How is Dutch EUV lithography monopolist ASML managing existing EUV systems being effectively sold out through 2027 and growing commitments for its roughly $400 million High NA machines, including its Eindhoven manufacturing expansion and JPMorgan reported target of 30% production

ASML is expanding capacity to meet strong EUV demand while helping customers move High-NA from qualification into production; the central challenge is converting orders into reliable, economical factory output.235 Europe’s weak equipment demand is a separate problem: owning a leading equipment supplier does not automatically translate into domestic chip-factory investment.48

Capacity and adoption

• Expanding supply: ASML broke ground on a new Eindhoven plant, while JPMorgan reported that it was targeting roughly 30% production growth in 2028 and exploring output above 110 EUV tools.237 “Nearly sold out for 2027” is more precise than completely sold out, and the reported 2028 capacity ambitions are not guaranteed deliveries.27

• Adoption is progressing at different speeds: Intel’s use on selected Panther Lake/18A layers represents actual production adoption, rather than only laboratory qualification; Reuters separately confirms that Intel is using High-NA in production while Samsung and TSMC have not yet adopted it for high-volume work.59 The reported million-wafer milestone indicates accumulated operating experience, but the testing-and-manufacturing total should not be interpreted as one million commercially saleable wafers.3

• Future commitments are not present output: The earlier retrieved GuruFocus report places Samsung’s planned DRAM adoption in 2028 and TSMC’s advanced-chip adoption from 2030.8 Taken together, these schedules imply a phased, application-specific transition—not an immediate replacement of conventional EUV across every customer or chip layer.5

Why High-NA must prove its economics

• Fewer patterning steps: The reported EXE:5200B capability—features approximately 1.7 times smaller than NXE can print in one exposure—could allow some patterns requiring multiple conventional-EUV exposures to use a single High-NA exposure instead.15 That is a resolution improvement, not a claim that entire chips become 1.7 times smaller or that production becomes 1.7 times faster.

• The relevant test is cost per good chip: At roughly $400 million per machine, the investment case depends on whether fewer processing steps, adequate throughput, availability and good-chip yield outweigh the higher equipment cost. Insufficient evidence in the retrieved material establishes that this economic threshold has been met across all three manufacturers. ASML’s work with customers on larger masks also shows that practical production constraints remain part of the adoption process.5

Europe’s manufacturing gap

• Heemskerk was criticizing inadequate investment: His explanation for currently selling no machines in Europe was a lack of investment and new fab construction, alongside stronger semiconductor investment elsewhere.48 This should not be broadened into a claim that Europe has no operating fabs or that ASML receives no European service revenue.

• Be careful with the regional metric: Earlier results described Europe’s second-quarter share as negligible or rounded to zero and identified South Korea and Taiwan as the leading markets, but described the measure as net system sales, not necessarily machine-unit deliveries.14 Insufficient evidence establishes the exact below-1% delivery share from a primary quarterly table.

• Policy implication: The mismatch illustrates the challenge for the European Chips Act’s production-share ambition: equipment-manufacturing expansion in Eindhoven is not itself additional European wafer-production capacity. The warning is about attracting and completing customer fabs, not ASML losing global demand.248

Why the shares could fall despite demand

• Market pressure and operating demand can diverge: The retrieved market report associated ASML’s Thursday decline with a broader semiconductor selloff driven by rising bond yields.13 Financially, higher discount rates can reduce the present value investors assign to future earnings even when orders remain strong.

• The 1.8% figure appears to be a trading snapshot: The earlier report quoted $1,713.41, down about 1.8%, whereas the explicit September 24 closing quote was $1,722.50, down 1.27%; those should not be presented as the same closing move.1516

• Valuation increases sensitivity, not proof of failure: The earlier GuruFocus report compared the trading price with a roughly $1,260 GF Value estimate, implying a premium of about 36%.8 That estimate is model-based, not an established intrinsic value. The investment concern is that disappointing throughput, yields or delivery timing could undermine optimistic earnings assumptions; insufficient evidence shows that an actual High-NA performance failure or delivery delay caused Thursday’s decline.