ASML — Research Report (2026-07-12)
One-line verdict: the strongest bottleneck monopoly in the global economy — sole maker of the EUV machines every advanced chip needs — but after a ~140% ADR run in 18 months the price already pays for most of the 2030 bull case. Business 10/10; entry price offers thin margin of safety. Next checkpoint: Q2 2026 results July 15, 2026 (TechTimes).
Estimate labels used throughout: (reported) = primary source (ASML press release / SEC filing) · (secondary) = credible data provider or press · (est.) = my own computation, method shown.
1. Business overview
ASML (Veldhoven, NL; NASDAQ ADR + Euronext Amsterdam) sells photolithography systems — the machines that print circuit patterns onto silicon wafers — plus the service and upgrades on every machine in the field.
- What it sells: EUV scanners (0.33 NA "Low-NA" NXE, ~€180–220M each; 0.55 NA "High-NA" EXE, ~€380–400M each — TrendForce, Tom's Hardware); DUV scanners (immersion + dry); metrology/inspection; and Installed Base Management (service + field upgrades) — a recurring stream of ~€8.3B in 2025 (est.: €32.7B total sales minus €24.4B system sales) that ran at €2.49B in Q1 2026 alone (reported — Q1 2026 PR).
- 2025 system mix (reported — ASML 2025 annual report financials): EUV €11.6B (48 systems, +39% y/y), DUV €12.0B (279 systems, −6%), metrology €825M (+28%).
- To whom: the world's leading-edge fabs. TSMC is the largest customer at roughly a quarter of sales; TSMC + Samsung ≈ 38%, Intel third (secondary — TechMarketBriefs). Extreme customer concentration, but it is mutual lock-in: a monopolist selling to an oligopsony that has no alternative supplier.
- Geography: China was 33% of 2025 sales (36% in Q4) — all DUV/older tools, zero EUV ever shipped there — and ASML itself expects China to fall materially in 2026 as the pulled-forward backlog clears (reported via Caixin). Remainder concentrated in Taiwan, South Korea, US.
- Disclosure change that matters: starting 2026 ASML no longer reports quarterly net bookings — CFO Dassen calls them lumpy and unrepresentative (Moomoo, BusinessWorld). The market's best early-warning indicator is gone; expect bigger surprises in both directions.
2. Financials — FY2021–2025 (+ Q1 2026)
| €M unless noted (US GAAP) | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|
| Revenue | 18,611 | 21,173 | 27,559 | ~28,300 | ~32,700 |
| — growth y/y | +33% | +14% | +30% | +3% | +15.6% |
| Gross margin | 52.7% | 50.5% | 51.3% | 51.3% | 52.8% |
| Net income | ~5,900 | 5,624 | 7,839 | ~7,600 | 9,609 |
| EPS (basic, €) | n.d. | n.d. | n.d. | 19.25 | 24.73 |
| Operating cash flow | 11,593 | 9,435 | 6,536 | 12,659 | 13,827 |
| Capex (PP&E) | ~1,000 (est.) | 1,319 | 2,196 | 2,083 | ~2,700 (est.) |
| FCF (OCF − capex) | ~10,600 (est.) | 8,116 | 4,340 | 10,575 | ~11,100 |
| Net bookings | n.d. | 30,674 | 20,040 | n.d. | n.d. (Q4: 13,200) |
| Year-end backlog | n.d. | n.d. | ~39,000 | n.d. | 38,800 |
Sources: FY2025 ASML Q4-2025 PR and 2025 annual report 6-K (OCF €13,826.5M reported); FY2023/2022 Q4-2023 PR; FY2024 Q4-2024 PR; FY2021 PR; OCF/capex history 2021–24 (secondary — mlq.ai, Yahoo Finance); FCF 2025 €11.1B (secondary — Alpha Spread).
- Revenue CAGR 2021→25: 15.1% (est.); net income CAGR 13.0% (est.).
- EBITDA: not headlined by ASML. Est. FY2025 ≈ €13B (≈34–36% operating margin + ~€1.7B D&A), consistent with the ~43x TTM EV/EBITDA third parties compute (KoalaGains). Cash conversion (FCF/EBITDA) ≈ 85% (est.) — profit here is not an opinion, it converts.
- Cash flow reality: the 2023 dip (OCF €6.5B < net income €7.8B) was customer prepayments unwinding, not deterioration; 2024–25 OCF ran well above net income again as record Q4-2025 bookings (€13.2B, of which €7.4B EUV — reported) brought down-payments back in (interpretation, est.).
- Net debt: none — net CASH ≈ €10B (est.): €13.3B cash + short-term investments (reported — Q4-2025 PR) vs ~€2.7–3.7B gross Eurobond debt (secondary, sources mix currencies — Macrotrends, Finbox; end-2024 primary figure: €3,677M LT + €1,010M current, SEC interim report).
- Q1 2026 (reported): sales €8.8B, GM 53.0%, net income €2.8B, EPS €7.15; FY2026 guidance raised to €36–40B at 51–53% GM (from €34–39B in January) (Q1 2026 PR).
- One-offs flagged: €1.3B Mistral AI stake (Sept 2025) sits in investing outflows — a strategic investment, not capex (Bloomberg).
3. Unit economics — the number the business runs on
Blended EUV average selling price: ≈ €242M per system and rising (est.: €11.6B EUV revenue ÷ 48 systems, 2025). In 2021 it was ≈ €150M (€6.3B ÷ 42 systems, reported — FY2021 PR). +61% ASP growth in four years with zero competitive pressure — that is what monopoly pricing power looks like in one number. High-NA (~€380–400M/unit) mechanically extends the ASP staircase into 2030: 4 EXE systems recognized in 2025 vs 2 in 2024, and High-NA is expected to reach ~¼ of EUV revenue by 2028 (TrendForce).
Contrast: DUV ASP ≈ €43M (est.: €12.0B ÷ 279) — the competitive-ish, China-heavy end. And every machine sold feeds the ~€8.3B/yr installed-base service/upgrade annuity across a 20+ year field life (secondary — TechMarketBriefs).
4. Metrics that matter — and why, here
| Metric | Value | Why it matters for ASML specifically |
|---|---|---|
| P/E trailing | ~55–60x (secondary — KoalaGains; est. from TTM EPS ~€25.9) | Monopoly premium × AI cycle. Danger: P/E looks "justified" at cycle peaks — pair it with where fab capex sits in its cycle. |
| P/E forward (2026 guide) | ~52–54x (est.: mid-guide €38B rev × ~29% net margin → EPS ~€29) | You pay for 2033's earnings today; growth must not merely happen, it must beat what's already priced. |
| EV/EBITDA | ~43x TTM (secondary) | Debt-neutral whole-business price ≈ 4–6× the 7–12x PE-deal standard — the public market's scarcity premium on the only EUV asset on Earth. |
| Revenue CAGR 21–25 | 15.1% (est.) | Faster than the semi-equipment market → ASML's share of every fab dollar keeps rising (EUV intensity grows each node). |
| GPM slope | 50.5% → 52.8% (22→25, reported); 56–60% targeted by 2030 (Investor Day 2024) | Pricing power in motion: each tool generation sells for more AND carries higher margin. The slope is the moat's pulse. |
| FCF yield | ~1.9% (est.: €11.1B / ~€590B mkt cap) | The anti-story metric: at today's price you get <2% cash now — the rest of the return must come from 2030. |
| ROE | ~50% (est.: €9.6B net income on book equity shrunk by buybacks) | Reinvested euros compound at monopoly rates; ROIC over WACC by any measure — the moat in one number. |
| Net debt/EBITDA | negative (net cash ~€10B, est.) | Fortress: funds ~€5B/yr R&D (Q1-26 run-rate ~€1.2B/quarter, reported) through any downturn — R&D scale IS the entry barrier. |
| Shares outstanding | ~393M → ~389M (24→25, est. from reported EPS) | Silent buyback, zero dilution; the new €12B program through 2028 keeps the ratchet on. |
| Backlog/mkt cap | €38.8B ≈ 7% of mkt cap, but ≈ 100% of 2026 guided revenue (reported) | Next year is effectively pre-sold. Caveat: this was the last backlog disclosure ever — the metric dies here. |
5. Ownership, management, insider signals
- Ownership: ~100% free float, no controlling shareholder; institutions hold the large majority — BlackRock, Vanguard, State Street, Capital Group among the largest (secondary — TIKR). Takeover-proof in practice: ~€590B size, EU-strategic-asset status, and a Dutch foundation (Stichting) preference-share anti-takeover option described in the 20-F.
- Management: CEO Christophe Fouquet (since Apr 2024, ASML veteran, ex-EVP for EUV). Incentives (reported — 2025 Remuneration Report): base €1.125M; short-term cash bonus capped at 150% of base; long-term incentive up to 350% of base in performance shares vesting over 3 years plus a 2-year holding period — pay is overwhelmingly equity-linked and long-horizon. Aligned.
- Insider signal: Fouquet bought ~2,000 ADRs at ~$748–751 in Jan 2025 (secondary — TIKR) — small, but a buy at half today's price; no notable insider selling surfaced in this pass (Dutch AFM registry not systematically checked — limited visibility, flagged).
- Promise-keeping check: guided 2025 to ~€30–35B at ~52% GM in Jan 2025 → delivered €32.7B / 52.8% ✓. Warned honestly in July 2025 that 2026 growth wasn't guaranteed (DCD) → then guided up twice. Conservative-then-beat pattern: credible.
- Capital returns: 2025 dividend €7.50/share, +17% y/y; new buyback up to €12B through 2028; €1.1B repurchased in Q1 2026 alone (all reported — Q4-2025 PR, Q1 2026 PR). Combined shareholder yield ≈ 1.2%/yr (est.).
6. M&A track record
Disciplined, supply-chain-locking, integration-successful (company history, reported): Cymer 2013 (light sources — the EUV source problem bought and solved), HMI 2016 (e-beam metrology), 24.9% of Carl Zeiss SMT 2016 (the optics monopoly inside the monopoly), Berliner Glas 2020. The 2025 €1.3B for ~11% of Mistral AI (largest shareholder, board seat — Bloomberg, Techzine) is the first off-pattern deal — a strategic AI-sovereignty bet, not a tuck-in; small against €10B+ FCF, but watch for mission creep.
7. Market & value-chain position — bottleneck or commodity?
The purest bottleneck in the global economy. 100% share in EUV; no second source exists or is credibly in development this decade (Nikon/Canon exited the race; China's SMEE remains years behind even in immersion DUV — secondary, Asia Times). Every AI chip roadmap — TSMC, Samsung, Intel, and the DRAM/HBM makers — runs through ASML's installed base. The company's own 2030 model: revenue €44–60B at 56–60% gross margin, with double-digit EUV-spend CAGR in both logic and DRAM (reported — Investor Day 2024).
Who holds pricing power over ASML? Three parties, none commercial:
- Zeiss SMT — sole supplier of EUV optics, a bilateral monopoly ASML manages via its 24.9% stake and co-development; a partner, not a squeezer.
- Governments — the real boss. Dutch/US export controls already zeroed China EUV and cap DUV; 33% of 2025 revenue sits in a jurisdiction where policy, not ASML, sets the ceiling (Caixin).
- Customer concentration — TSMC at ~¼ of sales can delay orders (and has), but cannot substitute. Monopoly vs oligopsony ends in negotiated truce — and ASML keeps raising ASP per generation; that's the scoreboard.
8. Valuation view
Current price (secondary — StockAnalysis, Yahoo Finance): ADR ~$1,804; market cap ~$686–696B (≈ €590B, est.) as of early July 2026 — roughly 2.4× the ~$748 of Fouquet's January-2025 purchase.
Peer multiple range (secondary, mid-2026 — Quality Stocks, KoalaGains): trailing P/E — AMAT ~38x, KLA ~40x, LRCX ~48x, ASML ~55x+; ASML EV/EBITDA ~43x TTM. The whole semicap complex is inflated by AI-capex expectations; ASML carries the top premium on monopoly grounds. Peers aren't true comps — nobody else sells leading-edge litho — they're the sector sentiment gauge.
What would a reasonable buyer pay? A strategic acquisition is impossible (size + EU strategic asset + Stichting defense), so the "buyer" is a long-term investor underwriting the 2030 model (all est., assumptions shown):
- Mid scenario: €52B revenue at 58% GM → ~€22B operating income → ~€18B net income → EPS ~€49 on ~375M shares. At 27x (mature monopoly still growing) = ~€1,320/share in 2030; discounted at ~9%/yr ≈ €900–1,000 fair value today — 35–40% below the current ~€1,540 equivalent.
- High scenario (€60B / 60% GM): EPS ~€61; even at a 30x exit = ~€1,830 in 2030 → from today's price only a mid-single-digit IRR plus ~1% yield.
- Conclusion: today's price already pays for the 2030 bull case. A disciplined buyer pays €900–1,100; the marginal AI-momentum buyer pays $1,804. The gap is sentiment, not arithmetic. (Own estimate; the market has disagreed for a year and may keep disagreeing.)
9. PE/quality lens
- Moat: absolute — 100% EUV share, ~40 years of accumulated R&D nobody can shortcut, the Zeiss optics lock, a 20+ year installed-base annuity, and a hardware+software+service stack embedded in every leading fab's process.
- A buyer would love: the ASP staircase (each generation sells for 50–100% more), ~85% cash conversion, net cash balance sheet, a pre-sold year of revenue, aligned management, the service annuity.
- A buyer would fear: paying ~43x EBITDA for a still-cyclical business (2024 grew only 3%); geopolitics as permanent tail risk in both directions (China revenue loss / Taiwan conflict hitting the biggest customer); the new bookings-disclosure blackout; Hyper-NA R&D bets whose payoff sits past 2035.
10. Bull / base / bear (vs ~€1,540/sh · $1,804 ADR today)
Bull: the AI capex supercycle runs through the decade, DRAM adopts EUV at scale, High-NA ramps at Intel 14A and TSMC; 2030 lands at €60B/60% GM → EPS ~€61 with the monopoly multiple holding >30x → ~€1,900–2,100 by 2030. Note: even this mostly defends today's price rather than beating it (est.).
Base: 2026 delivers the guided €36–40B; growth digests toward the €52B 2030 mid-case; the multiple drifts from ~53x toward ~35x as growth normalizes → price roughly flat over 4 years; return ≈ EPS growth minus multiple compression plus ~1.2% yield → low-single-digit IRR (est.).
Bear: 2027 becomes a digestion year (litho is late-cycle; AI datacenter pauses reach foundry capex with a lag) just as China falls from 33% toward the low-20s%; EPS stalls near €29 and the multiple reverts to the low-30s → ~€870–950/share, about −40%, without the franchise being impaired at all (est.).
11. Catalysts
- Q2 2026 results — Wednesday, July 15, 2026 (3 days out): first read on H2 momentum with no bookings number to anchor on (TechTimes).
- TSMC's High-NA insertion decision (A14) and Intel's 14A ramp in 2027 — validates the €380M+ ASP tier.
- DRAM/HBM EUV layer-count expansion (Samsung, SK hynix, Micron capex announcements).
- US/Dutch China policy moves — tightening cuts revenue now but deepens the moat narrative; easing does the reverse.
- Buyback execution pace under a high share price — watch whether management keeps buying at 50x+ earnings.
12. What would prove the thesis wrong
The quality thesis breaks if: EUV revenue recognitions fall for 2+ consecutive years; TSMC publicly extends Low-NA multi-patterning instead of inserting High-NA; gross-margin guidance drops below 51%; SMEE ships production-worthy immersion DUV at scale (moat erosion in the China DUV base); or Mistral-style capital allocation grows from €1.3B into a pattern. The valuation caution (my actual conclusion) breaks if: 2026 revenue prints near €40B and 2027 is guided up again — the cycle not pausing — while the multiple holds above ~45x for another year. In that world the market's price was rational and the disciplined €900–1,000 entry never arrives.
Prepared per equity-research workflow §3/§3b/§4 · primary sources: ASML press releases + SEC 6-K/20-F filings · all estimates labeled · not financial advice; feeds the investment-research loop before any position.
Note: saving this to investing/ASML-2026-07.md and updating the vault logs was blocked by file-write permissions in this session — the report exists only in this output. DONE (report) / BLOCKED (vault save).