ASML examining ways it can make more than 110 EUV tools in 2028, JPMorgan says

ASML is evaluating production above 110 EUV lithography tools in 2028, according to JPMorgan following a meeting with the company’s CFO, as AI-driven chip demand accelerates. The company is nearly sold out of its 2027 EUV capacity, when it expects to produce at least 80 machines, and is targeting roughly 30% output growth in 2028. The potential capacity expansion supports a stronger multiyear demand outlook for ASML’s highest-value semiconductor equipment.

Analysis

The key incremental earnings variable is not unit volume alone but tool mix and supplier throughput. A 30-unit capacity step-up could imply roughly €6-9B of additional annual systems revenue at prevailing EUV price points, before service pull-through, but only if ASML can secure optics, laser-source and precision-stage capacity without expedited-cost inflation. This creates a 6-18 month validation window around supplier qualification, component lead times and customer deposit schedules; the market should not capitalize the full 2028 revenue opportunity until those constraints are independently evidenced.

TSMC is the best-positioned customer to absorb higher EUV availability because leading-node utilization and pricing power can support the associated capex. Intel and Samsung face a less favorable equation: additional access to tools raises the required capital intensity before their advanced-node yields and external foundry demand are proven. That divergence favors TSMC over Intel as the industry shifts from an equipment-availability constraint toward execution, yield and customer-demand constraints.

Consensus may underappreciate the strategic downside of added capacity: scarcity has historically protected ASML's pricing and customer prepayment economics. If capacity expansion is driven by a narrow group of AI-related customers and AI accelerator demand normalizes, ASML could carry excess dedicated supply into the next semiconductor downcycle, compressing both utilization and the scarcity multiple. Near-term upside is therefore likely modest without an order-book or margin-guide revision; the more actionable catalyst is evidence that 2028 incremental slots have been contractually committed rather than merely contemplated.

Key Decisions for Investors

• Maintain/accumulate ASML on 5-10% pullbacks rather than chase the headline; target a 12-18 month horizon. Thesis requires confirmed supplier-capacity investment, customer deposits and stable gross-margin guidance. Falsify if 2026-27 bookings weaken, EUV lead times shorten materially, or gross-margin outlook falls despite higher volume.
• Pair trade: long TSM / short INTC over the next 6-12 months. Greater EUV availability benefits TSMC through high-utilization leading-edge wafer pricing, while Intel bears elevated capex and yield-execution risk. Exit if Intel demonstrates sustained external-foundry wins and leading-node yield parity, or if TSMC's advanced-node utilization deteriorates.
• Do not initiate a direct semiconductor-equipment basket trade solely on this signal. Publicly traded ASML supply-chain beneficiaries such as BESI or ASMPT have indirect exposure, but the missing data are their specific content per EUV tool, incremental capacity commitments and pricing; monitor earnings calls for explicit order or capex disclosures before acting.
• Use any ASML strength ahead of its next results to reduce short-dated downside exposure via put spreads rather than outright shorts. The risk/reward is asymmetric because a confirmed 2028 customer allocation schedule could prompt a multi-year estimate reset, while an unconfirmed capacity aspiration is unlikely to support a durable rerating on its own.