Disco Corporation
Capital-equipment sales (dicing saws, laser saws, grinders, polishers -- ~64% of sales) plus a razor-and-blade recurring stream of proprietary precision dicing blades, consumables and service (~30% of sales); vertically integrated (in-house blades, air spindles) with Japan-based manufacturing. Public: TSE Prime 6146, OTC ADR DSCSY; founded 1937.
Earnings, margins, COGS & capex
High-margin, cyclical semiconductor back-end equipment leader riding an AI/HBM advanced-packaging up-cycle. FYE Mar-2026 was a record 6th consecutive year (net sales Y436.9B, +11.1%; net income Y135.5B); growth normalized from the +27.9% FYE Mar-2025 spike, but Q1 FYE Mar-2027 is guided to re-accelerate (+18% sales, +24% net YoY). Margins are structurally high (~42% operating, high-60s% gross est.) on pricing power, in-house consumables and a weak yen (the Q1 guide assumes Y157/USD); the balance sheet is a fortress (net cash, ~79% equity ratio). DISCO discloses guidance only one quarter ahead, given semi-demand volatility.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~37¢ is cost of goods and ~21¢ operating expense, leaving ~42¢ of operating profit (~31¢ net).
Revenue trend
Margins
elevated on consumables mix + weak yen
stable-high
stable-high
fortress balance sheet, net cash, high ROE
COGS structure
COGS is heavily precision manufacturing: in-house air spindles, diamond dicing blades and hones, granite/precision stages, laser sources for laser/stealth dicing, plus skilled Japanese labor. Vertical integration on blades and spindles protects margin; the razor-and-blade consumables stream carries higher margin than the equipment itself, lifting blended gross margin into the mid-60s%.
Capex
Sustained capacity + R&D investment (new works buildings, R&D center) to meet AI-packaging demand; est. ~Y40-50B/yr, ~10-12% of sales. Self-funded from operating cash flow; no reliance on debt.
Latest earnings
Record year, but growth normalized from the +27.9% FYE Mar-2025 spike to double-digit +11.1%. The forward tell is the Q1 FYE Mar-2027 guide (+18% sales / +24% net income YoY), pointing to re-acceleration on generative-AI packaging demand.
DISCO forecasts only one quarter ahead (stated policy, given drastic semi-demand swings) -- there is no full-year guide. Q1 FYE Mar-2027 (Apr-Jun 2026): net sales Y106.1B (+18.0%), op profit Y42.0B (+21.8%), net income Y29.5B (+24.1%), shipment-value forecast Y132B; assumed FX Y157/USD. Annual dividend FYE Mar-2026 set at Y505.
- FYE Mar-2026 net sales
- Y436.9B (+11.1% YoY, record)
- FYE Mar-2026 operating margin
- 42.3%
- Q1 FYE Mar-2027 guide
- sales +18% / net income +24% YoY
- Annual DPS
- Y505/share (from Y413)
Growth drivers
- AI compute -> advanced packaging (CoWoS, chiplets, 2.5D/3D) drives dicing + grinding tool demand
- HBM stacking requires ultra-thin wafer grinding + precise dicing -- DISCO's sweet spot
- Recurring high-margin consumables (proprietary dicing blades) grow with installed-base utilization
- Weak yen inflates reported yen sales/margins on largely USD-linked global demand (guide assumes Y157/USD)
- Power-semiconductor (SiC/GaN) and glass-core substrate processing as adjacent TAM
Bull & bear
The purest listed way to own the AI advanced-packaging supercycle at a near-monopoly choke point: every HBM stack and every 2.5D/3D package must be ground thin and diced, and DISCO sells the tools plus the recurring blades to do it -- at ~42% operating margins off a fortress balance sheet, with a record 6th straight year just printed and next quarter guided to re-accelerate.
- ~70-80% share in precision dicing plus grinding leadership gives pricing power and an unavoidable position in the AI packaging flow
- Razor-and-blade consumables (~30% of sales) are recurring, high-margin and far less cyclical than equipment
- Record 6th consecutive year (FYE Mar-2026 +11.1%) and a Q1 FYE Mar-2027 guide of +18% sales / +24% net income YoY -> re-acceleration
- Net-cash, ~79%-equity balance sheet + dividend raised to Y505 signal management confidence and a downside cushion
- HBM4 ramp, CoWoS expansion and glass-core substrates extend the runway of grinding/dicing intensity per chip
A monopoly you pay ~67x trailing (~55-60x forward) earnings for -- the multiple already prices a multi-year AI-packaging supercycle, so any capex digestion, memory air-pocket, yen reversal (the guide leans on Y157/USD) or long-run bonding-architecture shift hits both estimates and the premium at once.
- ~67x trailing / ~55-60x forward P/E (Y9.1T mcap vs Y135.5B net income) demands sustained high-teens earnings growth for years
- Growth normalized from +27.9% (FYE Mar-2025) to +11.1% (FYE Mar-2026) -- the supercycle's torrid first pace has already cooled to double digits
- Cyclical WFE exposure -- an AI-capex pause or HBM oversupply would cut equipment orders sharply
- Weak-yen tailwind is doing part of the work; the Q1 guide assumes Y157/USD, so yen strength would deflate reported sales and margins
- Hybrid/bumpless bonding and panel-level packaging could, over time, reduce dicing steps or reshape grinding demand
- China demand + export controls add a geopolitical tail; US access is only via an illiquid OTC ADR
What it is worth
Peer-relative P/E cross-checked with a reverse-DCF/earnings-growth sanity read. Mcap ~Y9.1T (~$58-60B; ~108.4M shares x ~Y84,500, Jul 2026) vs FYE Mar-2026 actual net income Y135.5B implies ~67x trailing P/E; against an est. FYE Mar-2027 net income (~Y155-165B -- no company full-year guide, DISCO forecasts one quarter ahead only, and Q1 is tracking +24% net YoY) that is ~55-60x forward. A large premium to WFE peers (TEL/AMAT ~20-30x, Tokyo Seimitsu ~20-25x), nearer high-growth back-end names (Besi ~35-45x). The premium reflects monopoly-like dicing/grinding share + AI leverage but leaves little room for disappointment.
AI-capex digestion or a memory downturn cuts EPS and compresses the multiple toward ~30-35x -> ~Y5.0-5.5T (~$33-37B); ADR toward the low-$30s (toward its 52-week-low zone).
Earnings compound but growth stays low-teens; multiple normalizes to ~50-55x -> roughly current ~Y9.0-9.5T (~$58-62B); ADR ~$49-55.
AI/HBM capex accelerates, the +18%/+24% Q1 guide extends through the year, multiple holds ~55-60x on upgraded EPS -> ~Y10.5-11.5T (~$68-75B); ADR upside into the high-$50s+.
At ~55-60x forward the price embeds a multi-year AI-packaging supercycle with sustained high-teens+ earnings growth and margins holding near peak; the debate is duration of the up-cycle and multiple durability, not near-term earnings quality.
SWOT
Strengths
- Near-monopoly in wafer dicing (~70-80% precision-dicing share) and the leader in wafer grinding -- category dominance
- Recurring high-margin consumables (proprietary blades tuned to its machines) smooth the cycle
- Best-in-class margins (~42% operating, mid-60s% gross) and a net-cash, ~79%-equity balance sheet
- Deep process know-how in ultra-thin wafer handling, now mission-critical for HBM stacking
Weaknesses
- Highly cyclical exposure to semiconductor/WFE capex swings
- Revenue increasingly levered to a single theme (AI/HBM advanced packaging) -- concentration
- Reported strength partly FX-driven (weak yen); yen appreciation is a real earnings headwind
- Thin US investability -- only an illiquid OTC ADR (DSCSY); primary liquidity is on the TSE
Opportunities
- HBM4/HBM4E ramp and TSMC CoWoS capacity expansion multiply thinning + dicing steps
- Glass-core substrates and panel-level packaging open new grinding/processing TAM
- Power semis (SiC/GaN) and hybrid-bonding wafer prep extend the addressable process set
- Consumables installed-base monetization grows structurally with every tool shipped
Threats
- AI-capex digestion or a memory downturn compresses both volumes and the premium multiple
- Long-run packaging-architecture shifts (hybrid/bumpless bonding) could alter dicing/grinding intensity
- China mature-node demand + export-control regime risk to a meaningful revenue slice
- New entrants/rivals (Tokyo Seimitsu, Besi) in adjacent back-end steps
Moats, dependencies & bottlenecks
Moats
Category dominance / installed base (~70-80% precision-dicing share, grinding leadership) Decades of leadership; tools embedded in customers' qualified process flows raise switching costs.
Razor-and-blade consumables (proprietary dicing blades/hones tuned to DISCO machines) Recurring, high-margin, sticky; ties consumable revenue to the installed base.
HBM-critical grinding/dicing recipes; hard to replicate, but architecture shifts are the long-run risk.
Controls the highest-value components -> margin protection and quality control.
Applications engineering / customer intimacy at leading-edge nodes Co-develops processes with foundries/OSATs/memory makers, embedding DISCO early.
Dependencies
Equipment orders swing with the broader capex cycle.
Demand concentration The current growth engine; a single-theme dependence cuts both ways.
HBM tool demand tracks memory-maker investment and the memory cycle.
CoWoS/chiplet capacity expansion drives dicing + grinding orders.
Macro/operational Weak yen is a reported-margin tailwind (guide assumes Y157/USD); appreciation is a headwind.
Mature-node/power-semi China demand exposed to tightening controls.
Advantages
- Monopoly-like share and pricing power in dicing + grinding
- Recurring, high-margin consumables cushion the equipment cycle
- Fortress balance sheet (net cash, ~79% equity ratio) funds capex and dividends through downturns
- Direct, high-purity leverage to AI advanced-packaging demand
- Structurally high margins (~42% operating, mid-60s% gross)
Weaknesses
- Deep cyclicality tied to semiconductor capex
- Premium valuation (~67x trailing / ~55-60x forward) vulnerable to any estimate cut or multiple compression
- End-market/customer concentration in AI/HBM
- Long-run risk that packaging architecture (hybrid bonding) reshapes dicing/grinding intensity
- Limited US investability -- illiquid OTC ADR only
Bottlenecks
- Own production/capacity ramp during demand booms (lead times on high-value tools)
- Skilled precision-engineering labor and R&D throughput in Japan
- Supply of specialized components (laser sources, precision stages) not made in-house
Top signals & trends
Top signals
Points to re-acceleration after FYE Mar-2026 normalized to +11.1%.
Management confidence in sustained cash generation.
More thinning/dicing steps per package.
New opportunity but potential long-run substitution of some dicing steps.
HBM oversupply would slow memory-maker tool orders.
Tightening controls threaten a revenue slice.
Weak yen flatters reported results; reversal is a headwind.
Trends
Core structural driver of dicing + grinding demand.
More grinding + precise dicing per HBM stack.
Adjacent opportunity, but could reshape traditional dicing/grinding steps long-term.
New processing/grinding TAM as substrates evolve.
The perennial cyclical risk to equipment orders.
China-facing revenue and supply chains at risk.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Linear-motion / precision guides for tool stages.
Sensors / machine-vision components.
Laser sources feeding laser/stealth-dicing systems.
Fiber-laser source supplier for laser-based processing.
diamond blades/hones) DISCO makes its highest-value components internally -- a key margin protector.
Leading foundry; CoWoS/advanced-packaging buildout is a primary demand source.
HBM leader; grinding/dicing-intensive stacking (Asia-listed; context only).
Memory + foundry HBM/packaging (Asia-listed; context only).
US HBM/DRAM maker scaling advanced packaging.
Advanced packaging (Foveros/EMIB) and foundry.
Largest OSAT; heavy dicing/grinding tool user.
Major US-listed OSAT customer.
End-demand driver -- GPU/HBM packaging pull-through via foundry/OSAT partners.
Closest direct rival -- competes head-on in dicing saws, wafer grinders and probers; the clear #2 to DISCO's #1 (processing-systems revenue a fraction of DISCO's).
Advanced-packaging assembly, and the hybrid-bonding leader -- the architecture that could reshape back-end process demand.
Broad WFE with a growing advanced-packaging portfolio; adjacent overlap rather than a direct dicing/grinding rival.
Bonding/assembly equipment for packaging; adjacent back-end competitor.
Assembly/packaging equipment (Asia-listed; named for context only, not a buy/own call).
Japanese WFE leader; more ecosystem peer than direct competitor in dicing/grinding.