Deep DD Evolve — 2026-05-14

Stocks updated: 7004.T (Kanadevia), 6302.T (Sumitomo Heavy Industries), 6503.T (Mitsubishi Electric)

Research date: 2026-05-14 | Method: T-glass Supply Chain Bottleneck Analysis

Summary

Three stocks with oldest data_date (2026-04-12, never evolved) selected for deep evolution. All three had critical earnings catalysts in the past week — 7004.T had earnings announced May 12, 6302.T's Q1 results came out ~April 28, and 6503.T reported FY2026 record results April 28. Major new developments across all three.

7004.T — Kanadevia Corporation ⬆ UPGRADED

Previous conviction: LOW → New conviction: MEDIUM

What Changed

FY2026 earnings (reported May 12, 2026): Revenue hit a record ¥645.2B (+5.7% YoY) but operating profit collapsed 54.8% to ¥12.2B and net income fell 50% to ¥11.1B. The damage came from two one-time sources: (1) an unnamed overseas environmental EPC subsidiary accumulated technical cost overruns through Q1–Q3, and (2) ¥4.6B in extraordinary charges from the marine engine data falsification scandal and related restructuring.

FY2027 guidance is the real signal: Management guided OP ¥25.5B (+109% YoY), NI ¥21.0B (+89%), and raised the dividend 52% to ¥38/share from ¥25. The board's willingness to raise the dividend so aggressively signals genuine confidence that FY2026 losses were non-recurring. The stock surged +11% on May 13 to ~¥1,407.

Three structural upgrades since April 2026 data:

1. Nippon Steel Engineering merger MOU (Feb 5, 2026): Absorption merger with Kanadevia as surviving entity. Combined revenue >¥1T, creating Japan's largest engineering firm. Final agreement target: September 2026. Shareholder vote: November 2026. Completion target: April 2027. Decarbonization + environmental EPC focus aligns perfectly with Kanadevia's core.

2. TEPCO Kashiwazaki-Kariwa Unit 6 restart (Feb 9, 2026): TEPCO's first reactor restart since the 2011 Fukushima disaster. This is a landmark event for Japan's nuclear renaissance and directly increases spent nuclear fuel cask demand — Kanadevia's core business.

3. Portfolio cleanup accelerating: All-solid-state battery business transferred to Suzuki Motor (effective July 2026), marine engine stake diluted to Imabari Shipbuilding affiliate, valve subsidiary (VTEX) sold to KITZ. The conglomerate is clarifying its earnings profile around decarbonization + nuclear.

Updated Metrics (2026-05-14)

Supply Chain Verification

• ✅ TEPCO Tousou Mirai JV — confirmed active. Kashiwazaki restart drives cask demand. Plant operational status unverified (no public first-cask announcement found).
• ✅ NAC International — NRC Amendment 10 (Jan 2026). Selected as DOE ARPA-E SCALEUP subcontractor for Universal Canister System + deep borehole disposal ($40M program, Apr 7, 2026).
• ✅ Nippon Steel Engineering — merger MOU Feb 5, 2026. Confirmed. Source: argusmedia.com.

Risk Flags

• FCF guardrail triggered: FCF -¥22.6B in FY2026 (OpCF -¥3.5B + CapEx -¥19.1B). Override: FY2027 guidance implies strong recovery if one-time losses truly non-recurring.
• EV validation guardrail: Post-surge market cap ~¥237B. PE ~11x FY2027 guidance — reasonable but priced for delivery.
• EPC loss recurrence risk (AP01 analog): Q1–Q3 FY2026 EPS was deeply negative (¥-21.59, ¥-11.15, ¥-4.77) before massive Q4 swing (+¥103.71). Back-end loaded pattern.
• Marine engine compliance: Former chairman/president implicated in data falsification. Culture risk not fully resolved.
• Merger integration risk: 14-month timeline to April 2027 is compressed. Regulatory approvals required.

Inversion

Thesis breaks if: (1) Q1 FY2027 results (Aug 2026) show overseas EPC losses recurring — if OP is negative again, FY2026 was not truly one-time; (2) Nippon Steel Engineering merger is delayed past December 2026 or falls through; (3) Equity ratio deteriorates below 25%.

6302.T — Sumitomo Heavy Industries ➡ HOLD

Conviction: MEDIUM (unchanged)

What Changed

Q1 FY2026 results (Jan–Mar 2026, reported ~April 28): Revenue +5.8% YoY, OP +19.6% to ¥13.4B, NI +21.8% to ¥7.9B. Strong start to FY2026. Full-year guidance maintained: Revenue ¥1,090B, OP ¥60B, NI ¥34B.

Cryocooler recovery confirmed: Mechatronics segment OP surged +40% YoY in Q1, with SHI explicitly citing "semiconductor cryocooler demand recovery in US and China." New product launches in 2025 (CH-160D3 highest-capacity single-stage cryocooler, RJT-100 4K GM-JT). Global cryocooler market growing at 7.1% CAGR to $4.90B by 2030.

BNCT expansion: December 2024 orphan device designation for recurrent glioma and meningioma dramatically expands the addressable cancer indication beyond the original head/neck approval (2020). First overseas installation (China Hainan) targeted FY2025 — commissioning status unconfirmed but marks a commercial export milestone. Strong order backlog: ¥1,158.4B (+24% YoY).

New demand vectors: Liquid hydrogen zero-boil-off cryocooler (announced 2025, targeting hydrogen society), quantum computing pre-coolers (pulse-tube cryocoolers for dilution refrigerators, 5,000+ units/year needed by 2027), JAXA satellite applications.

Updated Metrics (2026-05-14)

Supply Chain Verification

• ✅ ITER Organization — construction ongoing (assembly phase, Apr 2026). First plasma revised to 2035. Cryocooler supply is part of ongoing integration.
• ✅ Semiconductor manufacturers (US/China) — Q1 FY2026 confirms active recovery. Cited explicitly in SHI Q1 results. Source: shi.co.jp/info/2026/...
• ✅ NeuCure BNCT hospitals — 3 Japan clinical sites confirmed. China Hainan in progress. Orphan device Dec 2024.
• ✅ KEK/J-PARC/JAXA — longstanding confirmed.
• ⬜ GE HealthCare/Siemens MRI — probable (world #1 cryocooler supplier, dominant MRI OEMs; no specific press release).

Risk Flags

• FCF negative (AP guardrail triggered): FCF -¥2.3B in FY2025 vs our prior +¥11B estimate. Override: elevated capex during mid-term plan investment phase; recovery expected if cryocooler demand sustains.
• Capacity expansion hangover (L&C segment): Logistics & Construction OP -45% FY2025, -40% Q1 FY2026. US tariff costs and crane demand normalization. Classic heavy equipment cycle hangover (AP02).
• Midterm plan shortfall: FY2026 OP guidance ¥60B is 25% below the mid-term plan target of ¥80B. Management credibility at risk if ¥80B target is revised down.
• ITER delay: First plasma pushed to 2035. Reduces near-term revenue urgency but does not remove the contract.
• PB guardrail: PB 0.96x is below book — opposite of guardrail concern. Not near cyclical peak valuation.

Inversion

Thesis breaks if: L&C segment OP remains negative for 3 consecutive quarters (structural deterioration, not cyclical); BNCT installs no new hospital contracts by December 2026; FCF remains negative through FY2026 full-year, forcing dividend cut.

6503.T — Mitsubishi Electric Corporation ➡ HOLD

Conviction: MEDIUM (unchanged, slightly strengthened)

What Changed

FY2026 record results (announced April 28, 2026): Revenue ¥5,894.7B (record, +6.8%), adjusted OP ¥538.4B (adj. margin 9.1% — record, vs prior 6.5% baseline), NI ¥407.7B (record, +25.8%). FY2027 guidance: Revenue ¥6,200B (+5.2%), adj. OP ¥590B (+9.6%). The margin expansion is structural — driven by portfolio restructuring, FA recovery, and defense segment growth.

Power semiconductor JV — decisive developments: MOU signed March 27, 2026 with Rohm Co., Toshiba Electronic Devices & Storage, Japan Industrial Partners, and TBJ Holdings to carve out all three companies' power semiconductor businesses into a new JV. President Uruma confirmed April 28: "We want to carve out the three companies' power semiconductor businesses and establish a joint venture." Critical unlock: Denso formally withdrew its competing Rohm acquisition bid on April 28, 2026 after failing to secure Rohm board endorsement. This removes the main obstacle. Combined entity would hold ~10-11% global power semiconductor market share, positioning as world #2 behind Infineon (~24%).

FA China recovery: Q1 FY2026 FA segment posted first profit increase in 9 quarters. Mitsubishi Electric established dedicated FA Business Headquarters in China (March 2025) and invested in ADT Technology Service (Suzhou) in January 2026 for MES/engineering solutions. FY2027 guidance explicitly cites FA systems as a key driver.

Defense: Japan's defense budget expansion (2% GDP target) driving defense electronics demand. FY2027 guidance cites defense systems as a primary growth area.

Updated Metrics (2026-05-14)

Supply Chain Verification

• ✅ Rohm + Toshiba power semi JV — MOU confirmed March 27, 2026. Denso withdrew April 28, 2026. Source: mitsubishielectric.com/en/pr/2026/pdf/0327.pdf
• ✅ Japan MoD / defense OEMs — Infrastructure segment. FY2027 guidance drives upward. Japan defense budget expansion ongoing.
• ✅ Data center optical devices — Growing revenue in Semiconductor & Device segment. New high-margin vector.
• ✅ Toyota — Long-standing keiretsu. Industry & Mobility segment.
• ⬜ Stellantis/Renault-Nissan — IGBT modules for EV inverters. Probable (global IGBT supplier position); no specific press release.

Risk Flags

• Peak earnings trap (AP01 triggered): FY2026 = record revenue + record margins + record NI. PE 32.85x on peak margins. FA and power semi are cyclical. Classic AP01 setup. Override: FY2027 guidance still shows +9.6% OP growth; not yet at plateau. Monitor closely.
• JV uncertainty: No definitive agreement, no valuation set, multi-jurisdiction regulatory review (China competition law critical). Deal could take 2-3 years or fail. Until definitive agreement, treat as optionality, not a catalyst.
• Restructuring cost: ¥100B "Next-Stage Support Program" (voluntary retirement) in FY2026 signals structural change; execution risk.
• Valuation stretch: PB 2.9x on a stock up 117-144% in 52 weeks. Much of re-rating captured.
• PB guardrail (triggered): PB 2.9x elevated for a partial-cyclical. Override: ROE now 10.1% (above Japan 8% TSE target), justifying premium. JV optionality adds further premium.

Inversion

Thesis breaks if: (1) JV definitive agreement not reached by December 2026 — if Rohm/Toshiba negotiations stall, the key catalyst evaporates; (2) FA China revenue growth reverses back to loss in Q2 FY2027 — one-quarter bounce, not recovery; (3) FY2027 adj. OP margin drops below 8% — structural margin improvement story is broken.

Sources

• Kabutan, IRBank, Nikkei (7004.T FY2026 earnings, May 12, 2026)
• SHI IR (6302.T Q1 FY2026, April 28, 2026): shi.co.jp/info/2026/6kgpsq000000odsm.html
• Mitsubishi Electric IR (6503.T FY2026, April 28, 2026): us.mitsubishielectric.com/en/pr/global/2026/0428_co/
• Mitsubishi Electric MOU press release (March 27, 2026): mitsubishielectric.com/en/pr/2026/pdf/0327.pdf
• Argus Media (Kanadevia merger, Feb 5, 2026): argusmedia.com/en/news-and-insights/...
• Japan Times (Denso withdrawal, April 26, 2026): japantimes.co.jp/business/2026/04/26/...
• ITER.org (construction progress, May 2026)
• Deep Isolation (NAC ARPA-E award, April 7, 2026): deepisolation.com/press/...
• Japan Nuclear Organization (Kashiwazaki restart): energi.media/news/...
• StockAnalysis.com (metrics, May 2026)
• Mordor Intelligence, GMInsights (cryocooler market size)

*Research conducted 2026-05-14. All financials from Tier 1-2 sources. Thesis status: 7004.T UPGRADED, 6302.T HOLD, 6503.T HOLD.*