JAPAN AI/SEMICONDUCTOR SUPPLY CHAIN: COMPLETE DEEP DUE DILIGENCE

Date: 2026-04-18

Methodology: Dual-approach (Direct + T-Glass backward trace), 13-layer value chain map, StockAnalysis.com verified financials

TABLE OF CONTENTS

1. Macro Context

2. Value Chain Map -- Japan's Position at Every Layer

3. Full Comparison Table

4. Per-Stock Detailed Analysis

5. Conviction Tiers

6. Japan vs Global Competitive Matrix

7. Sources

1. MACRO CONTEXT

The AI Capex Supercycle

The Big Five hyperscalers (Amazon, Microsoft, Google, Meta, Oracle) will spend over $600 billion on infrastructure in 2026 -- a 36% increase from 2025. Goldman Sachs projects total hyperscaler capex from 2025-2027 at $1.15 trillion, more than double the $477B spent from 2022-2024. Roughly 75% (~$450B) is targeted at AI infrastructure.

Nvidia captures ~90% of AI accelerator spending. Demand for B200/GB200 accelerators has lead times exceeding 30 weeks, with entire production runs consumed by hyperscaler allocations.

Japan's Strategic Position

Japan occupies a unique and critical position in the global semiconductor supply chain -- not as a leading-edge chip manufacturer, but as the dominant supplier of materials, equipment, chemicals, and niche components without which no advanced chip can be made. Key facts:

• Materials: Japanese firms control ~90% of silicon wafers (Shin-Etsu + SUMCO), ~91% of photoresists (TOK + JSR + Shin-Etsu + Fujifilm), and 100% of T-Glass (Nittobo)
• Equipment: Japanese firms hold ~90% of coater/developer market (TEL + SCREEN), ~58% of test equipment (Advantest), ~75% of dicing/grinding (Disco), and 100% of EUV mask inspection (Lasertec)
• Packaging materials: Ajinomoto controls >95% of ABF film for CPU/GPU packaging
• Government support: Japan's 2026 stimulus of JPY 1.23 trillion targets AI/semiconductors; Rapidus has received ~$16.3B in cumulative government support for 2nm chip development

TSMC in Japan (JASM)

• Fab 1: Operating since late 2024 at 28nm/16nm, ~55,000 wafers/month capacity. 60% local supplier content.
• Fab 2: Under construction, initially 6nm, now reportedly pivoting to 2nm for AI chip customers (Nvidia, AMD). Capex could exceed $25B.
• Fab 3: Under consideration for post-2030, embedding TSMC deeper into the Japan-Taiwan-US semiconductor triangle.
• Kyushu's IC production value hit JPY 1 trillion in 2024 for the first time in 16 years.

Japan Semiconductor Market Forecast

• 2025: ~USD 48.2B
• 2026: ~USD 54.3B (+11.9% YoY per WSTS)
• 2034 projection: ~USD 175.3B (15.8% CAGR)

2. VALUE CHAIN MAP -- Japan's Position at Every Layer

Layer 1: RAW MATERIALS (Silicon Wafers)

Combined 60% of global 300mm silicon wafers. No non-Japanese firm can replicate at scale.

Layer 2: SEMICONDUCTOR EQUIPMENT

Layer 3: PROCESS CHEMICALS (Photoresists & Specialty Chemicals)

Japanese firms collectively hold ~91% of global photoresist market. JSR was taken private by Japan Investment Corp (JIC) for ~$6.3B in 2024 to consolidate Japan's semiconductor materials strategy.

Layer 4: PHOTOMASK

HOYA dominates EUV photomask blanks. Without HOYA's ultra-flat quartz substrates, ASML's EUV machines cannot produce masks.

Layer 5: CMP (Chemical Mechanical Planarization)

Layer 6: SUBSTRATES (IC Package Substrates)

Ibiden is a primary substrate supplier for Nvidia data center GPUs. Plans to double AI-related substrate output by 2027-2028.

Layer 7: GLASS CLOTH / CCL (Copper-Clad Laminate)

Nittobo is THE critical bottleneck. Jensen Huang (Nvidia CEO) personally visited to secure supply. Apple, Google, Amazon, Microsoft all competing for allocation. Capacity fully booked through 2026; new capacity online 2027. 20% price hike imposed August 2025. Stock up +647% in 52 weeks.

Layer 8: PCB (Printed Circuit Board) & Plating

JCU is an asymmetric play -- its chemicals cost a fraction of the substrate but failure scraps the entire unit. Revenue tied directly to Ibiden/Shinko capex cycles.

Layer 9: CONNECTORS & OPTICAL INTERCONNECTS

Fujikura stock up +654% in 52 weeks -- the best performer on the Nikkei 225, driven by AI data center fiber demand.

Layer 10: PASSIVE COMPONENTS & POWER

Murata raised its MLCC-for-AI-servers CAGR forecast to 30%. AI servers require dramatically more capacitors per board than traditional servers.

Layer 11: POWER SEMICONDUCTORS

New Mitsubishi Electric fab in Kumamoto starts production April 2026 -- JPY 100B investment for power devices.

Layer 12: COOLING & THERMAL

Japan data center liquid cooling market: USD 248M (2024) to USD 1.48B (2032), CAGR 25%.

Layer 13: PLATFORM / SYSTEM INTEGRATION

SPECIALTY GASES

3. FULL COMPARISON TABLE

All data from StockAnalysis.com as of April 2026 unless noted. Market cap in JPY trillions. Revenue in JPY billions.

4. PER-STOCK DETAILED ANALYSIS

TIER 1: MONOPOLY / IRREPLACEABLE POSITIONS

#### NITTOBO (3110) -- T-Glass Monopoly

Thesis: Nittobo holds a genuine 100% monopoly on T-Glass (low-CTE glass cloth) used in every advanced ABF substrate for AI accelerators. Nvidia CEO Jensen Huang personally visited to secure supply. Apple, Google, Amazon, Microsoft all competing for allocation. Capacity fully booked through 2026; 20% price hike in Aug 2025 demonstrates pricing power. Plans to triple capacity by 2027.

Risk: (1) Taiwan Glass or AGY qualifying alternative T-Glass by 2028 could break monopoly premium. (2) Forward PE (38.9x) is HIGHER than trailing (25.0x), suggesting the market expects a normalization after capacity expansion. (3) Glass-to-glass substrate technology shift could reduce T-Glass demand long-term.

Monitoring Trigger: If Taiwan Glass qualifies for T-Glass at major substrate makers, reduce position. If Nittobo raises guidance above current capacity plan, add.

Supply Chain: Customers: Ibiden, Shinko Electric (direct), ultimately Nvidia/AMD/Intel (end users). No named competitors for T-Glass.

Key Dates: FY results (May), new Fukushima line start (2027), Nan Ya weaving partnership ramp.

52-Week Performance: +647%

#### LASERTEC (6920) -- EUV Mask Inspection Monopoly

Thesis: ~100% global market share in EUV photomask defect inspection. Every TSMC, Samsung, Intel EUV fab requires Lasertec tools. Backlog exceeds JPY 1.4T providing multi-year visibility. Operating margin of 48.7% is extraordinary. Transitioning to high-NA EUV (ACTIS A300 series) maintains technological moat.

Risk: (1) Scorpion Capital short report alleging fraud (not substantiated by subsequent audits). (2) KLA (US) developing competing actinic inspection tools -- timeline uncertain but credible threat by 2028+. (3) High beta (1.71) means sharp drawdowns in corrections.

Monitoring Trigger: If KLA announces commercial actinic patterned mask inspection tool, reassess. If Lasertec wins first high-NA EUV inspection tool orders, confirms next-gen moat.

Supply Chain: Customers: TSMC, Samsung, Intel, all photomask shops. Supplier: proprietary EUV light sources.

Key Dates: Q3 FY2026 results (Feb), high-NA tool qualification timeline.

52-Week Performance: +263%

#### TOKYO ELECTRON (8035) -- Equipment Behemoth

Thesis: #3 global semi equipment maker (after ASML and Applied Materials). Dominates coater/developer (90% share with SCREEN) and holds strong positions in etch and deposition. Zero debt, JPY 418B net cash. Revenue growing 33% YoY. Beneficiary of every fab expansion worldwide -- TSMC, Samsung, Intel, Rapidus.

Risk: (1) US-China export controls could restrict sales to Chinese fabs (~25-30% of revenue historically). (2) Cyclical -- equipment spending can drop 20-30% in downturns. (3) PE of 40x is elevated vs. historical range.

Monitoring Trigger: If China revenue drops below 15% of total due to tightened controls, reassess valuation. If TEL wins key GAA/backside power delivery tool-of-record positions, confirms next-cycle growth.

Supply Chain: Customers: TSMC, Samsung, Intel, Rapidus, JASM, Chinese fabs. Suppliers: various Japanese component makers.

Key Dates: FY2026 full-year results (May), SEMI equipment spending forecast updates.

52-Week Performance: +135%

#### DISCO (6146) -- Dicing/Grinding Monopoly

Thesis: ~75% global share in wafer dicing and grinding. Patented Stealth Dicing and Ultrasonic Dicing technologies. Zero debt, ROIC of 60%. Equipment + consumables model creates recurring revenue. Critical for SiC power semiconductors (growing with EV adoption) and advanced packaging (growing with AI).

Risk: (1) PE of 61.5x is extremely elevated. (2) Concentrated in a niche -- total market is smaller than broad equipment peers. (3) Any technological shift away from mechanical dicing (e.g., full laser processing) could erode position.

Monitoring Trigger: If operating margin drops below 35%, investigate whether competition is emerging. If SiC wafer adoption accelerates beyond forecasts, this is a tailwind.

Supply Chain: Customers: All major foundries, IDMs, and OSAT/packaging houses globally.

Key Dates: Quarterly results, SiC wafer market growth data.

52-Week Performance: +177%

#### SHIN-ETSU CHEMICAL (4063) -- Materials Conglomerate

Thesis: #1 global silicon wafer maker (~35% share), plus photoresist, silicones, and PVC. JPY 1.25 trillion net cash -- one of the strongest balance sheets in Japan. D/E of 0.05. Current ratio of 6.26. Operates as a cash machine with 25.6% operating margin. Silicon wafer demand is structurally growing with all fab expansions.

Risk: (1) PVC business is cyclical and China-exposed. (2) Silicon wafer pricing has been weak due to memory downturn, though recovering. (3) SUMCO competition in wafers keeps pricing disciplined.

Monitoring Trigger: If 300mm wafer ASP increases >5% QoQ, signals pricing power recovery. If PVC margins deteriorate, watch for earnings drag.

Supply Chain: Customers: TSMC, Samsung, Intel, all foundries (wafers); TOK, JSR (photoresist). Suppliers: polysilicon, silica sand.

Key Dates: Q1 FY2027 results (Jul), wafer pricing contract negotiations.

52-Week Performance: +77%

#### HOYA (7741) -- EUV Photomask Blanks

Thesis: >80% market share in EUV photomask blanks (ultra-flat quartz substrates). Without HOYA blanks, no EUV masks can be made, meaning no leading-edge chips. 24.2% ROE, zero-debt equivalent (D/E 0.04), JPY 539B net cash. Also a major player in semiconductor lithography optics, medical endoscopes, and eyeglass lenses.

Risk: (1) PE of 40.5x leaves little room for execution misses. (2) High-NA EUV could require different blank specs -- HOYA must stay ahead. (3) Diversified business means semi is not the only earnings driver.

Monitoring Trigger: If ASML high-NA EUV ramp accelerates, HOYA benefits. If competitor (AGC, Corning) qualifies EUV blanks, reassess moat.

Supply Chain: Customers: Toppan, DNP, Photronics (mask shops) serving TSMC, Samsung, Intel.

Key Dates: FY results (May), high-NA EUV adoption timeline.

52-Week Performance: +86%

#### MURATA MANUFACTURING (6981) -- MLCC King

Thesis: 40% global market share in MLCCs -- the most critical passive component. AI servers require 10-100x more capacitors than traditional servers. Murata raised its MLCC-for-AI-servers CAGR to 30%. Zero-debt equivalent (D/E 0.02), JPY 521B net cash. Diversified across automotive, smartphone, and industrial.

Risk: (1) Smartphone slowdown could offset AI server growth. (2) Samsung Electro-Mechanics and Yageo are investing heavily in high-cap MLCC. (3) Current ROE of 7.3% is low, suggesting margins are compressed -- should improve as AI mix increases.

Monitoring Trigger: If MLCC ASP for high-layer-count AI models increases, confirms pricing power. If smartphone demand weakens further, watch for inventory build.

Supply Chain: Customers: Nvidia, AMD, Apple, Samsung, all server OEMs. Raw materials: barium titanate, palladium.

Key Dates: Q1 FY2027 results (Jul), MLCC capacity expansion announcements.

52-Week Performance: +129%

#### TOKYO OHKA KOGYO (4186) -- Photoresist Leader

Thesis: #1 global photoresist maker with ~30% share. Investing JPY 20B in new Korea plant. EUV photoresist is a critical consumable -- every wafer exposure requires fresh resist. 17.2% ROE, low debt (D/E 0.11). JSR's privatization by JIC removes the #2 player from public markets, making TOK the only pure-play photoresist stock.

Risk: (1) Korea plant expansion won't generate revenue until 2030. (2) EUV resist qualification is highly sticky but competition from Fujifilm and Shin-Etsu exists. (3) PE of 33.6x is premium.

Monitoring Trigger: If Rapidus qualifies TOK resist for 2nm, signals next-gen positioning. If resist consumption per wafer increases with high-NA EUV, this is a structural tailwind.

Supply Chain: Customers: TSMC, Samsung, Intel (direct). JSR (now JIC-owned) is both competitor and potential consolidation partner.

Key Dates: FY results, Korea plant milestones, Rapidus resist qualification.

52-Week Performance: +222%

#### FUJIMI INC. (5384) -- CMP Slurry Monopoly

Thesis: 84-92% global share in silicon wafer lapping/polishing agents. ~50% share in front-end polysilicon CMP slurry. Every wafer goes through CMP. Consumable business model with recurring revenue. Reasonable PE (22.2x), decent ROE (12.6%), and 2.47% dividend yield -- the most attractively valued monopoly in this space.

Risk: (1) Small company (JPY 220B market cap) with limited analyst coverage. (2) Versum/CMC (now Entegris) and Cabot compete in certain CMP segments. (3) Lower growth than equipment names -- more steady than explosive.

Monitoring Trigger: If advanced node CMP steps increase (e.g., GAA transistors require more CMP), this is incremental growth. Watch wafer start data from SEMI.

Supply Chain: Customers: Shin-Etsu, SUMCO (wafer polish), all foundries (device CMP).

Key Dates: FY results (May), wafer start trends.

52-Week Performance: +68%

TIER 2: DOMINANT POSITIONS WITH STRONG FINANCIALS

#### ADVANTEST (6857) -- AI Test Equipment

Thesis: 58% share in semiconductor test equipment (ATE). The massive increase in AI chip complexity (larger dies, more I/O, higher speeds) drives longer test times and more sophisticated test equipment. ROE of 49.3% and ROIC of 69.4% are best-in-class. Revenue just crossed JPY 1 trillion.

Risk: (1) PE of 69.1x is the highest in the peer group. (2) Test equipment is late-cycle -- orders can drop sharply. (3) Teradyne (US) is a strong competitor with similar capabilities.

Monitoring Trigger: If HBM test revenue grows >50% YoY, confirms AI demand structural. If forward bookings decline 2 consecutive quarters, reduce.

Supply Chain: Customers: TSMC, Samsung, Intel, OSAT houses, Nvidia (indirectly).

52-Week Performance: +396%

#### SCREEN HOLDINGS (7735) -- Wafer Cleaning

Thesis: ~57% share in wafer cleaning equipment. Most attractively valued equipment name: PE 23.7x (forward 18.4x), ROE 19.8%, ROIC 31.9%, and 1.33% dividend yield. Zero debt. Cleaning steps increase with node shrinks -- structural growth tailwind.

Risk: (1) Smaller scale than TEL limits pricing power. (2) Lam Research and Tokyo Electron compete in cleaning. (3) Less visible than EUV/litho plays.

Monitoring Trigger: If cleaning equipment orders outpace overall WFE growth, confirms structural share gain. If JASM/Rapidus select SCREEN tools, confirms domestic positioning.

Supply Chain: Customers: All major foundries and IDMs.

52-Week Performance: +133%

#### IBIDEN (4062) -- Nvidia's Substrate Partner

Thesis: Primary ABF substrate supplier for Nvidia data center GPUs. Plans to nearly double AI-related substrate output by 2027-2028. Stock up 495% in 52 weeks reflecting AI substrate demand explosion. CoWoS and advanced packaging trends are structural tailwinds.

Risk: (1) PE of 69.8x is very expensive. (2) Negative free cash flow (-JPY 8B) due to massive capex -- execution risk on capacity expansion. (3) D/E of 0.55 is higher than most peers. (4) Altman Z-Score of 2.41 indicates some financial stress.

Monitoring Trigger: If Nvidia's CoWoS substrate demand increases with GB200/GB300, positive. If FCF turns positive post-capex cycle, reduces risk. If glass substrates mature as ABF replacement, major threat.

Supply Chain: Customers: Nvidia (primary), AMD, Intel. Suppliers: Ajinomoto (ABF film), Nittobo (glass cloth).

52-Week Performance: +495%

#### FUJIKURA (5803) -- Fiber Optic Data Center Play

Thesis: Best performer on Nikkei 225 (+654% in 52 weeks). Fiber optic cables are the physical backbone of AI data centers -- connecting GPU clusters requires massive fiber. 31.1% ROE, 36.4% ROIC. Revenue crossed JPY 1 trillion.

Risk: (1) PE of 63.6x prices in enormous growth expectations. (2) PB of 17.0 is extreme. (3) Fiber optic cable is ultimately a commodity -- differentiation is in specialty products. (4) Stock may have overshot near-term fundamentals.

Monitoring Trigger: If data center fiber orders decelerate, watch for multiple compression. If co-packaged optics reduces fiber per rack, structural headwind.

Supply Chain: Customers: Hyperscaler data centers (AWS, Azure, GCP), telecom operators.

52-Week Performance: +654%

#### JCU CORPORATION (4975) -- Hidden Gem

Thesis: Plating chemicals for IC substrates -- a picks-and-shovels play on Ibiden/Shinko capex. 40.4% operating margin, 68.9% gross margin, 31.9% ROIC. Essentially debt-free (D/E 0.01). PE of 18.2x is remarkably cheap for a company with monopoly-like customer switching costs and 40%+ margins. Current ratio of 8.56 shows fortress balance sheet.

Risk: (1) Tiny company (JPY 152B market cap) -- liquidity risk. (2) Revenue concentration with Ibiden/Shinko means customer-specific risk. (3) If substrate makers diversify chemical suppliers, margin pressure.

Monitoring Trigger: If Ibiden/Shinko capex accelerates beyond current plans, JCU is a direct beneficiary. If optical transceiver plating revenue reaches >10% of total, new growth vector.

Supply Chain: Customers: Ibiden (4062), Shinko Electric (6967), PCB makers. Suppliers: specialty chemical raw materials.

52-Week Performance: +98%

#### KOKUSAI ELECTRIC (6525) -- Batch ALD Leader

Thesis: 70% market share in batch ALD -- critical for high-aspect-ratio memory (3D NAND, DRAM). ALD steps increasing with each technology generation. Mid-term target: JPY 330B revenue, 30%+ operating margin. Recently IPO'd (2023), providing clean balance sheet.

Risk: (1) PE of 49.2x is expensive for current margins (18.6% operating). (2) Single-wafer ALD from TEL and Lam could take share. (3) Memory cycle dependency.

Monitoring Trigger: If 3D NAND layer count increases drive higher batch ALD tool demand, positive. If operating margin trajectory shows clear path to 30%+, confirms thesis.

Supply Chain: Customers: Samsung, SK Hynix, Micron, Kioxia (memory makers).

52-Week Performance: n/a (relatively new listing)

#### RENESAS ELECTRONICS (6723) -- Edge AI MCU Leader

Thesis: #3 global MCU maker. Strong position in automotive semiconductors, now pivoting to edge AI and IoT. Massive FCF generation (JPY 364B) despite net loss on GAAP basis (acquisition amortization). Forward PE of 21.2x is reasonable. FCF yield of 6.5% is best among peers.

Risk: (1) JPY 892B net debt from acquisitions (Dialog Semi, Intersil). (2) GAAP net loss reported. (3) Automotive downturn could hit core business. (4) Not a pure AI play.

Monitoring Trigger: If edge AI MCU revenue grows >30% YoY, confirms AI pivot. If debt/EBITDA drops below 2.5x, improving financial health.

Supply Chain: Customers: Toyota, Honda, all major automakers; industrial OEMs.

52-Week Performance: +92%

TIER 3: STRONG BUT LESS DIFFERENTIATED

#### TDK (6762) -- Passives Diversified

Thesis: Top 3 passive component maker. Battery business (via TDK Lambda) provides EV exposure. 9.5% ROE improving. Reasonable PE (26.1x). JPY 183B net cash.

Risk: Consumer electronics exposure. Lower margin than Murata. Battery competition from Chinese players.

52-Week Performance: +87%

#### NITTO DENKO (6988) -- Specialty Films

Thesis: Most attractive valuation among established names: PE 17.1x, forward 14.2x. Zero debt. 1.78% dividend yield. Optical films for displays + semiconductor tape. Boring but consistent.

Risk: Display market maturity. Not a direct AI accelerator play.

52-Week Performance: +34%

#### FUJI ELECTRIC (6504) -- Power Semiconductors

Thesis: Solid mid-cap power semiconductor play. PE 20.3x, ROE 11.8%, 1.54% dividend yield. Growing SiC portfolio for EVs and industrial.

Risk: Lower SiC market share than Infineon, STMicro, or Wolfspeed. Not a pure play.

52-Week Performance: +103%

#### NIPPON SANSO (4091) -- Semiconductor Gases

Thesis: Top 4 global industrial gas company. Semiconductor-grade gases are essential for every fab process. Stable, defensive growth. PE 21.8x, 1.0% yield.

Risk: Air Liquide, Linde, and Air Products are larger competitors. Gas is not a differentiator.

#### RESONAC HOLDINGS (4004) -- Turnaround Story

Thesis: Formerly Showa Denko. Semiconductor materials + automotive materials. Trading at forward PE of 27.4x with massive 52-week gain (+425%). SiC substrate and advanced packaging materials exposure.

Risk: D/E of 1.33 is the highest in peer group. Altman Z-Score of 1.51 signals financial stress. Only 4.3% operating margin. Turnaround execution risk.

52-Week Performance: +425%

#### KEYENCE (6861) -- Factory Automation Premium

Thesis: 50.9% operating margin -- highest of any company in this universe. Factory automation sensors and vision systems are critical for semiconductor manufacturing. JPY 1.33T net cash. Fortress balance sheet.

Risk: Premium valuation (PE 36.7x) with only +9% price gain suggests re-rating exhaustion. Not a direct semiconductor play -- more adjacent.

52-Week Performance: +9%

5. CONVICTION TIERS

HIGH CONVICTION -- Buy on Dips

MEDIUM CONVICTION -- Own but Size Appropriately

LOW CONVICTION -- Watch / Avoid at Current Prices

6. JAPAN vs GLOBAL COMPETITIVE MATRIX

Key Takeaways from the Competitive Matrix:

1. Japan has UNASSAILABLE positions (no credible alternative globally) in:

- T-Glass cloth (Nittobo)

- EUV mask inspection (Lasertec)

- EUV photomask blanks (HOYA)

- ABF film (Ajinomoto)

- Photoresist (TOK/JSR/Shin-Etsu collective)

- Wafer polishing agents (Fujimi)

- Wafer dicing/grinding (Disco)

2. Japan has DOMINANT but challengeable positions in:

- Silicon wafers (Shin-Etsu/SUMCO -- challenged by Siltronic)

- Coaters/developers (TEL -- no real challenger)

- Test equipment (Advantest -- Teradyne competition)

- MLCC (Murata -- Samsung E-M competition)

- IC substrates (Ibiden/Shinko -- Unimicron/AT&S competition)

3. Japan is WEAK in:

- Leading-edge foundry (Rapidus is aspirational, not proven)

- Memory (Kioxia is private and struggling)

- Power semiconductors (European firms lead in SiC)

- Semiconductor gases (European firms dominate)

- AI chip design (entirely US-dominated: Nvidia, AMD, Broadcom)

7. SOURCES

Financial Data

• StockAnalysis.com/quote/tyo/{TICKER}/statistics/ -- All 29 companies verified individually

Industry & Supply Chain Research

• Kaohsiung Times -- Japan's Semiconductor Industry Repositioning
• AInvest -- Japan's Strategic Bet on AI and Semiconductors 2026
• Deloitte -- 2026 Semiconductor Industry Outlook
• AMRO Asia -- Japan's Strategic Comeback in the Global Chip Race
• US Trade.gov -- Japan Semiconductors
• Brookings -- Renaissance of Japanese Semiconductor Industry
• TrendForce -- Japan Photoresist Investment for 2nm
• CERADIR -- Passive component supply chain analysis

T-Glass / Nittobo

• Nikkei Asia -- Nvidia courts Japanese supplier to ease AI bottlenecks
• Tom's Hardware -- Glass cloth could be the next great AI shortage
• DigiTimes -- Race for T-Glass heats up
• DigiTimes -- Nittobo to triple T-Glass cloth capacity by 2027
• Global Tech Research -- Nittobo: King of Electronic Fiber Glass
• SmartKarma -- Nittobo Scarcity Premium in AI Substrates

TSMC Japan / JASM

• Semiwiki -- TSMC Kumamoto: Pioneering Japan's Semiconductor Revival
• TrendForce -- TSMC Kumamoto Fab 2 pivot to 2nm
• Wikipedia -- Japan Advanced Semiconductor Manufacturing

Rapidus

• Bloomberg -- Japan Bets $16 Billion on Rapidus
• The Register -- Rapidus lands $1.7B for 2nm
• Rapidus Official -- Funding Announcement

Company-Specific

• Nikkei Asia -- Ibiden substrate expansion amid AI boom
• SemiAnalysis -- Disco Corporation world leader
• Senrigan -- Lasertec monopoly analysis
• Senrigan -- Disco insight report
• Hibiki Path Advisors -- Japan Pure Chemical AI stock analysis
• Substack/iamfabian -- JCU Corporation overlooked AI asset
• Fortune -- Fujikura hottest stock in Japan
• Gabelli -- Japan Technology Leaders research
• TrendForce -- Ajinomoto ABF expansion
• DigiTimes -- JSR acquisition and delisting
• Introl -- Hyperscaler CapEx $600B in 2026
• SEMI -- Global Equipment Billings $135B in 2025

DISCLAIMER: This report is for informational purposes only and does not constitute investment advice. All financial data is sourced from StockAnalysis.com and public sources as of April 2026. Past performance does not guarantee future results. Japanese stock prices and financials are denominated in JPY. Always conduct your own due diligence before making investment decisions.

Report generated 2026-04-18. Research methodology: Dual-approach (Direct search + T-Glass backward trace from Nvidia/TSMC/ASML value chain). 29 stocks analyzed across 13 value chain layers. All financial metrics verified against StockAnalysis.com.