The driver is unmistakable: AI.
The key insight: this surge is entirely driven by current-generation 0.33 NA EUV and DUV tools — the machines chipmakers need right now to meet AI demand. The next-generation upgrade cycle is a separate story.
In April 2026, TSMC — ASML's largest customer — dropped a bombshell. SVP Kevin Zhang told reporters the company has no plans to adopt High-NA EUV before 2029, bluntly calling the roughly €350 million per-tool price "very, very expensive" . TSMC believes its existing 0.33 NA EUV tools can still meet upcoming node requirements, including its A13 process
.
Then in August 2026, Samsung confirmed it would skip High-NA EUV entirely at its 2 nm and 1.4 nm nodes, targeting its first 1 nm-class node (SF1A) around 2030 for adoption . Samsung executives cited immature ecosystem readiness — photoresists, masks, and inspection — as the key reason
.
The critical nuance: This caution is not about ASML's current business. It is about the next upgrade cycle, which has now been pushed out to 2029–2030. ASML itself has said it expects "the first memory and logic products manufactured on High-NA EUV systems within the next few months" from other early adopters — likely Intel and memory manufacturers .
The Q2 2026 SEC 13F filings (quarter ended June 30, 2026) reveal a dramatic divergence among the largest institutional investors.
On the bullish side:
On the bearish side:
Interpretation: This is not a disagreement about the current business. Bulls like Arrowstreet are betting that AI-driven demand for current-generation tools will sustain growth and that High-NA adoption is simply a timing question. Bears like Maverick may be concerned about the High-NA overhang, peak-cycle valuation, or the multi-year narrative gap before the next major upgrade catalyst materializes.
Wall Street remains broadly constructive despite the High-NA uncertainty.
Share buyback: ASML announced in January 2026 a €12 billion share repurchase program through 2028 (~$14.32 billion), signaling management confidence in long-term cash generation . In Q1 2026 alone, ASML bought back roughly €1.1 billion of its own stock
.
The conflict is timeline-based. Analysts see the current AI-driven order book as strong enough to support near-term earnings, while the High-NA delay pushes the next major catalyst to 2029–2030, creating a potential air pocket in narrative momentum — but not in this year's financials.
| Factor | Bullish Signal | Bearish Signal |
|---|---|---|
| 2026 revenue guidance | €43–45B, second upgrade; record backlog | Guidance depends on sustained AI CapEx |
| High-NA EUV adoption | First High-NA chips from others imminent; TSMC/Samsung will eventually upgrade | Top two customers pushing adoption to 2029–2030, delaying the next big cycle |
| Institutional 13F positioning | Arrowstreet, Sei, T. Rowe, Invesco adding | Maverick Capital near-full exit |
| Analyst consensus | "Moderate Buy," average target ~$1,970; JPMorgan at $2,400 | 3 Sell ratings; one downgrade to Hold in mid-August |
| Capital returns | €12B buyback program through 2028 | — |
ASML's story in 2026 is a tale of two timelines: the AI-powered present, which is booming, and the High-NA future, which is delayed. Investors are placing their bets accordingly.