SCREEN Holdings (7735.T) — Valuation Using Japanese Sources Only: Cyclical Semicap at Full Multiples

by psind707

Published: May 5, 2026 | Price: ¥10,390 (May 1, 2026) | Rating: HOLD with moderate negative bias | PWFV: ¥7,179 | Time horizon: 12-24 months

TL;DR

SCREEN Holdings is a high-quality cyclical semicap business with dominant cleaning equipment market share (38-45% global) and a fortress balance sheet (~¥250B net cash, debt-free). But the stock has rallied 31% from January lows to ¥10,390 on hope, not fundamentals. FY2026 will be down 11-14%, the dividend was just cut from ¥308 to ¥280, and China dependency cliff risk remains.

PWFV: ¥7,179 — implying ~31% downside. This is a HOLD/trim, not a buy at current levels. Better entry around ¥7,000-8,000 after FY26 full-year print (late May 2026).

This report uses Japanese sources only as a methodology test. Unlike Rakuten where Japanese sources were materially more bearish than English coverage, for SCREEN the picture is roughly similar across both languages — Japanese sources add depth and granularity without directional bias.

Spot Setup

• Price: ¥10,390 (May 1, 2026, Matsui Securities)
• Market cap: ~¥2.49T (240M shares outstanding)
• 52-week range: roughly ¥7,000–13,000+ (highly cyclical)
• Stock rallied ~31% from January low (~¥7,940)
• Pre-cyclical-correction peak (2024): equivalent of ~¥13,000+ post 2:1 split (Sep 2023)
• Next earnings: FY2026 (March 2026) full year results due late May 2026

What SCREEN Holdings Does (per Japanese sources)

The company is an 80-year-old Kyoto-based semiconductor equipment specialist (founded 1943, originally photolithography for printing). Per Nikkei and Strainer:

• Single-wafer cleaning equipment: Global #1 with 38-45% market share
• Batch cleaning equipment: 47% global share (#1)
• Spin scrubber: 59% global share (#1)
• Cleaning is ~30% of total semiconductor manufacturing process steps
• Customers: TSMC, Samsung, SK Hynix, Intel, all major foundries/memory makers

Per Nikkei Business: "the precision required is like removing a single grain of pollen from a wafer the size of Tokyo Dome."

Segment Breakdown (FY2025)

SPE took 100%+ of operating profit in FY2025.

Recent Earnings Trajectory

FY2025 (March 2025) — Peak Year

• Revenue: ¥625.3B (+23.8% YoY)
• Operating profit: ¥135.7B (+44.1%)
• Net income: ¥99.5B (+40.9%)
• OPM: 21.7%
• 4th consecutive year of record results
• China was 47% of revenue at peak

FY2026 (March 2026) — Cyclical Decline

• Q1 FY26 (Apr-Jun 2025): Revenue +1%, OP -12%
• Q3 FY26 cumulative: Revenue -7.5%, OP -23.0%
• Q3 alone OPM compressed from 23.2% → 20.5%
• Full year guidance: Revenue -0.7%, OP -13.8%, Net income -11.5%
• Annual dividend cut: ¥308 → ¥280 (-9% cut, first cut signaling cycle caution)

Why the Decline (per Nikkei July 2025)

米中貿易摩擦を見込んだ中国メーカーによる駆け込み需要が前期で一巡し、反動減

Translated: "Pull-forward demand from Chinese makers anticipating US-China trade friction completed in prior year; rebound decline this year."

China fell from 47% (FY2025) → 34% (Q1 FY26) → ~30% expected (FY26). CEO Goto stated "the China ratio decline is as expected; will stabilize in the 30% range."

Strategic Position (Japanese Sources)

Global Market Share

Dominant niche player — undifferentiated cleaning is its sole core competency, but it's the global leader.

Long-Term Vision (per IR Day Dec 2025)

FY2033 (March 2033) targets: Revenue ¥1T+, Operating margin 20%+

Currently at ¥625B run-rate; needs 5% revenue CAGR over 8 years to reach ¥1T. Production capacity at Hikone planned to expand to ¥600B SPE revenue equivalent by FY2027 (+20% vs FY2024).

Balance Sheet (per Q3 FY26 disclosure)

Fortress balance sheet — far better than most peers globally. The pristine financials are rarely emphasized in English coverage but are the bull's strongest argument.

WACC Inputs

Current Valuation Multiples

Three-Scenario Valuation

Bull Case (25% probability) — Cycle Inflection 2027+

• Revenue 2030E: ¥850B (5-6% CAGR, partial path to ¥1T target)
• OPM: 22% (back to peak)
• Net income 2030E: ¥130B
• Apply 22x P/E (peak-cycle multiple)
• Equity value: ¥2,860B
• Per share: ¥11,917

Trigger: AI demand sustains advanced node investment + HBM expansion + US/Japan domestic fab buildouts (Rapidus, TSMC Arizona, Samsung Texas) compensate for China share loss.

Base Case (45% probability) — Modest Growth, China Stabilizes

• Revenue 2030E: ¥720B (3% CAGR)
• OPM: 19% (slightly below peak due to mix shift)
• Net income 2030E: ¥100B
• Apply 18x P/E (long-term average)
• Equity value: ¥1,800B
• Per share: ¥7,500

Trigger: China stabilizes at 25-30% of revenue; advanced node + memory demand offsets gradual China decoupling; long-term ¥1T target tracks slightly behind plan.

Bear Case (30% probability) — China Share Keeps Falling, Cycle Prolonged

• Revenue 2030E: ¥520B (-3% CAGR)
• OPM: 14% (capacity overhang from S3-5 expansion)
• Net income 2030E: ¥55B
• Apply 12x P/E (cycle-trough multiple)
• Equity value: ¥660B
• Per share: ¥2,750

Trigger: BIS export controls keep tightening; China share falls to 15-20%; competition from local Chinese cleaning equipment makers; fixed cost overhang from new capacity hits margins.

Probability-Weighted Fair Value

PWFV = ¥11,917 × 0.25 + ¥7,500 × 0.45 + ¥2,750 × 0.30 = ¥7,179

vs spot ¥10,390 = ~31% downside to PWFV

Reconciliation with Japanese Analyst Consensus

Per Minkabu (April 10, 2026):

• 14 analysts: 5 strong buy, 3 buy, 6 neutral, 1 strong sell
• Consensus target: ¥11,280 (+4.93% from then-spot)

My PWFV (¥7,179) is materially below consensus (¥11,280). Why?

The Japanese sell-side is essentially pricing my bull scenario as base case:

• Assumes cycle has bottomed
• Assumes China stabilization
• Assumes ¥1T target is achievable
• Applies peak-cycle multiples

I'm pricing in:

• 30% bear probability (US-China decoupling + capacity overhang + multiple compression)
• More modest base case (¥720B revenue 2030 vs ¥850B+ in bull)
• Cycle-aware multiples (18x base, not 22x)
• Acknowledging FY26 already shows decline

This is a battleground: the divergence reflects reasonable disagreement about the secular vs cyclical nature of the China decline.

Critical Risks (Japanese-Source Emphasized)

1. China Share Cliff Risk

Per Note.com analysis (December 2025):

中国向け売上が全体の半分程度を占める...米中貿易摩擦の激化リスク

If BIS controls escalate further, even the 30% China share could fall to 15-20%, leaving SCREEN with significant capacity overhang.

2. US BIS Export Controls

• Dec 2024: 24 equipment types newly controlled, 3 software types, HBM included
• Jan 2025: Technical adjustments
• Sep 2025: TSMC China VEU (Validated End-User) reportedly being revoked
• Each tightening reduces SCREEN's addressable Chinese market

3. Cyclicality

• FY2024: +44% YoY operating profit growth
• FY2026: -14% YoY decline (guided)
• This is a 58-point swing in one year — not a defensive stock
• Stock peaked at ~¥22,000 (post-split equivalent) in 2024, fell to ~¥7,000 (~-68%), now bounced to ¥10,390

4. Capacity Overhang from S3-5 Expansion

S3-5 building completed Jan 2024; ¥600B production capacity by FY2027 target. If demand stays at FY26 ¥520B SPE level, the new capacity becomes fixed cost burden, margin headwind through 2027-2028.

5. Dividend Cut Signal

DPS cut from ¥308 (FY25) to ¥280 (FY26) at midpoint of guidance. Even with strong balance sheet, management chose to signal caution by reducing payout. Meaningful insider signaling.

6. Competition from Lam Research

Lam Research is increasingly focused on cleaning equipment. SCREEN's 38-45% global share in single-wafer cleaning has been gradually challenged.

Bull Catalysts

1. AI/HBM-Driven Memory Equipment Demand

Per Note.com analysis:

生成系AIは高い演算能力をもつGPUと、それに最適なHBMを合わせた高性能モジュールの拡大が期待され、2023年〜2027年の年間平均成長率は+52%

HBM CAGR forecast 2023-2027: +52% per year. SCREEN cleaning equipment is critical for HBM stack manufacturing.

2. Advanced Node Investment Continues

TSMC, Samsung, SK Hynix all expanding 3nm/2nm capacity. SCREEN single-wafer cleaning is essential for these advanced nodes.

3. Long-Term ¥1T Revenue Target

If achieved:

• ¥1T × 20% × (1-0.27 tax) = ¥146B net income
• × 20x P/E = ¥2,920B equity
• Per share: ¥12,167 (modest upside vs current ¥10,390)

But discounted back to today at 10% over 8 years = ¥5,675. Long-term DCF suggests current price already prices in success.

4. Domestic Japan Demand (Rapidus, Kioxia)

Rapidus 2nm fab in Hokkaido starting volume production targeted 2027. Kioxia memory expansion. SCREEN's home turf advantages.

5. Defensive China Strategy Working

CEO Goto's claim that China ratio "stabilizes in 30%" range — if this holds, China is no longer the marginal swing factor it once was. Risk has been somewhat priced in.

Comparable Company Multiples

SCREEN trades roughly in line with global semicap peers — not particularly cheap on multiples; not expensive either.

Investment Conclusion

Rating: HOLD with moderate negative bias

The asymmetry here goes the wrong way: only ~15% upside to bull case, but 73% downside to bear case. The math says HOLD or trim, not buy at ¥10,390.

Sizing Discipline

• Pass at ¥10,390 — too close to bull case; insufficient margin of safety
• Initiate small at ¥8,000 (within base case)
• Add aggressively below ¥6,000 (margin of safety vs base PWFV)
• Stop loss: ¥5,500 if cycle deepens beyond bear scenario
• Trim above ¥12,000 (tactical resistance + bull case fully priced)

Single-Leg Options Note

• Liquid options exist on TSE for 7735, but spread is wide
• Better expression: stock at ¥6,000-8,000 range entry
• For tactical short bias: avoid — semicap can squeeze hard on AI/HBM news flow

What the FY2026 Full Year Print (Late May 2026) Needs to Show

To support bull/base case:

• FY2027 guidance for revenue growth ≥5% YoY
• OPM guidance ≥18%
• China share stabilizing at 30%
• Order book inflection visible
• AI/HBM equipment specific commentary on backlog

Bear signals to watch:

• FY2027 guidance below FY2026
• Further dividend cut
• Margin guidance below 17%
• China share falling below 25%
• Capacity utilization commentary indicating overbuild

Honest Methodology Reflection

This is the second name where I've stress-tested English-source vs Japanese-source equity research workflow:

For SCREEN Holdings, Japanese sources provided:

• Note.com / CommercePick: Critical decomposition of China dependency risk with explicit probability framing
• Strainer: Detailed market share data not in English coverage
• Nikkei Business + Nikkei: CEO/CFO commentary in Japanese with more nuance than translated press releases
• YH Research / SEAJ Japan: Industry-specific market data on Japanese semicap companies
• EE Times Japan: Equipment-segment-by-segment competitive landscape
• Buffett-Code Japan / IRBank: Granular financial analysis tools

Unlike Rakuten where Japanese sources were materially more bearish than English, for SCREEN the picture is roughly similar across both languages — Japanese sources are equally analytical and rigorous. The difference is depth and granularity, not directional bias.

Why? SCREEN is a B2B equipment supplier, not a consumer-facing company. There's less promotional spin in their IR. The English coverage from semicap-specialist sell-side analysts is reasonably faithful to the underlying reality.

But Japanese sources still added value by:

1. Providing market share specifics (38-45% single wafer cleaning)
2. Explicit China revenue trajectory (47% → 34% → 30%)
3. CEO Goto's specific commentary on China stabilization
4. Long-term ¥1T target framework
5. Capacity expansion specifics (S3-3, S3-4, S3-5)

Summary Verdict

SCREEN Holdings is a high-quality cyclical semicap business trading at full cyclical multiples on declining earnings. The fortress balance sheet (¥250B+ net cash) + dominant cleaning equipment market share (38-45%) + AI/HBM long-term tailwind = real bull case. But:

• Stock has rallied 31% from January lows on hope, not fundamentals
• FY26 numbers will be down 11-14%
• Dividend just cut (signal)
• China dependency cliff risk real
• US BIS controls keep tightening
• Capacity expansion creates fixed-cost overhang

Better entry: ¥7,000-8,000 range. Current ¥10,390 leaves no margin of safety.

This is cleaner than Rakuten (where the conglomerate complexity creates more uncertainty), but the risk-reward is similarly unattractive at current levels. Wait for the FY2026 full year print (late May 2026) — if FY27 guidance is muted, expect another leg down to ¥7-8K range. That's the trade entry point.

Sources Used (Japanese Only)

• 株式会社SCREENホールディングス IR (screen.co.jp) — FY26 Q1/Q2/Q3 disclosures, IR Day Dec 2025
• 日本経済新聞 (nikkei.com) — Multiple articles on quarterly earnings + China commentary
• 日経ビジネス (business.nikkei.com) — Series on SCREEN technology and TSMC relationship
• みんかぶ (minkabu.jp) — Stock data, analyst consensus, earnings analysis
• Yahoo!ファイナンス (finance.yahoo.co.jp) — Financial summaries, AI earnings briefings
• IRBank (irbank.net) — Long-term financial database
• Buffett-Code Japan (buffett-code.com) — Risk factors and detailed disclosures
• ストレイナー (strainer.jp) — Strategic profile and market share specifics
• 株予報Pro (kabuyoho.jp) — Target prices and ratings
• かぶたん (kabutan.jp) — Earnings reaction and short-term commentary
• ニュースイッチ (newswitch.jp) — Production capacity expansion news
• ITmedia EE Times Japan (eetimes.itmedia.co.jp) — Industry market share data
• Note.com (note.com) — Independent analyst critical decompositions (yuki_semicon, ijnrdx)
• Wikipedia 日本語版 — Company history and structural facts
• ORIX Insight (go.orixrentec.jp) — Production base details
• Y-SKT Coating Magazine — Market share and BIS controls analysis
• 化学情報センター (chem-3.com) — Comprehensive equipment maker share data
• 兼松PWS (pwsj.co.jp) — Market structure analysis
• IRライブラリ direct downloads (PDFs in Japanese)

Research notes for personal use, not investment advice. Verify all figures independently before acting. Sources used are exclusively Japanese-language to test the methodology framework established in the Rakuten valuation comparison earlier today.