# ASML Holding (ASML) — InvestMoat Analysis

_Last analyzed: July 16, 2026_
_Asset class: equity · Canonical page: https://investmoat.com/stocks/asml_

## Scores

| Dimension | Score (0–100) |
| --- | --- |
| Moat durability | 95 |
| Growth trajectory | 77 |
| Valuation | 68 |
| **Composite** | **80** |
| **Recommendation** | **Accumulate** |

Scores are computed deterministically from this asset’s data by the InvestMoat formula (see https://investmoat.com/llms.txt for methodology). Scores are not directly comparable across asset classes.

## Key stats

- **Ticker:** ASML
- **Market Cap:** ~$677B

## Moat

The only company in the world capable of producing EUV and High-NA EUV machines required for advanced chips below 7nm.

### The Strategic Bottleneck Moat

ASML is the **Sole Provider** of the world's most complex machines:

- **EUV and High-NA Monopoly:** Extreme Ultraviolet (EUV) lithography is required for every advanced chip from Apple, NVIDIA, and AMD. High-NA EUV — at $380M per system — extends this monopoly to 1.4nm and below, the only viable path to future AI silicon. ASML is the only company that can build either.
- **Technological Barrier:** Developing EUV took 20+ years and billions in funding from TSMC, Intel, and Samsung co-investment. A competitor would need at minimum a decade and the cooperation of the entire semiconductor ecosystem to come close — and still wouldn't have ASML's 30+ years of machine performance data.
- **Service Ecosystem and AI Demand Flywheel:** Once a machine is installed, ASML generates recurring service revenue for 20+ years. Q2 2026 Installed Base Management revenue hit €2.8B as the installed base expands and customers pay for high-margin performance upgrades to extract more output — the upside surprise in Q2 came primarily from these high-margin service sales. AI-driven chip demand accelerated capacity buildouts, prompting a second full-year 2026 guidance raise to €43–45B (from €36–40B in Q1, +16% at the midpoint); ASML is now nearly fully booked on Low-NA EUV for 2027 and is lifting Low-NA capacity ~30% in 2026, with another ~30% under investigation for 2028. Memory-related system sales are guided to grow >75% in 2026 as DRAM/HBM customers expand aggressively.

**Moat verdict:** ASML is the most AI-resilient business in the portfolio. Physical hardware monopolies, regulatory protection, and manufacturing expertise rooted in physics cannot be disrupted by software AI — and the AI infrastructure boom is actively accelerating demand for ASML's machines. High-NA EUV cements this monopoly for the next decade.

## Growth

Q2 2026 revenue of €9.3B beat consensus (~€8.8B, +21% YoY) with gross margin of 54.0% and net income of €2.9B (EPS €7.59), driving a second full-year guidance raise to €43–45B — a +16% lift at the midpoint from the €36–40B set only a quarter earlier, and roughly +36% YoY on 2025's ~€32B. Q3 2026 is guided to €11–12B at a 55–57% gross margin. Korea was the top market at 43% of Q2 system sales on the memory/HBM ramp; China held at ~20% for the year. The Q2 upside came primarily from high-margin Installed Base Management (€2.8B). ASML expects memory-related system sales to grow >75% in 2026, is nearly fully booked on Low-NA EUV for 2027, and is raising Low-NA capacity ~30% in 2026 with another ~30% under study for 2028. The counter-signal for the score: two guidance raises in six months pull the €44–60B 2030 target forward — 2026 already reaches the low end — so while near-term growth is explosive, the blended 3–5yr CAGR flattens off a much higher base and the 2027–28 digestion air-pocket, if AI capex plateaus, is now deeper.

- **Revenue CAGR estimate:** 13–18%
- **Primary type:** TAM expansion
- **Margin trend:** expanding
- **Key risk (high):** Hyperscaler AI capex digestion, now off a higher base: the two 2026 guidance raises lift the run-rate to a ~€44B/€11–12B-per-quarter pace, so if any two of Microsoft, Meta, Alphabet, Amazon, or Oracle cut FY2027 capex guidance by >15% in their H2 2026 earnings, TSMC and SK Hynix would defer EUV deliveries into 2027–28 against tougher comparisons — a deeper air-pocket than a quarter ago because more 2030 demand has been pulled forward. That would compress the forward multiple from ~40× toward the semi-cyclical norm of ~25×, a scenario made more probable by SOX trading well above its 200-day MA and CAPE at dot-com-peak readings
- **Drivers:**
  - EUV / High-NA + Capacity — Low-NA nearly fully booked for 2027; +~30% Low-NA capacity in 2026, another ~30% studied for 2028; 4–5 High-NA systems to be recognised in 2026 (Intel 18A milestone) (accelerating)
  - Memory / HBM — Memory-related system sales guided >75% growth in 2026 as DRAM/HBM customers expand aggressively on firm DDR/HBM pricing (accelerating)
  - Installed Base Management — €2.8B in Q2 2026; high-margin output upgrades compounding on the 20-year installed base and the main source of the Q2 beat (accelerating)
  - Non-China AI Capex — Korea 43% of Q2 system sales (memory/HBM ramp); China structurally reset to ~20% of the year under export controls (stable)
- **Score derivation:** Base ~80 (2026 revenue +~36% YoY but blended 3–5yr CAGR ~13–18% as the raised guidance pulls the €44–60B 2030 target forward, flattening the out-years off a higher base) + margin now expanding (GM 53%→54%, FY guide 55%, Q3 55–57%, tracking to the 56–60% 2030 target ahead of schedule) + accelerating trajectory (memory system sales >75%; 2027 Low-NA nearly fully booked) − high cyclicality/concentration risk (deeper 2027–28 digestion air-pocket from demand pull-forward; TSMC/Samsung/SK Hynix triopoly; China structurally ~20%)

## Valuation

At ~$1,823 (July 16, 2026, a fresh all-time high) ASML popped ~5% on the Q2 print — but that is a muted reaction to a +16% guidance raise, so on fundamentals the stock arguably de-rated: the earnings base rose faster than the price. Against scenarios lifted to reflect the €43–45B guide (base $1,950, bear $1,300, bull $2,750), the live price now sits between bear and base for a valuation score of ~68, modestly better than June's ~65 despite the higher absolute price. The margin of safety is partial rather than gone — the raised guidance and a 2027 order book that is nearly full restore some cushion — but the stock still discounts flawless execution of the capacity ramp while the AI-capex digestion risk stays live. Forward P/E of ~40× sits just below ASML's 5-year premium of ~45×.

| Multiple | Value | Note |
| --- | --- | --- |
| Trailing P/E (GAAP) | ~62× | TTM EPS ~€25 (post-Q2 2026) |
| Forward P/E (NTM) | ~40× | 2026 EPS tracking ~€34–35 on €44B/55% GM |
| PEG Ratio | ~1.8× | fwd P/E ÷ ~22% EPS CAGR |
| Price / Sales (NTM) | ~14× | ~€620B mkt cap / €44B NTM rev |
| Price / FCF | ~80× | ~€7.5B FCF (TTM est.) |

At ~40× forward P/E after the Q2 pop, ASML trades just below its own 5-year average of ~45× and at a premium to the semiconductor sector median of ~25×, reflecting the unmatched EUV monopoly. Because the €38B→€44B guidance raise outpaced the ~5% share move, the forward multiple actually compressed versus June — the PEG of ~1.8× is marginally cheaper than a quarter ago while still requiring ~22% EPS CAGR through 2027 to hold with no multiple compression, which the High-NA ramp, memory (+75%) mix, and a nearly-full 2027 order book make plausible. The gap between trailing (~62×) and forward (~40×) still signals sharp earnings acceleration as High-NA and memory contribute — structurally bullish, but offset by the rising probability of an AI-capex digestion event compressing the whole semi-equipment multiple toward the cyclical norm of ~25×, an air-pocket made deeper by the demand that has been pulled forward. _(as of July 2026 (price ~$1,823, post-Q2))_

## Price scenarios

### Bear — $1,300

AI capex plateaus into 2027 just as ASML ships at a ~€44B run-rate — the pulled-forward demand becomes a 2027–28 air-pocket, EPS stalls near €32 and the multiple compresses to ~28× on earnings uncertainty. Floor is higher than a quarter ago because the 2027 Low-NA book is nearly full.

- Hyperscaler AI infrastructure spending decelerates materially in H2 2026–H1 2027, reducing TSMC and SK Hynix capital expenditure plans and deferring EUV deliveries into a tougher post-raise comparison base
- US/EU tighten DUV immersion export controls further, removing the remaining ~20% China revenue with no near-term geographic offset
- High-NA EUV gross margins stay below 50% through 2027, capping blended gross margin near 54% and pushing 2027 EPS consensus toward €32 as the memory ramp normalises

### Base — $1,950

The raised 2026 guidance (€43–45B, 54–56% GM) is delivered; memory system sales grow >75%; the nearly-full 2027 Low-NA book and ~30% capacity add carry revenue toward €50B in 2027, sustaining ~22% EPS growth and a ~40× NTM multiple.

- 2026 revenue lands in the €43–45B guided range with gross margin at 54–56%; Korea/memory demand more than offsets the structural ~20% China cap
- High-NA scales toward 4–5 recognised systems in 2026 and a steeper 2027 ramp (Intel 18A in volume), improving ASP mix and pushing gross margin toward 56%
- 2027 revenue guidance of ~€48–52B issued at H2 2026 earnings, backed by the near-full Low-NA order book, the ~30% capacity expansion, and sovereign fab programmes (US CHIPS Act, EU Chips Act)

### Bull — $2,750

AI and memory demand sustain 20%+ EUV revenue growth into 2028; High-NA scales on Intel 18A / 14A and leading-edge logic; margin expansion above 57% drives an earnings re-rating toward 42× NTM as the €44–60B 2030 target is pulled forward to ~2028.

- Sovereign foundry programmes across the US, Europe, Japan, and Middle East add a new demand cohort on top of TSMC/Samsung/SK Hynix, pulling the €44–60B 2030 revenue target forward to ~€55B by 2028
- High-NA EUV reaches 20+ annual shipments by 2028 at improving unit margins; blended gross margin inflects above 57% ahead of the 2030 target
- AI accelerator and HBM demand for 1.4nm/1nm nodes materialises faster than consensus, doubling ASML's addressable High-NA opportunity and sustaining 25%+ EPS growth through 2028

---

InvestMoat is an open-source research and education framework. Nothing here is financial advice. Past performance does not guarantee future results.