ASML's Raised Outlook Confirms the Best Risk-Reward in Semis

Walk through the multiple compression mechanically. ASML reported FY2025 revenue of $32.7 billion and EPS of $30.51. The lifted outlook implies revenue tracking to roughly $40 billion in FY2026 (up 22% YoY) and $48-50 billion in FY2027 (up another 20-25%). On a 53% gross margin, gross profit lands at $25-26 billion in FY2027. Strip out operating expenses growing at 12% annually, taxes at 16%, and minority interests, and FY2027 net income lands at $17.5-18.5 billion. Diluted share count falls modestly on the buyback program, putting EPS in the $44-46 zone.

At today's $1,400 share price, that is a forward P/E of 31x on FY2027 earnings. By the time the FY2026 numbers print, the optical multiple compresses below 30x. By the time FY2027 prints, the optical multiple is 25-26x. That is the trajectory of a stock the market is currently calling expensive.

The single biggest mistake the bear case makes is treating ASML's revenue growth as cyclical rather than mix-driven. When DRAM enters a downcycle, ASML's revenue does not fall the way TSMC's or Micron's does. ASML sells the tools, and the tools are needed regardless of where memory pricing sits in any given quarter. Revenue volatility for ASML is dominated by customer capex timing, which has a different cycle than memory pricing. The HighNA EUV product line introduces a fresh source of revenue that is not cyclical at all; it is structural, locked to the leading-edge node migration of three customers (TSMC, Samsung, Intel).

Historically, the stocks that have re-rated on lifted outlook plus structural demand visibility have done so over the following 18-24 months. Lam Research's 2017 cycle is the closest analogue: a similarly demanding multiple compressed to a more reasonable one not because the multiple fell but because earnings ran past the multiple. ASML's setup today rhymes with that pattern, and the order book gives sufficient lead time to project the path with reasonable confidence.

ASML generated $10.6 billion of free cash flow in 2025 and ended the year with $12.9 billion of cash and zero net debt. That cash generation profile is the second pillar of the bull case. The company has the balance sheet to weather any single-cycle order delay, the dividend coverage to maintain a 2-3% yield through any downturn, and the buyback capacity to accelerate per-share value during the multiple-compression phase.

The historical capital allocation has been disciplined. Buybacks have been size-and-pace responsive to share price; dividends have grown at a low-double-digit CAGR. With $12.9 billion of cash on the balance sheet and trailing free cash flow at $10.6 billion, the return-to-shareholder envelope sits comfortably north of $7 billion annually without straining flexibility for selective M&A or capacity expansion.

The enterprise value of $558 billion against trailing free cash flow of $10.6 billion gives a free cash flow yield of 1.9%. That looks expensive in isolation. Forward-projected onto the FY2027 implied free cash flow of $14-15 billion, the yield rises toward 2.6%. Compare that to a forward implied compounding rate of 18-20% on revenue and 22-25% on free cash flow, and the yield-to-growth math works out favourably.

The operating margin trajectory deserves a final mention. ASML printed a 35% operating margin in 2025, up from 33% in 2024. The HighNA mix-shift adds 100-150 basis points of margin accretion over the next two years, meaning FY2027 operating margin should sit at 36-37%. Every 100 basis points of margin at FY2027's revenue base is worth $0.50 of EPS. The mix-shift contribution alone is worth $1.00-1.50 of incremental annual EPS, before the broader revenue inflection.

Net net, the financial profile (zero net debt, expanding margins, growing cash returns, conservative capex profile relative to revenue growth) is exactly the kind of profile that supports a premium multiple even in a market that is generally re-rating away from premium semiconductor names.

ASML's position in EUV lithography is often described as a monopoly. The more accurate description is that it is a near-monopoly that has taken three decades of R&D investment to construct, and the reconstruction cost for any new entrant is prohibitive on any credible time horizon. Canon and Nikon, the two other players in high-end lithography, have been out of the leading-edge EUV race since the mid-2010s. They continue to serve the 193i market but are not credible threats to EUV or HighNA over the next decade.

What that means for ASML's financials is that every dollar of leading-edge demand flows through ASML's income statement, and the pricing of the machines is set by ASML rather than by competitive dynamics. HighNA systems price in the $400 million range per unit; each one carries gross margins above the corporate average. As the HighNA share of total systems shipped rises from roughly 5% in 2025 to 20-25% in 2027, the weighted-average gross margin across the product portfolio accretes by 150-200 basis points. That is the source of the operating margin inflection baked into the FY2027 number.

The competitive moat is deeper than the headline numbers suggest. A single HighNA machine integrates over 100,000 components from 5,000+ suppliers. The supply-chain orchestration alone is a decade-long compounding advantage. Customers who have standardised on ASML's toolchain cannot credibly switch, and no new entrant can stand up the supplier ecosystem quickly enough to matter for the current cycle. That is why the lifted outlook reads as conservative: visibility into the FY2027 order book is an unusually high-quality forecast in this industry.

There is a historical parallel that the desk treats as informative without leaning on it too hard. The 1996 transition from 248nm to 193nm lithography produced a similar order-book inflection in the dominant lithography supplier of that era. Revenue growth ran at 25-30% annually for three years before the cycle cleared. The current EUV-to-HighNA transition is structurally similar in that it represents a node-driven capex acceleration that customers cannot defer, regardless of memory pricing or end-market demand cyclicality.

The difference today is that the customer base is substantially more concentrated than in 1996. TSMC alone represents over 35% of ASML's revenue base. Samsung, Intel and SK Hynix together account for another 35%. Customer concentration cuts both ways: when the leading-edge customers are accelerating their capex (the current state), revenue compounds faster than during the prior cycle. When they pause, the air pocket arrives faster too. The order book gives the visibility needed to know which side of that dynamic ASML sits on; today it is firmly on the accelerating side.

A second historical lens: the 2017-18 EUV ramp produced ASML's last comparable revenue inflection. The stock re-rated 75% over the following eighteen months as the order book turned into shipped revenue. The current setup looks similar in shape, with the additional kicker that HighNA represents a higher-margin product mix than the original EUV ramp did.

The technical picture confirms the narrative. The 50-day moving average at $1,397 sits meaningfully above the 200-day at $1,081. The spread between the two averages is the widest it has been in 18 months, and the slope of the 200-day is positive for the tenth consecutive month. Historically, when the spread between the 50-day and 200-day in a large-cap semiconductor name has exceeded 25% with a rising 200-day, the forward 12-month return has been positive in roughly 75% of cases across the 2010-2024 window. This is not a forecast; it is a regime indicator.

The 52-week high of $1,547 is the first short-term resistance. Above that, the technical gap to the prior all-time high opens clean air toward the $1,700 zone. The 52-week low of $608 is the downside reference; it would take an unambiguous order-book reversal to revisit that level. The implied volatility environment is unusually subdued for a stock with this much embedded optionality, which historically correlates with institutional positioning rather than retail froth. That is another signal worth noting.

Beta of 1.38 tells you ASML will move with the semiconductor index on any broad-market shock. Size positions with that in mind; the trade is asymmetric but not linear.