268A — Rigaku Holdings Corporation
Thesis
The Bull Case
Rigaku is a bet on a physics problem. As chips go three-dimensional — GAA transistors stacking vertically, NAND memory towers exceeding 200 layers, HBM packages bonding multiple dies through silicon vias — the optical metrology tools that have dominated semiconductor inspection for decades are hitting fundamental physical limits. X-rays can see what light cannot: through opaque layers, into buried structures, with angstrom-level precision. Rigaku, the world's #1 X-ray diffraction instrument maker with 75 years of accumulated IP, is the company that fills the gap.
The "lab to fab" narrative is the right way to think about this stock. Rigaku is not a boring lab instrument company that happens to sell some semi tools. It's becoming a semi equipment company with a legacy lab business attached. The market hasn't fully absorbed this yet. Dong/Shim's key insight — that semiconductor exposure is ~60% of operating profits when you reclassify GaN/SiC analytical tools and semi-related service revenue — reframes the entire investment case. The market still prices Rigaku as a lab instrument company.
The X-ray S-curve argument is structurally sound. Every next-gen chip architecture (GAA, CFET, 3D DRAM, HBM hybrid bonding, co-packaged optics) requires more X-ray inspection steps. This isn't a cyclical demand story — it's a penetration story within a growing market. X-ray tool intensity per fab rises from 3-4 tools (FinFET era) to 5-8 (GAA) to 8-14 (CFET). That's a 2-4x increase in tools per fab before counting new fab builds. Only ~5% of wafers in fabs that have adopted X-ray are actually inspected with it. Early innings.
The investment case rests on the semiconductor process control segment, which grew 19% in FY2025 to become 30% of revenue and 42% of operating profit — headed to 33% and 49% by FY2026. But the true semi exposure is higher than reported numbers suggest. Dong/Shim correctly identifies that GaN/SiC analytical tools are reported in the "Analytical" bucket but serve semiconductor fabs, and ~1/3 of service/parts revenue comes from semi customers. When you re-slice by end market, semiconductor is already 60%+ of operating profits, not ~30% of revenue. This is the single most important insight for understanding the stock — the market still prices Rigaku as a lab instrument company. Memory applications tripled in a single year. Asia ex-China revenue went from ¥1.1B to ¥9.0B. Seven new products are in various stages of ramp, targeting a combined $1B addressable market by 2030. If Rigaku captures 50%+ share as management targets, the semi segment alone would generate ¥50B+ in revenue at 30%+ margins — more than half of today's entire company revenue.
The Bear Case
The stock has already priced much of this in at ~30-39x trailing P/E. Carlyle's 41% overhang creates persistent selling pressure — their CJP IV fund has a 2030-2033 liquidation window, meaning years of selling ahead. FY2025 showed margin compression that management needs to reverse. The CEO is a PE-placed operator with minimal domain expertise in analytical instruments. And the IPO was 100% secondary — the company received $0 in proceeds. The LBO debt (~¥62B) sits on the public company's balance sheet while the dividend exceeds free cash flow.
The problem is everything else. FY2025 was an investment year that saw net income drop 17% despite revenue growth. Two of five "outside" directors are Carlyle affiliates. The CEO's personal ownership is 0.032% (~¥86M) — no evidence of personal share purchases. Analysts only number 4-6, and the stock trades 28-33% above consensus targets.
Conviction Statement
Medium-High on the business, cautious on the entry point. The thesis is structurally sound — X-ray inspection is physics-mandated for 3D chips, and Rigaku is the dominant pure-play supplier. But the valuation demands a margin of safety that hasn't always been available. The opportunity is knowing exactly what you want to buy and at what price, so you can act decisively when the inevitable Carlyle block trade or cyclical wobble creates the entry point.
Fair value range: ¥2,160-2,520 (30-35x FY2027E EPS of ¥72). Bull case: ¥3,000+ if semi metrology accelerates to 25%+ growth. Bear case: ¥1,200-1,400 if the cycle turns or new products underperform.
Held position — current stance (2026-05-02 pre-sell review, see Decision log): Pink holds 268A. By 2026-05-02 the stock had run to ¥2,667 (+39% in ~30 days from the ¥1,915 deep-dive level; new 52-week high ¥3,100), pushing it above base-case fair value and into a worse asymmetry: roughly +30% upside to bull case vs −55% downside to bear case. None of the wiki-defined invalidation triggers had fired and the thesis was intact, so the discipline call was TRIM IF OVERSIZED, otherwise HOLD — not sell the thesis, just right-size the run. Hard stop on a close below ¥2,000; soft re-evaluate stop below ¥2,300.
Snapshot
One-liner: Rigaku is the world's #1 X-ray diffraction instrument maker turning into a semiconductor metrology growth company — X-ray inspection is physics-mandated for 3D chips (GAA, CFET, 3D NAND, HBM, advanced packaging) where optical tools go blind, and Rigaku is the dominant pure-play supplier; the catch is a Carlyle PE exit overhang (~41%) and a price that has already run hard.
| Field | Detail |
|---|---|
| Full Legal Name | Rigaku Holdings Corporation (リガクホールディングス株式会社) |
| Ticker | 268A (TSE Prime Market); OTC: RGAKF |
| Sector | Information Technology — Electronic Equipment, Instruments & Components → semi-equipment / metrology |
| Headquarters | 3-9-12 Matsubara-cho, Akishima-shi, Tokyo 196-8666, Japan |
| Year Founded | 1951 (by Dr. Yoshihiro Shimura) |
| IPO Date | October 25, 2024 (100% secondary; company received $0) |
| FY End | December 31 |
| Employees | ~2,226 (FY2025); guided to 2,351 in FY2026 |
| Countries Served | 136+ |
| Website | rigaku-holdings.com / rigaku.com |
| ISIN | JP3969750003 |
Price / valuation context (two timestamps — verify live before any PT math):
- mid-Mar 2026 (deep-dive): ~¥1,915 | mkt cap ~¥425B (~$2.8B) | EV ~¥463B | P/E (TTM) 38.7x | fwd P/E (FY26E) 32.3x | EV/EBITDA 21.2x | P/B 4.8x | div yield ~1.0% | 52-wk range ¥641–¥2,076
- 2026-05-02 (pre-sell checklist): ¥2,667 (+39% vs the deep-dive level) | new 52-wk high ¥3,100 | −14% below ATH | next earnings then dated 2026-05-13 (the canonical monitoring table lists May 14 — minor 1-day source discrepancy, both retained) | ex-div 2026-06-29
- Analyst coverage thin: 4–6 analysts, consensus Buy (3 Buy / 1 Hold / 0 Sell), avg PT ¥1,462–1,500 (range ¥950–¥2,000); at ¥2,667 the stock traded ~+78% above consensus target (vs +28–33% above at the deep-dive level)
- Conviction: Medium-High on the business, cautious on entry. Fair value ¥2,160–2,520; bull ¥3,000+; bear ¥1,200–1,400. Held; trim-if-oversized at ¥2,667, hard stop ¥2,000.
Business
What Rigaku Does
Rigaku builds machines that shoot X-rays at materials and measure how those X-rays scatter, reflect, or fluoresce — revealing crystal structure, elemental composition, layer thickness, and defects at the atomic scale. Think of it as the "quality inspector's microscope" for advanced manufacturing, except instead of visible light, it uses X-rays that can penetrate deep into 3D chip structures that optical tools cannot see.
If a semiconductor fab is a kitchen making the most complicated recipes in human history, Rigaku makes the food safety inspectors. Their instruments don't build the chips — they verify that every layer, every film thickness, every crystal structure is exactly right. As chips get more complex, there are more things to inspect, and the older inspection methods (optical, electron-beam) can't see through the new 3D structures. That's where X-ray steps in.
The company was founded in 1951, spent 70 years building the global #1 position in X-ray diffraction instruments (~29% market share), went private via Carlyle LBO in 2021 for ~¥100 billion, and re-IPO'd on the TSE Prime Market in October 2024 — Japan's second-largest IPO that year at ~¥129 billion raised.
Before Rigaku existed, X-ray analysis was done with hand-built lab equipment at universities. Rigaku commercialized it — first with rotating anode X-ray generators (1952, world's first), then with increasingly sophisticated automated instruments for industrial and semiconductor applications. The company has a 75-year headstart that no competitor can replicate.
The Three Businesses
Rigaku reports three segments, but the story is really about how the mix is shifting.
General Purpose Analytical Instruments (43% of FY2025 revenue, 32% of operating profit)
The legacy bread-and-butter business. XRD systems for universities, XRF analyzers for mining companies, thermal analysis equipment for materials labs. Revenue was ¥40.4B in FY2025 (-2% YoY), with operating margins of 15.4%. It's a stable, globally diversified business serving academia, pharma, battery materials, and industrial QC. The growth rate will be low single digits — management is guiding +3% for FY2026 without assuming a US academic spending recovery — but it provides a steady base of recurring instrument sales and service contracts.
The regional dynamics within this segment are interesting. Europe grew 27% on academic R&D in new materials and sustainability, while US academia got hammered by budget cuts (-13%) and China dropped 23% from a supplementary budget reversal. Life sciences was the growth pocket (+25% from pharma and chemical demand). Management is pivoting resources toward industrial applications — semi components, EV battery materials, pharma — to reduce academic cyclicality.
Key products:
- SmartLab SE: Flagship multipurpose XRD system with AI-powered sample handling (+55% efficiency) and pattern recognition (+35% accuracy over manual analysis). The workhorse for university labs and industrial R&D.
- XtaLAB Synergy: Single-crystal X-ray diffractometer for molecular structure determination. Dominant in academic crystallography.
- NEX QC II Series: Benchtop XRF analyzers launched at Pittcon 2026 for industrial quality control — petroleum, coatings, recycling.
- ZSX Primus series: Wavelength-dispersive XRF for elemental analysis in mining, cement, metals.
- MiniFlex: Benchtop XRD. Lower price point, high volume. Academic and industrial screening.
- Icon Series (4th gen): Handheld Raman spectrometers for field analysis and security screening.
Competitors: Bruker, Malvern Panalytical (Spectris), Shimadzu, Thermo Fisher. Rigaku leads in XRD but faces established competition across the product range.
Semiconductor Process Control Equipment (30% of revenue, 42% of operating profit)
The crown jewel and the entire reason to own this stock. Revenue grew 19% to ¥27.8B with 29% operating margins, and the segment is guided to ¥33.3B (+20%) with 31.4% margins in FY2026. This is where Rigaku's X-ray technology meets the AI semiconductor investment cycle.
The demand drivers are structural, not cyclical. Advanced logic below 2nm requires gate-all-around (GAA) transistor architectures that are inherently three-dimensional — X-ray small-angle scattering is one of the only techniques that can measure the critical dimensions of these buried structures. High-bandwidth memory (HBM) for AI accelerators stacks multiple DRAM dies with through-silicon vias, creating inspection challenges that X-ray excels at. Advanced packaging (CoWoS, chiplets) creates new metrology steps at the packaging level.
FY2025 application breakdown:
| Sub-segment | Revenue (¥100M) | YoY Change | Notes |
|---|---|---|---|
| Memory | 72 | +213% | HBM, 3D NAND — exploded from ¥2.3B to ¥7.2B |
| Advanced Logic/Foundry | 58 | -29% | Geographic shift from Americas to Asia, not demand decline |
| WFE-related | — | +31% | Next-gen development projects |
| Power Devices | 23 | — | SiC, GaN metrology |
| Components/Services | 32 | — | Recurring |
The geographic shift was dramatic: Asia ex-China went from ¥1.1B to ¥9.0B (8x), while Americas doubled from ¥3.5B to ¥7.2B. Investment gravity moved east — to where the fabs are being built.
Key products:
- XTRAIA XD-3300: High-resolution microspot XRD for HBM, 3D DRAM, and 2nm+ logic node characterization. Entered mass production July 2025. The product that directly rides the AI/HBM wave.
- ONYX 3200: Film thickness, composition, and bump structure metrology for BEOL and advanced packaging. Launched December 2025 — first unit shipped to a "major global foundry." Detects silver content as low as 2% in SnAg solder bumps with precision of 4 parts per 100,000. Targets ¥1.5B in FY2026 sales, ¥3.0B in FY2027.
- MF-3400: Fourth-generation combined XRF/XRD for light element thin film measurement. Target market: $200M+, >70% share. GEN4 shipments started, major logic and memory customers adopting. Highest-TAM product in the portfolio.
- TXRF 3760 / V310 / XHEMIS TX-3000: TXRF contamination monitoring tools. KIOXIA is a confirmed customer.
Why customers choose Rigaku: No one else offers a full-stack inline semiconductor X-ray metrology platform (XRD + XRF + XRR + SAXS on integrated tools). 35+ years of fab tool history and qualification advantages. Switching costs are enormous (12-24 months of requalification). 95% repeat order rate. X-ray tools are <0.3% of total fab equipment spend, so price isn't a switching lever — reliability and workflow lock-in are.
The JEP (Joint Evaluation Program) model is the sales engine. JEPs are paid trials where Rigaku places tools at customer fabs before purchase. The customer runs the tool in their production environment, develops measurement recipes, and evaluates performance against their specific process requirements. This creates deep lock-in before the sale even happens — once engineers have invested months building recipes on a Rigaku platform, switching to a competitor means starting over. JEPs expanded from ~2 in 2024 to ~10 in 2025, and 5 of 14 FY2025 JEPs completed (implying conversion to purchase orders). The expanding JEP pipeline is the most concrete leading indicator that semiconductor revenue growth will continue.
Components & Services (27% of revenue, 26% of operating profit)
The recurring revenue base. Revenue was flat at ¥25.9B, with service revenue growing 8% on an expanding installed base. The drag came from EUV multilayer mirror weakness (-¥1B) and unfavorable US service margins from tariff impacts. Excluding EUV, the underlying business grew 4%. Operating margin compressed from 23.9% to 19.8%, but management expects the EUV headwind to lighten. The "razors and blades" model — once a tool is installed in a fab, the customer buys X-ray tubes, maintenance, and software upgrades for 10-15 years.
The Product Roadmap
This is the section that matters most for the forward thesis. Rigaku has seven products in various stages of development and ramp, each targeting a specific metrology gap created by the 3D semiconductor transition.
Products entering full production in FY2026:
The XTRAIA XD (high-resolution X-ray diffraction) measures thin films in high-k/metal gate structures for both logic and memory. Target market: $100M+, with >70% share ambition. Six customers certified, three JEPs (Joint Evaluation Programs — essentially paid trials at customer fabs) continuing. Most mature new product.
The XTRAIA MF-3400 is the fourth-generation combined XRF/XRD system for light element thin film measurement. Target market: $200M+, >70% share. GEN4 shipments started, major logic and memory customers adopting. Highest-TAM product in the portfolio.
The XTRAIA CD-3200T uses transmission small-angle X-ray scattering (T-SAXS) to measure high-aspect-ratio structures in DRAM and NAND memory. Target market: $100M+, >50% share. Adopted by major memory and WFE companies. A 3D DRAM variant (3300T) is in development, extending the platform.
Products entering customer evaluation in FY2026:
The ONYX 3200 is the one that excites me most. It measures light element (aluminum) dipole layers in GAA/CFET transistor structures — a measurement that conventional XPS (X-ray photoelectron spectroscopy) physically cannot do in a production environment. XPS requires ultra-high vacuum and only measures the surface; ONYX 3200 works at atmospheric pressure and measures through multilayer stacks. Genuinely new capability, not an incremental improvement. Shipped to two advanced logic customers in Q1 2026, expanding to memory. Target market: $100M+ (new market). If GAA adoption proceeds as planned, every leading-edge fab will need this measurement.
The AXI (Advanced X-ray Inspection) targets HBM/CoWoS/TSV packaging defect inspection — the hottest area in semiconductor manufacturing right now. Target market: $300M+, >30% share. Alpha machines in demo, beta demos start Q3 2026, full production targeted for 2027. Largest potential SAM in the portfolio.
In development for 2027: XTRAIA CD-3200G (next-gen logic below 1nm — GAA/CFET 3D structures) and a Hybrid Metrology platform combining X-ray and optical techniques with a co-development partner. Both targeting new markets with $100M+ SAM each.
Total estimated SAM across all products: ~$1 billion by 2030. Rigaku's target: 50%+ share. Combined with the existing base, management's goal is ¥100B in semiconductor segment revenue by 2030 — roughly 3.6x today's level.
The new product revenue share within the semi segment is projected to reach 41% in FY2026, up from 24% in FY2025 — a 2.2x increase. The AI semiconductor revenue ratio is expected to reach 71%, up from 60%. These aren't abstractions — they directly drive the 2-point gross margin improvement that underpins the FY2026 guidance.
The Technology Foundation
All of Rigaku's products exploit the interaction between X-ray photons and matter. The key techniques:
X-Ray Diffraction (XRD) — Rigaku's flagship. When monochromatic X-rays hit a crystalline material, the regularly spaced atoms act as a diffraction grating. At specific angles, scattered X-rays constructively interfere according to Bragg's Law (nλ = 2d sinθ), where d is the spacing between atomic planes. By scanning a detector across angles, XRD reveals crystal structure, phase identity, lattice strain, film thickness, and alloy composition. Think of identifying a building by its echo pattern — the spacing of floors and walls creates a unique acoustic signature, and XRD reads the atomic-scale equivalent.
X-Ray Fluorescence (XRF) — elemental analysis. An incident X-ray beam knocks inner-shell electrons out of atoms. When outer-shell electrons drop down to fill the vacancy, they emit secondary X-rays at energies unique to each element — like a fingerprint. XRF identifies which elements are present and in what concentrations. In fabs, XRF measures metal barrier/seed layer thickness and detects contamination. Total Reflection XRF (TXRF) achieves extreme surface sensitivity — detecting trace metals down to ~10⁸ atoms/cm² on cleaned wafers.
Small-Angle X-Ray Scattering (SAXS) — nanostructure profiling. SAXS measures X-ray scattering at very small angles (0.1-5°), capturing information about nanoscale periodic structures in the 1-100nm range. Critical Dimension SAXS (CD-SAXS) can non-destructively measure 3D profiles of FinFETs, GAA nanosheets, and 3D NAND channel holes — including sidewall angle, line width, and pattern depth — all from a single measurement. The frontier where Rigaku is investing heavily.
X-Ray Reflectometry (XRR) — thin-film characterization. By analyzing interference fringes in X-rays reflected at grazing angles, XRR determines film thickness (sub-nm to hundreds of nm), density, and interface roughness for both crystalline and amorphous materials. Critical for gate dielectric, barrier layer, and multilayer stack deposition monitoring.
X-Ray vs. Optical Metrology
| Dimension | Optical (KLA, Onto) | X-Ray (Rigaku, Bruker) |
|---|---|---|
| Throughput | Very high (seconds/measurement) | Historically slower; improving |
| Penetration | Surface/near-surface | Deep bulk & sub-surface |
| Strengths | Speed, cost, mature fab deployment | Non-destructive depth profiling, material-specific, model-independent 3D structural info |
| Limitations | Struggles with high-aspect-ratio 3D features | Lower throughput; X-ray source brightness constraints |
| Competitive dynamic | Dominant today | Gaining share as devices go 3D |
The crucial insight: X-ray and optical metrology are complementary, not competitive. Rigaku doesn't need to displace KLA — it needs more measurement steps to require X-ray techniques. Every new 3D NAND layer, every HBM stack, and every GAA transistor generation increases the "X-ray metrology intensity per wafer."
How a Rigaku Instrument Works — Step by Step
A Rigaku X-ray analytical instrument has four core subsystems:
[X-ray Source] → [Beam Conditioning Optics] → [Sample Stage] → [Detector + Analysis Software]
Step 1: X-ray Generation. The X-ray source is the heart of the instrument. Rigaku's key advantage is its proprietary rotating anode technology. In a rotating anode, a focused electron beam bombards a spinning metal target (typically copper, molybdenum, or chromium). The rotation spreads the heat load over a larger area, allowing much higher power (and therefore brighter X-rays) than a sealed tube. Rigaku's FR-X rotating anode produces up to 10x the flux of a microfocus sealed tube, with a focal spot as small as 70μm.
This is where Rigaku's moat lives. Building a reliable, high-brightness rotating anode requires extreme precision engineering — the target spins at thousands of RPM while being bombarded by a focused electron beam, all under high vacuum. Rigaku has been doing this for 75 years. Competitors mostly buy their X-ray sources or use lower-power sealed tubes.
Step 2: Beam Conditioning. Raw X-rays from the source are polychromatic and divergent. Multilayer mirrors and crystal monochromators select a specific wavelength and focus the beam. Rigaku makes its own optics (via Rigaku Innovative Technologies), allowing tight integration between source and optics.
Step 3: Sample Interaction. The conditioned beam hits the sample. Depending on the technique — XRD measures diffraction off crystal planes via a goniometer sweep; XRF/TXRF captures fluorescent spectrum from element-specific emissions; XRR analyzes interference fringes from thin film interfaces at grazing incidence.
Step 4: Detection and Analysis. Modern detectors (silicon drift, hybrid photon counting) count individual X-ray photons and measure energy. Software converts raw data into crystal structures, compositions, film thicknesses, and contamination maps. Rigaku is integrating ML algorithms for real-time defect analysis in semiconductor applications.
How Semiconductor Metrology Works in a Fab
Wafer Start → Epitaxy → [HRXRD: verify crystal quality, strain, composition]
→ Deposition → [XRR: verify film thickness, density, roughness]
→ Lithography/Etch → [CD-SAXS: verify critical dimensions, profile]
→ Metal fill → [XRF: verify barrier/seed thickness, composition]
→ CMP → [TXRF: verify surface contamination levels]
→ Packaging/Bumping → [XRF: verify SnAg solder composition]
→ HBM Stacking → [X-ray: verify TSV fill, bump integrity]
Rigaku's inline tools (XTRAIA family) are installed directly on the production floor, connected to the fab's MES. Each measurement feeds back into process control — if a film is too thin, the fab adjusts the deposition recipe in real time. Once a Rigaku tool is the "tool of record" for a process step, switching requires 12-24 months of requalification, recipe migration, and risk.
Key Technical Metrics
| Metric | Why It Matters | State of the Art |
|---|---|---|
| Throughput (wafers/hour) | More measurements per shift = better process control | Inline tools: 20-40 wph |
| Precision (repeatability) | Must detect sub-angstrom changes in film thickness | <0.1Å for XRR |
| Sensitivity | Lower detection limits = finding smaller defects | TXRF: ~10⁸ atoms/cm² |
| Measurement speed | Faster = more affordable inline deployment | Improving with brighter X-ray sources |
| Depth penetration | Deeper = characterize taller 3D structures | X-ray: >10μm (vs. optical ~1-2μm) |
The key bottleneck is X-ray source brightness. Brighter sources enable faster measurements, higher throughput, economically viable inline deployment at more process steps. Rigaku's vertical integration in X-ray source manufacturing is the competitive edge — they control the single most important component.
Competitive Position
Market Structure
The X-ray analytical instruments market is an oligopoly. Top 3 players control >60% of XRD. Semiconductor X-ray metrology is effectively a Rigaku-Bruker duopoly. Rigaku is the only publicly traded pure-play X-ray analytical instruments company.
| Company | Ticker | Revenue | XRD Share | Semi Overlap | Pure-Play? |
|---|---|---|---|---|---|
| Rigaku | 268A (TSE) | ¥94B ($630M) | ~29% (#1) | Dominant in inline fab tools | Yes — 100% X-ray |
| Bruker | BRKR (NASDAQ) | ~$3.4B | ~22% (#2) | Growing; direct competitor | No — diversified (NMR, mass spec) |
| Malvern Panalytical | SEPJF (Spectris) | ~$500M | ~10-12% | Acquired Freiberg for semi XRD | No — Spectris portfolio |
| Shimadzu | 7701 (TSE) | ~$3.8B | ~5-8% | Limited | No — chromatography-focused |
| Thermo Fisher | TMO (NYSE) | ~$42B | ~15% | Minimal inline | No — XRD is <1% of TMO |
Rigaku's R&D is 100% focused on X-ray technology. Competitors split attention across multiple product lines.
Competitive Moat
Strongest moats:
- Qualification lock-in (semiconductor): 12-24 month requalification cycle makes switching prohibitively expensive. Once "tool of record," they hold the position for the fab's life.
- Vertical integration: Rigaku manufactures its own X-ray sources, optics, and detectors. Most competitors buy from third parties. This is 75 years in the making and extremely difficult to replicate.
- Installed base: 7,000+ instruments across 70+ countries. Each unit generates service and consumable revenue for years. 95% repeat order rate.
- IP/patents: Extensive portfolio covering X-ray source design, optical configurations, measurement algorithms, and application software. Invented the world's first commercially available rotating anode X-ray generator (1952).
- No substitutes for X-ray in 3D inspection: Optical and electron-beam tools cannot penetrate sealed 3D structures. X-ray is physics-mandated, not just preferred.
- Application know-how: Hundreds of measurement recipes developed over decades for specific semiconductor process steps.
Absent moats: No meaningful network effects. Limited pricing power in general analytical instruments (competitive market); stronger pricing power in semiconductor metrology (less competition, higher switching costs).
Porter's Five Forces
| Force | Rating | Comment |
|---|---|---|
| Supplier power | Low | Vertically integrated; makes own critical components |
| Buyer power | Low-Moderate | Fabs have limited alternatives for inline X-ray metrology |
| Threat of substitutes | Low | Optical is complementary, not substitute; e-beam is partial competitor |
| Threat of new entrants | Low | 12-24 month qualification cycles, deep IP, 75-year head start |
| Competitive rivalry | Moderate | Oligopoly with rational pricing; Bruker is the main rival |
Industry Structure & Cycle Position
The analytical instruments industry is mature and stable (3-5% CAGR), but the semiconductor metrology sub-segment is in a secular growth phase driven by the 3D transition. This creates a bifurcated cycle dynamic:
- Analytical instruments (43% of revenue): Modestly cyclical, tied to academic budgets, pharma R&D, and industrial capex. Currently in a soft patch (China -23%).
- Semiconductor metrology (30% of revenue): Highly cyclical around semiconductor capex, but benefiting from a structural upgrade cycle (AI, HBM, advanced packaging). The last semiconductor capex downturn was 2023 — we are in the mid-to-late expansion phase.
- Services (27% of revenue): Counter-cyclical buffer — grows with installed base regardless of new tool orders.
Where are we in the cycle? Third year of recovery. AI capex is the primary driver. The risk is that AI capex moderates or semiconductor inventory builds — but the structural demand for X-ray metrology (more 3D structures = more X-ray measurement steps) provides a floor that didn't exist in previous cycles.
Leading indicators to watch:
- HBM production volumes (SK Hynix, Samsung, Micron)
- Advanced packaging capacity expansion (TSMC CoWoS, Intel EMIB)
- Rigaku's semiconductor order book (disclosed quarterly)
- New fab announcements and CHIPS Act funding disbursements
The disruption threat is low. X-ray inspection is physics-constrained — there is no Moore's Law equivalent that could allow a startup to leapfrog Rigaku with a software-only solution. The moat is in the hardware.
Emerging Threats
Bruker (BRKR): Most credible near-term threat. Expanding semiconductor metrology line with resources of a $3.4B company. But Bruker's semiconductor X-ray business is still much smaller, and qualification cycles protect Rigaku's installed base.
Malvern Panalytical (Spectris): Acquired Freiberg Instruments for semi XRD in 2024. Longer-term threat starting from a small base.
Nova Ltd. (NVMI): Israeli metrology company expanding from optical into X-ray. Worth watching for combined optical/X-ray platform.
Applied Materials (AMAT): E-beam metrology (PROVision platform) competes for some process control budget. Higher resolution but much lower throughput.
AI-enhanced optical metrology: ML advances could extend the useful range of optical methods (KLA, Onto), partially reducing X-ray need in some applications. Moderate long-term risk.
Chinese entrants: Could eventually compete in lower-end analytical instruments. X-ray source technology at the cutting edge remains a high barrier.
Value Chain Position
[X-ray Sources/Optics] → [Instrument Manufacturer ★] → [Semiconductor Fabs / Labs] → [Chips / Drugs / Materials]
↑ ↑
Rigaku makes these Rigaku does this
in-house (vertical (final assembly,
integration) software, recipes)
Rigaku sits at the instrument manufacturing layer and also controls the upstream component layer through vertical integration. This is the source of the moat.
Barriers to entry:
- 12-24 month fab qualification cycles
- Deep IP portfolio covering X-ray sources, optics, measurement algorithms
- 75 years of accumulated domain expertise
- Vertical integration in X-ray source manufacturing
- Installed base lock-in through recipes, data continuity, service contracts
End Markets & TAM
| Market | Size | CAGR | Source |
|---|---|---|---|
| Analytical X-ray instrumentation (all end markets) | ~$2.1B | ~4.8% | Market Research Future |
| Semiconductor X-ray metrology & inspection | ~$393M | ~6.3% | Verified Market Research |
| Broader semiconductor metrology & inspection | ~$5.5B | ~5.1% | Straits Research |
| Industrial X-ray | ~$3.4B | ~7.1% | Data Bridge |
SAM: ~$2.5B (analytical X-ray + semiconductor X-ray metrology). Current market share: ~20-25% of SAM, with #1 in XRD (~29%) and 50-70% in semiconductor TXRF.
Management's $1B SAM target by 2030 across all new semiconductor products. At 50%+ share, the semi segment alone would generate ¥50B+ in revenue at 30%+ margins.
End-Use Applications
| End Market | Applications | Growth Driver |
|---|---|---|
| Semiconductor manufacturing | Wafer-level metrology, thin-film measurement, contamination detection, bump/via inspection in advanced packaging | AI chip demand, HBM, GAA, 3D packaging |
| Materials science | Crystal structure analysis, phase identification, residual stress measurement | R&D, battery development, new materials |
| Life sciences / Pharma | Protein crystallography, small-molecule structure determination | Drug discovery, structural biology |
| Industrial quality control | Elemental analysis, coating thickness, cement/mining composition | Manufacturing quality, compliance |
| Power semiconductors | GaN/SiC substrate inspection | EV adoption, electrification |
Secular Tailwinds
| Tailwind | Mechanism | Magnitude | Duration |
|---|---|---|---|
| AI/HBM scaling | More HBM stacks = more X-ray metrology steps per wafer | High | 5-10yr+ |
| GAA transistors (2nm and below) | New architecture requires XRD for nanosheet characterization | High | 5-10yr+ |
| 3D NAND >200 layers | Optical can't see the bottom; X-ray can | High | 5-10yr+ |
| Advanced packaging (CoWoS, chiplets) | 2.5D/3D packaging needs bump/TSV metrology | Medium-High | 3-5yr+ |
| Fab buildout (CHIPS Act, Japan strategy) | New fabs need metrology tools | Medium-High | 3-5yr |
| Power semiconductor transition | GaN/SiC adoption drives inspection demand | Medium | 5-10yr+ |
| Battery materials R&D | XRD is essential for cathode/anode characterization | Low-Medium | 5-10yr+ |
These tailwinds are structural (5-10+ year horizon), not cyclical. Each technology node increases the "metrology intensity per wafer."
Business Model Economics
Revenue split: ~70% equipment sales, ~30% recurring services/components. The 30% recurring base provides stability — once a Rigaku tool is installed, the customer buys X-ray tubes, maintenance, and software upgrades for 10-15 years. This is the "razors and blades" model. The installed base is large enough that services should grow mid-single digits organically as Rigaku ships more semi equipment and creates service annuities.
Gross margins run ~56%. This reflects high-precision, low-volume products with significant engineering IP embedded. The semiconductor segment carries higher margins than general-purpose instruments (estimated 29-31% operating vs. ~15% for GP). The semiconductor segment also has a stronger services tail because fabs run 24/7 and cannot tolerate downtime — they pay premium rates for guaranteed uptime support.
Average selling prices (estimated):
- Major XRD/XRF systems: $100K-$500K
- Single-crystal systems: $300K-$800K
- Benchtop units: $30K-$80K
- Production-grade semiconductor metrology: $500K-$2M+
Customer concentration: Not disclosed at the individual level. Semiconductor revenue is distributed across multiple global fabs. General-purpose revenue is diversified across academic, industrial, and pharma customers in 136 countries. No single customer appears to represent >10% of total revenue, though semiconductor customer concentration is higher within that segment.
Industry structure: The XRD market is an oligopoly — top 3 players (Rigaku, Bruker, Thermo Fisher) hold >60% share. The inline semiconductor X-ray metrology market is effectively a Rigaku-Bruker duopoly. This is a structurally attractive market with high barriers to entry. Pricing is rational — competition is primarily on technology and application support, not price. X-ray tools are <0.3% of total fab equipment spend, making cost competition largely irrelevant.
Quarterly seasonality: The business has enormous operating leverage and is structurally Q4-loaded. Year-end capital equipment budgets at universities and semiconductor fabs drive a natural back-end concentration. Q4 FY2025 generated 37% of annual revenue and 62% of net income. This makes Q1-Q3 results misleading as standalone data points — do not panic-sell on soft interim quarters.
Pricing power: Strong in semiconductor metrology (mission-critical, no alternatives for TXRF, high switching costs). Moderate in general-purpose analytical instruments (competition from Bruker, Shimadzu exists, but differentiation on performance matters). The semiconductor segment's pricing power is a key reason operating margins are nearly double the GP segment.
Geographic Revenue Mix (FY2025)
| Region | Revenue (¥B) | % of Total |
|---|---|---|
| Japan | 27.4 | 29% |
| Asia ex-China | 23.0 | 24% |
| Americas | 16.8 | 18% |
| China | 14.2 | 15% |
| Europe | 11.4 | 12% |
Roughly 71% of revenue comes from outside Japan. Asia ex-China (Taiwan, Korea — TSMC, Samsung, SK Hynix territory) is the largest overseas region, reflecting semiconductor metrology demand.
Geographic dynamics by segment (FY2025 GP Instruments):
| Region | Revenue (¥B) | YoY Change |
|---|---|---|
| Japan | ¥11.5B | +3% |
| Americas | ¥7.7B | -13% (US academic budget cuts) |
| Europe | ¥8.1B | +27% (academic R&D in new materials) |
| China | ¥7.1B | -23% (supplementary budget reversal) |
| Asia (ex-China) | ¥5.7B | N/A |
China analytical revenue fell 23% — a geopolitical headwind that warrants monitoring but represents only ~8% of total revenue. The Americas academic weakness (-13%) reflects US government budget uncertainty.
Operations Footprint
| Facility | Location | Function |
|---|---|---|
| Tokyo Plant (HQ) | Akishima, Tokyo | Corporate HQ + manufacturing + R&D |
| Yamanashi Plant | Hokuto, Yamanashi | Primary manufacturing — XRD, thin-film, semiconductor tools. New 23,000m² building completed June 2025, doubled capacity. Another 50% increase planned by 2027. |
| Rigaku Americas Corp. | The Woodlands, TX | Americas HQ, sales, service |
| Applied Rigaku Technologies | Austin, TX | XRF instrument manufacturing |
| Rigaku Innovative Technologies | Auburn Hills, MI | X-ray optics manufacturing |
| Rigaku BioScience Lab | Cambridge, MA | Life science applications |
| Rigaku Semiconductor Instruments | Israel (fka XwinSys) | Semiconductor inspection tools |
| Rigaku Europe | Neu-Isenburg, Germany | European sales & service |
| Rigaku Czech | Prague | European operations |
| Rigaku Polska | Wroclaw, Poland | European operations |
| Rigaku Asia Pacific | Singapore | Asia-Pacific sales & service |
| Rigaku Technology Taiwan | Taiwan | Near TSMC/semiconductor ecosystem. Full ops starting Oct 2025. |
Rigaku is asset-moderate — it owns key manufacturing facilities but the real value is in engineering IP and the precision manufacturing know-how required to build X-ray instruments. The Yamanashi expansion is significant: it doubled semiconductor tool production capacity and is being expanded further (another 50% by 2027) to meet AI-driven demand.
Corporate Structure
Rigaku Holdings Corporation (268A.T — listed entity)
├── Rigaku Corporation (Japan — operating subsidiary)
│ ├── Rigaku Americas Holding Inc. (Delaware)
│ │ ├── Rigaku Americas Corporation (TX)
│ │ ├── Applied Rigaku Technologies (TX)
│ │ ├── Rigaku Innovative Technologies (MI)
│ │ ├── Rigaku Reagents (TX)
│ │ └── Rigaku Analytical Devices
│ ├── Rigaku Semiconductor Instruments Ltd. (Israel, fka XwinSys)
│ ├── Rigaku Europe SE (Germany)
│ ├── Rigaku Polska Sp. z o.o. (Poland)
│ ├── Rigaku Czech s.r.o. (Czech Republic)
│ ├── Rigaku Asia Pacific Pte. Ltd. (Singapore)
│ ├── Rigaku Technology Taiwan Co., Ltd.
│ └── Other subsidiaries
└── Atom Investment L.P. (Carlyle's vehicle — ~40.7% shareholder)
Clean holding company structure. No layered shell entities, no undercapitalized affiliates, no complex cross-holding webs.
Smart Factory & Automation
Rigaku is integrating ML algorithms into its TXRF and WDXRF tools for real-time defect analysis. The Yamanashi Plant is deploying AI-driven robotics for smart factory automation, targeting a 40% throughput improvement by 2026. This is not just a cost play — faster manufacturing cycles mean faster delivery to customers who are capacity-constrained on metrology tools during the current AI buildout. SmartLab SE's AI-powered sample handling achieves +55% efficiency improvement and +35% accuracy improvement over manual analysis, demonstrating Rigaku's ability to integrate software intelligence into hardware platforms.
Strategic Collaboration
JEOL Ltd. (TYO: 6951) — Joint development of XtaLAB Synergy-ED, a turnkey electron diffraction platform for nanocrystal molecular structure determination. JEOL is a leading electron microscope maker; the partnership combines Rigaku's crystallography software with JEOL's electron beam hardware. This collaboration signals that Rigaku can bridge the electron beam and X-ray worlds — potentially relevant for future hybrid metrology platforms.
Key Acquisitions
| Target | Location | Strategic Logic | Status |
|---|---|---|---|
| XwinSys Technology | Israel | Enter semiconductor metrology. Deal value not disclosed — unusual for a strategically important acquisition. | Now Rigaku Semiconductor Instruments — driving growth |
| MILabs B.V. | Netherlands | Advanced X-ray imaging/CT, life science imaging (PET, SPECT, CT, Optical Tomography) | Integrated |
| Osmic, Inc. | USA | X-ray optics (vertical integration of critical component) | Now Rigaku Innovative Technologies — core to the moat |
| Newton Scientific | USA | Detector technology | Integrated |
All were technology tuck-ins at modest prices. No "empire building," no transformative M&A at high multiples. Disciplined, strategy-driven capital allocation.
Financials
Key Metrics
| Metric | FY2022 | FY2023 | FY2024 | FY2025 | FY2026E (Guidance) |
|---|---|---|---|---|---|
| Revenue (¥B) | 62.7 | 79.9 | 90.7 | 94.2 | 101.0 |
| Revenue growth | — | +27.4% | +13.5% | +3.9% | +7.2% |
| Gross profit (¥B) | ~37.8 | ~46.7 | 55.3 | 52.9 | — |
| Gross margin | 60.3% | 58.4% | 61.0% | 56.1% | — |
| Operating profit (¥B) | 6.3 | 15.3 | 18.6 | 16.7 | 19.4 (adj: 21.5) |
| Operating margin | 10.1% | 19.1% | 20.6% | 17.7% | 19.2% (adj: 21.3%) |
| EBITDA (¥B) | — | — | — | 21.8 | 25.3 |
| EBITDA margin | — | — | 25.9% | 23.0% | 25.1% |
| Net income (¥B) | 0.9 | 10.9 | 13.6 | 11.4 | 12.5 (adj: 13.9) |
| Net margin | 1.5% | 13.7% | 15.0% | 12.1% | 12.4% |
| EPS (¥) | 4.06 | 48.44 | 60.44 | 50.19 | ~56 |
| DPS (¥) | — | — | 6.00 | 18.80 | 19.00 |
FY2022's 1.5% net margin was depressed by LBO-related costs. FY2023-24 showed normalized profitability with 60%+ gross margins and 19-21% operating margins.
FY2025 was a deliberate investment year. Gross margin fell 500bp from FY2024. The EBITDA margin bridge shows where the money went: manufacturing capacity expansion (-0.9pts), EUV weakness in parts & service (-0.8pts), semi product/regional mix shift (-0.6pts), higher R&D (-0.7pts), FX losses (-0.5pts). Semi revenue growth contributed +0.7pts, but it was overwhelmed by investment costs.
FY2026 guide implies the payoff: Gross margin expands 2.2pts from new product mix and pricing, semi revenue growth adds 0.6pts, partially offset by continued SGA growth (-0.9pts) and FX headwinds (-0.4pts). EBITDA margin recovers to 25.1%.
Quarterly Detail — This Matters
FY2024-FY2025 Quarterly Data (¥M)
| Quarter | Revenue | Op. Income | Net Income | Op. Margin |
|---|---|---|---|---|
| Q1 2024 | 19,537 | 3,161 | 2,185 | 16.2% |
| Q2 2024 | 23,242 | 5,512 | 4,323 | 23.7% |
| Q3 2024 | 19,801 | 3,177 | 2,482 | 16.0% |
| Q4 2024 | 28,072 | 6,777 | 4,625 | 24.1% |
| Q1 2025 | 20,614 | 2,836 | 1,918 | 13.8% |
| Q2 2025 | 20,142 | 2,883 | 1,861 | 14.3% |
| Q3 2025 | 18,740 | 843 | 494 | 4.5% |
| Q4 2025 | 34,697 | 10,147 | 7,128 | 29.2% |
Q4 FY2025 generated 37% of annual revenue and 62% of annual net income. This is extreme back-end loading — year-end budget flush for capital equipment combined with the Yamanashi capacity coming online. Enormous operating leverage: when revenue scales (Q4), the fixed cost base is absorbed and margins spike. When it doesn't (Q1-Q3), margins compress.
Q3 FY2025 was particularly ugly at 4.5% operating margin — the worst quarter. This created noise and likely scared some investors. But this is the characteristic pattern of a capital equipment company in an investment cycle.
Management says FY2026 will moderate from 84% Q3→Q4 swing to ~40%, but this will always be a Q4-loaded business. Don't panic-sell on soft interim quarters — evaluate on a full-year basis.
Incremental Margin Analysis (Last 8 Quarters)
| Q1 YoY | Q2 YoY | Q3 YoY | Q4 YoY | |
|---|---|---|---|---|
| Δ Revenue | +1,077 | -3,100 | -1,061 | +6,625 |
| Δ Op. Income | -325 | -2,629 | -2,334 | +3,370 |
| Incremental EBIT Margin | Neg | Neg | Neg | 50.9% |
Q1-Q3 showed negative operating leverage — front-loaded investment costs. All the value creation happened in Q4, where incremental EBIT margin hit 50.9%. This is the characteristic pattern of a capital equipment company in an investment cycle. The fixed costs are front-loaded; the revenue payoff is back-loaded and lumpy. When the semi metrology ramp delivers, normalized incremental EBIT margins should be 35-45%.
TTM Margin Progression
| Period | Revenue (¥B) | Gross Margin | Operating Margin |
|---|---|---|---|
| TTM ending Sep 2024 | 66.8 | 59.7% | 19.4% |
| TTM ending Dec 2024 | 90.7 | 61.0% | 20.3% |
| TTM ending Mar 2025 | 91.7 | 60.6% | 19.7% |
| TTM ending Jun 2025 | 88.6 | 58.9% | 17.4% |
| TTM ending Sep 2025 | 87.6 | 56.5% | 14.9% |
| TTM ending Dec 2025 | 94.2 | 56.1% | 17.7% |
The TTM data shows margins troughing in the Sep 2025 trailing period and recovering in Q4 as the Yamanashi capacity started delivering. FY2026 should confirm the inflection.
Cash Flow & Balance Sheet
| Metric | FY2024 | FY2025 | FY2026E |
|---|---|---|---|
| Operating cash flow (¥B) | 14.6 | 9.4 | — |
| Capex (¥B) | 5.9 | 5.8 | ~7.8 (7.7% of rev) |
| Free cash flow (¥B) | 8.7 | 3.6 | — |
| FCF margin | 9.6% | 3.8% | — |
| Cash on hand (¥B) | — | 24.3 | — |
| Total debt (¥B) | — | 62.2 | — |
| Net debt (¥B) | 33.6 | 37.9 | — |
| Net debt / EBITDA | ~1.4x | ~1.7x | — |
| ROIC | ~15% | 9.7% | — |
| ROE | — | 13.4% | 15.0% (guided) |
| Goodwill (¥B) | 51.7 | 51.9 | — |
| Goodwill / Equity | 63% | 59% | — |
FCF was thin at ¥3.6B due to working capital absorption (timing of receivables, inventory build for semiconductor ramp) and elevated capex. Once the investment phase passes (likely FY2027+), FCF should normalize significantly higher. Lake Cornelia estimates normalized FCF margins of 15-20% at scale.
The debt is Carlyle's fingerprint. ¥62.2B in total debt is legacy LBO financing from 2021. Since the IPO raised zero capital for the company, this debt was not paid down with IPO proceeds. The company is still servicing acquisition financing that funded Carlyle's buyout. Net debt/EBITDA of 1.7x is manageable and declining.
The balance sheet carries ¥52B of goodwill from Carlyle's LBO — 59% of shareholders' equity. Not a red flag per se (reflects the premium Carlyle paid for a quality business), but impairment risk exists if earnings deteriorate significantly.
Dividend exceeds free cash flow. FY2025 dividend payout: ~¥4.2B (221.79M shares x ¥19). Free cash flow: ¥3.6B. The company also did a ¥4B share buyback (4.28M shares, all cancelled January 2026). Combined shareholder returns of ¥8.3B on ¥3.6B FCF is obviously unsustainable without either FCF recovery or balance sheet deterioration. This means the company is funding shareholder returns partially from debt capacity or working capital.
The share buyback serves Carlyle's interest in supporting stock price during their exit window. It's shareholder-friendly but also strategically convenient for the controlling PE sponsor.
Adjusted ROE has deteriorated from 20.9% to 14.9% — partly reflecting the growing equity base, partly lower profitability.
R&D investment: ~8.5% of revenue target for FY2026 (~¥8.6B). Combined R&D + capex intensity of ~16% of revenue reflects a company in full investment mode. This should normalize as capacity comes online and starts generating returns.
Financial Health Indicators
| Metric | Value | Assessment |
|---|---|---|
| Altman Z-Score | 2.62 | "Gray zone" — moderate financial stress risk. Below 2.99 "safe" threshold. |
| Current ratio | 2.27 | Adequate — short-term liquidity is not a concern |
| Debt/Equity | 0.70 | Moderate — legacy LBO debt but manageable |
| Interest coverage | Adequate (not separately disclosed) | No stress |
| Net debt / EBITDA | 1.7x | Manageable — declining as EBITDA grows |
| FCF yield | 0.8% | Very low — suppressed by investment phase |
| Equity ratio | 47.7% | Improving |
The Altman Z-Score in the "gray zone" reflects the LBO debt burden more than operational weakness. As EBITDA grows and debt declines, this should improve toward the safe range. The current ratio is comfortable. No refinancing walls have been flagged.
ROIC vs. WACC
ROIC (TTM): 9.7%. Estimated WACC: ~7-8% (Japan risk-free rate ~1%, equity risk premium ~5-6%, beta ~1.2). Spread: +2-3%. Rigaku is creating value, but the spread is narrow. As the semiconductor segment scales (higher margins, lower incremental capital), ROIC should expand significantly. The Lake Cornelia analysis projects ROIC above 15% by FY2028. The mid-teens ROIC estimated for FY2024 was compressed by the investment year.
The narrow ROIC spread is the strongest financial argument for waiting for a cheaper entry. At 9.7% ROIC vs. 7-8% WACC, you're paying a premium for a company creating only modest incremental value per dollar of invested capital. The bet is that this spread widens materially as the semiconductor mix increases — higher margins on semiconductor tools combined with lower incremental capital (Yamanashi is already built) should drive ROIC toward 15%+ within 2-3 years.
Industry landscape
A pure-play X-ray metrology bet on the AI / advanced-packaging semiconductor cycle. Industry-wide detail (X-ray vs optical metrology, the 3D-transition demand drivers, the merchant TAM) is treated at first principles in the primer — see xray-semi-metrology-primer. Sector hubs: semi-equipment · metrology · japan-semi · cross-cutting ai-infrastructure.
Merchant landscape in one line: the inline semiconductor X-ray metrology market is effectively a Rigaku–Bruker duopoly (Rigaku #1 in XRD at ~29% share); X-ray and optical metrology are complementary, not substitutes — the thesis is more measurement steps requiring X-ray as devices go 3D, not displacing KLA/Onto. SAM across all new semi products ~$1B by 2030, management targeting 50%+ share. (Full competitive set, Porter's Five Forces, moat, and TAM tables live under Business → Competitive Position.)
Management
Leadership Team
Jun Kawakami — President & CEO (since January 31, 2023)
Education: University of Tokyo (BA Economics, 1987); Kellogg School of Management, Northwestern (MBA, 1994)
Career path: Booz Allen Hamilton (strategy consulting) → GE Aviation → GE Yokogawa Medical → GE Healthcare Japan (President & CEO, 2011-2017) → Arteria Networks Corporation (President & CEO, led to IPO December 2018) → Carlyle Japan (Senior Advisor, April 2020) → Rigaku (Board March 2021, CEO January 2023)
How he got the role: Carlyle brought him in. He joined the board in March 2021 (coinciding with Carlyle's acquisition) and was appointed CEO in January 2023, replacing transitional CEO Toshiyuki Ikeda who had held the role for only 18 months. Kawakami was a Carlyle advisor before becoming CEO — he was not an industry hire discovered organically. He was a Carlyle network appointment. His path: GE Japan → Carlyle Senior Advisor → Carlyle portfolio company board member → Carlyle portfolio company CEO.
Prior track record:
- GE Healthcare Japan: Led a large subsidiary of a Fortune 50 company. Medical imaging (CT, MRI) shares engineering DNA with X-ray analytical instruments. However, specific revenue or profitability figures under his tenure are not publicly available — GE Healthcare Japan is a private subsidiary, so the claim that he "drove growth" is unverifiable.
- Arteria Networks IPO (Dec 2018): IPO'd at ¥1,250/share, raising ¥25.2B (~$220M). Stock opened below offering (¥1,190) but has appreciated over time to ~¥1,978. Completed in challenging market conditions (Q4 2018 global selloff). ARTERIA Networks (TSE: 4423) continues to trade and has performed reasonably well. This is his most credible verifiable accomplishment.
- Rigaku IPO (Oct 2024): Priced at top of range (¥1,260), raised ¥129B (~$860M). First-day performance was poor (-10% to ¥1,130), but stock has since tripled from April 2025 lows.
Assessment: Kawakami is a credible, well-pedigreed professional manager. He is not a visionary or founder-operator — he's a PE-installed CEO whose job is execution. That is the right archetype for this stage: the company has the technology and products; it needs someone to scale the semiconductor business, manage costs, and navigate the Carlyle exit. His GE and Arteria track records support this.
BUT: His background is in healthcare IT and broadband telecom, not analytical instruments or semiconductor metrology. He is fundamentally a Carlyle-placed operator, not a domain expert. No lawsuits, regulatory actions, or bankruptcies found.
Ownership: 0.032% (~¥86M). Minimal. No evidence of personal share purchases with his own money. For a company this size trading at premium multiples, investors should want to see the CEO backing the thesis with personal capital.
Kiyoshi Ogata — Senior Executive Vice President
Education: Ph.D. Physics, University of Tokyo (1986). Previously at Hitachi before joining Rigaku in 2009.
Ogata is the technical anchor and the most domain-credible member of the leadership team. With a physics PhD from Tokyo and 17 years at Rigaku, he understands the X-ray science at the deepest level. He oversees both the X-ray Instruments Division and Semiconductor Metrology Division — essentially all of Rigaku's technology and products. His long tenure (vs. Kawakami's 3 years) means the technical organization has stability even as the CEO changed. This is a valuable counterweight to the Carlyle-installed CEO — Ogata ensures that commercial strategy does not override engineering quality.
Kent Heath — CEO, Rigaku Americas
Background: 30+ years in semiconductor and analytical instruments. Former VP & GM at Bruker Corporation (Tribology, Stylus and Optical Metrology business unit). 10 years at Motorola Semiconductor.
Hiring a senior executive directly from Bruker (Rigaku's closest competitor) is a strong signal. It means Rigaku is serious about competing aggressively in semiconductor metrology in the Americas, and they wanted someone who knows the Bruker playbook inside and out. Heath's Motorola Semiconductor background also brings customer-side perspective.
Hikaru (Akira) Shimura — Chairman
Son of founder Dr. Yoshihiro Shimura. Led Rigaku as President & CEO from 1971 to 2021 — 50 years. Co-invested with Carlyle in the 2021 take-private (~20% stake). Transitioned to Chairman. Holds ~11-15% personally.
Shimura Name Discrepancy: Major shareholder filings use "Akira Shimura" (志村 晶), not "Hikaru." The kanji 晶 can be read either way. Almost certainly the same person. He classified his holding as "pure investment" in filings — a notable disclosure for a Chairman who has been with the company 54 years. He has been selling since the IPO — from 14.16% down to ~12%.
His ~¥67-85B position is likely the vast majority of his net worth. This is the strongest alignment signal among insiders, even though the "pure investment" classification suggests he views it as a financial exit, not a permanent holding.
Akihiko Miki — CFO
OPAQUE. Despite extensive searching in both English and Japanese, almost no biographical information is publicly available. Not even a LinkedIn profile. For a ~¥400B market cap company, this level of CFO opacity is concerning. Cannot verify qualifications or track record.
Other Named Executive Officers
The following appear in public filings though detailed backgrounds are limited: Adam Chong, Kozaburo Fujimoto, Kent Heath, Michael Hippler, Shunji Hiratsuka, Markus Kuhn, Jeff Li, Kazuhiko Omote, Tetsuya Ozawa, Shinji Tomita, Chihiro Yamada, Kenji Yoshioka.
Average Management Tenure
~2.9-3.1 years. Reflects the Carlyle-driven leadership refresh that began in 2021.
Leadership Transition Timeline
| Date | Event |
|---|---|
| Jan 2021 | Carlyle acquires ~80% of Rigaku; Shimura retains ~20% |
| Mar 2021 | Kawakami joins the board |
| Jun 2021 | Toshiyuki Ikeda succeeds Shimura as CEO |
| Jan 2023 | Kawakami replaces Ikeda after only 18 months |
| Oct 2024 | IPO on TSE Prime Market |
Three CEOs in two years (Shimura → Ikeda → Kawakami). Standard PE playbook: transitional CEO to bridge from founder to professional management, then install the long-term operator. The rapid turnover could be a yellow flag in other contexts, but here it reflects planned PE-driven succession.
Insider Ownership & Skin in the Game
| Shareholder | Shares | % Ownership | Notes |
|---|---|---|---|
| Atom Investment L.P. (Carlyle) | 95,140,800 | 41.48% | Down from 75.5% pre-IPO |
| Hikaru/Akira Shimura | 27,476,600 | 11.98% | Classified "pure investment." Has been selling. |
| Japan Trustee Services Bank (Trust) | 17,954,200 | 7.83% | Institutional custody |
| Japan Custody Bank (Trust) | 5,229,700 | 2.28% | Institutional custody |
| Indus Capital Partners | 4,995,600 | 2.18% | Japan-focused hedge fund; crossed 5% in Jan 2025 |
| Goldman Sachs International | 4,160,350 | 1.81% | |
| Employee Stock Ownership Plan | 1,974,249 | 0.86% | |
| Jun Kawakami | ~71,000 | 0.032% | ~¥86M. Likely granted, not purchased. |
Combined insider/founder/sponsor ownership: ~55-60%. Deep alignment on paper. But the CEO has no verified personal share purchases, and the Chairman classifies his holding as "pure investment."
Stock option program: 37,126 option grants outstanding as of December 2024, representing 5,470,154 dilutive shares (~2.5% dilution). At current price (~¥1,880), the full option pool has a face value of approximately ¥10.3B. The options cover directors, auditors, executive officers, and subsidiary executives. This is the primary vehicle for management "ownership" — upside participation without personal capital at risk. Classic PE-backed structure that prioritizes alignment with the sponsor over genuine skin-in-the-game.
Executive ownership is option-based, not purchased. There is no evidence that Kawakami or any other named executive has bought shares with personal money on the open market. For a company this size trading at premium multiples, investors should want to see the CEO backing the thesis with personal capital.
Compensation transparency is poor by US standards. Specific compensation figures for CEO Kawakami, CFO Miki, or other executives were not found in English-language sources. Japanese-listed companies only disclose individual compensation above ¥100M/year. No information was found about what metrics drive incentive compensation (this is a critical gap), golden parachute or change-of-control provisions, or severance terms.
What we can infer: Carlyle typically structures CEO compensation with significant equity incentives tied to share price. Kawakami almost certainly has stock options or restricted shares that vest over time, aligning him with Carlyle's exit valuation goal. The Arteria Networks precedent suggests his model is "manage the IPO, grow the value, participate in the upside." No unusual perks (personal jet, family on payroll, related-party leases) have been disclosed or reported.
Holdings Concentration
| Name | Role | Rigaku Holdings | Other Known Holdings | Where Is the Majority? |
|---|---|---|---|---|
| Carlyle (CJP IV) | Sponsor | ~¥173B ($1.15B) | Diversified PE portfolio | One of many investments, but largest holding in CJP IV |
| Hikaru Shimura | Chairman | ~¥67-85B ($450-570M) | Not publicly known | Likely the vast majority of his wealth |
| Jun Kawakami | CEO | Undisclosed | Previously held Arteria Networks equity | Unknown — Carlyle advisory suggests diversified interests |
| Takaomi Tomioka | Director (Carlyle) | Indirect via fund | MD at Carlyle; boards of Orion Beer, Solasto, Sunsho | Diversified across Carlyle portfolio |
Shimura's wealth concentration in Rigaku is the strongest alignment signal. Carlyle is aligned but will exit. Kawakami's personal stake is unknown — this is the main gap.
Institutional ownership is remarkably low (<5% ex-Carlyle/Indus). This reflects the stock's youth, limited analyst coverage, and Carlyle overhang deterring institutions.
Board & Governance
| Name | Title | Independent? | Background |
|---|---|---|---|
| Jun Kawakami | President & CEO | No | See above |
| Kiyoshi Ogata | Senior EVP | No | See above |
| Takaomi Tomioka | External Director | NO — Carlyle | Co-Head of Carlyle Japan. 30yr PE. Led Carlyle's acquisition of Rigaku. Also on boards of Orion Beer, Solasto. He is literally running Carlyle Japan. Over 19 years at Carlyle, previously at GE Capital (1998-2003). |
| Mitsuo Hirose | External Director | Forensic DD says NO (Carlyle Operating Executive); later DD says Yes (former J&J Japan President) | Listed as Operating Executive at Carlyle Japan LLC + multiple Carlyle portfolio company boards (Orion Breweries, Sunsho Pharmaceutical, Business Breakthrough). Also President & Representative Director at Maverick Transnational Inc. Should not be classified as independent despite "External" title. |
| Andrea Knoblich | External Director | Yes | Former COO APAC, BNY Mellon. Prior: Shinsei Bank, Deutsche Bank, Morgan Stanley, Mitsubishi Trust. American Univ. BA, LSE Diploma. Genuinely independent — no direct Carlyle connection found. |
| Tomoaki Taguchi | External Director | Yes | Limited public background — independence status cannot be verified. |
| Takako Ebata | External Director | Yes | Former Member, Japan House of Representatives (2009-2012). Former Head of Govt Affairs, J&J Japan. Yokohama National Univ., MIT Sloan MS. Policy and government affairs expertise. |
| Ryota Isogai | Auditor | — | |
| Yutaka Kamisawa | Auditor | — | |
| Tomoyoshi Matsuo | Auditor | — |
Key governance findings:
-
At least 2 of 5 "External Directors" are Carlyle affiliates (Tomioka and Hirose). Tomioka is literally the Co-Head of Carlyle Japan — there is no ambiguity about his allegiance. Hirose is listed as an Operating Executive at Carlyle Japan LLC and sits on boards of multiple Carlyle portfolio companies. Calling either "independent" is misleading. This means Carlyle controls the CEO appointment and has 2 board seats, giving it dominant influence over governance even without a formal board majority.
-
True independence is weaker than numbers suggest. Only Knoblich has a clearly independent profile. Taguchi's background is insufficiently disclosed. Even in the best case, truly independent directors may constitute only 3 of 7 total board members.
-
Tomioka is the key person to track. As long as he's on the board, Carlyle is actively managing its exit. His departure = Carlyle is done or nearly done selling.
-
No anti-takeover provisions. No dual-class shares, no poison pill, no staggered board. With Carlyle at 41% and Shimura at 12%, the combined ~53% insider block makes hostile takeover impossible anyway. When Carlyle exits, the company becomes fully contestable.
-
IPO was 100% secondary. Zero new shares issued. All ~¥129.1B went to selling shareholders. The company received no capital from its own IPO. This is the defining feature of a PE exit transaction dressed as an IPO.
-
Related-party transactions between Rigaku and Carlyle are NOT disclosed in English-language sources. PE sponsors commonly charge portfolio companies advisory fees. The absence of disclosure is a yellow flag.
-
Conflict of Interest — Tomioka: Having the Co-Head of Carlyle Japan on the board while Carlyle holds 41.5% and is actively managing its exit creates a structural conflict. Every board decision about capital allocation, dividends, share buybacks, and M&A is influenced by someone whose primary obligation is to generate returns for Carlyle LPs, not minority public shareholders.
-
Indus Capital Partners crossed the 5% threshold in January 2025 and holds ~5.4% through multiple funds. Whether they will push for governance changes remains to be seen.
Management DD Verdict Summary
Two management DDs were run with different scopes, reaching different conclusions. Both are valid.
Forensic DD (March 3, 2026) — Overall Grade: Yellow-Red
| Dimension | Rating |
|---|---|
| Skin in the Game | Red — CEO has no verified personal purchases; Chairman classifies as "pure investment" and is selling |
| Capital Allocation | Red — IPO was 100% secondary (company got $0); LBO debt remains; dividend exceeds FCF |
| Governance | Red — 2 of 5 "External" directors are Carlyle affiliates; CEO was placed by controlling shareholder |
| Compensation | Yellow — largely opaque |
| Shell/Cross-Holdings | Yellow — standard PE fund vehicle, but related-party fees undisclosed |
| Litigation | Green — clean across all executives |
Key forensic findings: IPO was purely a Carlyle exit mechanism. Board stacked with Carlyle affiliates. CEO is a PE-placed operator. LBO debt sits on the public company's balance sheet. No evidence CEO has purchased shares with personal capital.
Standard DD (March 17, 2026) — Overall Grade: B+
| Dimension | Rating |
|---|---|
| Skin in the Game | Green — Carlyle (41%) + Shimura (~18%) = ~59% combined. Deep alignment. |
| Capital Allocation | Green — Disciplined M&A, demand-driven capex, no dilution. Grade: B+ |
| Governance | Green — 57% independent board, no anti-takeover provisions |
| Shell/Cross-Holdings | Green — Clean structure. No red flags. |
| Compensation | Yellow — insufficient disclosure |
| Litigation | Green — clean |
Key standard findings: Professional, competent, Carlyle-aligned. Hiring from Bruker (Heath) signals competitive intent. Kawakami's IPO track record is credible. Clean corporate structure — no shell entities, no asset shuffling, no revenue circularity.
Reconciliation
The forensic DD was harsher because it focused on what we can't verify — compensation, related-party transactions, CEO's personal purchases. The standard DD focused on what we can see — combined ownership alignment, clean structure, disciplined M&A. Both are valid perspectives. The right synthesis: trust the business, be skeptical of the governance, demand a margin of safety.
Items That Could Not Be Verified
This list matters because it defines the information risk premium investors should demand:
- Specific executive compensation (total pay, incentive metrics, golden parachutes)
- Whether Carlyle charged management/advisory fees to Rigaku (and whether these terminated at IPO)
- CFO Akihiko Miki's background and qualifications
- Independence status of director Tomoaki Taguchi
- Jun Kawakami's personal shareholding and whether he has purchased any shares with his own money
- Specific terms of the XwinSys acquisition (deal value was not disclosed — unusual for a strategically important acquisition)
- Any consulting or IP licensing agreements between Carlyle entities and Rigaku
- Specific goodwill/intangible assets breakdown from the Carlyle LBO on the balance sheet
- What metrics drive incentive compensation (revenue vs. ROIC vs. share price)
- Change-of-control provisions and golden parachute terms
Japanese corporate governance operates under the Companies Act and TSE listing rules. Disclosure is less granular than SEC requirements, particularly around individual executive compensation (only disclosed above ¥100M/year), insider transactions (reported through TSE timely disclosure, not real-time Form 4), and related-party transactions (disclosed in annual securities reports / Yuho, not in a separate proxy filing). This is a structural limitation for non-Japanese investors and should be factored into position sizing. It is not, by itself, a red flag — it applies to all TSE-listed companies.
Management Bottom Line
This is a Carlyle-controlled exit vehicle with a genuine underlying business. The X-ray technology platform and semiconductor metrology growth story are real — Ogata provides the technical credibility, and the ONYX 3200 launch validates execution capability. However, the governance structure is designed to serve Carlyle's exit, not minority shareholders. The IPO was purely a sell-down. The board includes Carlyle affiliates. The CEO is a PE-placed operator. The LBO debt sits on the public company's balance sheet. And the 41.5% overhang will create persistent selling pressure over the next 1-3 years.
Would I trust these people with my capital? Conditionally. The business is good and the technical team is credible. But you are not investing alongside management — you are investing behind a PE sponsor that is actively exiting. Size the position for the overhang risk. The investment thesis needs to be strong enough to absorb 95 million shares of Carlyle selling into the market over the coming years. If the semiconductor metrology growth materializes, the business can absorb it. If it doesn't, there is no natural buyer support at these prices.
Capital Allocation Track Record
| Area | Grade | Detail |
|---|---|---|
| M&A | B+ | Technology tuck-ins (XwinSys, MILabs, Osmic, Newton Scientific). Disciplined, strategy-driven. No "empire building." |
| Capex | A- | Yamanashi expansion tied directly to customer demand. Not speculative. FY2026 capex at 7.7% of revenue (~¥7.8B). |
| Buybacks | N/A | ¥4B buyback in FY2025 (4.28M shares cancelled Jan 2026). Constrained by LBO debt. Serves Carlyle's interest in supporting stock price during exit window. |
| Dividends | B | ¥19/share FY2026E, ~1% yield, ~25-30% payout. Conservative and growing. |
| Equity Issuance | A | IPO was secondary (no dilution to shareholders). Share count stable post-IPO at ~222M. |
| Debt Management | C+ | ¥62B from LBO. Declining as EBITDA grows, but limits flexibility. |
| R&D | A- | ~8.5% of revenue target for FY2026 (~¥8.6B). Heavy for an instruments company, reflects the semiconductor bet. |
| Overall | B | Disciplined and investment-focused, but constrained by LBO debt and PE exit dynamics. |
Catalysts & Risks
Catalysts
Near-term (0-12 months):
- 1H FY2026 results (August 2026): First datapoint showing whether FY2026 guidance is on track. If semi revenue accelerates above +21% guide, the stock re-rates.
- ONYX 3200 customer wins: Each new foundry/memory customer builds conviction. First shipment to "major global foundry" already confirmed. Shipped to two advanced logic customers in Q1 2026.
- Analyst coverage expansion: Only 4-6 analysts cover. Additional initiations increase institutional awareness and liquidity.
- Carlyle secondary offering clarity: Removes uncertainty, increases free float.
Medium-term (1-3 years):
- Semiconductor segment becoming >40-50% of revenue — the tipping point for re-rating from "analytical instruments" to "semiconductor metrology growth company."
- ONYX 3200 reaching ¥3.0B annual sales (FY2027 target) — adds ~3% to total revenue from a single product.
- AXI packaging inspection tool entering production (2027) — captures the $300M+ SAM opportunity.
- Production capacity +50% by 2027 — enables growth without further capex step-up.
- Carlyle full exit — removes overhang, triggers potential index inclusion. Increases free float for institutional investors.
Risks
| Risk | Likelihood | Impact | Mitigant | Can It Be Closed? |
|---|---|---|---|---|
| Carlyle 41% overhang | Near-certain | Medium | Orderly exit via block trades over 2-4 years. Fund needs to liquidate by 2030-2033. | Yes — resolves as Carlyle exits (expected 2026-2028) |
| Semiconductor cycle peaks | Medium | High | Services/components provide ballast. X-ray adoption is earlier in curve vs. optical when semi last cycled down. | No — structural industry risk. But X-ray penetration story should grow through cycles. |
| Margin recovery fails | Medium | High | FY2026 1H results (August 2026) are the checkpoint. If adj. EBITDA margin not tracking toward 25%, reassess. | Yes — closes with FY2026 earnings delivery |
| China revenue decline | Medium | Medium | China is ~15% of revenue. Manageable through geographic diversification. Europe grew +27% in FY2025, offsetting. | No — geopolitical risk is structural. But exposure is small enough to absorb. |
| Competition from Bruker | Low-Medium | Medium | 35+ years of fab qualification history + vertical integration. Hiring from Bruker (Kent Heath) shows intent. | No — ongoing competitive dynamic, but Rigaku's installed base and moat are strong |
| ONYX 3200 / AXI underperform | Medium | High | If these don't gain traction, the $1B SAM target by 2030 shrinks significantly. | Partially — early customer data should clarify within 12 months |
| Valuation / stock ahead of fundamentals | Medium-High | Medium | Trading 28-33% above analyst consensus. Resolves with earnings growth or stock correction. | Yes — resolves with earnings growth catching up |
| LBO debt burden | Low-Medium | Low | 1.7x net debt/EBITDA, positive FCF, deleveraging. No refinancing wall. | Yes — reaches <1x by FY2027-2028 |
| FX headwinds | Medium | Medium | ~71% of revenue from outside Japan. Yen strengthening compresses reported results. | No — structural for all Japanese exporters |
| Q4 revenue concentration | Medium | Medium | 62% of net income in Q4. If shipment delays occur, annual results blow up. | Partially — management guides moderation to ~40% swing in FY2026 |
| Low float / volatility | Medium | Medium | Free float is only 38% (~84M shares). Stock can be volatile on small volume. | Yes — improves as Carlyle exits and institutional investors accumulate |
Bear case scenario: Semi capex slows in late 2026-27. Margins don't recover because analytical weakness offsets semiconductor growth. Carlyle dumps a large block at a discount. Stock de-rates from 36x to 20x. EPS falls to ¥40-50, stock trades at ¥800-1,000. That's a 50%+ drawdown.
Thesis invalidation triggers:
- Semiconductor segment growth turns negative for two consecutive quarters
- Operating margin fails to recover above 19% by FY2027
- Carlyle announces secondary sale of >20% of outstanding in a single year
- A competitor achieves a breakthrough in X-ray source technology that narrows Rigaku's moat
- ONYX 3200 fails to gain customer qualifications at leading-edge fabs
- A major fab defects from Rigaku to Bruker or a new entrant
What the thesis invalidation would look like in practice: The worst outcome is a scenario where semiconductor capex slows in late 2026 or 2027 (typical cycle length suggests a pullback), margins don't fully recover because analytical instrument weakness offsets semiconductor growth, Carlyle dumps a large block at a discount creating a negative signaling spiral, and the stock de-rates from 36x to 20x earnings. In that scenario, EPS might fall to ¥40-45, and at 20x, the stock would trade at ¥800-900 — a 55-60% drawdown from current levels. This is the scenario you need to be able to stomach before sizing the position.
Technology Roadmap & R&D
R&D spending is guided at ~8.5% of revenue for FY2026 (~¥8.6B). This is heavy for an instruments company but appropriate given the semiconductor opportunity. Key focus areas:
-
Next-gen X-ray sources: Brighter sources = faster measurements = higher throughput inline tools. This is the single most important technology bottleneck. Whoever builds the brightest, most reliable X-ray source wins the semiconductor metrology market.
-
CD-SAXS commercialization: Critical dimension measurement for GAA and 3D NAND is an emerging technique that could open a significant new market. The XTRAIA CD-3200T and CD-3200G target this space.
-
AI-powered software integration: Beyond the SmartLab SE's AI sample handling, Rigaku is developing ML algorithms for real-time defect analysis in semiconductor tools. The goal is fully automated metrology workflows where the tool identifies, classifies, and escalates defects without human intervention.
-
Hybrid metrology platform: Combining X-ray and optical techniques in a single platform with a co-development partner. This could be a game-changer if it offers the speed of optical with the depth of X-ray. Target: $100M+ new market by 2027+.
-
Power semiconductor inspection: GaN/SiC substrates for EVs require specialized XRD characterization. This is a smaller market but growing fast with the EV transition.
-
Manufacturing capacity: Yamanashi Phase 2 targets another 50% capacity increase by 2027, on top of the doubling already completed. This takes semiconductor tool production capacity to roughly 3x 2022 levels.
Combined R&D + capex intensity of ~16% of revenue in FY2026 reflects a company in full investment mode. This should normalize as capacity comes online, but for now it depresses near-term FCF and margins.
Technical Picture (as of March 2026)
| Indicator | Value | Signal |
|---|---|---|
| Price position | 91.5% of 52-week range | Near highs — extended |
| 52-week return | +63% | Strong momentum |
| Distance from ATH | ~6% below ¥2,076 | Limited upside to resistance |
| RSI (14-day) | 60.66 | Neutral — not overbought, room to run |
| Price vs. 200-day MA | +98% above | Strongly bullish |
| 6-month performance vs. Nikkei 225 | +75% outperformance | Very strong relative strength |
| Nearest support | ¥1,500-1,600 (prior consolidation) | -17% to -22% below |
| Nearest resistance | ¥2,076 (52-week/ATH) | +6% above |
The stock made a massive move from ¥641 (April 2025) to ¥2,076 (February 2026) — 224% in 10 months. Currently consolidating slightly below ATH. Could be forming a bull flag for the next leg up, or a distribution top after a parabolic run. The next decisive move will be determined by 1H FY2026 results.
Technical verdict: Trend is strongly bullish, but the easy technical money has been made. RSI at 60 is neutral — not warning of reversal but not screaming "buy." Best entry on a pullback to ¥1,500-1,700 (support zone) or on a confirmed breakout above ¥2,076 with volume. Buying at ¥1,915 means buying in the middle of the range between support and resistance — not optimal.
The Carlyle Exit — Critical Question
This is the most important near-term risk factor for the stock. It deserves detailed treatment.
Carlyle's Fund Structure
Atom Investment L.P. is registered in the Cayman Islands — a standard PE fund vehicle, not a red-flag shell structure. It is the investment vehicle of Carlyle Japan Partners IV (CJP IV), a 2020-vintage fund that raised JPY 258 billion (~$2.3 billion). CJP IV is now fully invested, and Carlyle has since raised Carlyle Japan Partners V (CJP V) at JPY 430 billion. Other CJP IV portfolio companies include Totoku, Uzabase, Iwasaki Electric, Seiko PMC, among 12 total investments.
Fund recycling dynamics create pressure to exit. CJP IV is fully invested, CJP V is raising. Returning capital to CJP IV LPs requires liquidating positions. This isn't a choice — it's a structural obligation.
Carlyle's Hold Period Pattern
| Company | Acquired | IPO'd | Exit Strategy | Hold Period |
|---|---|---|---|---|
| Orion Breweries | 2019 | Sept 2025 | Sold entire stake at IPO | ~6 years |
| Rigaku | 2021 | Oct 2024 | Sold 35% at IPO, retained 41.5% | ~3.5 years to IPO |
| Solasto | — | Listed | Carlyle remains involved post-listing | Ongoing |
Carlyle Japan has completed 18 exits total, 7 via IPO.
For Orion Breweries, Carlyle sold its entire stake at IPO. For Rigaku, it sold only ~35% at IPO and retained 41.5%. This suggests either (a) Rigaku was more strategically important to hold, or (b) the market could not absorb a larger sale at IPO pricing. Given that the stock debuted below IPO price (-4.4% on day 1, then fell to ¥641 by April 2025), option (b) seems more plausible.
IPO Details
| Detail | Value |
|---|---|
| IPO Date | October 25, 2024 |
| Offer Price | ¥1,260/share (top of ¥1,230-1,260 range) |
| Shares Offered | ~89.1M shares (secondary sale by Atom Investment LP and Shimura) |
| Gross Proceeds | ~¥129.1B (~$860M) |
| Market Cap at IPO | ~¥284.1B (~$1.89B) |
| Lead Underwriters | Bank of America, JPMorgan, Morgan Stanley, Nomura |
| Day 1 Performance | Fell ~10% to ¥1,130 (poor debut) |
| New Shares Issued | ZERO. 100% secondary. Company received $0. |
This was the second-largest IPO in Japan in 2024. And the company got nothing from it.
Lockup & Selling Timeline
- 180-day lockup expiry: ~April 22, 2025
- Stock hit all-time low (¥641) on April 9, 2025 — just before lockup expired, reflecting market fear of impending selling
- Current holding: 95,140,800 shares (41.48%)
- At ¥1,953, this position is worth ~¥179B ($1.2 billion)
- Selling this position will take multiple transactions over many quarters given daily volume constraints
Exit Scenario Analysis
Most likely Carlyle exit path:
- Block trades to institutional investors (common in Japan)
- Secondary public offerings (like the IPO itself)
- Gradual market selling over 2-4 years
- Strategic sale of entire remaining stake to a single buyer (least likely given size)
Timeline: CJP IV is a 2020 vintage fund. Typical PE fund life is 10 years with extensions. This means CJP IV needs to be substantially liquidated by 2030-2033. Given the ~¥179B position, Carlyle needs to start meaningful selling within the next 1-3 years. Every quarter that passes increases the urgency.
Carlyle's Total Economics
At current price (~¥1,953), Carlyle's estimated total proceeds:
- IPO proceeds: ~¥129B (sold ~35% of position)
- Remaining stake value: ~¥179B
- Total implied: ¥308B on a ~¥100B investment
- ~3x gross MOIC if fully exited at current prices
This is a very successful PE investment. Carlyle has strong incentive to manage the exit orderly — dumping at a discount would destroy returns on the remaining position.
Valuation / DCF
Current (as of mid-March 2026)
| Metric | Value |
|---|---|
| Market cap | ¥425B (~$2.8B) |
| Enterprise value | ¥463B |
| P/E (TTM) | 38.7x |
| Forward P/E (FY2026E) | 32.3x |
| EV/EBITDA | 21.2x |
| EV/EBITDA (NTM) | ~17.8x |
| EV/Revenue | 4.9x |
| P/FCF (TTM) | ~119x (depressed FCF year) |
| FCF yield | 0.8% |
| P/B | 4.8x |
| PEG | 1.58 |
| Dividend yield | ~1.0% |
| 52-week range | ¥641 – ¥2,076 |
Segment Operating Profit Detail (FY2025)
Understanding the segment-level economics is critical because the stock's future depends on the semiconductor segment's growing share of profit.
| Segment | Revenue (¥B) | Op. Profit (¥B) | Op. Margin | % of Total OP | FY2026E OP Margin |
|---|---|---|---|---|---|
| General Purpose Analytical | 40.4 | 6.2 | 15.4% | 32% | ~16% |
| Semiconductor Process Control | 27.8 | 8.1 | 29.0% | 42% | 31.4% |
| Components & Services | 25.9 | 5.1 | 19.8% | 26% | ~21% |
| Total | 94.2 | 16.7 (adj) | 17.7% | 100% | 19.2% |
The semiconductor segment earns nearly 2x the operating margin of GP instruments and contributes 42% of profit on only 30% of revenue. As this segment grows to 33%+ of revenue in FY2026, overall margins expand mechanically — before any operating leverage kicks in. This is the simple math behind the FY2026 margin recovery story. By FY2027, if the semi segment reaches 40% of revenue (which the growth trajectory implies), it could contribute >50% of operating profit. At that point, you have a very different company than what the current multiple appears to be pricing.
Multiples vs. Peers
| Company | EV/EBITDA (NTM) | P/E (NTM) | EV/Rev | Growth (NTM) |
|---|---|---|---|---|
| Rigaku | ~17.2x | ~30x | 4.3x | +7.2% |
| Bruker | ~18x | ~28x | 4.5x | +5-7% |
| KLA | ~18x | ~24x | 10x | +10-12% |
| Onto Innovation | ~22x | ~28x | 8x | +8-10% |
Rigaku trades roughly in line with Bruker on EV/EBITDA. KLA and Onto — more directly comparable as semiconductor metrology companies — trade at similar or higher multiples with faster growth. Rigaku needs to prove itself as a "semiconductor metrology growth story" to earn a re-rating toward the KLA/Onto range.
Implied Expectations
At ~30x forward P/E, the market is pricing in:
- Successful margin recovery (operating margin back to ~19%)
- Continued semi metrology growth at 15-20% annually
- No significant Carlyle selling pressure
- Revenue CAGR of ~10%
Not unreasonable, but limited margin of safety. Stock trades 28-33% above consensus analyst target (~¥1,462-1,500). Either analysts need to catch up or the stock is ahead of fundamentals.
Analyst Coverage & Sentiment
| Metric | Value |
|---|---|
| Analysts covering | 4-6 (very thin) |
| Consensus | Buy (3 Buy, 1 Hold, 0 Sell) |
| Average price target | ¥1,462-1,500 |
| Target range | ¥950 (low) — ¥2,000 (high) |
| Stock vs. consensus | +28-33% above average target |
Coverage is thin — a clear information asymmetry opportunity for investors willing to do primary research. Even the highest analyst target (¥2,000) offers only ~5% upside at ~¥1,915.
Institutional ownership is remarkably low (<5% ex-Carlyle/Indus) for a ¥425B+ market cap Prime Market company. As Carlyle exits and coverage builds, institutional penetration should increase — a potential positive catalyst.
Forward Earnings Model
| FY2025 (actual) | FY2026E | FY2027E | FY2028E | |
|---|---|---|---|---|
| Revenue (¥B) | 94.2 | 101.0 | ~115 | ~130 |
| Revenue growth | +3.9% | +7.2% | +14% | +13% |
| Semi metrology growth | +19% | ~20% | ~25% | ~20% |
| Operating margin | 17.7% | 19.2% | ~21% | ~22% |
| Net income (¥B) | 11.4 | 12.5 | ~16 | ~19 |
| EPS (¥) | 50.19 | ~56 | ~72 | ~86 |
Target Price Framework
| Scenario | FY2027E EPS | Multiple | Target Price | Return from ¥1,915 |
|---|---|---|---|---|
| Bear | ¥60 | 20x | ¥1,200 | -37% |
| Base | ¥72-100 | 25-35x | ¥2,160-2,520 | +13% to +32% |
| Bull | ¥100-150 | 30-35x | ¥3,000-4,500 | +57% to +135% |
Simple DCF sanity check:
- If Rigaku earns ¥100 EPS in FY2027 (base case) and trades at 25x, that is ¥2,500 — 31% upside.
- If Rigaku earns ¥60 EPS in FY2027 (bear case) and trades at 20x, that is ¥1,200 — 37% downside.
- Asymmetry: +31% upside vs. -37% downside. Not great. The risk/reward improves significantly on any pullback to ¥1,500-1,700.
Decision Log
2026-05-02 — Pre-sell review (held position): TRIM IF OVERSIZED, otherwise HOLD into the 5/13 print
Trigger. The 2026-04-25 weekly Japan-semi screen (KB/brief/weekly-semi-review.md) flagged 268A as a held Stage-2 chart masking a −45% 0Q EPS revision into the 2026-05-13 print and recommended "review stop." This forced a discipline review.
Updated facts (2026-05-02, yfinance):
- Price ¥2,667 — vs ¥1,915 at the deep-dive (2026-04-05 wiki state) = +39% in ~30 days.
- 52-week range ¥696–¥3,100 — new 52-wk high ¥3,100 (vs the deep-dive's ¥2,076 ATH); now −14% below the new ATH.
- Earnings date 2026-05-13 (11 days out at time of review). (The canonical Key Monitoring Dates table lists "May 14, 2026" for Q1 FY2026 — a 1-day source discrepancy; both retained.)
- Q1 revenue consensus ¥18.5B (range ¥18.0–18.8B).
- EPS-revision trajectory: the −45% 0Q cut is real but stale. 0Q EPS ¥10.48 (90d ago) → ¥5.78 (60/30/7d) → ¥6.29 current = −40% over 90d, stabilized 60d ago, +8.8% in the last 7d. FY (0Y) EPS ¥63.59 (90d) → ¥61.74 current (−2.9%, +4% off the bottom). +1Q EPS ¥12.89 → ¥12.82 (+15% in last 7d). Every period turned up in the last 7 days — the shape of an estimate-cut bottoming, not a fresh deterioration. The screen flag was lagging.
Why it's a TRIM call, not a SELL. None of the wiki-defined invalidation triggers had fired: semi-segment growth still +19% YoY (FY2025), op margin recovery too early to judge (FY26 guide 19.2%), no Carlyle >20% secondary announced, no competitor X-ray-source breakthrough, ONYX 3200 on track (first "major global foundry" + 2 advanced-logic customers shipped), no major fab defection. The thesis is intact. What changed is the price, which moved the asymmetry against the holder:
| Metric | At deep-dive (¥1,915) | At review (¥2,667) |
|---|---|---|
| vs base-case FV (¥2,160–2,520) | −11% to +13% | +6% to +24% (above midpoint) |
| vs bull-case FV (¥3,000+) | −36% | −11% (within striking distance) |
| vs bear-case FV (¥1,200–1,400) | +37% to +60% | +91% to +122% |
| vs analyst consensus (~¥1,500) | +28–33% above | +78% above |
Risk/reward at ¥2,667 ≈ +30% upside (to bull) vs −55% downside (to bear) — materially worse than at ¥1,915. The anchor question "would you buy it today at ¥2,667?" answers itself: no (the deep-dive's own framework said best entry ¥1,500–1,700 or a confirmed breakout >¥2,076 with volume).
Behavioral read. The most active trap is regret avoidance disguised as risk management — "lock in the 39% gain before giving it back." That alone does not justify selling the thesis; it justifies re-checking size and the sell triggers. Anchoring to ¥3,000+ as a "deserved" target and recency bias on the stale −45% headline are secondary risks, both countered above.
Execution (sized to position weight — Pink to confirm the %):
| If position is | Action | Reason |
|---|---|---|
| ≤3% of portfolio | HOLD into print | Right-sized for the asymmetry; thesis intact; no fresh sell trigger |
| 4–6% | Trim 25–33% | Lock part of the run; cut single-name event risk into 5/13; keep the core |
| ≥7% | Trim to 5% | Rebalance discipline regardless of thesis |
- Hard stop: close below ¥2,000 (−25% from review; below prior 52-wk-high-turned-support; deeper fallback aligns with the ¥1,500–1,700 support zone).
- Soft stop / re-evaluate: close below ¥2,300 (post-breakout retest area).
- Timing: any trim should happen before the 5/13 print, not after — the print is the binary event; trimming after is information-disadvantaged "selling on news."
- Friction notes: ex-div 2026-06-29 (not relevant pre-print). TSE Prime liquidity fine; ~38% float should absorb a trim. Holding-period / tax handling flagged for Pink (short- vs long-term depends on entry date and broker).
Open question for Pink: current 268A position size as % of portfolio — dictates which branch above applies.
(Folded 2026-06-02 from 268a-sell-checklist.md, dated 2026-05-02. Topics it carried: ai-infrastructure, optical-components, japan-semi, green-finance.)
2026-04-07 — Discord datapoint: Onto co-development + Samsung 2nm/SiPho rumor
From Collyer Bridge in market chatter Discord (Apr 1, 2026):
- "2nm or sipho ramp would see Samsung place major orders for rigaku kit" — confirms unreported channel of demand for Rigaku tools tied to Samsung's leading-edge ramp. Adds another customer leg beyond TSMC.
- "There might be some co-development with Onto. Referred to in the earnings slides... Hybrid engine combining x-ray and optical, co-development with optical partners." — Rigaku's earnings slides apparently flagged a co-development with ONTO on a hybrid x-ray/optical metrology engine. Bob in the same thread thinks it's specific to 3D DRAM, not 3D NAND.
- Counterpointresearch.com Hybrid Bonding article was the source link.
Why it matters: A Rigaku-Onto hybrid x-ray/optical tool would be an entirely new product class, exactly the kind of "lab-to-fab" expansion the bull case requires. If real, it adds optionality to the semi process control segment and could materially expand the addressable opportunity beyond the standalone Rigaku tool roadmap. Watch the next Rigaku earnings slides for explicit confirmation — the Discord discussion suggests it's already disclosed but not yet picked up by analysts.
Checklist Scores
Two separate checklists were run — one at ¥1,953 (March 9) and one at ¥1,915 (March 17). They reached different conclusions.
First Checklist (March 9, 2026 — Price: ¥1,953): PASS
| Question | Answer |
|---|---|
| Can I state the thesis clearly? | Yes |
| Do I understand the business? | Yes |
| Are the financials healthy? | Mixed — great gross margin but LBO debt, declining profits, thin FCF |
| Is the valuation reasonable? | No — 30x+ P/E, 33% above analyst targets, 40% above DCF |
| Am I falling into a behavioral trap? | Yes — FOMO and narrative seduction |
| Do the technicals support buying now? | No — near ATH, limited upside |
| Have I sized appropriately? | N/A — not buying |
Verdict: Great business at a terrible price. Wait for ¥1,200-1,500.
Key reasoning: The stock nearly tripled while earnings fell — classic momentum/narrative divergence. At ¥1,953, analyst consensus of ¥1,462 implies 25% downside to fair value. Alpha Spread DCF suggests ~40% above fair value. Would not buy at 10-15% higher — absolutely not. At ¥2,200+, paying 35x+ forward for a mid-single-digit grower with declining profits and an LBO-encumbered balance sheet.
Second Checklist (March 17, 2026 — Price: ¥1,915): BUY (Scale In)
| Question | Answer |
|---|---|
| Can I state the thesis clearly? | Yes — physics-mandated X-ray inspection for 3D chips |
| Do I understand the business? | Yes — XRD, XRF, TXRF from first principles |
| Are the financials healthy? | Yes, with caveats — positive FCF, manageable debt, margins must recover |
| Is the valuation reasonable? | Borderline — 39x TTM, justified if semi thesis holds |
| Am I falling into a behavioral trap? | Partially — narrative seduction and recency bias are moderate risks |
| Do the technicals support buying now? | Mixed — strong trend, neutral RSI, but near ATH |
| Have I sized appropriately? | Yes — 2-3% of portfolio, scaled in over tranches |
| Do I have a clear exit plan? | Yes — base target ¥2,500, thesis-break triggers defined |
Verdict: Scale in with discipline. Tranche 1: 0.5-1% at ~¥1,915. Tranche 2: 1% on pullback to ¥1,500-1,700. Tranche 3: 0.5-1% on breakout above ¥2,076 or strong 1H FY2026 results.
Key reasoning: Rigaku sits at the intersection of growth compounder and special situation. The market is pricing it as a slow-growth scientific instruments company, but the semiconductor segment (30% of revenue, >50% of profits) is growing 20%+ annually. The "unlock" is the market recognizing the semi segment IS the company. Physics-based moat (no substitutes for X-ray in 3D inspection). 57% combined insider ownership. The main risks (valuation, Carlyle overhang, execution) argue for scaling in rather than going full-size.
Reconciling the Two Checklists
The first was written with a pure valuation lens — is this cheap? No. The second incorporated the secular thesis more fully — is this a good risk/reward for a growth compounder? Borderline yes, if you scale in. Both agree: don't chase. The optimal entry is on a pullback or Carlyle block trade.
Behavioral Traps Audit
This stock has several active traps that require conscious management:
FOMO — The stock nearly tripled from its 52-week low of ¥641. It's natural to feel left behind. The run-up is backed by fundamental improvement (Q4 blowout, FY2026 guidance), but the magnitude of the move means much of the easy money has been made.
Narrative seduction — "The only company that can solve the 3D chip inspection bottleneck" is a compelling narrative. The science supports it (X-rays do penetrate where optical/e-beam cannot), but the revenue numbers need to catch up to the story. Semiconductor is still only 30% of revenue. The narrative is ahead of the numbers, but not wildly so.
Confirmation bias — The Dong/Shim write-up and Lake Cornelia expert analysis are very bullish. It's tempting to anchor on the most optimistic scenario. The bear case (margin compression persists, semiconductor cycle peaks) is real and should be given equal weight.
Recency bias — Q4 FY2025 was a blowout (¥34.7B revenue, 62% of annual net income). It's tempting to annualize Q4 and project massive growth. But the business is heavily back-end loaded, and Q4 may have included one-time catch-up deliveries from the Yamanashi ramp.
Authority bias — Carlyle's involvement adds credibility, but PE firms are not infallible. Carlyle's objective is to maximize exit price, not to be right about the long-term thesis.
Assessment: Multiple traps are partially active. The thesis is fundamentally sound, but the stock has run hard and the most bullish scenarios may be front-running reality. Disciplined entry (scale in, don't chase) is the appropriate response.
Dong/Shim External Thesis Tracking
Source: "Lab to Fab: Catching Tokyo's Hidden Gem" by Jianshu Dong & Miru Shim, Feb 13, 2026. Written at ¥1,160, $1.7B market cap.
Their targets:
| Scenario | 1-Year | 5-Year | Return |
|---|---|---|---|
| Base | ¥2,700-3,700 | ¥5,500-9,100 | 2-3x / 5-8x |
| Upside | — | ¥13,200+ | 13x+ |
| Downside | ¥900-1,300 | — | 0.9-1.1x |
Dong uses 30x PE for base case, 35x for upside. His EPS estimates: ¥91 in 2027, ¥187 by 2031 (base); ¥101/¥261 (upside).
Key Dong/Shim claims to track:
- Semi fab is 60% of operating profits (2026E) — the killer insight. Reclassifies GaN/SiC tools and ~1/3 of service revenue as semi-related, boosting true semi mix well above reported ~30% of revenue.
- X-ray adoption S-curve is early innings — only ~5% of wafers actually inspected. Tool intensity per fab: 3-4 (FinFET) → 5-8 (GAA) → 8-14 (CFET).
- ~40% market share in semi X-ray (Yole + Dong estimates). 95% repeat order rate. X-ray is <0.3% of total fab equipment spend.
- New product cycle in 2H 2026-1H 2027 is the inflection. JEPs expanded from ~2 in 2024 to ~10 in 2025, transitioning to mass production orders.
- Management quote (Q4 2025 earnings, Feb 13 2026): "AI-related applications will account for roughly 70% of the Semiconductor business in 2026... estimated total SAM is approximately $1.0 billion, and our target is to capture over 50% share."
- Revenue target: ¥100B by 2030. Dong's base case has ¥132B by 2028 and ¥195B by 2031 — well above management guidance, driven by 21% CAGR in semi fab.
Where Dong adds value we didn't have:
- The semi profit reclassification is genuinely differentiated. Reframes Rigaku as a semi equipment play at a lab instrument multiple.
- The S-curve framing is more rigorous — maps tool intensity per fab across architecture transitions with specific tool counts per node.
- Direct management quotes from Q3/Q4 2025 earnings calls that we didn't incorporate.
Where we add value Dong doesn't cover:
- Carlyle overhang — Dong doesn't mention it at all. Glaring omission for a stock with thin institutional ownership. Our forensic DD covers PE exit timeline, Tomioka as key person, block trade risk.
- Margin compression detail — FY2025 net margins dropped from 19% to 12%. Real concern that Dong acknowledges but treats as transient.
- Analyst targets well below market — suggests either massive analyst lag or market ahead of fundamentals.
Where I'd push back on Dong:
- 13x+ upside case assumes 60x PE — aggressive for any Japanese company. ASML trades at ~35x. The base case at 30x is more defensible.
- No discussion of competition (Bruker, potentially KLA entering X-ray). Dong treats Rigaku as a near-monopoly, but the moat is real but not unassailable.
- Revenue forecast is nearly 2x management guidance by 2031. Either management is wildly conservative or Dong is wildly optimistic.
- The thesis was written at ¥1,160. At that price, the asymmetry was fantastic — downside was 0.9-1.1x while upside was 2-8x+. At higher prices, a lot of the easy money has been made.
How Our Analysis Compared to Dong/Shim
| Dimension | Our Deep-Dive (at ¥1,953) | Dong/Shim (at ¥1,160) |
|---|---|---|
| Fair value | ¥2,160-2,520 (30-35x FY2027E) | ¥2,700-3,700 1Y base |
| Bull case | ¥3,000+ | ¥9,100+ (5Y base) |
| FY2027 EPS | ~¥72 (implied) | ¥91 (base) / ¥101 (upside) |
| Semi mix framing | ~1/3 of revenue | 60%+ of operating profits (reclassified) |
| Carlyle overhang | Flagged as #1 risk | Not discussed |
| Margin concern | FY2025 compression flagged | Acknowledged; expects recovery |
| Conviction | Medium-High | Very High |
Bottom line on Dong/Shim: The single best English-language deep-dive on Rigaku. If you're going to own this stock, read this report. The thesis is strongest at lower prices. At ¥1,160 this was a screaming buy with 3-8x upside. At ¥1,953 it's a good hold with maybe 50-100% upside over 2-3 years if the semi ramp plays out. The Carlyle overhang (which Dong ignores) is the near-term risk that could give you a better entry.
Source: Jianshu Dong & Miru Shim, "Lab to Fab: Catching Tokyo's Hidden Gem — Rigaku Investment Thesis," Feb 13, 2026.
Position Sizing & Entry Plan
Conviction level: Medium-High. Physics-based secular thesis. Dominant pure-play. But valuation is not cheap and Carlyle overhang is real.
Target position size: 2-3% of portfolio. Scale to 4-5% on conviction-building events.
| Parameter | Value |
|---|---|
| Entry strategy | Scale in over 2-3 tranches |
| Tranche 1 | 0.5-1% at ~¥1,900 |
| Tranche 2 | 1% on pullback to ¥1,500-1,700 (or Carlyle block trade) |
| Tranche 3 | 0.5-1% on breakout >¥2,076 or strong 1H FY2026 |
| Maximum loss | 15-20% from average entry |
| Base target | ¥2,500 |
| Bull target | ¥3,000-4,500 |
| Stop/reassess | Below ¥1,200 |
Exit criteria:
- Sell if FY2026 revenue misses below ¥95B
- Sell if semiconductor demand weakens materially (negative growth for 2 quarters)
- Trim at ¥2,500+ (base case target)
- Full harvest at ¥3,000+
Expected holding period: 12-24 months for base case. 3-5 years for bull case.
Portfolio correlation: Rigaku is a Japanese semiconductor equipment play. If you own other Japanese semi equipment names (Tokyo Electron, Disco, Advantest, Lasertec), there is meaningful correlation. The X-ray niche is differentiated, but all Japanese semi equipment stocks trade together on macro and sector sentiment.
Key-person risk: Low. Management is professional (not founder-dependent). Kawakami is replaceable. Ogata provides technical continuity. No government incentives tied to specific individuals.
Dilution risk: Low. Share count stable at ~222M since IPO. No ATM programs, shelf registrations, or convertible instruments. The primary dilution risk is the 5.47M dilutive shares from the stock option program (~2.5%).
Information risk: Moderate. As a recently IPO'd Japanese mid-cap, English-language disclosure is improving but still limited vs. US-listed peers. Only 4-6 analysts. This creates both risk (less visibility) and opportunity (information edge for those doing the work). Japanese-listed companies offer less transparency than US-listed equivalents — no DEF 14A proxy statements, no Form 4 real-time insider trading disclosure, no CD&A compensation breakdowns. Equivalent disclosures exist in the yukashoken hokokusho (annual securities report) filed with EDINET, but they require Japanese language access.
What To Do Next
The right approach to Rigaku is patient and systematic:
-
Don't chase. The stock has tripled from its lows. The easy money has been made. A great business at a rich price is still a mediocre investment.
-
Wait for catalysts. Specifically: (a) a Carlyle block trade that depresses the stock 10-20%, creating a forced-selling entry; (b) 1H FY2026 results that confirm the margin recovery and semi acceleration, justifying a higher price; or (c) a general market correction that takes the stock to ¥1,500-1,700 without fundamental deterioration.
-
Size for the overhang. Carlyle has 95 million shares to sell. That is a fact. Every block trade will create a temporary headwind. The investment needs to be strong enough to absorb this. At 2-3% of portfolio, the position can weather 20% drawdowns without breaking the portfolio.
-
Use quarterly noise as opportunity. Q1-Q3 will look ugly because of the Q4-loaded earnings pattern. Soft interim quarters are buying opportunities, not sell signals — as long as the full-year trajectory is intact.
-
Track the JEP pipeline. The number and conversion rate of Joint Evaluation Programs is the single best leading indicator of future semiconductor revenue. If JEPs continue expanding from 10 in 2025, the revenue pipeline is building.
-
Watch Tomioka. As long as Carlyle's Co-Head of Japan is on the board, Carlyle is actively managing the exit. When Tomioka steps down, Carlyle is done or nearly done selling. That's when the overhang lifts.
Key Monitoring Dates
| Date | Event | What to Watch |
|---|---|---|
| May 14, 2026 | Q1 FY2026 results | Semi revenue trajectory; margin recovery |
| August 2026 | 1H FY2026 results | THE critical checkpoint. If semi revenue >+21% and EBITDA margin tracking 25%, thesis is working. |
| Any time | Carlyle block trade | Creates short-term pressure but buying opportunity if fundamentals intact |
| Feb 2027 | FY2026 full-year results | New product ramp validation. ONYX 3200 sales vs. ¥1.5B target. |
Sources
Fragments folded into this canonical page (consolidated 2026-06-02; originals archived to _migration-archive/2026-06-02/268A/): 268a-sell-checklist.md (2026-05-02 pre-sell review → folded into Thesis/Snapshot/Decision log).
Filings detail (sanctioned second file, kept separate): 268a-filings — FY2025 earnings, Q3/Q4 segment breakdowns, FY2026 guidance, full product roadmap. Source PDF: src-268a-rigaku-thesis-jianshu-dong-2026-02.pdf (Dong/Shim external thesis, Feb 2026).
Company & IR
- Rigaku Holdings IR
- Rigaku Holdings About
- Rigaku Holdings Group Structure
- Rigaku Semiconductor Metrology
- FY2025 Results Presentation
- Medium-Term Business Plan (Feb 2025)
- Integrated Report 2025
Carlyle / IPO
- Carlyle Announces Rigaku Listing
- Carlyle to Acquire Majority Stake (2021)
- Carlyle Raises JPY 258B for CJP IV
- Carlyle Raises JPY 430B for CJP V
- Carlyle Announces Orion Breweries Listing
- Rigaku IPO — Davis Polk
- Nikkei — Carlyle Invests $970M
Product Launches
- ONYX 3200 Launch (Dec 2025)
- XTRAIA XD-3300 Mass Production (Jul 2025)
- NEX QC II Launch at Pittcon (Mar 2026)
- Yamanashi Expansion (Jun 2025)
- XwinSys Renamed to Rigaku Semiconductor Instruments
Leadership
- Jun Kawakami CEO Announcement
- Carlyle Names Kawakami Senior Advisor
- Kent Heath CEO Americas
- Takaomi Tomioka — Carlyle
- Arteria Networks IPO
- Hikaru Shimura — MarketScreener
- Mitsuo Hirose — MarketScreener
- Takako Ebata — Wikipedia
Market Data & Analysis
- StockAnalysis — 268A Statistics
- StockAnalysis — RGAKF
- Simply Wall St — Rigaku
- TipRanks — FY2025 Results
- Lake Cornelia — Expert Write-Up
- Alpha Spread — Income Statement
- Kabutan — 268A Major Shareholders
External Thesis
- Jianshu Dong & Miru Shim, "Lab to Fab: Catching Tokyo's Hidden Gem — Rigaku Investment Thesis," Feb 13, 2026 (source PDF: src-268a-rigaku-thesis-jianshu-dong-2026-02.pdf)
Industry Research
- Market Research Future — Analytical X-Ray Market
- Semiconductor Engineering — 3D Metrology
- Bragg's Law — Chemistry LibreTexts
- NIST CD-SAXS program documentation
Filings
See 268a-filings for detailed FY2025 earnings, Q3/Q4 segment breakdowns, FY2026 guidance, and the full product roadmap analysis.
Related Research
- xray-semi-metrology-primer — first-principles primer on X-ray metrology in semiconductor manufacturing
- ai-infrastructure — cross-cutting topic hub on AI infrastructure buildout
- JP — Japan semi supply chain buy checklist
- 6855-jem-deep-dive — Japan Electronic Materials (probe cards) — same AI/HBM cycle
- 7741-deep-dive — HOYA (EUV blanks) — the other TSE semi equipment compounder
Topics
- japan-semi
- semi-equipment
- metrology
- carlyle-pe
- x-ray-technology
Source updates (auto-maintained)
Drop/3. Backend (Mar 11, 26) - Rigaku Investment Thesis [Feb 13, 2026]
At ¥1,160, the author forecasts semi fab profits reaching 60% of total in 2026 and 80%+ by 2031, with base-case EPS of ¥90 in 2027 implying a ¥2,700 PT (2-3x) at 30x, and ¥5,700 by 2031 (5x); bear case floors at ¥900–1,200 on ¥60–70 EPS at 15-17x.
Relevant to your thesis: Corroborates the wiki's core mispricing argument — reported ~30% semi revenue understates true exposure once GaN/SiC and service revenue are reclassified — and sets higher PT targets than the wiki's ¥2,160–2,520 fair value range.
Source: dropfile://3. Backend/OSAT-Test/Rigaku/Rigaku Investment Thesis [Feb 13, 2026].pdf