Web Research
The Bottom Line from the Web
The web reveals two facts the filings soften: China's domestic chip-equipment substitution surged to 35% in calendar 2025 from 25% a year earlier — exceeding Beijing's own 30% target — and Chinese authorities now reportedly require local sourcing for at least 50% of equipment in new capacity additions, a quantification that sits uncomfortably with Disco's undisclosed China exposure (sell-side estimates 20–30%). At the same time, Besi's hybrid-bonding momentum is no longer a slide-deck threat: Q1 2025 orders rose 8.2% YoY to €131.9M driven by hybrid bonding, Applied Materials took a 9% stake in Besi in 2025, and Besi's Investor Day raised long-term hybrid-bonding revenue targets to €1.5–1.9B with 40–55% operating margins — the architectural shift that would compress wafer-thinning steps per HBM stack is being capitalized in real time. Set against this, the FY2026 print (year ended March 2026) was a record sixth consecutive year of profit growth, with revenue ¥436.9B (+11.1%) and op income ¥184.9B (+10.9%), and analyst consensus PT ¥77,150 sits 21% above the May 18, 2026 close of ¥63,500.
What Matters Most
Close 2026-05-18 (¥)
Consensus PT (¥)
▲ 21.4% vs close
P/E (TTM)
1. China domestic substitution hit 35% in 2025 — the biggest under-disclosed risk just got sized
China's domestically manufactured share of semiconductor equipment used in-country rose from 25% in 2024 to 35% in 2025, beating Beijing's own 30% target, with etching and thin-film deposition substitution already above 40% and AMEC's 5nm etcher entering validation on TSMC advanced lines. By end-2025, Chinese authorities reportedly required at least 50% local-equipment sourcing for new capacity additions. Disco does not disclose China revenue, but Asia ex-Japan is roughly 75% of sales and sell-side estimates put China at 20–30%. [TrendForce, 2026-01-12]
Material: every percentage point of China share that flips to NAURA, AMEC, or a domestic dicing/grinding entrant erodes Disco's installed base — and consumables royalties — at compounding cost. The dicing/grinding category has been slower to substitute than etch/deposition, but the 50% new-capacity rule changes the marginal-tool decision regardless of technology readiness.
2. Hybrid bonding is no longer a slide — Besi's order book and AMAT's 9% stake price it in
Per a Jul 2025 SemiVision teardown, Besi's Q1 2025 orders rose 8.2% YoY to €131.9M, "largely driven by hybrid bonding orders," and Applied Materials became Besi's largest shareholder at 9% in 2025, deepening their D2W hybrid-bonding partnership. Besi's 2025 Investor Day elevated hybrid bonding to a core pillar with €1.5–1.9B long-term revenue targets at 40–55% op margins; the Financial Times projected hybrid-bonding revenues from €36M in 2023 to €476M by 2026 — roughly one-third of Besi's business. Samsung is expected to decide on Besi's hybrid bonding for HBM4 by mid-2026. [Substack SemiVision 2025-07-11; The Elec 2025]
The wallet-per-stack question is unresolved. Disco still owns pre-bond grinding to under 30 µm and post-bond dicing; but if Samsung commits hybrid bonding for HBM4 at scale, the count of grind/dice steps per stack falls relative to 16-layer microbump alternatives. Worth watching: Besi's mid-2026 Samsung decision is a direct read-through to Disco's HBM grinder content per stack.
3. CEO succession remains publicly unnamed — 17-year founder-family tenure, no designated successor
Per Disco's own Corporate Outline (last updated end-Dec 2025), Kazuma Sekiya remains Representative Executive Officer, President, and CEO with a 17.1-year tenure and direct ownership of 1.94% (~$861M at current prices). The remaining executive bench is thin: EVP Noboru Yoshinaga (3.9 yrs), CFO Takao Tamura (Managing Exec Officer), Managing Exec Officer Naoki Abe (Purchasing), and Yutaka Nishimura (Hiroshima Works). Chairman Hitoshi Mizorogi has held the chair role since the May 2017 reshuffle. No public successor has been named in any English-language IR material, Yuho summary, or analyst report surfaced by search. [Disco Corporate Outline; Simply Wall St Management; Disco News 2017-05-24]
Key-person risk in a ¥7T market-cap company. Founder family controls the role but no aggregate Sekiya-family stake (beyond Kazuma's 1.94% directly) was disclosed in any English search hit; the Yuho would be the canonical source.
4. Gohara Plant Phase 1 confirmed — ¥33B, construction starts Feb 2026, for consumables (not equipment)
Disco's April 17, 2025 press release confirms Hiroshima Works Gohara Plant Phase 1 on the former Kure City Sports Center site in Gohara-cho, Kure City: 11-story seismically-isolated steel + reinforced-concrete building, 13,179 m² footprint, 133,570 m² total floor space, ¥33B investment, construction start February 1, 2026. Production scope: precision processing tools (abrasive consumables — blades and wheels), not equipment. The plant is the first of three planned phases; phases 2 and 3 sizing is not yet disclosed. [Disco News 2025-04-18; MarketScreener 2025-04-17]
Bullish read: the largest single capex commitment in Disco's history is targeted at consumables capacity — confirming management's view that recurring-revenue blade/wheel volumes will scale with the installed base. Combines with existing Kure Plant and Kuwabata Plant to add geographic diversity (concern about both Hiroshima plants being ~10 km apart and exposed to common-mode disaster was explicitly raised in a 2021 disclosure).
5. BIS January 15, 2026 final rule reset — and AMAT's $252M penalty in February 2026
On January 15, 2026, BIS replaced its presumption-of-denial policy for advanced AI chip exports (Nvidia H200, AMD MI325X class) to China with case-by-case licensing under new compliance conditions. Crucially, on February 12, 2026, Applied Materials agreed to a $252M settlement with BIS — the second-highest BIS penalty ever — for illegal exports of semiconductor manufacturing equipment to China through 2020. [Visual Compliance 2026-02-11; BIS press release 2026-02-12]
Two-sided signal: licensing liberalizes for advanced compute, but enforcement against equipment makers escalates. Disco's back-end tools have historically been less restricted than front-end, but the AMAT case shows BIS is willing to pursue settlements years after the underlying conduct. No public BIS action against Disco was found.
6. FY2026 (year ended March 2026) was a record beat — but growth has decelerated
Disco's April 22, 2026 release reports consolidated FY2026 net sales of ¥436.8B (+11.1% YoY), shipment value ¥442.8B (+10.3%), operating income ¥184.9B (+10.9%), and net income ¥135.5B (+9.4%) — record highs across the board for the sixth consecutive period. Q4 EPS came in at ¥395.42 versus the ¥378.89 forecast (+4.4% beat). However, growth has slowed meaningfully from the five-year average of 21.1% to 9.4%. Op margin 42.3%, net margin 31.0%. FY2026 dividend ¥505 (record). [Japan IR FY2026 Financial Summary; tipranks; stockanalysis.com]
Beat is real, but Q1 FY2027 guidance is only one quarter ahead (¥106.1B sales) with management citing "market volatility" — a soft signal that visibility into AI capex through 2H FY2027 is thinner than the headline narrative suggests.
7. Bear case is anchored to a DCF that's roughly half the market price
Multiple aggregators (Perplexity Finance, Morningstar) cite an analyst DCF fair value of approximately ¥34,244 versus a May 18, 2026 close of ¥63,500 — i.e., the stock trades at more than 1.85× the most-quoted intrinsic value, with the bear thesis hinging on multiple compression as growth decelerates from 21% to 9%. The consensus 12-month target of ¥77,150 implies the sell-side discounts this bear DCF as out of regime for an AI capex-cycle name. [Perplexity Finance 6146.T; Yahoo Finance Analyst Targets]
8. Power semiconductors / EV demand explicitly flagged as headwind in the FY2026 release
Management's FY2026 commentary calls out: "power semiconductors were affected by a slowdown in EV demand," offset by data-center / generative-AI capex and a moderate recovery in PC/smartphone. The Q4 FY2025 earnings-call topics include "Impact of EV demand slump vs AI growth on sales" and "Factors for the YoY decline in power semi sales" — confirming the EV-related power-semi business is now in shrinkage, with the AI mix carrying the entire growth story. [Japan IR; Quartr Q4 FY2025 summary]
9. Stock has round-tripped — from ¥81,000 high to ¥22,640 trough to ¥63,500 today
The 52-week range is ¥31,890–¥81,000 (post-October 2024 3-for-1 split). Per the Dividend Hike Substack (Sep 2025), shares hit ¥22,640 in early 2025 before rebounding to ¥45,300 by September 2025. After peaking at ¥81,000 in late-February 2026, the shares pulled back ~22% to the ¥63,500 May 18 close; the stock is still up roughly 32% from the ¥48,170 year-end 2025 close but has lagged TOPIX since the late-February peak even as fundamentals beat. No single corporate-specific catalyst for the July 2024 / early 2025 drawdown was identified in search — the narrative is sector-wide AI-related derating plus a chip-cycle rotation. [MarketScreener TOPIX components; stockinvest.us; Dividend Hike 2025-09-17]
10. No fraud, no SEC action, no Stanford securities class action linked to Disco 6146
Forensic searches across litigation, SEC inquiry, short-seller report, accounting restatement, and class-action databases returned no results for Disco Corporation 6146. Search hits for "DISCO" in the Stanford Securities Class Action Clearinghouse map to CS Disco, Inc. (US legal-tech SaaS, NYSE:LAW) — a different company. Auditor relationship (KPMG AZSA) was not corroborated in English-language web sources. [Stanford SCAC; BIS press releases; SEC search]
Clean forensic record on the web. The absence of class actions and SEC inquiries — combined with no analyst short reports surfaced — reinforces the operational-quality read from the financial filings.
Recent News Timeline
What the Specialists Asked
Governance and People Signals
Leadership snapshot
Three observations from the leadership table that the filings don't surface:
- EVP Yoshinaga has only 3.9 years on the executive officer roster, despite being the most-senior non-family executive. The internal bench depth behind Sekiya is thin.
- The outside directors are mostly new — three of six have under 3 years tenure, with two under one year. The board has been refreshed recently, which is governance-positive but reduces institutional continuity in the event of a CEO event.
- Three female outside directors (Oki, Matsuo, Ahmadjian, Kobayashi — actually four) is strong by Japanese norms and signals genuine engagement with TSE Prime governance code.
Top institutional holders
Top six (including CEO direct stake) hold ~17.2% — a wide free float. No activist or strategic block above 5% other than Nomura AM's passive position.
Insider activity
No Form 4 / insider trading filings exist for Disco 6146 (no US listing other than thinly-traded ADR). Per Disco's February 17, 2026 release, the company filed a Notice of Partial Amendment to the Terms for Issuing Stock Options (Share Acquisition Rights) to DISCO's Executive Officers — terms not surfaced in the search snippet, but the amendment is the only insider-related corporate action in the trailing six months.
Industry Context
China domestic substitution — the dominant structural overhang
The single most thesis-altering web finding is the TrendForce / Jiemian / Commercial Times reporting (January 12, 2026) that:
- China's domestic equipment share rose 25% → 35% in calendar 2025
- Etching / thin-film deposition substitution >40%
- AMEC 5nm etcher in TSMC validation (signal that Chinese kit is no longer below frontier)
- NAURA order backlog booked through Q1 2027
- Local equipment order value +80% YoY
- End-2025 informal rule: at least 50% local-equipment sourcing for new capacity additions in China
Dicing and grinding remain a slower-to-substitute category — there is no documented domestic Chinese champion at Disco's precision tier — but the marginal-tool decision in new Chinese fabs no longer defaults to Disco. Combined with the BIS's revised January 15, 2026 rule (case-by-case licensing for advanced compute, with the AMAT $252M settlement showing escalating enforcement), the geopolitical context has hardened materially in the last 90 days.
Hybrid bonding capitalization
Hybrid bonding is past the "demonstration" phase and into "capacity capitalization":
- Besi Q1 2025 orders +8.2% YoY to €131.9M, hybrid-bonding led
- AMAT took 9% stake in Besi (became largest shareholder, deepening the 2020 D2W partnership)
- Besi Investor Day 2025: long-term hybrid-bonding revenue target €1.5–1.9B at 40–55% op margins
- FT projection: hybrid-bonding revenue €36M (2023) → €476M (2026)
- Samsung HBM4 hybrid-bonding decision expected by mid-2026
- Bonding pitches projected to compress from 40 µm to under 1 µm for chiplet-class interconnects
The mechanism that matters for Disco: hybrid bonding requires wafer thinning to under 30 µm before bonding (Disco-advantaged step), but reduces the count of dicing operations per stack relative to micro-bump-era HBM. The net per-stack wallet is therefore an empirical question — and the answer arrives in the back half of 2026 as Samsung's first hybrid-bonded HBM4 stacks ship.
Sector benchmarking — Disco is lagging peers in 2026
Disco rallied to a ¥81,000 late-February peak and has since corrected ~22% to ¥63,500, lagging TOPIX over the five months following the peak even as fundamentals beat. The market is voting with its feet on either the China substitution risk, the hybrid bonding architectural shift, or both — at a time when broader Japanese equities are working. This is the variant-perception signal that the headline EPS beat does not capture.