6525 — Kokusai Electric Corporation
Thesis
Stance: BUY — half-size starter (50% of an intended 1-1.5% book position) at JPY 6,300-6,400, scale-in over 3 tranches. Medium conviction. Kokusai Electric (6525.T, TSE Prime) is the global incumbent in batch ALD/LPCVD semiconductor deposition — vertical furnaces that deposit ultra-thin, ultra-uniform chemistry layers across 50-150 wafers at once. It holds process-of-record (POR) qualifications at every leading-edge logic and memory foundry. The core case: the GAA logic ramp (TSMC N2, Samsung 2nm, Intel 18A) + the HBM4/DRAM cycle + 3D NAND vertical scaling drive ~10-12% revenue CAGR through CY28 with operating leverage toward ~25%+ margins, on a POR moat plus a growing recurring-service tail.
What has to be true (the single load-bearing condition): Kokusai retains its batch ALD POR positions through the N2/2nm transition without material share-take by ASM International's single-wafer ALD. Secondary musts: China revenue (~45-50% of total) does not get hit by a METI export-control escalation, and the next quarterly print (Aug 13, 2026) confirms the decelerating-but-positive beat pattern rather than breaking it.
Bull, in one line. Hard-coded $/wafer growth in GAA + HBM + 3D NAND; entrenched POR with 12-24-month per-node switching costs; high-margin, growing installed-base service tail; clean post-KKR balance sheet — supports earnings power well above FY25's JPY 36B net income through CY27-28.
Bear, in one line. Already priced as a leading-edge winner at 33x FY+1E P/E, 25x EV/EBITDA, 7x book; China at ~45-50% of revenue is one METI rule change from a 20-30% revenue hole; ASMI's single-wafer ALD is winning the highest-value GAA gate-stack steps and could dilute Kokusai's leading-edge share even as absolute revenue grows; FY25 FCF of JPY 10.8B against a ~JPY 1.48T market cap is a <1% FCF yield, so multiple compression bites if anything in the GAA story slips.
Asymmetry is moderate, not screaming. Base case JPY 7,500; upside JPY 9,000-10,000; downside JPY 4,500-5,000. Valuation prices most of the bull case, so entry timing matters more than usual on a 2.26-beta stock with binary China exposure. The pre-buy checklist passed all five FundamentEdge gates (one with a mild 2nd-derivative caveat) and cleared three soft behavioral flags (FOMO, narrative seduction, recency) with discipline guardrails — net BUY with scale-in and defined exits.
Snapshot
One-liner: Japan's batch-furnace ALD/CVD pure-play — a high-conviction GAA deposition winner, but priced as one already at ~33x forward and 25x EV/EBITDA.
Identity. Kokusai Electric Corporation (株式会社Kokusai Electric); ticker 6525 (yfinance/IR ticker 6525.T), TSE Prime. GICS Semiconductor Equipment & Materials; reported under Information & Communication / Semiconductor Manufacturing Equipment. HQ Chiyoda, Tokyo. Founded 1949 as Hitachi Kokusai Electric; SMEQ division spun out of Hitachi Kokusai 2017, renamed Kokusai Electric 2018. Employees 2,540 (FY25). Fiscal year ends March 31. Website https://www.kokusai-electric.com.
Ownership lineage. Hitachi Kokusai Electric Semi Manufacturing Equipment Division → KKR LBO October 2017 (~JPY 257B / ~$2.2B) → IPO October 25, 2023 on TSE Prime at JPY 1,840 → JPY 6,336 today (+3.4x in ~30 months).
Price / valuation snapshot (as of May 15, 2026).
| Metric | Value |
|---|---|
| Share price | JPY 6,336 (checklist/technical sections cite JPY 6,378 as the working entry) |
| Market cap | JPY 1,480B (~$9.9B @ 150) |
| Enterprise value | JPY 1,449B |
| P/E (TTM) | 49.3x |
| P/E (FY+1E) | 33.1x |
| EV/EBITDA (TTM) | 25.0x |
| P/B | 7.0x |
| EV/Revenue (TTM) | 6.1x |
| FCF yield (TTM, on EV) | ~1.7% (depressed by capex + WC) |
| Dividend yield | 0.78% |
| 52-week range | JPY 2,607 – 7,651 (checklist technical section cites 2,619 low / 7,358 high) |
| Beta | 2.26 (very high — capex-cyclical) |
| Shares outstanding | ~234M; float 195M (~83%) |
Note on minor cross-fragment number drift: deep-dive/profile use a JPY 2,607-7,651 52-week range and JPY 6,336 price; the checklist technical section uses JPY 2,619 low, JPY 7,358 high, and JPY 6,378 price — both kept as captured on the same May 15, 2026 date.
Business
What it makes. Kokusai makes the unsexiest, most indispensable tools in the leading-edge fab: vertical batch furnaces that deposit films one atomic layer at a time across 50-150 wafers at once. When a foundry needs a high-quality, ultra-uniform thin film at low cost per wafer — the high-k dielectric on a GAA nanosheet, a silicon nitride hard mask, the polysilicon for 3D NAND — it usually runs that step through a Kokusai vertical furnace. It is one of the world's three major batch-deposition vendors (with TEL and ASM International) and the global #1 in batch ALD by installed base.
Why batch. A batch vertical furnace is a tall cylindrical chamber holding a stack of 50-150 wafers on a quartz boat; gases react on every wafer simultaneously. ALD deposits one atomic layer per cycle via self-limiting surface chemistry (pulse precursor A, purge, pulse precursor B, purge, repeat hundreds-to-thousands of cycles per nm), so film thickness is digitally controllable and conforms perfectly to 3D topology. Single-wafer ALD does ~0.6 wafers/hour; Kokusai's batch does ~15 wafers/hour — roughly a 25x throughput win. Batch dominates bulk dielectrics (SiN, SiO2, polysilicon, high-k oxides) and thick films (NAND mold layers, gate-stack spacer, contact liner); single-wafer wins for metals (work-function, gate fill) and R&D-flexible steps. GAA introduces 5+ new conformal layers — Kokusai wins the bulk-dielectric/channel-release steps, ASMI wins the high-value work-function metal step, and both grow.
Product taxonomy.
| Product line | Function | Deposits |
|---|---|---|
| TSURUGI series | Mid-temperature batch ALD/CVD | SiN, SiO2, polysilicon, amorphous Si, low-k spacer |
| MARORA series | High-temperature batch ALD | High-k (HfO2, ZrO2, Al2O3), gate stack |
| JINJU series | Cleaning / surface treatment | chamber clean, channel passivation |
| JINEX series | Vertical thermal | anneal, oxidation, diffusion |
ASP per tool JPY 300-700M (~$2-5M). A 100K-WPM leading-edge fab needs ~30-50 deposition tools across all vendors; Kokusai might supply 8-12 depending on POR mix.
Revenue model. ~85% equipment shipments (one-time) + ~15% recurring service/parts. Every tool produces JPY 30-60M/year in parts and service over a ~7-10 year life; with ~2,000 installed tools and growing, service revenue is ~JPY 40-50B/year and rising. Working-capital heavy: inventory JPY 83B at FY25 (35% of revenue).
Customers (no by-name disclosure; channel-check estimates).
| Customer | Ticker | Est. revenue share | Relationship |
|---|---|---|---|
| Samsung (Memory + Foundry) | 005930.KS | 15-20% | POR — DRAM HKMG/high-k, NAND gate stack/polysilicon, 3nm GAA dielectric |
| TSMC | TSM | 10-15% | POR — N2/A16 gate stack, capacitor metal liner; co-development |
| SK Hynix | 000660.KS | 10-15% | POR — DRAM high-k ALD, HBM TSV liner |
| China fabs (SMIC, CXMT, YMTC, ChangXin) | private | 25-30% combined | Workhorse trailing-edge SiN/SiO2/poly furnaces |
| Micron | MU | 5-10% | DRAM HKMG |
| Kioxia | 285A/285A | 285A.T | 5-10% |
| Intel Foundry | INTC | <5%, growing | 18A / 14A GAA — POR being established |
Top-3 concentration ~40-50%; top-5 ~65-75%; no single customer >20%. The China cluster (~25-30%) is the dependency that matters.
Geographic mix (FY25): China ~45-50%, Korea ~15%, Taiwan ~15%, Japan/SE Asia ~15%, US/EU ~5-10%.
Moat. (1) POR lock-in — once qualified for a process step, switching is a 12-24-month requalification per node. (2) Installed base + service — high-margin recurring revenue and an information advantage. (3) Process IP — characterized as a 75-year accumulation of recipe IP (gas chemistries, temperature profiles, chamber geometries) yielding POR-grade films; the profile alternatively frames this as decades of Hitachi-origin recipes. (4) Cost-per-wafer advantage in batch workloads, structurally hard for single-wafer competitors to dislodge on thick/conformal films. Within select GAA steps, Korean/Taiwanese channel checks (2024-25) place Kokusai at ~50-70% share of specific high-k POR steps at Samsung Foundry and TSMC.
Footprint. Manufacturing concentrated single-site at the Toyama plant (a tail risk — earthquake/labor disruption has no domestic alternative); R&D at Kokubunji R&D Center (Tokyo); local service/application hubs in Korea, Taiwan, China, US, Singapore (no manufacturing). 100%-owned sales/service subs: Korea, Taiwan, Shanghai/China, Singapore, USA. No JVs of consequence; organic-only strategy. Amicable but non-exclusive Hitachi IP-licensing/service relationship persists.
Financials
Income statement & margins (JPY billions; FY ends March 31).
| FY22 | FY23 | FY24 | FY25 | FY26E | |
|---|---|---|---|---|---|
| Revenue | 245.4 | 245.7 | 180.8 | 238.9 | ~265 |
| Revenue growth YoY | n/a | +0.1% | -26.4% | +32.1% | +11% |
| Gross profit | 107.1 | 100.8 | 75.0 | 101.7 | ~114 |
| Gross margin | 43.6% | 41.0% | 41.4% | 42.6% | ~43% |
| Operating income | 70.7 | 56.1 | 30.7 | 51.3 | ~62 |
| Operating margin | 28.8% | 22.8% | 17.0% | 21.5% | ~23% |
| Net income | 51.3 | 40.3 | 22.4 | 36.0 | ~45 |
| Net margin | 20.9% | 16.4% | 12.4% | 15.1% | ~17% |
| Diluted EPS (JPY) | 215.3 | 168.8 | 93.7 | 152.5 | ~191 |
The FY24 (Mar'24) trough was the memory capex air-pocket — DRAM/NAND customers paused. FY25 was a sharp recovery as DRAM/HBM and leading-edge logic restarted; margins still below the FY22 peak. Revenue 3-year CAGR is -0.9% (cycle-distorted); forward consensus +11% FY26E, +12% FY27E.
Cash flow & balance sheet (JPY billions).
| FY22 | FY23 | FY24 | FY25 | |
|---|---|---|---|---|
| Operating cash flow | 73.6 | 51.4 | 33.7 | 46.2 |
| Capex | 3.3 | 6.1 | 15.6 | 27.7 |
| Free cash flow | 70.3 | 45.3 | 18.1 | 10.8 |
| FCF margin | 28.6% | 18.4% | 10.0% | 4.5% |
| Total debt | 123.2 | — | — | 60.2 |
| Net debt | 13.6 | — | — | 13.5 |
| Net debt / EBITDA | 0.17x | — | — | 0.21x |
| ROIC (rough) | ~22% | ~16% | ~9% | ~13% |
| Inventory | 51.6 | — | — | 83.2 (35% of revenue) |
FCF cratered to JPY 10.8B (4.5% margin) on Toyama capex + inventory build — investment-cycle compression, not earnings degradation. The JPY ~16B working-capital build + JPY ~24B incremental capex vs FY22 fully explain the FCF differential; normalized mid-cycle FCF should be JPY 40-60B (3-4% FCF yield). Balance sheet is now clean: debt halved JPY 123B → JPY 60B since the KKR LBO; net debt JPY 13.5B against ~JPY 65B EBITDA; interest coverage ~35x. Note a cross-fragment discrepancy on FY25 net deleveraging: the deep-dive Core-Four read cites "debt paydown JPY 31.5B net," while its mgmt-DD-style capital-allocation note cites "Debt paydown JPY 91.5B (FY25) vs JPY 60B new issuance" — both retained; the JPY 123B → JPY 60.2B total-debt move is the reconciling anchor.
Capital intensity. Capex JPY 27.7B FY25 = 11.6% of revenue vs JPY 3.3B FY22 = 1.4% — investment-cycle elevated for Toyama, normalizes to 3-5% post-expansion (CY27+). Incremental revenue per dollar of capex not yet measurable; FY27-28 is the read-out.
Quarterly EPS surprise tape (yfinance earnings dates, last 8 quarters).
| Quarter | EPS Est. | Reported | Surprise |
|---|---|---|---|
| Q1 FY26 (May'26) | 28.78 | 31.16 | +8.3% |
| Q4 FY25 (Feb'26) | 26.59 | 31.12 | +17.0% |
| Q3 FY25 (Nov'25) | 28.88 | 37.64 | +30.3% |
| Q2 FY25 (Aug'25) | 33.48 | 29.08 | -13.2% |
| Q1 FY25 (May'25) | 38.87 | 42.89 | +10.4% |
| Q4 FY24 (Feb'25) | 35.95 | 34.65 | -3.6% |
| Q3 FY24 (Nov'24) | 27.44 | 20.30 | -26.0% |
| Q2 FY24 (Aug'24) | 29.15 | 56.76 | +94.7% |
Beat/miss 5 beats / 3 misses; average surprise +14.7%, median +9.4%. Last four quarters are all beats with moderating magnitude (30% → 17% → 17% → 8%) — second derivative positive but decelerating; consensus has caught up. Next-quarter (Aug'26) consensus JPY 43.18 (deep-dive) is the decisive print.
Incremental margins (annual; quarterly disclosure in Japan is less granular).
| FY23 v FY22 | FY24 v FY23 | FY25 v FY24 | |
|---|---|---|---|
| Δ Revenue (JPY B) | +0.3 | -64.9 | +58.1 |
| Δ Gross profit | -6.3 | -25.8 | +26.7 |
| Incremental GM | n/a | 39.8% | 46.0% |
| Δ EBIT | -14.6 | -25.4 | +20.6 |
| Incremental EBIT margin | n/a | 39.1% | 35.5% |
FY25-vs-FY24 incremental gross margin of 46% sits above the reported 42.6% GM — the new revenue dollar is higher-quality (mix shift to higher-ASP leading-edge tools). Incremental EBIT 35.5% vs reported 21.5% implies meaningful operating leverage; sustainable steady-state incremental EBIT ~30-35%, suggesting peak-cycle margins could reach ~26-28% if revenue scales another 25-30%. This is the most encouraging quantitative read in the analysis.
ROIC vs WACC. Estimated WACC ~7% (low Japan rates; 2.26 beta offset by high equity weight + cash). ROIC ~13-22% across the cycle, above WACC every year — Kokusai creates value; the question is the multiple the market pays. R&D ~12-14% of revenue (no disclosed split), heavy on high-k, work-function metal, and conformality engineering for sub-2nm.
Industry landscape
Semiconductor wafer-fab equipment (WFE) is structurally cyclical (~2-3 years up, 1-2 down). Trough was 2H24; mid-cycle now (mid-2026); peak expected 2027-28. WFE ~$110B 2024 → $125-140B 2026-27; deposition ~25% of WFE = ~$28-35B; batch ALD/LPCVD SAM ~$8-12B with Kokusai estimating ~30-40% share. GAA lifts deposition $/wafer +30-40% vs FinFET; at a GAA node deposition is ~30-35% of WFE spend (up from ~25% at FinFET), of which batch is ~25-30% — i.e. ~8-10% of total WFE flows through batch tools. The cycle is unusual in three ways: AI capex is structurally additive, leading-edge is more concentrated than ever (TSMC + Samsung + Intel + a few HBM specialists), and China is now ~a third of the market. Competitive set: AMAT (~$26B, #1 broad-line deposition), Tokyo Electron 8035 (~$16B, #1 CVD broad), Lam Research LRCX (~$15B, #1 etch), ASM International ASM.AS (~$3B, #1 single-wafer ALD — the biggest GAA threat), Kokusai (~$1.6B, #1 batch ALD), Wonik IPS 240810.KQ (~$0.6B Korean second-source, Samsung-favoured), NAURA 002371.SZ (~$3.5B China, trailing-edge, ~5 years behind leading-edge).
See sector page: passives-mlcc
Management
Verdict: Overall Management Grade B+ — trustworthy, competent, conservative; not exceptional. No red flags identified. A career-Hitachi team that KKR kept in place through the 2017 LBO and 2023 IPO. Stable, low-drama, conservatively aligned. The trust question is less about character than whether a hand-picked KKR management team has the strategic agency to navigate a step-change in geopolitical exposure.
Leadership.
- Fumiyuki Kanai (金井 史幸) — President & CEO since the 2018 carve-out (~8 years in title). Career semiconductor-equipment executive within the Hitachi group; engineering background, deep batch-deposition knowledge. Internal succession — KKR retained the operating CEO rather than parachuting an outsider. Led the operational standalone transition, the 2017 KKR LBO, the COVID memory air-pocket, the Oct 2023 IPO, Toyama capacity expansion, and the FY25 JPY 18.5B buyback. Low-key, sober, non-promotional, engineering-driven communicator. No litigation/enforcement disclosed.
- Toshihiko Tanaka (田中 俊彦) — Director, EVP, CFO (~8 years). Hitachi corporate-finance lineage; hired at carve-out to manage the LBO financial architecture and IPO. Led the oversubscribed Oct 2023 IPO (priced JPY 1,840), oversaw deleveraging JPY 123B → JPY 60B over FY22-25, executed the FY25 JPY 18.5B buyback. No litigation disclosed.
- Akira Kuribayashi (栗林 明) — EVP, COO / Sales & Marketing. Long-tenure Hitachi Kokusai sales executive; 20+ year relationships at Samsung, SK Hynix, TSMC. No litigation disclosed.
- Other EVPs head R&D, Manufacturing, China, and Korea-Taiwan regions — all career insiders, no external hires at EVP level since 2018. Strength: institutional knowledge and customer relationships. Weakness: limited external perspective.
Skin in the game — Yellow. yfinance shows "Insider ~15.4%," but this conflates KKR residual (~30-50M shares, ~13-21%, ~$1.2-2.0B) and Hitachi residual (~10-25M, ~4-11%, ~$400M-1B) with management; true management+board direct ownership is single-digit % (likely <2% combined). Kanai owns ~50-200K shares (<0.1%, ~$2-8M); Tanaka ~30-100K (<0.05%, ~$1-4M); other EVPs ~30-80K each. No open-market insider buying disclosed in the last 12 months (Japan TDnet would catch it). No 10b5-1 equivalent. KKR has been a steady seller — secondary offerings and small disposals. Alignment is conservative-Japanese-norm, not founder-style; on par with TEL/ASMI/AMAT executives.
Shell & cross-holdings scan — GREEN. No related-party transactions of consequence in the latest Yuho or the 2023 IPO prospectus. Operating subs (Korea, Taiwan, Shanghai, Singapore, USA) all 100%-owned; no JVs of consequence, no insider-controlled counterparties, no nominee directors, no layered ownership. The KKR Holdings SPV (Hitachi Kokusai Electric Holdings) is a clean equity vehicle, not an asset-shuffler. Legacy Hitachi IP-licensing is disclosed and standard; Toyama property owned, Tokyo HQ leased on commercial terms. No METI/JFTC/BIS/OFAC enforcement; subject to METI export-control regulation as a compliant operator. PwC Japan audit, standard opinion.
Compensation — Yellow. CEO total comp ~$1.3M (~JPY 200M), materially below peers (TEL CEO ~$3-4M, ASMI ~$3-5M, AMAT ~$25M) and cash-heavy (cash+bonus ~65-75%, stock-based ~25-35% — mostly time-vested RSUs with some performance-vested). Performance metrics inferred as operating income, EPS, ROIC; PSU hurdles not granularly disclosed (Japanese filers aren't required to). No US-style golden parachutes (CIC entitlement ≤1x base); no family on payroll/unusual perks. Mixed signal — low comp limits self-enrichment risk but also limits upside alignment for ambitious capital allocation.
Capital allocation — B+ (Green). FY25 buyback JPY 18.5B executed at ~JPY 5,000-6,500 (at ~28-40x P/E — a capital-return buyback, not an opportunistic "buy cheap" one; Capital Allocation Timing graded Neutral). No M&A under KKR or post-IPO (organic-only). No post-IPO dilution — share count flat ~234M, no shelf/ATM/convertible (dilution discipline grade A). Dividend 0.78% yield, payout ~25-30%. Capex ramped JPY 3.3B → 27.7B FY22-FY25 for Toyama.
Credibility — High, 8/10 (Green). Tracked statements FY24-FY26: 6/6 follow-through (small, tailwind-biased sample). Guidance style conservative/sandbagger lean (5 beats/3 misses, avg +14.7%). Weasel-language frequency low. Through the FY24 memory trough (one quarter missed by 26%) management did not panic-cut, issue equity, lay off, or change strategy — quiet competence in a trough.
Board & governance — Yellow-Green. ~8 directors: ~3 executive, ~1-2 KKR-affiliated outside (per shareholder agreement), ~3-4 independent (ex-METI officials, ex-Japanese industrial CEOs, governance specialists; independent audit-committee chair). ~50% independent (Japan-standard for post-IPO TSE Prime), trending up as KKR exits. No dual-class shares, no poison pill, no staggered board. No notable activist proposals. Better than Japan median, not US best-in-class; improves mechanically as KKR rolls off.
Bottom line. A team you can trust to execute a known plan competently — exactly right for a thesis that depends on executing the GAA-cycle roadmap the foundries have already locked in. Not the team you'd expect to make a bold contrarian pivot or defend the stock in a drawdown via personal buying, so less well-suited to a thesis requiring management to outmanoeuvre an existential geopolitical or competitive threat.
Catalysts & risks
Catalysts — secular tailwinds (the bull engine).
- GAA logic ramp (2025-2028+) — +30-40% deposition $/wafer vs FinFET; ~$500M-1B incremental Kokusai-addressable TAM from the GAA node alone; biggest single driver, durable 3-5 years per node then the pattern continues to CFET.
- HBM cycle (2024-2030) — every HBM stack uses HKMG; HBM4 from 2026, HBM5 by 2028-29; ~$200-300M incremental TAM; durable 5-7 years.
- 3D NAND vertical scale (2025-2030) — 200→400→500+ layers, each increment adds deposition cycle time; ~$150-250M incremental TAM; durable 5+ years (Kioxia, SK Hynix, Micron roadmaps).
- China trailing-edge demand — ~$700M of current revenue; continues unless controls tighten dramatically; durable 2-5 years.
- Service mix growth toward 25%+ of revenue.
Near-term catalysts (0-12mo): Q2 FY26 earnings August 13, 2026 (guide direction; ~3 months out); TSMC N2 risk-production ramp checkpoint; Samsung 2nm progress; Intel 18A; any TSMC/Samsung/Intel order signal; KKR sell-down events; China export-rule reviews (fall/spring 2026). Leading indicators: TEL/ASMI/AMAT quarterly commentary (Kokusai usually +1 quarter behind), SEAJ monthly Japanese WFE orders, Chinese fab utilization/capex.
Structural change already in place. GAA POR positions awarded (Samsung 3nm mid-2024, TSMC N2 late-2024/early-2025, Intel 18A 2025) — sticky 5-7-year revenue streams; HBM4 architecture locked batch ALD for HKMG; China export controls stabilized at current state; KKR overhang reducing. What analysts may miss: the durability of the batch ALD POR — models often assume gradual single-wafer share loss, but per-node requalification slows that.
Risks (the bear case), with likelihood.
- China export-control escalation — High likelihood, significant impact (-20-30% revenue). The dominant risk and structurally unclosable. Kokusai's SiN/SiO2/poly furnaces are exactly the workhorses Chinese trailing-edge fabs need; shipments are governed tool-by-tool under METI controls. Current rules permit substantial China shipments; risk is escalation (lower process-node threshold, e.g. controls extended to 28nm DRAM; or expansion to more tool categories), not the status quo. Mitigant: ~50% non-China, geographic diversification, Korea/TW/US capacity push.
- ASMI single-wafer ALD share-take at the GAA gate stack — Medium likelihood, moderate impact (-10-15% to long-term TAM growth). Not catastrophic to the batch core but a share-of-share headwind; mitigant is batch cost advantage + POR moat + heavy next-gen batch R&D; manageable per node.
- Quarterly miss + multiple compression — High likelihood at any given quarter (-15-25% stock impact). With FCF yield <1% and the multiple pricing the bull case, an in-line/miss print compresses 33x → 22-25x. Execution is the only response.
- WFE cycle normalization post-CY28 — High likelihood, significant impact (prior cycles swung Kokusai revenue ±35-40%, operating margin 15-29%).
- Single-site Toyama manufacturing — Medium; earthquake/disruption Low-Medium — hard to close without $1B+ capex into a second site; insurable only.
- KKR float overhang — Medium-High — closable; resolves as KKR exits over 2026-27.
- JPY strengthening — Medium, moderate impact — partially hedgeable.
- Valuation compression — High in the current setup — a market risk, not company risk; unmanageable.
Emerging/longer-dated threats. NAURA indigenous China WFE (gradual trailing-edge share-take CY28-32); Wonik IPS Korean second-source at Samsung (slow erosion, bounded by capacity + recipe IP); selective deposition (ASMI/AMAT leading, Kokusai investing not leading — possible 2030s share shift if it scales).
Bear scenario trigger stack: tighter China tooling controls (-10-15% revenue) + a quarterly miss (33x → 22x) + KKR accelerating its sell-down → combined JPY 4,000-4,500 (-30 to -37% from current), probability ~20-25%.
Dilution risk: Low (~233M shares stable since IPO; FY25 buyback reduced count; no convertibles/shelf; KKR is a seller of existing shares, not dilutive). Key-person risk: Moderate (Kanai is the face; internal succession depth via long-tenure EVPs, none externally named).
Valuation / DCF
Multiples (May 15, 2026): P/E TTM 49.3x, P/E FY+1E 33.1x, EV/EBITDA TTM 25.0x, P/B 7.0x, EV/Revenue 6.1x, FCF yield ~1.7%, dividend 0.78%.
Peer comparison.
| Company | P/E FY+1E | EV/EBITDA TTM | P/B |
|---|---|---|---|
| Kokusai 6525 | 33.1x | 25.0x | 7.0x |
| Tokyo Electron 8035 | ~25x | ~18x | ~5x |
| ASM International ASM.AS | ~32x | ~22x | ~7x |
| Applied Materials AMAT | ~22x | ~17x | ~7x |
| Lam Research LRCX | ~24x | ~18x | ~12x |
Kokusai trades at a premium to AMAT/LRCX/TEL but in line with ASMI — both smaller, more pure-play, more GAA-levered. The premium is defensible if and only if the GAA POR narrative plays out. Versus its own ~2.5-year post-IPO range of 18-50x P/E TTM it sits at the high end but not the peak; ~2x the TSE Prime market average (~16x P/E).
Implied expectations at 33x FY+1E (JPY 191 EPS implied): revenue +~10-12%/yr through CY28, operating margin 21.5% → ~24-25% by FY28, net income compounding ~12-15%/yr, ~50% of forward earnings retained into the multiple. Realistic only if GAA POR holds.
Quick DCF (checklist): JPY 45B FY26E NI → 12% growth → 25% terminal margin → 9% discount → fair value ~JPY 7,000-8,500 (slight upside from spot).
Scenarios / price targets.
- Base case JPY 7,500-8,000 (+18-26% to 12-month) — premium to mean consensus JPY 7,082, predicated on full GAA POR retention + one more positive memory upgrade cycle. (Deep-dive narrows base to JPY 7,500.)
- Upside JPY 8,500-10,000 if revenue compounds ~15% and margins reach 26%+ by FY28.
- Downside JPY 4,500-5,000 (~-25 to -37%) if revenue growth slows to ~5% (modest China hit) or margins stay flat → multiple compresses to 22-25x (≈ JPY 4,200-4,800); deep-dive downside JPY 4,500 (-29%) on China escalation + ASMI share-take; bear-stack JPY 4,000-4,500.
Asymmetry is moderate. Conviction Medium — high-quality business, but the valuation prices most of the bull case and entry timing matters more than usual on a 2.26-beta stock with binary China exposure.
Analyst sentiment. 10 analysts covering; 3 Strong Buy / 5 Buy / 1 Hold / 1 Strong Sell (8 Buy-or-better). Mean target JPY 7,082 (+12% from JPY 6,336); range JPY 3,900-10,000 (the JPY 3,900 outlier is the China-escalation + GAA single-wafer share-take bear). Recommendation mean 1.73 (Buy-leaning).
Ownership. Institutions ~52.4% (172 filers); insiders ~15.4% (conflates KKR/Hitachi residuals); float 195M / 234M (~83%). Holders: KKR residual ~10-15% (steady seller, tracked via Form 18), Hitachi residual ~5-10% (held ~20% at IPO), Japanese passive ~10%, foreign passive ~5-8% (Vanguard/BlackRock/State Street), foreign active ~10-15% (sector specialists — likely Fidelity/T Rowe/Capital/Baillie Gifford). KKR overhang is the biggest known supply pressure, expected to resolve over 2026-27.
Decision log
2026-05-15 — Pre-buy checklist verdict: BUY — half-size starter (50% of an intended 1-1.5% book position) at JPY 6,300-6,400, scale-in over 3 tranches. Conviction Medium. Defined max loss -25% (JPY 4,800 stop). All five FundamentEdge gates pass (one mild caveat on the decelerating second derivative — the next print is decisive). Three soft behavioral flags raised (FOMO — stock +143% off the 52w low; narrative seduction on "POR for GAA"; recency — FY25 +32% is one recovery year off the FY24 trough), mitigated by half-size + scale-in + defined exits.
Single most important thing that must go right: Kokusai retains its batch ALD POR positions through the N2/2nm transition without material share-take by ASMI single-wafer ALD. Expected holding period 18-36 months (covers the GAA ramp into HVM and one HBM cycle).
Entry plan (scale-in over 3 tranches): Tranche 1 (50%) at JPY 6,300-6,400 now; Tranche 2 (25%) at JPY 5,800-6,000 on a pullback OR on a clean break above JPY 6,800 with volume; Tranche 3 (25%) reserved for thesis confirmation (TSMC N2 POR confirmation, a Q2 FY26 earnings beat, or a stress-test pullback to JPY 5,200-5,500).
Exit / re-evaluation triggers (defined upfront): hit JPY 7,500-8,000 → trim 30-50%; hit JPY 9,000+ → trim to half-size; ASMI grabs visible POR share → trim 50%; China tightens export controls on Kokusai tools → trim 50%; consecutive quarterly misses with downward guide → exit fully; break below JPY 5,200 on no specific news → re-evaluate; KKR sell-down completes and stock breaks higher → consider adding.
Technical read (2026-05-15): confirmed uptrend, price ~JPY 6,378 above 50d MA (JPY 6,091, +4.7%) and 200d MA (JPY 4,994, +28%), golden cross intact; RSI(14) 42.5 (neutral); price 13% below 52w high, 143% above low; volume below average. Consolidating JPY 6,000-6,800. Technical verdict: Neutral with a slight wait bias — scale-in acceptable, no high-conviction entry urgency.
Pre-buy scorecard: thesis clear Yes; understand business Yes; FundamentEdge gates Yes (mild 2nd-deriv caveat); mgmt incentives aligned Partial; financials healthy Yes; valuation reasonable Yes (margin of safety thin); behavioral traps 3 mild flags; technicals neutral; sizing appropriate Yes; clear exit plan Yes. Recommendation: BUY — half-size starter, scale-in, max loss -25%; wait for the Aug 13 print before the second tranche unless the stock pulls back to JPY 5,800 first.
Portfolio-fit note. High correlation to other semi-cap names; if Pink already holds 2308 (TSMC), 6809 (Keyence), 4062 (Ibiden), 4971 (MEC), adding Kokusai concentrates Japan/Asia semi-cap exposure.
2026-05-15 — Management DD verdict: Overall grade B+. No red flags. Would trust this management with capital, conditionally — competent at executing a known plan, not suited to a thesis requiring a bold contrarian or geopolitical pivot.
2026-05-15 — Deep-dive conviction: Medium. Base case JPY 7,500-8,000; downside JPY 4,500. High-quality business; valuation prices most of the bull case; entry timing matters on a 2.26-beta stock with binary China exposure.
2026-05-30 — Consolidation (W3). Four research fragments (mgmt-dd, deep-dive, buy-checklist, profile) merged into this single thesis-first canonical page. Open data flag carried forward: the SemiAnalysis "Going Vertical — Gate-All-Around 3D" piece (Oct 15, 2023, https://semianalysis.com/2023/10/15/going-vertical-gate-all-around-3d/) — the authoritative Kokusai+GAA write-up — is misrouted in Pink's SA mirror (the local file contains Fabricated Knowledge content instead); re-fetch recommended. STF Research substack coverage check ran 2026-05-15: no Kokusai coverage found (two posts could not be loaded due to an unstable browser bridge — re-attempt recommended).
SECTOR-LABEL FLAG (2026-05-30). This consolidation was tasked under sector slug passives-mlcc, which is carried in frontmatter and the industry pointer per the reorg spec. That label is factually wrong for this entity: 6525 / Kokusai Electric is a semiconductor batch-ALD/CVD deposition-equipment maker, not a passives/MLCC company. No fragment was a wrong-entity stub — all four correctly describe Kokusai; the mis-assignment is the sector tag alone. Flag for Pink to re-sector to a semiconductor-equipment / semi-cap page before relying on the passives-mlcc cross-link.
Sources
Fragments folded into this canonical page (consolidated 2026-05-30):
6525-mgmt-dd.md— Management Due Diligence (swarm-research / mgmt-dd, 2026-05-15)6525-deep-dive.md— Deep Dive, full write-up (swarm-research / deep-dive, 2026-05-15)6525-buy-checklist.md— Pre-Buy Checklist (swarm-research / checklist, 2026-05-15)6525-profile.md— Company Profile (swarm-research / profile, 2026-05-15)- (A
6525-filings.mdstub exists in the folder but is an empty placeholder — not merged.)
External sources cited across the fragments:
- Yahoo Finance (yfinance) — financials, earnings dates, ownership, technicals
- Kokusai Electric IR — https://www.kokusai-electric.com/en/ir/ ; FY25 results presentation (May 2026)
- SemiAnalysis local mirror (
~/Dropbox/Wafflebun/KB/wiki/semianalysis/) — Clash of the Foundries (Mar 2024), Wafer Wars — Latest Restrictions (Oct 17, 2023), China AI & Semiconductors Rise (2023); plus the misrouted "Going Vertical — GAA 3D" piece (Oct 15, 2023) flagged for re-fetch - TEL, ASMI, AMAT IR — competitive context
- METI export-control public filings (referenced via SA)
- Japanese disclosures — Yuho annual securities report, TDnet timely disclosure; PwC Japan audit
- STF Research substack (https://stfbutnou.substack.com/) — coverage check 2026-05-15, no Kokusai coverage found
Topics referenced across fragments: ai-infrastructure, japan-semi, supply-chain-security, green-finance.
Consolidation queue (merged 2026-05-30)
These fragment files were folded into this canonical page and remain live pending Pink's archive confirmation.
- [ ]
6525-mgmt-dd.md - [ ]
6525-deep-dive.md - [ ]
6525-buy-checklist.md - [ ]
6525-profile.md