ASML reported record third-quarter 2026 net sales of €11.662 billion on October 14 and guided fourth-quarter sales to €14.2-15.2 billion. The midpoint of that range, €14.7 billion, is about 26% above the third quarter, at a gross margin that management expects to hold near 57%. Chief executive Christophe Fouquet added a sentence that matters more than the quarter: supply as currently planned for both 2027 and 2028, he said, will not fully meet the demand of ASML's customers. This piece works from ASML's own Form 6-K filed with the US Securities and Exchange Commission, not from secondary coverage.

The numbers

  • Q3 net sales: €11,662 million, of which installed base management, which ASML reports as a separate line, was €3,289 million.
  • Q3 gross margin: 57.5%, gross profit about €6,700 million. Net income: €4,960 million; basic earnings per share €12.95.
  • Systems sold: 113 new lithography systems and 8 used, against 86 new and 5 used in Q2.
  • Q4 guidance: net sales €14.2-15.2 billion, gross margin around 57%, R&D around €1.2 billion, SG&A around €0.4 billion.
  • Full-year 2026 guidance: net sales €44-45 billion, gross margin around 56%.
  • Buyback: about €4.1 billion of shares repurchased in Q3 under the 2026-2028 programme. Interim dividend of €1.88 per share, payable November 4, 2026.
  • Cash and short-term investments at quarter end: €12,735 million, up from €7,582 million at the end of Q2.

The filing does not disclose bookings for the quarter, which means the strongest claim in the release, excess demand, cannot be checked against an order figure. We return to that below.

What the arithmetic shows that the headline does not

The headline figure is the guide. Our calculation: the Q4 midpoint of €14.7 billion is 26.0% above Q3's €11.662 billion, and the range spans roughly 22% (€14.2 billion) to 30% (€15.2 billion). That follows a Q3 that was itself up 25.0% on Q2's €9.326 billion. Two consecutive quarters of about 25% growth is unusual for a company of this size and is the numerical expression of the capacity story.

The margin story is more subtle than it first looks. The Q4 gross margin guide of about 57% is slightly below Q3's 57.5%, so the guide implies growth without further margin expansion. But the margin did expand sharply in the previous step, from 54.0% in Q2 to 57.5% in Q3, a gain of 3.5 percentage points. A reasonable reading is that the third quarter captured the benefit of volume and mix, and that management is guiding to a plateau and not to a decline. The full-year margin guide of about 56% is consistent with that, since the first half was lower.

A rough decomposition also helps. Subtracting installed base management sales from total sales gives about €8,373 million for Q3 and €6,564 million for Q2, which approximates system sales. That is an increase of about 28%, against a rise in system units from 91 (86 new plus 5 used) to 121 (113 plus 8), or about 33%. Dividing out gives about €69 million per system in Q3 and €72 million in Q2. This is our back-of-envelope arithmetic, not ASML's disclosure, and it blends new and used systems and different tool types, so it should not be read as a price. It does suggest that Q3 growth came from volume, with the average revenue per tool roughly flat to slightly lower, probably reflecting product mix.

One further figure: net income of €4.96 billion against €4.1 billion of buybacks means ASML returned roughly 83% of the quarter's profit through repurchases alone, before the dividend, and still ended the quarter with €5.2 billion more cash than a quarter earlier. That is a company generating cash faster than it can find a use for it, which is typical of a supplier with pricing power and limited ability to expand supply in the short run.

Why ASML is the constraint

The filing itself does not explain ASML's role, so this section is general industry context, not drawn from the 6-K. ASML builds the lithography machines that print circuit patterns onto silicon wafers, and it is the only commercial supplier of extreme ultraviolet (EUV) systems, the tools used for the most advanced chip layers. The same company makes deep ultraviolet immersion systems used on many other layers. A leading-edge chip fab cannot be built or expanded without them, so the availability of ASML tools sets a ceiling on how fast the industry can add advanced capacity, regardless of how much money its customers want to spend.

The 6-K gives a useful clue about the nature of the constraint. Management said customers are raising short- and medium-term demand forecasts, and some want to place orders for 2029 and 2030. ASML previously planned a 30% capacity increase in 2027 for low-NA EUV and DUV immersion systems, with a further 30% under investigation for 2028. Fouquet now says supply as planned for 2027 and 2028 will not fully meet demand, and that ASML is investigating ways to increase 2028 capacity beyond the 30% it previously communicated. The wording is careful: capacity is being studied, not committed. Such tools are built from large numbers of specialised components and take considerable time to assemble and ship, so adding capacity involves suppliers as well as ASML's own factories.

Who gets allocation, and what customers can do

When a single supplier cannot meet demand, allocation becomes a commercial and strategic decision. The 6-K does not describe how ASML allocates slots, and we will not speculate on specific customers. What the filing does show is the direction of travel: customers are pushing orders further out, to 2029 and 2030, in an attempt to secure future slots. That is a bargaining-power shift toward the supplier. It supports pricing, helps explain the stable margin, and gives ASML unusual visibility into future revenue.

Customers' options are limited. They can order earlier and further ahead, accept slower expansion, push more output from existing tools, or invest in alternatives. For the industry, the practical consequence is that capital spending by chipmakers does not convert into new advanced capacity as quickly as budgets suggest. A fab operator may have the money and the building, and still wait for the tools.

The risks, including the counter-argument

The strongest argument against reading this as a pure win for ASML is that a sold-out supplier has an incentive problem for its customers. If delivery slots are scarce, customers have reason to over-order, placing orders for more than they need to secure a place in the queue and cancelling later. That is double-ordering, a familiar feature of shortages in semiconductors and elsewhere. It would make reported demand look stronger than true end demand, and the correction when it unwinds can be abrupt. We are presenting this as a risk, not a prediction: nothing in the filing indicates double-ordering is occurring.

The second risk is that the evidence for demand exceeding supply is largely management's statement, since ASML did not disclose bookings in the release. The CEO's words are specific and attributed, and the Q4 guide is a numerical commitment, but a reader cannot verify the 2027-28 claim from the filing. The third is cyclical: semiconductor equipment demand has historically followed customers' capital spending cycles, which follow end-market demand for chips, and ASML's visibility into 2029 and 2030 is based on customer intentions, not binding orders as far as the release shows. Export controls and trade restrictions are an additional general uncertainty for any supplier with global customers, and the 6-K excerpt we read does not address them.

A final measurement caution. The 26% figure is a step from a record quarter to a guided quarter, and company guidance is a forecast, not a result. ASML's history of guiding prudently is not something the filing speaks to, so we do not characterise it either way.

Bottom line

The quarter confirms strong demand for the tools behind advanced chips, and the guide extends it. What is new is the framing: ASML is telling investors the limit is its own output, through 2028. If that is right, the pace of AI-related chip capacity is set less by customers' budgets than by a single supplier's factory throughput. If it is exaggerated by defensive ordering, the shortage could prove shorter than advertised. The number to look for next, which the October release did not provide, is bookings.