Nova’s small beat hides a mix shift the Street is underpricing: margins held while logic/foundry did the work

The actionable read from this print is that the stock should not be judged on the +0.8% revenue surprise alone, because the real variant perception is in the quality of the revenue and the margin guide. What was priced in was modest upside to a $220.9 million Street revenue estimate and a $2.13 EPS estimate, and Nova delivered actual revenue of $222.6 million and EPS of $2.14. That is not enough by itself to force a broad estimate reset. What surprised was that the company sustained 57.6% GAAP gross margin in Q4 FY2025 after revenue slipped -0.9% QoQ, then pointed to approximately 58% non-GAAP gross margin for Q1 FY2026 while guiding non-GAAP EPS to $2.13 to $2.25. The stock reaction should therefore hinge less on whether the quarter was a beat, because +0.8% and +0.5% are not thesis-changing, and more on whether investors believe a logic/foundry-heavy mix can keep Nova inside its 57% to 60% gross-margin model while advanced-node intensity rises.

The reason that distinction matters is that Nova’s revenue path has already shifted from recovery to digestion, and the market may be misreading digestion as exhaustion. Revenue grew from $194.8 million in Q4 FY2024 to $222.6 million in Q4 FY2025, a +14.3% YoY increase, but sequential momentum cooled from +9.5% in Q1 FY2025 to +3.1% in Q2 FY2025, +2.1% in Q3 FY2025, and -0.9% in Q4 FY2025. That slowdown is visible and was likely embedded in a Street estimate of $220.9 million. The counterpoint is that gross margin moved from 57.3% in Q1 FY2025 to 57.8% in Q2 FY2025, 56.7% in Q3 FY2025, and 57.6% in Q4 FY2025, so profitability did not unravel as the topline flattened. In process control, that is the signal: a vendor exposed to leading-edge transitions can see lumpy tool timing without immediate margin compression if the application mix is valuable enough.

That financial trajectory also reframes the Q1 guide, because the guidance is not just a sequential revenue call, it is a margin credibility test. CFO Guy Kizner said the company expects Q1 revenue “between $222 million and $232 million,” with GAAP EPS from $1.90 to $2.02 and non-GAAP EPS from $2.13 to $2.25. He also committed to gross margin language that matters more than the range itself: “At the midpoint of our first quarter 2026 estimate, we anticipate the following: gross margins of approximately 56% on a GAAP basis and approximately 58% on a non-GAAP basis.” That wording matters because Q4 blended gross margin was 57.6% on a GAAP basis and 59.6% on a non-GAAP basis, so management is not promising a step-up in profitability despite a guide that holds revenue around the same zone. The thesis is not that Nova is accelerating sharply next quarter; the thesis is that the company is being paid for growth optionality while the financials increasingly look like a process-control franchise with margin support even during order timing pauses.

The product-mix detail is the best evidence that the margin support is not accidental. Kizner said product revenue distribution was approximately 75% from logic and foundry and 25% from memory, and that product revenue included 3 customers and 4 territories each contributing 10% or more to product revenue. That concentration cuts both ways, but the important point for portfolio managers is that the quarter’s 57.6% GAAP gross margin was achieved with memory only 25% of product revenue. If memory had been the hidden source of the beat, the read-through would be more cyclical and less durable. Instead, Nova is showing that logic/foundry demand can carry the revenue base while memory is a smaller contributor. The HBM exposure is real but not large enough to dominate the story: Gabriel Waisman framed it as “out of the 20% of product revenue that I indicated, about 1/4 to 1/3, depending on the quarter is high-bandwidth memory and the rest is logic.” The wording matters because it caps the HBM narrative; Nova is not simply an HBM derivative trade, and the print argues that leading-edge logic and foundry process complexity remain the more important source of sustained gross margin.

That mix is especially relevant for TSMC, the named customer read-through in the data pack, because Nova’s logic/foundry skew and geographic mix point to process-control content tied to leading-edge foundry rather than broad memory capex. The customer implication is not that TSMC’s capex is up or down, because this data pack does not give TSMC capex; it is that Nova’s approximately 75% logic/foundry product revenue mix and 29% Taiwan revenue share in 2025 are consistent with continued metrology intensity at major foundry customers. China is the other named regional exposure that portfolio managers should not hand-wave away: Kizner said China was 33% of 2025 revenue, Taiwan was 29%, Korea was 16%, U.S. was 9%, and other territories were 13%, while Waisman said China should “represent around 30% of our sales.” The second-order implication is that any tightening in China tool demand would matter directly to Nova’s revenue base, but the current quarter does not show a China-only dependency problem because product revenue also had 4 territories each at 10% or more.

The balance sheet and cash conversion are what make the China debate investable rather than binary. Kizner said Nova ended 2025 with more than $1.6 billion in cash, cash equivalents, bank deposit and marketable securities, and generated $218 million in free cash flow during 2025. Those figures matter because the annual revenue base was $880.6 million, up 31% year-over-year per Waisman, with annual operating margin of 29% on a GAAP basis and 33% on a non-GAAP basis. In other words, the company is not funding advanced-node participation by sacrificing cash generation or moving outside its stated operating-margin model range of 28% to 33%. The market often penalizes metrology names when sequential revenue decelerates because tool timing can signal a pause in customer ramps. Here, the 2025 cash and margin numbers argue the opposite: Nova has enough earnings quality to let investors wait for the next advanced-node revenue leg without underwriting a balance-sheet stretch.

The peer comparison reinforces why the right debate is multiple quality, not just next-quarter revenue. In the latest reported Process_Control peer set, NVMI shows $235.3 million revenue, 57.7% gross margin, and +10.3% revenue YoY. That places its gross margin below KLAC at 61.1% and 6861.T at 83.5%, but above LSRCY at 55.5%, ONTO at 50.1%, BRKR at 47.3%, TMO at 40.7%, and 6951.T at 41.9%. Revenue growth is also not an outlier at +10.3%, sitting below 6861.T at +17.9% and KLAC at +11.5%, but above ONTO at +9.5%, TMO at +6.2%, LSRCY at +5.4%, BRKR at +2.7%, and 6951.T at -17.9%. The portfolio takeaway is that Nova is not the fastest grower or the highest-margin company in the group, but the combination of 57.7% gross margin and +10.3% revenue YoY gives it a differentiated profile relative to smaller and mid-cap process-control peers. That is why a small Q4 beat can still matter: it supports a premium-quality argument if the market was bracing for margin slippage as growth normalized.

The call tone, however, gives one reason not to overpay for the beat immediately. The tone history shows Q4 FY2025 sentiment improved to 0.52 from 0.47 in Q3 FY2025, while uncertainty fell to 69.5 from 99.6 and qa_evasiveness fell to 72.2 from 104.3. That is cleaner delivery, not more promotional delivery. The conflicting signal is guidance_tone, which dropped to 0.39 from 0.68, and tone_confidence, which fell to 0.40 from 0.49. The interpretation is that management was more precise and less evasive, but less expansive about near-term guidance. That matches the numbers: Q4 beat the Street by +0.8% on revenue and +0.5% on EPS, but Q1 guidance of $222 million to $232 million does not ask investors to model a sharp sequential break higher. If the stock was priced for a larger forward guide, this is where disappointment would come from.

The gate-all-around discussion is the bridge between the muted guide and the longer-duration thesis. The transcript points to an accumulated revenue target of $500 million for 2024 to 2026, and Waisman said Nova is “on track with getting to the $500 million mark as an accumulated revenue for '24 to '26.” That quote earns attention because it is a milestone statement, not a generic demand comment. It gives investors a concrete yardstick for whether advanced-node adoption is converting to Nova revenue. The key is that the $500 million marker sits across 2024 to 2026, while this Q4 FY2025 print already shows annual revenue of $880.6 million and product revenue mix skewed approximately 75% logic/foundry. If gate-all-around revenue remains on track, then the current sequential deceleration from Q3 FY2025 revenue of $224.6 million to Q4 FY2025 revenue of $222.6 million is more likely a timing pause inside an advanced-node buildout than a sign that process-control intensity has peaked.

That said, the thesis would break if gross margin starts validating the bears’ timing concerns. The company’s own Q4 accounts showed GAAP EPS of $1.94 and non-GAAP EPS of $2.14, while the Street-comparison print uses EPS actual $2.14 versus estimate $2.13. Those are different bases and should not be blended, but both tell the same investment story: the quarter was only a slight EPS beat, and the burden of proof now moves to Q1 margin and revenue conversion. Operating expenses also need monitoring. Kizner said Q4 operating expenses were $67.5 million on a GAAP basis and $62 million on a non-GAAP basis, and guided Q1 operating expenses to approximately $65 million on a GAAP basis and approximately $60 million on a non-GAAP basis. That opex decline is constructive, but if revenue comes in near the low end of $222 million while non-GAAP gross margin slips below approximately 58%, the margin-defense thesis weakens quickly.

What to watch next is therefore specific. For Q1 FY2026, the confirming setup is revenue within or above the $222 million to $232 million guided range, non-GAAP EPS within or above $2.13 to $2.25, GAAP EPS within or above $1.90 to $2.02, and gross margin close to approximately 56% GAAP and approximately 58% non-GAAP at the midpoint. The thesis strengthens if China stays around 30% of sales without crowding out Taiwan’s 29% 2025 mix, if logic/foundry remains near the approximately 75% product-revenue level, and if management keeps the 2024 to 2026 gate-all-around accumulated revenue marker tied to the $500 million level. It breaks if Q1 lands at the low end of $222 million with gross margin below the approximately 58% non-GAAP guide, if customer or territory concentration rises beyond the current 3 customers and 4 territories at 10% or more of product revenue, or if the next call’s guidance_tone stays near 0.39 while tone_confidence remains near 0.40. The next quarter does not need a large revenue beat to validate Nova; it needs to prove that the Q4 message was real, that a 57.6% GAAP gross margin quarter with -0.9% QoQ revenue was not a one-off, and that advanced-node metrology content can keep earnings quality intact as the growth rate normalizes.