The setup
The Q2 dive was about a capacity decision. ASML said it would add 30% to low-NA EUV and DUV immersion capacity in 2027 and was studying another 30% for 2028, and the open question was whether customers would put money behind that capacity or only forecasts.
They put money behind it, or at least invoices.
The income statement is the less interesting half of the answer, and it was very good. Total net sales were €11.66B (~US$13.06B, +55.2% on the year), up 25.0% on the quarter and inside the €11.0–12.0B guide; gross margin was 57.5% against a 55–57% guide; net income was €4.96B (~US$5.55B), €12.95 per basic share. CEO Christophe Fouquet then went further than the July plan: “we now expect that our supply as currently planned for both 2027 and 2028 will not fully meet the demand of our customers.” ASML is “investigating ways to increase capacity for 2028 beyond the 30% increase we previously communicated,” and some customers want to place orders “for 2029 and even 2030.”
Management statements about demand are cheap. A customer paying, or being billed, for a tool it will not receive until after September 2027 is not, and the balance sheet says that happened at a scale ASML has not shown before. The ADS closed at $1,649.90 on October 14, down 3.95% from $1,717.76.
The numbers
| Total net sales | 11,662.0 · +55.2% YoY · +25.0% QoQ |
| Net system sales · Installed Base Management | 8,373.2 · 3,288.8 (28.2% of sales) |
| Gross profit · gross margin | 6,699.8 · 57.5% (Q2: 54.0%) |
| R&D · SG&A | 1,174.4 · 352.6 |
| Income from operations · operating margin | 5,172.8 · 44.4% (Q2: 37.1%) |
| Gain on equity investments | 509.7 (none in the four prior quarters shown) |
| Net income · basic EPS | 4,960.4 · €12.95 (Q3’25: €5.49) |
| Lithography units sold (new · used) | 121 (113 · 8) |
| Operating cash flow · free cash flow | 11,797.6 · 11,351 |
| Non-current contract liabilities | 9,300.6 (Q2: 3,570.3) |
| Share buyback · cash and short-term investments | 4,071.0 · 12,735 |
| Incremental gross margin QoQ (derived) | 71.3% |
| Incremental operating margin QoQ (derived) | 73.5% |
| Contract-liability increase as share of OCF (derived) | 48.6% |
Reported lines from ASML’s Q3 2026 Form 6-K (press release, US GAAP statements and investor presentation, ). The three copper rows are derived from those disclosures, not company-stated.
The operating line did more of the work than the sales line, and the reason is in the mix. Sales rose €2.34B on the quarter and gross profit rose €1.66B, an incremental gross margin of 71.3%; R&D actually fell, from €1.28B to €1.17B, so 73.5% of the extra sales reached operating income, which more than doubled on the year, up 109.6%.
The company says where the margin came from. Installed Base Management “came in above guidance … as a result of the continued strong demand for productivity upgrades. This drove the gross margin above guidance at 57.5%.” Decode that: IBM, the service and field-upgrade business, was €3.29B, up 67.6% on the year, and an upgrade sells wafer capacity on a tool the customer already owns. When the factory is the constraint, upgrades are the fastest supply ASML can ship. That makes the margin beat a symptom of the shortage, which is a better kind of beat than a pricing one — and also, under the house rule that mix is the softest margin to capitalize, one that should print twice before anyone extrapolates it.
Net income needs one adjustment. It includes a €509.7M gain on equity investments, 8.9% of pre-tax income, which appears in none of the four prior quarters shown; equity investments on the balance sheet went from €1.32B to €2.24B. ASML does not name the holding; one candidate is its Mistral AI stake (unverified — the release does not identify the holding). The 44.4% operating margin is the clean read. The 42.5% net margin is flattered.
The systems mix moved toward DUV. EUV was 49% of system sales on 17 units, against 57% on 16 units in Q2; immersion DUV (ArFi) was 39% on 40 units, against 29% on 23. China was 23% of system sales, up from 14% — on the rounded shares in ASML’s chart, roughly €1.93B against roughly €0.92B. Keep that number in mind for the counter-argument.
The rest of this dive is for paid subscribers.
The headline numbers are above. The 6 sections below carry the mechanism, the peer read, the valuation work, and the dated tests that decide it.
- What’s actually driving it
- How peers read this
- Valuation reality check
- What we’d watch from here
- The trade
- Bottom line
Already a subscriber? Sign in.
What’s actually driving it
The tell is one line: non-current contract liabilities. A contract liability is cash a customer has paid, or been billed for, before ASML delivers; the non-current part is what ASML expects to recognize more than twelve months after the balance-sheet date, which for a September 27, 2026 balance sheet means tools landing after September 2027. The line was €3.57B a year ago, €3.37B at year-end, €3.59B in March and €3.57B in June. In Q3 it went to €9.30B, up €5.73B (~US$6.42B), 2.6 times in one quarter.
The quarter-on-quarter jump is the headline; the series is the finding. A single quarter tells you a phase, not a decision, until you have the history, so the right benchmark is not June but the prior peaks. At year-ends from 2018 through 2025 this line topped out at €5.63B (~US$6.30B) in 2024, and the 2022 figure was €5.27B. Q3’s €9.30B is 1.65x the 2024 high. And this is only the long-dated portion — the current portion of contract liabilities is not split out in the quarterly statements, so the full prepayment build may be larger (current portion not disclosed quarterly). The release carries no bookings number. The bookings showed up anyway, one line further down.
Read it as a commitment is an allocation claim before it is a cost, from the other side of the table. A chipmaker that prepays for a post-2027 EUV or immersion slot is not mainly buying a tool. It is taking that slot away from everyone who did not prepay, and Fouquet’s statement that supply will fall short in 2027 and 2028 is what makes the slot worth paying for early.
The cash flow statement shows the same thing and needs the same care. Operating cash flow was €11.80B, of which €7.01B came from working capital, and the rise in non-current contract liabilities alone equals 48.6% of the quarter. Free cash flow of €11.35B (~US$12.71B) is mostly prepaid growth, a customer loan in a cash-flow costume: real money, earned on a schedule ASML has not met yet, and not a run-rate to capitalize. ASML spent €4.07B of it on buybacks and repaid €1.00B of debt in the same quarter. That is a company confident the tools will ship.
Memory is having its biggest year. Memory sales through nine months were €10.3B, against €8.6B in 2024, the best full year in ASML’s five-year chart; memory was 46% of Q3 system sales and logic 54%. The memory makers are buying lithography for capacity, not only for node shrinks.
The guide steps up again, and it needs the factory. Q4 is guided to €14.2–15.2B (~US$15.9–17.0B) at about 57% gross margin, with IBM around €3.7B. That leaves system sales of €10.5–11.5B, 25.4% to 37.3% above Q3’s €8.37B; the €14.7B midpoint is 26.1% above Q3 and 51.3% above Q4 2025. Full-year 2026 is now €44–45B (~US$49.3–50.4B), against €43–45B in July. The top end held. The floor rose.
Here is the bear argument at full strength, and it is not weak. Prepayments in a shortage are partly insurance — customers reserve slots because they fear being shut out, not because they have fabs ready for the tools — and ASML’s own balance sheet shows how that ends. This line fell from €5.27B to €4.83B in 2023 and from €5.63B to €3.37B in 2025, each time within a year of the peak, as prepaid slots were delivered and fresh orders slowed. Add the 23% China share, up from 14% in one quarter, and the Q3 jump could be a few customers front-running capacity and export rules at once, which is exactly the kind of demand that does not repeat. We still land on the long side, for one reason: the earlier peaks came without management saying two future years were already short, and this one arrived in a single quarter alongside exactly that statement. If that is right, the prepayments are a queue, and queues do not unwind while the factory is the constraint. What the print does not settle is how much of the €5.73B came from how few customers. ASML does not say, and nobody can tell from the outside.
How peers read this
For the rest of wafer-fab equipment — Applied Materials, Lam Research, KLA and Tokyo Electron — lithography is the first tool ordered for a new fab, so prepayments for post-2027 delivery are an early signal for etch, deposition and process control a few quarters later. A 46% memory share points in particular at the deposition- and etch-heavy memory spend Lam is most exposed to.
For the memory makers — Micron, SK hynix and Samsung — the read cuts both ways. Record memory litho spend supports the view that the current tightness lasts, but it is also the capacity that ends it once the tools run. South Korea fell from 43% to 31% of system sales on the quarter while China rose from 14% to 23%, so the Q3 tools went to a different mix of fabs than Q2’s.
For TSMC, Taiwan was 32% of system sales, up from 30%. The 12-inch High-NA photomask initiative announced with TSMC this quarter, which ASML says enables a more than 40% productivity increase for dual-mask exposures, says High-NA is being engineered into production flows rather than tested.
Valuation reality check
At $1,649.90 — about €1,473.78 at the ECB rate — and 383.1M weighted basic shares, ASML is worth roughly US$632.08B. Trailing-four-quarter basic EPS is €35.04, so the stock trades at about 42.1x trailing earnings. Annualizing Q3’s €12.95 gives €51.80 and about 28.5x, but Q3 carries the €509.7M equity gain, so the clean annualized multiple is somewhat higher.
A 3.95% fall on a quarter at the high end of its sales guide, with margin above it, a higher full-year floor and a statement that two future years are supply-constrained is the setup the library calls a record can sell off on expectations, not flaws. The business did nothing wrong. The price had already assumed a strong 2027, and the debate has moved from demand to conversion: how fast ASML adds output, and whether €9.30B of prepayments turns into revenue on schedule.
What we’d watch from here
- Non-current contract liabilities at Q4 vs €5.63B. The end-2024 high is the line between a standing order book and a one-quarter event. Above it, the prepayments are a queue; below it, the jump was a few customers buying slots. This is the test the page is written to (see the call below).
- The 2028 capacity number. ASML is “investigating” more than 30% for 2028. A firm figure by the June 10, 2027 Capital Markets Day would show the prepayments have turned into a supply plan.
- China share. China went to 23% of system sales from 14%. A sustained rise ahead of any export-rule change would look like pull-forward rather than end demand.
- The Q4 system-sales step. The guide implies €10.5–11.5B of systems. A miss against that would mean the ramp, not demand, is the constraint.
The trade
This is a long-biased view on a coverage name, kept evidence-led. At about 42.1x trailing earnings the stock already prices a constrained 2027, so the next leg depends on evidence that 2028 is constrained too: the contract-liability line holding and a firm 2028 capacity number. A holder who wants to stay exposed through that window while being paid for volatility could overwrite part of an existing position into the Q4 print, since the balance sheet published with it is the test. We have not priced specific strikes here.
The risk to the long view sits in ASML’s own filings: the two previous prepayment peaks, 2022 and 2024, each fell back within the following year, sharply after 2024.
Bottom line
Q2 announced the capacity. Q3 shows customers committing to it — €5.73B of new advances for tools due more than a year out, taking the line to 1.65x anything in the 2018–2025 series, while management says 2027 and 2028 supply will fall short. The stock fell anyway, because the question is no longer demand. It is conversion. Our call: non-current contract liabilities on the Q4 2026 balance sheet stay above €5.63B, the end-2024 high. Wrong if they fall below it.