Shares fell 3% on a TSMC delay to high-NA EUV rollout. The real story is a process node workaround that bypasses the most profitable tool in the ASML catalogue.
ASML at 41x forward earnings is priced like the only game in town. And to be fair, in traditional EUV lithography, it is. The problem is that the premium multiple is not really about current-generation EUV. It is about high-NA EUV, the next-generation tool at roughly $380 million per unit that is supposed to drive the revenue step-up into the 2027-2029 window.
On 22 April 2026, TSMC delayed its high-NA EUV rollout. The same day, separate reporting described a TSMC process that delivers smaller, faster chips without requiring high-NA. ASML shares fell 3 percent. The market reaction was too mild. The implication is that the forward revenue curve the stock is priced on just got flatter, and probably for longer than one quarter.
Consider the high-NA math. Each high-NA EUV tool sells at roughly 2.3x the price of a standard EUV unit. ASML had guided the market toward shipping into the low double digits of high-NA units by 2027 and scaling from there. Every one of those tools carries gross margin several hundred basis points above the corporate average because the installed-base service component compounds at a premium. The entire second-half multiple expansion thesis is built on this.
Now consider the counter-scenario. If TSMC delays high-NA by twelve to eighteen months and Samsung follows, the 2027 unit count collapses from the low double digits toward zero. That is not a revenue rounding error. That is five to seven billion dollars of tool revenue moving out one or two years, with the associated service revenue pushed even further right.
The bulls will say the revenue is deferred, not lost. That is technically correct and analytically insufficient. Deferred revenue at a 41x multiple is worth less than current revenue at the same multiple because the discount rate applies. Push a revenue dollar out two years and at 10 percent equity cost of capital it is worth 17 percent less. That alone justifies a double-digit percentage re-rating lower.
The deeper issue is the TSMC process workaround. If TSMC is demonstrating smaller, faster chips without requiring high-NA, the medium-term demand for the tool is not deferred. It is structurally reduced. The industry has always overestimated the speed of lithography transitions, but it has rarely demonstrated a clear process node workaround that gets customers most of the way to the next node at a fraction of the capex. That is a new kind of risk.
TickerXray Report
Get the complete ASML report with all 12 quantitative models, AI-generated investment thesis, and real-time data.
ASML's 2025 numbers are excellent. Revenue up 15.6 percent. Operating income up 25 percent. Operating margin at 36 percent. Net income at $9.6 billion. A 41x forward multiple is extravagant on any ordinary business. It is justifiable only if the forward order book shows meaningful acceleration into 2027.
It does not. The current order book is dominated by standard EUV and DUV replenishment. High-NA orders have been trickling in at a pace that, before the TSMC delay, was running behind the guidance trajectory. After the delay, the pace becomes harder to defend.
Capex at ASML has also been volatile. It jumped from $900 million in 2021 to $2.1 billion by 2024 as the company built out high-NA production capacity. That capacity is now at risk of being underutilised in the 2026-2027 window. Underutilised capex translates directly to depressed incremental margins.
The bull case rests on two pillars. First, the demand for advanced lithography is structural and will eventually absorb high-NA even if the timing slips. Second, ASML has no real competition, so any deferred revenue eventually becomes ASML revenue.
Both are true in the very long run. Neither is true enough to justify 41x forward earnings in the next eighteen months. Pricing power does not rescue a multiple if the forward revenue curve flattens. Historically, semiconductor capital equipment stocks have derated sharply in the window between a deferred upgrade cycle announcement and the eventual delivery of that cycle. The KLA and Applied Materials drawdowns of 2018-2019 are the textbook examples. Both names lost 30-40 percent in that interval even though their long-term thesis stayed intact.
ASML at the current $1,450 area is pricing a high-NA ramp that TSMC just pushed out and that a competitive process may shrink permanently. The right fair value range is closer to 30-32x forward, or roughly $1,050 to $1,200. We see downside risk to $1,100 in the next twelve months, with a further leg lower possible if a second foundry customer signals a similar delay. We're trimmers above $1,400 and would get interested again closer to $1,100. This is not a permanent bear call on ASML. It is a call that the next eighteen months contain more downside than upside, and that the market has not finished pricing it in.
Full forensic analysis of ASML
+ 6 more models included
150,000+ stocks covered
Global coverage across 60+ exchanges. Every report includes all 12 quantitative models and AI analysis.
View plansEvery report runs 12 quantitative models and generates an AI investment thesis. From Piotroski scores to manipulation detection -- get the full picture in seconds.
12 forensic models
Piotroski, Altman, Beneish, DuPont & more
AI investment thesis
Synthesized outlook on every stock
Manipulation detection
Spot red flags before they hit the news
150,000+ tickers
Global coverage across 60+ exchanges
Expected return
Forward return projections for every stock
Real-time data
Live prices, insider trades, news sentiment
Free accounts get 1 report per month. Pro gets unlimited.
Three weeks after we flagged High-NA risk, the Q1 2026 backlog print arrived. The bookings number was stronger than feared but the customer concentration deteriorated meaningfully. The thesis updates rather than reverses; ASML remains the best risk-reward in semiconductor equipment but the multiple deserves a sharper haircut.
ASML trades at 41.5x forward earnings; Applied Materials at 36.5x. The premium has logic but not as much as the gap implies. One of these companies is materially mispriced.
At 41x forward earnings the multiple looks demanding. The order book and HighNA mix shift make 41x look like the floor, not the ceiling, of where this stock should trade.