Verified

  • ASML Holding N.V. (NASDAQ/Euronext: ASML) reported Q3 2026 total net sales of €11.662B, up from €9.326B in Q2, with a 57.5% gross margin and net income of €4.96B. The company called it "a record total net sales of €11.7 billion." Source: ASML Q3 2026 results press release, SEC Form 6-K exhibit.
  • ASML sold 113 new and eight used lithography systems in Q3, against 86 new and five used in Q2. Installed Base Management (service and field upgrades) was €3.289B of the total. Source: same.
  • CEO Christophe Fouquet, in the release: "In light of their demand visibility, some customers are expressing an interest in placing orders with us for 2029 and even 2030." And: "we now expect that our supply as currently planned for both 2027 and 2028 will not fully meet the demand of our customers." Source: same.
  • The plan that falls short is the one ASML set out in July: a 30% increase in low-NA EUV and DUV immersion capacity in 2027, and a further 30% under investigation for 2028. ASML now says it is investigating ways to go beyond that 30% for 2028. Source: same (footnote 2).
  • ASML guided Q4 2026 net sales of €14.2B to €15.2B and full-year 2026 sales of €44B to €45B. The release gives no bookings figure. Source: same.
  • ASML's Q3 investor presentation splits €8.4B of net system sales by technology: EUV 49% and ArF immersion (ArFi, the most advanced deep-ultraviolet tool) 39%. In Q2, on €6.6B of system sales, EUV was 57% and ArFi 29%. By unit, ASML sold 17 EUV and 40 ArFi systems in Q3, against 16 EUV and 23 ArFi in Q2. China took 23% of Q3 system sales by ship-to location, up from 14% in Q2. Source: ASML Q3 2026 investor presentation, page 9, SEC Form 6-K exhibit.
  • On the call, CFO Roger Dassen said ASML expects to ship roughly 65 low-NA EUV systems in 2026, "nearly 50% more than the 44 systems that we shipped last year." He said net income included about €500M from a higher valuation of ASML's investment in Mistral AI, and that ASML continues to expect the China region at around 20% of 2026 total net sales. Source: ASML Q3 2026 earnings call transcript, The Motley Fool.
  • Dassen said customers are asking ASML to accept purchase orders for 2029 and 2030, and that ASML has not yet accepted them while it compares total signaled demand with its own capacity plans. His reason: "we do not want now to accept you know, orders from customers A, B, and C, and then to tell customer d who has not, you know, given orders yet for 2029 and 2030. Sorry. We are sold out." Source: same.
  • Dassen said price talks with customers "have indeed led to pricing increases for what we call the most critical products." He said there is no 2026 impact, some impact in 2027, and a sustained impact from 2028. He described the down payment on an ASML order as one that is "enough to hurt." Source: same.
  • ASML plans to introduce the NXE:3800F low-NA EUV system, specified at 260 wafers per hour, in small quantities from early 2027, and a next-generation DUV immersion system running at 330 wafers per hour. Source: same.
  • Independent coverage matches the release: CNBC and Tom's Hardware, citing Reuters both report the 2029 and 2030 order interest and the 2027 to 2028 shortfall.

Beneath the label

The headline everyone ran is "record quarter, AI demand." Both halves are true. Neither is the useful part.

Start with what made the quarter a record. Net system sales rose by about €1.8B from Q2 to Q3. Using ASML's own percentage split, ArFi revenue went from roughly €1.9B to roughly €3.3B, and EUV from roughly €3.8B to roughly €4.1B. That is our arithmetic from rounded percentages, so treat it as a shape, not a line item. The shape is clear anyway: most of the step-up came from immersion DUV, not EUV. EUV units went from 16 to 17. ArFi units went from 23 to 40.

China's share of system sales moved from 14% to 23% in the same quarter. ASML still guides China to 20% of full-year net sales, so this may be timing. But a record quarter driven by immersion tools, with China's share rising, is a different picture from "EUV for AI accelerators."

Then read the guidance language. ASML did not say demand is strong. It said its own planned supply for 2027 and 2028 will not cover what customers want, after it had already planned a 30% capacity increase for 2027. That makes ASML, not the fabs, the stated constraint.

The third piece is the one with structure. ASML is choosing not to accept 2029 and 2030 orders yet. It has raised prices on its most critical products. It describes its down payments as painful by design. Those three moves together are what a supplier does when it stops taking orders in sequence and starts allocating.

The read

The verified facts, restated. Record sales, mostly from immersion DUV. Supply for two years out is already short by ASML's own account. Customers want to book a third and fourth year out. ASML is pricing up and asking for heavy deposits, while declining to lock in the outer years until it can see total demand.

The read: lithography has moved from an order book to an allocation book. Under an order book, whoever signs first gets the slot. Under an allocation book, the supplier looks at everyone's signaled demand and decides the split. Dassen said as much: ASML does not want to tell a long-time customer it is sold out because others signed earlier. The practical effect is that a chipmaker's 2029 and 2030 tool count now depends on ASML's view of its demand, not only on its willingness to pay.

What this would mean if it holds:

First, fab roadmaps become negotiations. A chipmaker planning a new advanced-node fab for 2029 cannot assume tools arrive on its schedule. Its tool allocation becomes a strategic input, like power or water, and is no longer a procurement detail.

Second, pricing power shows up in margin, not volume. ASML says the price increases bite from 2027 and fully from 2028. Hypothesis: if ASML's capacity grows about 30% a year and prices rise on the critical tools, revenue per tool rises faster than unit output. That is the signature of a supplier that is capacity-bound and knows it.

Third, the near-term capacity lever is the installed base, not new tools. Installed Base Management, which includes the productivity upgrades ASML is selling, was €3.3B in Q3. ASML says those upgrades provide "significant wafer capacity." Reading: until new capacity and the faster NXE:3800F ramp, the quickest extra wafer capacity likely comes from upgrading scanners that are already installed, because it needs no new cleanroom and no new tool slot. Fabs with large installed bases gain the most from that.

The limit on this read: the 2029 and 2030 interest is customer intent, not booked orders, and ASML gives no bookings figure. If the AI capex cycle turns before those orders are accepted, the queue shortens without anyone cancelling anything.

Watch

  1. Whether ASML starts accepting 2029 and 2030 orders, and how it describes the split. Any language about allocation, priority, or "fair share" confirms the read. A plain first-come statement would weaken it.
  2. Q4 2026 results in January 2027. The Q4 guide implies €14.2B to €15.2B. Watch the EUV versus ArFi split and the China share. If immersion keeps carrying the growth, the "EUV for AI" framing needs to be split in two.
  3. Down payments on the balance sheet. Dassen said non-current liabilities include customer down payments and are "a bit of a lagging indicator" of order intake. Rising deposits with no bookings disclosure is the clearest hard number available.
  4. The 2028 capacity number. ASML says it is looking beyond a 30% increase. A firm figure, with or without new tool throughput from the NXE:3800F, tells you how wide the gap is.
  5. ASML's Capital Markets Day on June 10, 2027. ASML says it will update its longer-term views there. That is the most likely venue for a 2029 to 2030 capacity plan.

If items 1 and 3 move together by mid-2027, lithography is an allocated input. If ASML accepts outer-year orders on a first-come basis and deposits flatten, this was a strong cycle, not a structural shift.

Sources