EUV Shipment Guidance: What the 60+ Unit Target Actually Signals
ASML's 2026 guidance of 60+ EUV system shipments represents a ~25% increase over 2025 delivery volumes — but the more analytically significant number is EUV revenue, which reached €4.1B in Q1 2026 alone, up ~28% year over year despite only a modest ~3% rise in unit volume [1]. That divergence between units and revenue is the central dynamic procurement and strategic planning teams should track: High-NA systems carry substantially higher average selling prices than conventional EUV scanners, and even two recognized High-NA units in a single quarter materially move the revenue line [1][2].
The guidance hike reflects two converging demand vectors. First, memory fabs — historically the later adopters of EUV relative to leading-edge logic foundries — are now a structural demand pillar rather than a cyclical supplement. DRAM manufacturers scaling to 1z nm and beyond require EUV for an increasing number of layers, and as HBM capacity expansion accelerates across SK Hynix, Samsung, and Micron, EUV tool allocation becomes a direct upstream constraint on AI accelerator supply chains. Second, AI chip demand is sustaining wafer start intensity at TSMC's N5-class nodes, where NVIDIA's Blackwell family — the B100 and B200 at ~1,600mm² dies and the GB200 Superchip at ~3,200mm² — consumes disproportionate EUV layer counts per wafer start relative to smaller dies at the same node [2].
For context on how wafer economics interact with these dynamics: TSMC N5/N4 wafers run ~$19k per 300mm wafer (range: $16k–$21k), while N3 sits ~$20k (range: $18k–$27k). The per-wafer EUV tool intensity — measured in passes through the scanner — is higher at N3 than N5, which is why a relatively flat unit shipment count can translate into a growing revenue line as node mix shifts toward more EUV-intensive processes.
For a more detailed treatment of how EUV ASP dynamics interact with fab capex strategy, see our prior analysis: ASML Pricing Power and EUV Economics: How Lithography TCO Is Reshaping Fab Capex Strategy.
EUV already accounts for ~47% of ASML Q1 2026 net system sales, with High-NA units amplifying ASP
Source: ASML Q1 2026 earnings, as reported in research source [1]
High-NA Lithography: The Ramp Economics Nobody Is Pricing Correctly
High-NA EUV (the TWINSCAN EXE:5000 platform) represents a meaningful step-change in both capability and capital intensity. The machines are physically larger, require more complex infrastructure, and carry ASPs estimated by industry observers to be meaningfully above conventional EUV — some public estimates place High-NA tools at roughly 2x to 3x the ASP of a standard EUV scanner, though ASML does not publish per-unit pricing.
The critical planning question for fabs considering High-NA adoption is not whether the tool delivers the patterning improvement — it demonstrably does for critical layers at sub-2nm design rules — but whether the economics of transitioning a production line justify the infrastructure overhead. High-NA has a narrower depth of focus, which creates yield management challenges for thick resist processes. Early customer deployments are accordingly oriented toward R&D and pilot-line work rather than high-volume production, which is why recognized High-NA revenue at two units in Q1 2026 is a data point about early adoption trajectory rather than a production ramp.
For foundry equipment allocation purposes, what matters is that High-NA and conventional EUV compete for the same installation and service engineering bandwidth within ASML's field organization. As the High-NA installed base grows, service revenue — already up ~25% year over year at €2.5B in Q1 2026 [1] — will compound further. Installed base management is becoming a larger share of ASML's revenue mix, which changes the long-run cost-of-ownership calculus for fabs that are already EUV customers versus greenfield entrants.
Geopolitical Constraints and Foundry Equipment Allocation Rebalancing
China accounted for over 40% of ASML's revenue during 2024–2025, a share driven by stockpiling behavior ahead of anticipated export-control expansion [4]. That cycle has turned. Chinese fabs that front-loaded DUV purchases and secured whatever EUV-adjacent equipment was available before controls tightened are now in an absorption phase — deploying and qualifying installed tools rather than acquiring new ones.
The geopolitical constraint on ASML is asymmetric in an important way: EUV systems have been export-restricted from China since 2019 under Dutch government license requirements, meaning the China demand that existed was always DUV-weighted. The incremental tightening affects DUV immersion systems at leading-edge dimensions, not EUV per se. The net effect for ASML's 2026 guidance is that management has stated total net sales will not fall below 2025 levels [4], implying that Taiwan, the United States, and Korea collectively absorb the China demand displacement — a reallocation that directly tightens equipment queues for those customers.
For TSMC, Samsung, and SK Hynix — the three customers with the most aggressive EUV roadmaps — this means lead times and allocation priority become more competitive, not less. Equipment lead times for EUV systems run in the range of 12–18 months from order to installation under normal conditions; as the queue tightens with demand consolidating among fewer geographically eligible customers, that figure warrants upward pressure in planning models. Procurement teams modeling AI accelerator production capacity through 2027 should treat EUV tool availability as a binding upstream variable, not a background assumption.
For context on how export control dynamics are reshaping capital allocation across the supply chain, see: Export Controls, Capex Reallocation, and the Rewiring of the Allied Supply Chain and China's AI Chip Bifurcation: How Export Controls Are Accelerating a Parallel Semiconductor Ecosystem.
Europe's EUV Vacuum: A Structural Problem, Not a Cycle Problem
The most analytically underweighted data point in ASML's 2026 demand picture is the company's own disclosure that it sold "absolutely nothing" in Europe in 2026 [5]. ASML is headquartered in Eindhoven; its most advanced tools are manufactured in the Netherlands; and yet no European fab is buying EUV systems.
The reason is structural. The current wave of European fab investment — including GlobalFoundries' Fab 1 expansion in Dresden targeting 22nm FD-SOI, eNVM, and BCD processes, and the STMicroelectronics/GlobalFoundries joint facility in Grenoble — is concentrated in mature and specialty nodes that do not require EUV lithography [5]. The European Chips Act was designed in part to attract leading-edge investment, but the economics of building a leading-edge fab in Europe without an anchor customer of sufficient scale have not closed. Intel's Magdeburg fab, which was intended to be a leading-edge EUV customer in Europe, has faced well-documented delays and scope uncertainty.
ASML's call for EU support to create demand for its advanced tools is not a sales pitch — it is a statement about the structural mismatch between European semiconductor policy ambitions and the actual investment decisions being made by fabs. EUV demand requires a customer willing to commit to sub-7nm production at scale, and no European fab is currently on that roadmap with an operational timeline that would generate near-term equipment orders.
The policy implication is that Europe's semiconductor sovereignty goals, to the extent they depend on indigenous advanced-node manufacturing, face a multi-year gap that cannot be bridged by subsidy alone without an anchor demand commitment from a hyperscaler, automotive OEM, or defense program at sufficient wafer volume.
China's ~40% revenue share in 2024–2025 has materially declined in 2026; Europe's EUV demand is effectively zero
Source: Research sources [4][5]; China figure is 2024–2025 average per ASML disclosures as cited in [4]
Memory Capex and the EUV Demand Floor
One dimension of ASML's demand resilience that deserves more attention is the memory fab capex cycle. DRAM manufacturers are scaling aggressively on HBM capacity — a direct consequence of AI accelerator demand — and HBM production at 1z nm and 1-alpha nm process nodes is EUV-dependent. SK Hynix and Samsung are both in active HBM4 qualification and capacity ramp cycles; Micron is investing in US-based DRAM capacity partly subsidized under the CHIPS Act. Each of these programs requires incremental EUV tool purchases.
The following table summarizes approximate wafer pricing at the nodes most relevant to leading-edge memory and logic EUV demand:
| Process Node | Fab | Approx. Wafer Price ($/300mm) | EUV Required? |
|---|---|---|---|
| N3 | TSMC | ~$20k ($18k–$27k) | Yes (multiple layers) |
| N5/N4 | TSMC | ~$19k ($16k–$21k) | Yes |
| SF3 | Samsung | ~$15k ($13k–$17k) | Yes |
| SF5 | Samsung | ~$13k ($11k–$15k) | Yes |
| N7 | TSMC | ~$10k ($8k–$11k) | Limited/optional |
| SF7 | Samsung | ~$8k ($7k–$9k) | Limited/optional |
The EUV intensity — number of scanner passes per wafer — rises sharply as nodes shrink below 7nm. This means that memory capex commitments at leading-edge DRAM nodes translate almost linearly into incremental EUV tool demand, providing a demand floor that is structurally less cyclical than pure logic foundry demand (which is more sensitive to consumer electronics end-markets).
For a deeper look at how memory capex dynamics interact with AI accelerator production economics, the HBM Market Analysis tool provides a live view of HBM supply-demand dynamics, and our analysis of SK Hynix and Samsung: HBM4 Readiness, 1c DRAM Scaling, and What the 2026 Capacity Race Really Means covers the specific capex commitments in detail.
Strategic Implications for Equipment Allocation and Planning Horizons
For procurement and strategic planning teams, the synthesis of ASML's 2026 posture yields several actionable observations.
First, EUV tool lead times should be treated as a long-duration variable in any AI chip production model. The combination of a ~25% unit shipment ramp, High-NA adoption absorbing installation capacity, and demand consolidating among fewer geographically eligible customers creates a tighter allocation environment than the headline unit guidance implies.
Second, the installed base management revenue line — up ~25% year over year at €2.5B in Q1 2026 [1] — is a leading indicator of EUV productivity at existing fabs. As fabs optimize throughput on installed scanners, the marginal wafer output per tool increases, which partially offsets the need for additional units. This dynamic is relevant to modeling when TSMC or Samsung will need to order the next tranche of tools versus when they can extract more from existing installations.
Third, the Europe situation is a warning, not an anomaly. Any region that lacks an anchor leading-edge fab customer will find itself structurally absent from ASML's order book, regardless of policy intent. The gap between fab subsidy programs and actual EUV demand is measurable in tool orders — and right now, that gap in Europe is total.
For readers modeling AI hardware supply chains end-to-end, the Fab Site Explorer provides node-level capacity and equipment allocation data that can be layered against ASML's shipment trajectory to stress-test production assumptions through 2027.
References & Sources
[1] ASML Q1 2026 Earnings Results — Net system sales, EUV revenue breakdown, High-NA unit count, installed base management revenue. Primary earnings disclosure.
[2] Silicon Analysts, ASML EUV Shipments 2026: Memory Fab Adoption, High-NA Ramp, and What the Guidance Hike Really Signals — prior analysis covering 60+ unit guidance context and memory as structural EUV pillar. [/analysis/asml-euv-shipments-60-units-2026-memory-demand]
[3] ASML press release, ASML reaches agreement for delivery of minimum of 15 EUV lithography systems — forward-looking statement disclosures regarding EUV order and shipment commitments.
[4] Trefis, What To Expect From ASML In 2026? — China revenue share (~40% in 2024–2025), management guidance that 2026 net sales will not fall below 2025 levels, demand absorption by Taiwan, US, and Korea.
[5] Tom's Hardware, ASML says it sold 'absolutely nothing' in Europe in 2026 — ASML's direct disclosure on zero European EUV demand; GlobalFoundries Dresden Fab 1 expansion details; absence of leading-edge EUV-eligible fabs in current European investment pipeline.
[6] ASML: The EUV Lithography Giant Navigating Challenges — geopolitical constraints on EUV shipments, China export restriction context, ASML competitive positioning.