ASML: cash surged, but how much can repeat?

The 60-second read

Five things to know

1. The operating ramp is real. Sales rose 55.2% and operating income more than doubled; the improvement survives removing the investment gain.
2. Surging cash is not all repeatable. Working capital supplied €7.009B, or 59.4% of operating cash; future delivery obligations matter.
3. Upgrades drove margin strength. Service and field-option sales reached €3.289B; gross margin exceeded the previous guidance ceiling.
4. The stock already prices growth. Our 2027 earnings route supports about $1,650, close to the pre-release quotation rather than far above it.
5. HOLD, with measurable checkpoints. Watch Q4 revenue of €14.2–15.2B, gross margin near 57% and cash conversion after timing effects.

Three scenarios, one probability-weighted number

Illustrative 12-month equity values; probabilities and assumptions are ours.

Signposts: what would change our mind

These are our review thresholds. January dates are monitoring deadlines, not asserted issuer release dates. Capacity and investment-gain persistence require qualitative review alongside the numbers.

The tape

A dated pre-release reference, not an invented earnings-day move.

ASML enters this report with substantial expectations already embedded in its share price. The reference quote is the October 13 Nasdaq close of $1,717.76. The collected daily-price series ends on that date, before the October 14 release. This packet therefore makes no claim about the stock’s reaction to the new quarter. The quote endpoint’s percentage change and the historical endpoint’s daily change are inconsistent; neither is used as a since-print return. A subsequent market move would change the entry valuation, not the accounting facts below.

The share price is a claim on future output rather than on the bank balance. Even a powerful quarter can disappoint a valuation that already assumes another large capacity expansion. We therefore separate three clocks: the September accounting period, the October publication, and the forward earnings used in our valuation. None can substitute for another. Historical earnings-day returns are descriptive context only; their signs do not determine our rating.

The print

Revenue accelerated; the quality of the acceleration matters.

Revenue reached €11.662 billion, up 55.2% year over year and 25.0% sequentially. Gross margin reached 57.4%, with operating income of €5.173 billion. This is meaningful operating leverage: sales grew faster than the combined research and administrative cost base. It is not solely a financial-investment story. Operating income itself more than doubled from the year-ago quarter.

The correct comparison with management’s previous promise is nuanced. July guidance called for third-quarter revenue of €11–12 billion and gross margin of 55–57%. Revenue finished within that range, above its midpoint; gross margin exceeded the upper end. Calling the entire report an unqualified revenue-guidance beat would overstate the result. Quarterly analyst estimates were unavailable from the collected provider endpoint, so we do not substitute the company’s guidance for Wall Street consensus or manufacture a consensus beat percentage.

The full-year revenue midpoint rises only modestly, while the gross-margin outlook strengthens. That distinction matters because annualizing the exceptional quarter would produce a different earnings story from management’s actual annual range. The next report needs both shipment execution and continued upgrade demand. A higher capacity plan is not itself recognized revenue, and customer interest in future machines is not a completed sale.

What customers bought

One reportable segment; no invented product profit allocation.

ASML reports one consolidated segment. Systems generated €8.373 billion, while service and field options generated €3.289 billion. Installed Base Management is the company’s label for that service-and-options category. It rose 67.6% year over year and represented 28.2% of total sales. The label should not be read as a pure subscription stream: productivity upgrades can be substantial, uneven transactions. Management attributes the margin outperformance to strong productivity-upgrade demand.

The system count increased to 121, including 113 new and eight used machines. That provides a physical cross-check on the revenue ramp, but it is not an average-selling-price model: technology mix, revenue recognition and upgrade sales differ across periods. High-NA EUV, conventional EUV, immersion and other tools have very different economics. The correct inference is stronger throughput, not a claim that every machine became more profitable by the same percentage.

The longer-term opportunity is the interaction between customer capacity and lithography intensity. More demanding chip designs can require additional exposures, and upgrades can add wafer capacity without a complete new fab build. The announced TSMC photomask initiative targets a productivity increase exceeding 40% for dual-mask exposure. That is a technology objective, not a realized increase in consolidated revenue or margin. We give it strategic relevance but no separate dollar value in the base case.

Follow the profit

Operating improvement is real; below-the-line gains still need separation.

Net income of €4.960 billion includes a €509.7 million gain on equity investments. That gain accounts for 8.9% of pretax income. The cash-flow statement reverses €508.8 million related to equity investments, supporting the conclusion that most of this contribution is noncash in the quarter. It should not be capitalized as recurring machine earnings. The €134.7 million equity-method contribution is another separate source of profit and should not be confused with cash dividends received.

For an illustrative sensitivity, applying the quarter’s effective tax rate to the investment gain would reduce net income to approximately €4.533 billion. That is our computation, not management’s adjusted EPS and not a claim about the gain’s actual tax treatment. The resulting profit would still be well above the prior year. The analytical conclusion is therefore two-part: the operating advance survives normalization, but the headline net-income growth exaggerates the recurring improvement.

The shrinking share base helps per-share results, but its contribution is small beside the operating change. Basic weighted shares fell about 1.1% year over year. Repurchases are a capital-allocation decision rather than an explanation for the doubling of operating income. Our forward model holds the current basic denominator constant; it does not assume that every remaining authorization will be executed at a favorable price.

Earnings quality

The same nine checks, with unavailable measures identified.

Operating cash flow reached €11.798 billion, but €7.009 billion came from changes in operating assets and liabilities. That is 59.4% of the cash inflow. Removing that contribution mechanically leaves €4.789 billion of operating cash and €4.341 billion after both forms of capital expenditure. These are diagnostic sensitivities, not alternative GAAP measures and not estimates of exactly repeatable cash flow. They show why annualizing the reported cash figure would be a serious mistake.

Long-term contract liabilities rose by €5.730 billion sequentially to €9.301 billion. This supports the interpretation that customers are helping finance future deliveries, but the summary does not provide a complete current contract-liability bridge. We therefore do not assert that every euro of the working-capital inflow was a customer deposit. Receivables also increased, and other liabilities can contribute. The liability represents an obligation to deliver future value; it is not free distributable profit.

The nine-quarter record reveals the seasonality: a large year-end inflow can be followed by a cash-consuming first quarter. The current trailing cash total is stronger than the annual history, but remains affected by customer payment timing. We do not compute a Beneish or Altman score from incomplete current liability and retained-earnings detail. A numerical score built from guessed inputs would add confidence without information. The useful questions are whether receivables convert, the advance-funded equipment is delivered, and margins persist when the cash cycle normalizes.

Balance sheet and capital returns

A strong liquidity buffer does not make a growth valuation cheap.

Cash and short-term investments total €12.735 billion, equivalent to about $14.17 billion at the selected exchange rate. That is only 2.1% of the quoted equity market value. The balance sheet supports resilience and investment; it does not provide a meaningful valuation floor near the share price. Loans to strategic partners and equity stakes are not treated as immediately available cash.

The latest summary aggregates current liabilities. Consequently, cash less disclosed noncurrent debt is only an upper bound on net cash. Our enterprise-value strip uses that bound and explicitly labels it as such; the actual enterprise value would be higher if additional current borrowing is present. More importantly, none of our three equity valuation routes adds gross cash on top of an earnings value. That avoids both an unsupported debt assumption and double-counting assets required to operate the business.

Third-quarter repurchases consumed €4.071 billion and dividends €721.3 million. The next interim dividend is €1.88 per ordinary share, payable November 4. The authorized 2026–2028 buyback envelope is up to €12 billion, but authorization is not a commitment to buy at any price. Capital return should be graded against the value of shares retired and the cash needed to expand manufacturing. Repurchases made during an exceptional cash-inflow quarter are not evidence that the same distribution pace is sustainable indefinitely.

Three routes to value

Our assumptions are explicit; the market already prices a strong ramp.

Our rating is HOLD, with fair value $1,650 against the dated $1,717.76 reference. The three routes span roughly $1,435–$1,695. We use the operating earnings route as the central anchor, cross-check it against cash generation and a later-cycle earnings case, and round rather than imply false precision. High uncertainty requires a 20% margin of safety: approximately $1,320 before we would consider a BUY with the thesis intact. That is a decision threshold, not a forecast that the stock must fall there.

At our 30× multiple, the reference price requires about €51.47 of earnings per share. At a 35% net margin and the same share count, that translates to approximately €56.3 billion of annual revenue. This is a reverse valuation, not a prediction. It illustrates that the price already assumes substantial growth beyond the €44–45 billion annual sales guidance. A strong semiconductor market alone is insufficient; ASML must convert that demand into profitable output.

The cash-flow route is deliberately less enthusiastic than the earnings routes. Its terminal value contributes approximately 82% of total value, making both the discount rate and the long-run growth assumption important. A permanent 4% terminal growth rate is already demanding for a large equipment company. The higher earnings values depend on investors continuing to award a premium multiple to the technology position. These methods share economic assumptions; agreement is a cross-check, not three independent proofs.

The bull case deserves respect: capacity remains scarce, productivity upgrades support margins, and the installed base can generate growth even when new-machine output is constrained. Our restraint comes from the price, the customer concentration and the time needed to deliver the expansion. Conversely, a HOLD is not a prediction of operating deterioration. It means the base case offers insufficient prospective return to justify a fresh purchase at this reference valuation.

Wall Street context

Pre-release targets are context, not an updated consensus on this print.

The collected target consensus is $2,330.75, with a $2,300 median and a $2,100–2,623 range. These numbers are materially above our value. The latest dated record is RBC Capital at $2,100 on October 7; July records include Argus at $2,100 and Wells Fargo at $2,500. None of those dates reflects today’s financial release. We will not call a target fresh merely because the data endpoint was fetched today. The difference in valuation is principally how much future expansion receives credit and what multiple that expansion deserves.

Annual provider estimates were collected with analyst counts, but the response does not declare a currency. Several net-income and EPS estimates also imply inconsistent share denominators. We preserve these observations as research inputs rather than silently converting them into a USD target. The base-case model is therefore independently specified in euros and translated once. Quarterly consensus access failed, so no quarterly consensus comparison is asserted. This limits the precision of a Street-beat narrative, not our ability to compare the actual filing with management’s own prior guidance.

Management scorecard

Execution improved; promises remain conditional.

Management earned credit for delivering within the top half of its sales range and exceeding the margin ceiling. The reported explanation is unusually useful: productivity upgrades lifted Installed Base Management sales and the gross margin. That gives investors a concrete mechanism to monitor rather than a generic AI demand narrative. The lower end of annual revenue guidance also rose. We grade those outcomes positively while keeping future capacity intentions separate from completed execution.

The annual report describes ordinary shares with one vote each and no issued cumulative preference shares. A preference-share foundation has a protective option, which matters for takeover governance even though it is not current ordinary-share dilution. Nasdaq valuation uses the ordinary-share economic interest without introducing an unsupported ADS conversion ratio. The auditor is PwC, whose annual opinion states that it has served since 2024. An audit opinion is assurance over historical statements, not insurance against a future capacity shortfall.

A particularly important supplier relationship is Carl Zeiss SMT. ASML owns 24.9% of its holding entity and relies on it for optical columns. The annual report classifies that entity as a variable interest entity but says ASML is not the primary beneficiary; it uses equity-method accounting. The relationship includes financing and prepayment obligations. It should therefore be understood as a concentrated strategic dependency, not described as either wholly owned supply or a nonexistent VIE exposure.

1. How much of the current working-capital inflow is refundable or tied to specific delivery milestones?
2. What portion of upgrade-led margin strength should persist when new-system mix changes?
3. Which supplier constraints govern the next increment of capacity beyond the existing plan?
4. How should investors distinguish expressions of interest from firm, executable orders?
5. What return threshold governs buybacks while capacity and strategic supplier financing also require cash?

Risks and counterarguments

Ranked by our assessment, not by false numerical precision.

The strongest bull argument is that conventional equipment-cycle valuation underestimates a durable bottleneck. If customers continue to increase lithography spending, upgrades monetize the installed base and capacity comes online on schedule, both earnings and the appropriate multiple could exceed our assumptions. The bull scenario does not require the current investment gain to recur. It requires stronger recurring sales and margins, with a market still willing to pay a premium for them.

The strongest bear argument is that today’s visibility reflects synchronized customer optimism near the top of a capital-spending cycle. Commitments can slip, export permissions can change, and machinery built for an aggressive ramp can leave inventory and cash tied up. A high-margin quarter followed by lower utilization would compress earnings and the multiple together. Customer advances soften near-term funding needs, but they do not remove the obligation to build and deliver the tools.

Our position sits between those outcomes. We accept the operating improvement and the strategic strength, but do not treat either as permission to ignore valuation. The wide scenario range is intentional: this is a high-quality business with uncertain timing and a demanding stock price. Position size, time horizon and existing exposure matter to an investor’s decision; the published HOLD is a company-level research judgment, not a personalized trade instruction.

What happens next

Dates are verified events or clearly labeled review checkpoints.

The most informative catalyst is the next set of financial statements, not the next headline about AI spending. We want to see the revenue ramp accompanied by the promised gross margin, a manageable receivables balance, and cash flow that remains credible after the timing benefit is stripped out. The company’s new annual framework can validate a higher revenue base, but a rising target without stronger execution evidence will not by itself raise our valuation.

The signposts near the top of this packet are deliberately numeric and dated. Green does not mean automatic BUY, and red does not mean automatic SELL; each triggers a reassessment of the assumptions. A lower share price with unchanged economics improves the prospective return. A lower price caused by weaker earnings power may not. That distinction is why the entry threshold is conditional on the operating thesis remaining intact.

Data, sources and method

Primary financial statements govern; provider mappings are checked.

All financial-statement amounts are euros in millions unless otherwise indicated. We compute growth, margins, cash conversion, per-share liquidity and valuation in code from the saved dataset. Free cash flow equals operating cash flow less both property/equipment and intangible purchases, matching the company’s disclosed definition. Provider cash-flow fields can exclude intangible purchases, so they are not substituted. Ending receivable and inventory days use a standardized 91.25-day quarter; they are monitoring ratios, not exact collection or manufacturing lead times.

The model translates euro earnings and cash flows to dollars at $1.11238 per euro, a collected October 14 FX observation. It is a valuation translation assumption, not the exchange rate used to prepare historical statements. All routes use 383.1 million basic weighted shares. FMP’s market-cap denominator can differ from that weighted share count, so market-cap multiples and per-share models are labeled separately. We do not back-solve an undisclosed current share count and present it as a filing fact.

Quarterly provider data lag the new release and include discrepancies in weighted shares, stock compensation and certain balance-sheet fields. Actual SEC tables supersede those fields. The current release provides one reportable segment and no category-profit allocation; missing disaggregated profit is explicitly unavailable. The historical reaction chart uses eight prior release dates independently established in the actual releases. The ninth, current session remains unfinished and is explicitly unavailable. A previous Q2 packet was not present, so no numerical prior signpost score is fabricated.

Glossary: Installed Base Management means service and field-option sales. Working capital is the operating asset and liability timing that separates profit from cash. Equity-method income is ASML’s share of an investee’s earnings, not necessarily a contemporaneous cash dividend. A terminal value estimates cash generation beyond the explicit model period. Margin of safety is the discount we require to our assessed value before a BUY. Scenario probabilities are subjective research judgments, not measured frequencies or guarantees.