ticker: 3436.T layer: L0-Wafer verdict: red verify_date: 2026-05-28 #

SUMCO Corporation (3436.T) — Graham Screen #

Screened: 2026-05-28 | Price: 3,348 JPY (EUR ~20.80 at 161 JPY/EUR) | Source: yfinance + web

1. Price & Returns #

Note: Stock was at 937 JPY in Jan 2026 trough. The 126% YTD rally has already happened — this is NOT a trough-entry opportunity at current price.

2. Valuation Multiples #

3. Balance Sheet (FY2025, Dec 31) #

Debt trajectory: 110B (2022) → 190B (2023) → 312B (2024) → 354B (2025). The Yokkaichi expansion was entirely debt-funded. LT debt (312B) exceeds working capital (294B) — Graham F2 fails.

4. Five-Year Income & FCF #

All figures JPY. Revenue peaked 2022 and has not recovered. Capex peaked 2023-2024 (Yokkaichi expansion); 2025 shows first meaningful step-down to 111B. FCF was negative for three straight years (2023-2025), total FCF burn approximately -350B JPY. Capex is now declining — key thesis condition for improvement.

5. Dividend History (5Y) #

NEVER zero — technically "uninterrupted" since 2016. But the cut from 81 JPY (FY2022) to 20 JPY (FY2025) is a 75% economic reduction. Year-end 2026 dividend is undecided. In the spirit of Graham F4 — which tests for earnings quality and dividend reliability — this is a FAIL: the dividend behaves like an economically impaired payout that tracks a deeply cyclical P&L.

6. Insider Transactions (12 Months) #

Insider ownership: 3.9%. Institutional ownership: 62.6%. Japanese corporate governance norms mean insider ownership is structurally low for large industrials; no unusual open-market purchases or sales detected in data available from yfinance. No Form 4 equivalents surfaced via public searches. Neutral signal.

7. Analyst Coverage #

Current price (3,348 JPY) is 37% ABOVE the analyst mean target (2,448 JPY) and 63% above median (2,050 JPY). The stock's 126% YTD rally has overshot analyst consensus by a wide margin. This is a significant red flag: the crowd has priced in a recovery that analysts have not yet validated with target upgrades.

8. Recent Earnings (Q4 2025 / Q1 2026) #

Q1 FY2026 (reported May 12, 2026):

• Revenue: 101.4B JPY (-1.0% YoY)
• Operating loss: -5.2B JPY (vs +6.0B Q1 2025)
• Net loss: -8.4B JPY (EPS -24.22 vs consensus -14.73 — 64% negative surprise)
• EBITDA margin: 23.1%
• Stock fell ~3% post-print

Q2 FY2026 guidance:

• Revenue: 112.0B JPY (+10.5% QoQ)
• Operating loss: -2.5B JPY (narrowing but still loss)

Utilization rates:

• Blended: ~85%
• Leading-edge only: ~90%
• Legacy nodes (200mm, non-AI 300mm): well below average

Key management comments:

• Capex peaked; no new major commitments beyond Yoshinogari fill
• LTA prices stable, spot recovering gradually
• H2 2026 expected to improve on AI demand
• 200mm Miyazaki plant closing by end 2026 (capacity rationalization)

9. News Last 90 Days #

• May 12, 2026 — Q1 miss: EPS -24.22 vs -14.73 consensus. Stock dipped post-earnings despite the recent 18% single-day surge on May 9 (earnings anticipation buying).
• Feb 2025 — Miyazaki plant closure announced for 200mm production by end 2026. Rationalization signal: excess low-end capacity being retired.
• LTA renegotiations — None underway. Existing LTAs held; volume delivery extensions in 2024 weighed on FY2024 revenue. LTA price increases expected to flow through 2025-2026.
• AI demand — TSMC leading-edge wafer demand robust; this benefits SUMCO's 300mm epitaxial line. But AI demand is ~15-20% of total volumes; the rest remains in inventory digestion.

10. Cycle Context #

The silicon wafer industry entered a structural downcycle in late 2022 after the memory/logic inventory build peaked. Key cycle markers:

• 2022 peak: SUMCO at 441B revenue, 81 JPY dividend, FCF positive
• 2023-2025 trough: Utilization dropped to ~65-70% at worst; wafer shipments -14% in 2023, further weakness in 2024
• 2026 position: Partial recovery. Leading-edge 300mm for AI/HBM recovering (~90% util). Legacy 200mm + non-AI 300mm still sluggish, customer inventories still elevated for mature logic (28nm/40nm).
• 2027-2028 outlook: Management guides depreciation burden to ease significantly entering 2027 (Yokkaichi capex fully digested). This is the true FCF inflection point — not 2026.

The trough in price was Jan 2026 (937 JPY). The trough in fundamentals is 2025-early 2026. We are past the price trough; we are in early-innings of the fundamental recovery. The YTD +127% has consumed most of the obvious mean-reversion opportunity.

Graham's 7 Filters #

Score: 1/7 PASS (F1 only)

Value-Trap vs Cyclicality Test (shr-017) #

Each fail needs to be evaluated:

• F2 FAIL: Debt is capex-funded, structured. Not distress, but constrains FCF recovery.
• F3 FAIL: Pure cyclicality — profitable 2022-2024, trough loss 2025. NOT permanently impaired.
• F4 FAIL: Dividend maintained throughout but economically impaired. Signals management prudence (didn't zero it), not terminal decline.
• F5 FAIL: Revenue declining with cycle, not structural demand destruction.
• F6/F7 FAIL: This is the core problem. Even on normalized EPS the stock only barely passes F7 (22.8x × 2.06 = 47.0 — double the threshold). On forward EPS it's catastrophically expensive: 68x × 2.06 = 140x.

Conclusion: Fails are cyclical in nature, NOT value-trap indicators. But the valuation (P/B 2.06x, normalized P/E 22.8x) reflects that the market already knows this is a cyclical trough. The discount is gone.

Graham Intrinsic Value #

Normalized EPS used: 147 JPY (average 2022-2024; excludes trough year). Forward EPS: 49 JPY.

Break-even growth (normalized EPS): 7.2% — requires sustained 7%+ earnings CAGR from normalized earnings base to justify current price. Break-even growth (forward EPS): 29.8% — entirely unreachable with forward EPS of only 49 JPY.

The only scenario where SUMCO is fairly valued is if you use normalized earnings AND believe in sustained 7-10% EPS CAGR from cycle recovery. This is plausible (wafer duopoly, AI tailwind), but it is a growth-stock thesis, not a Graham defensive thesis.

Per shr-003: trailing vs forward P/E gap is vast (N/A vs 68x). The market is not pricing a profitability inflection at low P/E — it is pricing 2027-2028 recovery at an already-elevated multiple.

shr-020 Red Flag Check #

Active reds: 4 of 10 categories. Most damaging is valuation — stock has already run 127% and sits above every analyst target.

Verdict #

SUMCO is a high-quality structural asset (global #2 silicon wafer, duopoly with Shin-Etsu, 50%+ combined 300mm market share) in a cyclical trough that is partially clearing. The thesis hook is correct: capex is past peak, utilization is recovering for leading-edge, and the 2027-2028 FCF inflection is plausible.

But the thesis has already been priced in.

The stock was a strong buy at 937 JPY (Jan 2026 trough). It is now at 3,348 JPY — a 257% rally from that low, putting it 37% above the analyst consensus target of 2,448 JPY. Graham IV on normalized earnings requires 7.2% perpetual growth to justify current price (no margin of safety). On forward EPS, break-even requires 30% growth — structurally impossible with a duopoly wafer company.

Graham filter score 1/7 is forgivable for a cyclical — but P/B at 2.06x is not trough P/B. Classic Graham cyclical entry requires P/B approaching 1.0x or below during the earnings trough. At 937 JPY, P/B would have been ~0.58x — genuinely Grahamian. At 3,348 JPY and P/B 2.06x, the market has fully re-rated it to a recovery story.

This is a stock to own at the trough, not after a 257% move from the trough.

VERDICT LINE: RED — Passed the price trough by 3,400 JPY. No margin of safety at 3,348 JPY. Watchlist for re-entry only if a second leg down returns P/B toward 0.80-1.00x (entry zone ~1,300-1,600 JPY). Revisit after Q4 2026 results when FY2027 FCF trajectory becomes visible. Not suitable for current EUR satellite portfolio at current price.