet / EDINET — no fresh disclosure in window
• Consensus from 17-broker average per QUICK / Bloomberg compile (Nomura, Daiwa, SMBC Nikko, Mizuho, Mitsubishi UFJ Morgan Stanley, Goldman, Morgan Stanley, JPM, Citi, Jefferies, CLSA, Macquarie, BofA, Bernstein, Tokai Tokyo)
• SemiAnalysis mirror at ~/Dropbox/Wafflebun/KB/wiki/semianalysis/ — no May 2026 posts on 4062 (prior framing: sole-supplier 2023 → Unimicron share-loss 2024 → JEDEC-exceeding package size 2025)
• STF Research (@stfbutnou, paid sub): "Uyemura: Another Invisible Beneficiary Behind Every Substrate" (May 11, 2026, fully read); "The Substrate Wars: A Reality Check + Investor's Playbook" (fetch failed, re-try); "AI Supply Chain: Shortage Intensifying Toward 2027" (Apr 7, 2026, Founding-Member tier, locked); "Kitagawa Seiki: Hidden Champion" (fetch failed); "Finding the Last Cheap AI Stock in the Japanese Rally" (Mar 2026, referenced — STF did NOT pick Ibiden)
• Related primers: packaging-glass-substrate-primer (2026-05-03), cu-wiring-resin-sub-layer-primer (2026-05-12), ai-server-pcb-primer (2026-05-03)
• Prior canonical page 4062 (created 2026-05-12, updated 2026-05-15)

---

Consolidation queue (merged 2026-05-30)

The following research fragments were folded into this canonical page on 2026-05-30. They stay live pending Pink's archive confirmation.

• [ ] 4062-profile.md
• [ ] 4062-buy-checklist.md
• [ ] 4062-mgmt-dd.md
• [ ] 4062-deep-dive.md
• [ ] 4062.md

Source updates (auto-maintained)

Drop/2. Frontend (Jun 1, 26) - GS Memory

Goldman Sachs (31 May 2026) raises HBM/DRAM/NAND price forecasts and argues the memory upcycle extends into 2028, driven by AI server demand, supply discipline, and LTAs — with HBM TAM reaching USD 168bn by 2028E; Ibiden is mentioned only tangentially as part of the AI supply chain context, not covered directly.

Relevant to your thesis: Structurally bullish for Ibiden's substrate demand runway (Blackwell/Rubin cycle sustained longer), reinforcing the revenue-growth PASS but not addressing the valuation-multiple FAIL that keeps the verdict at WATCH.

Source: dropfile://2. Frontend/Memory/GS Memory.pdf

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

Ibiden (4062) is the substrate fabricator that consumes MicroThin Cu foil and ABF resin — it owns the process integration and customer qualification slot, while the chemistry moat sits upstream with Mitsui Kinzoku, MEC, and Ajinomoto.

Relevant to your thesis: Reinforces the existing wiki's upstream-supplier framing (Uyemura, MEC, Taiyo as the alpha-rich plays) and the STF Research read that Ibiden's quality is real but the value capture is split with chemistry suppliers.

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