Not Nvidia. Not Micron. This Dutch Company Controls a Key Bottleneck of AI Chip Production.

ASML reported record third-quarter revenue of €11.7 billion, net income near €5 billion, and gross margin of 57.5%, up from 54% in Q2. Management says planned production capacity for 2027 and 2028 will fall short of customer demand, with some customers seeking delivery in 2029 and 2030; it is exploring capacity expansion. ASML forecasts 2026 revenue of €44 billion–€45 billion and gross margin around 56%, supported in part by €3.3 billion in Q3 installed-base services revenue.

Analysis

ASML’s scarcity value is real, but the investment signal is stronger for long-run pricing power than for near-term earnings acceleration. Customer capacity requests extending years out are not equivalent to firm orders or recognized revenue: tool complexity, supplier throughput, installation and customer fab readiness can all defer conversion. If ASML expands capacity to capture demand, the trade-off is higher execution and investment risk; watch gross margin and delivery performance rather than treating backlog commentary as automatically accretive.

The second-order exposure runs through customers’ return on capital. TSMC, Samsung and Intel may secure scarce tools, but delayed tool deliveries or weak fab utilization could strand other capacity and postpone advanced-node output. High-NA adoption is also a productivity-and-yield test, not merely a technology milestone; slower customer qualification would shift demand later without eliminating the long-term need. ASML’s service business should soften cyclicality, but cannot fully offset a pause in new-system spending.

Over days, bullish coverage may reinforce positioning; over 1–3 months, order conversion, capacity plans and margin guidance matter more. Over 6–18 months, the key question is whether AI-related fab investment becomes durable across customers rather than concentrated in forecasts. Contrarian risk: the market may price the equipment bottleneck as permanent while underweighting customer capex discipline, export restrictions and the time required to qualify new tools.

Key Decisions for Investors

• Prefer a staged ASML long on weakness over chasing an article-driven gap. The thesis is durable equipment scarcity and service attachment; avoid sizing it as a near-term AI revenue proxy. Reassess if order intake/backlog conversion weakens or delivery delays push out revenue.
• Track the next earnings updates for firm orders versus customer forecasts, 2027–28 capacity commitments, gross-margin trajectory and service growth. Treat capacity expansion as positive only if it does not materially dilute margins or impair delivery execution.
• Monitor TSMC, Samsung and Intel capex plans and High-NA qualification/productivity evidence. Broad customer deferrals, weaker fab utilization, or a shift in announced adoption schedules would undermine the multi-year demand case.
• No direct NVDA or MU hedge is warranted from this signal alone: ASML’s order visibility does not establish the timing or profitability of downstream AI-chip demand. Revisit relative exposure if customer capex plans diverge from AI infrastructure spending.