Critical Dimension
In this edition of the Smart Investor newsletter, we spotlight the company guarding the nanometer frontier. But first, let’s review the latest Smart Portfolio developments.
1
Portfolio News and Updates
❖❖ JPMorgan (JPM) is reportedly close to participating in the financing of Japan’s $550 billion U.S. investment pledge, a trade-linked commitment announced in 2025 to support U.S.-based industrial and infrastructure projects in exchange for lower tariffs on Japanese exports.
Under the 2025 agreement, Japan has already announced more than $100 billion of projects, including power plants and industrial facilities, while later plans include small modular nuclear reactors and additional gas-fired power facilities. Yet only about $2.2 billion in financing has reportedly been committed so far, because yen-funded Japanese lenders face major costs when raising long-duration U.S. dollar funding.
This is where JPM comes in: as the largest U.S. bank, it has the balance sheet, funding access, and structuring expertise to support complex cross-border project finance. The report reinforces JPMorgan’s position in high-value global banking – not just lending capital, but enabling large strategic transactions where trade policy, infrastructure, energy security, and institutional finance converge.
1
❖❖ NVIDIA’s (NVDA) continues to deepen its role as the hardware backbone of sovereign AI infrastructure. Japan is planning to buy 27,500 next-generation Rubin chips to support a national AI initiative focused on robotics. The project will be overseen by newly created Noetra, backed by ¥387.3 billion (approximately $2.4 billion) in government funding through March 2027, with Sony, SoftBank, Toyota-backed Preferred Networks, NEC, Fujitsu, and others helping build and operate the venture.
Noetra plans to launch an initial AI model by March 2027, followed by regular updates, with a longer-term goal of developing a robotics-specific foundational model within a few years. A 140-megawatt data center is slated to go online in June 2028. The initiative underscores how sovereign AI is moving beyond language models into physical AI – systems that can train robots, industrial machines, and automation platforms to adapt to real-world environments. The order of this magnitude reinforces Rubin’s expected role as the next major infrastructure platform for governments and national champions seeking domestic AI capabilities.
❖ In parallel, Bristol Myers Squibb announced it is buying NVIDIA’s latest-generation DGX SuperPOD based on Vera Rubin systems, becoming the first life-sciences company to purchase the new architecture. The deal expands NVIDIA’s reach in life sciences, where AI infrastructure is increasingly becoming a core R&D asset rather than an experimental tool. Bristol Myers said AI is already helping reduce the time needed to produce medicines for clinical testing by 20-30%, with potential for that to reach 50% in coming years. For NVIDIA, the significance is broader than one customer: the company’s most advanced systems are now being adopted not only by hyperscalers and sovereign AI programs, but also by major pharma companies with compute-intensive discovery pipelines. That widens Vera Rubin’s addressable market across high-value verticals where speed, model scale, and energy efficiency can directly affect commercial outcomes.
1
❖❖ Alphabet’s (GOOGL) Google is reportedly developing a new server chip, internally codenamed “Frozen v2,” designed to run Gemini models far more efficiently. According to media reports, the chip would embed parts of Gemini’s architecture directly into the silicon, reducing the amount of calculation and data movement needed to answer queries.
The reported target is significant: Google engineers believe Frozen v2 could serve six to ten times more tokens per unit of power than the company’s latest TPUs. If achieved, that would directly address one of the biggest constraints in AI today – the cost, power intensity, and capacity limits of inference at scale. The chip is expected to be a specialized branch of Google’s custom-silicon portfolio, not a replacement for its broader TPU roadmap, which already includes TPU 8t for training and TPU 8i for inference.
The strategic read-through is Google’s full-stack advantage. By co-designing hardware, software, and model architecture around Gemini, Alphabet can potentially improve AI economics from inside the system rather than simply adding more capacity. Frozen v2 is reportedly targeted for deployment around 2028, making it a longer-term effort, but one that fits directly into Google’s push to ease compute bottlenecks and improve AI profitability.
1
❖❖ Taiwan Semiconductor Manufacturing, aka TSMC (TSM) delivered a record Q2 that reinforced its position as the manufacturing backbone of the AI buildout, with revenue, margins, guidance, and capex all moving higher.
Revenue reached $40.2 billion, up 12% sequentially and 33.7% year-over-year in U.S. dollar terms (the previously reported 36% increase was in NT dollars), while earnings surged 77.4% year-over-year to $4.31 per ADR. Gross margin expanded to 67.7%, above guidance, and operating margin reached 60.3%, helped by strong utilization, cost improvements, and a richer advanced-node mix.
The growth driver was unmistakable. High-performance computing revenue rose 20% quarter-over-quarter and accounted for 66% of total revenue, as AI and cloud demand continued to absorb leading-edge capacity. Advanced technologies, defined as 7nm and below, represented 77% of wafer revenue, with 5nm at 33%, 3nm at 30%, 7nm at 11%, and the first meaningful 2nm contribution already reaching 3%.
The outlook strengthened as well. TSMC guided Q3 revenue to $44.6-45.8 billion, implying another 12% sequential increase at the midpoint, and raised its full-year 2026 revenue growth outlook to slightly above 40% in U.S. dollar terms. Capex guidance was lifted to $60-64 billion from $52-56 billion, with 70-80% directed toward advanced process technologies. TSMC is expanding 3nm and advanced packaging capacity across Taiwan, Arizona, and Japan.
The higher spending underscores how far demand still exceeds available supply. CFO Wendell Huang later reinforced the message, saying TSMC is accelerating its Arizona expansion to $265 billion, backed by strong structural AI demand, while 2nm is emerging as a new Q3 revenue driver.
The buildout is not frictionless: management expects the steep 2nm ramp to dilute gross margin by 3-4 percentage points in the second half, while overseas fabs could create an additional 2-4 percentage-point drag over time. Still, the quarter confirms the thesis: AI demand is broadening across GPUs, CPUs, XPUs, power management, and sensors, while TSMC remains the scarce capacity layer nearly every leading-edge customer still needs.
1
❖❖ GE Aerospace (GE) delivered another strong beat-and-raise quarter, continuing a pattern that has become almost routine for the jet-engine leader.
Adjusted revenue rose 24% year-over-year to $12.6 billion, well above expectations, while adjusted EPS increased 22% to $2.02, also topping the consensus. Free cash flow jumped 43% to $3.0 billion, with conversion above 140%, and total orders rose 17% to $16.5 billion. The first-half picture was just as strong, with orders up 49%, revenue up 27%, adjusted EPS up 24%, and free cash flow up 31%.
The quarter was powered by Commercial Engines & Services, where revenue climbed 27% to $9.7 billion. Services revenue grew 26%, helped by higher shop-visit volume, LEAP internal shop visits up more than 50%, and spare parts revenue up over 25% as material availability improved. Equipment revenue rose 30%, with total engine deliveries up 26% and LEAP deliveries up 24%.
One of the key messages was that demand is no longer the issue. GE’s commercial services backlog stands at roughly $170 billion, total backlog exceeds $210 billion, more than 95% of Q3 spare-parts revenue is already in backlog, and planned removals/off-wing engines exceed the full-year shop-visit guide by more than 40%. Management’s focus is now on converting that demand into output through FLIGHT DECK, supplier coordination, AI-enabled process improvements, and faster shop turnaround times.
That visibility drove a broad guidance raise. GE now expects high-teens adjusted revenue growth, adjusted EPS of $7.65-7.85, operating profit of $10.55-10.75 billion, and free cash flow of $8.9-9.2 billion.
There is a caveat, however: margin pressure. Operating margin fell 130 basis points to 21.7%, and Commercial Engines & Services margin contracted 160 basis points to 27.3%, reflecting installed engine growth, GE9X investments, and inflation. Still, those lower-margin engine deliveries expand the future services base, while LEAP durability upgrades and improving throughput support the long-term cash flow story.
1
❖❖ Bank of New York Mellon (BNY) delivered back-to-back record quarters, with another strong Q2 showing that the custodian bank is extracting more growth and profitability from its massive capital-markets infrastructure.
Revenue reached a record $5.7 billion, up 13% year-over-year, while EPS rose 27% to $2.45, above the $2.22 consensus. Fee revenue increased 11% and net interest income (NII) jumped 20% to $1.4 billion, while expenses rose only 7%, producing roughly 600 basis points of positive operating leverage. That pushed pre-tax margin to 40% and ROTCE to 31%.
The quarter was not just helped by markets. BNY’s commercial and platform operating models are clearly gaining traction, with Q2 marking the 14th consecutive quarter of year-over-year sales growth and the second consecutive record revenue quarter. Average deal size in the first half rose more than 20%, roughly 10% of sales-generating clients were entirely new to BNY, and the number of clients using three or more BNY platforms has increased by more than 60% over the past three years.
That client momentum showed up across the business lines. Securities Services revenue rose 15%, with pre-tax income up 28%, supported by asset servicing, corporate trust, ETF growth, and higher activity. Market and Wealth Services revenue rose 12%, with clearance and collateral management benefiting from higher collateral balances and clearing volumes. Investment and Wealth Management revenue grew 8%.
With growth broadening and expenses well controlled, management had enough visibility to significantly raise its 2026 outlook: revenue excluding notable items is now expected to grow 10-11%, NII 12-13%, and operating leverage by approximately 400 basis points.
BNY is still exposed to the normal rhythm of capital markets: Q3 is seasonally slower, deposit and NII comparisons get tougher, and fee revenue remains tied to market levels and client activity. However, with roughly 75% of fees recurring and digital assets, AI, payments, collateral, and platform initiatives gaining traction, BNY’s growth story now looks more durable than it has in years.
1
❖❖ Morgan Stanley (MS) reported a stellar quarter that showed why it remains one of the clearest beneficiaries of stronger capital-markets activity and the accelerating AI infrastructure cycle.
Net revenues rose 27% year-over-year to a record $21.3 billion, while EPS surged 64% to a record $3.46. ROTCE reached 26.6%, and the firm’s first-half efficiency ratio improved to 65%, reflecting strong operating leverage as client activity broadened across advisory, underwriting, trading, and wealth.
Institutional Securities – MS’s largest revenue contributor this quarter – produced the biggest earnings swing, with revenue rising 44% to a record $11.0 billion and pre-tax income more than doubling to $4.3 billion. Investment Banking revenue jumped 58% to $2.4 billion, supported by higher M&A completions, stronger IPO activity, follow-ons, convertibles, and record fixed-income underwriting. Equities revenue surged 69% to a record $6.3 billion, helped by strong client engagement, active markets, and notable strength in Asia, while Fixed Income revenue rose 13% to $2.5 billion.
The AI buildout is becoming a major capital-markets opportunity. Morgan Stanley has been central to several of the largest and most creative AI infrastructure financings, from data centers to GPU-backed lending, and management framed the cycle as still early, estimating 2026 data center capex at about $850 billion, 2027 at $1.3 trillion, and the broader investment cycle only 10-15% complete.
The strength also flowed into the Wealth Management division, where revenue reached a record $8.9 billion and net new assets hit a record $148 billion. Just over half of that came from stock-plan clients tied to recent IPOs, showing how MS is turning corporate relationships into long-term wealth-management assets. Across Wealth and Investment Management, client assets reached $10 trillion.
Capital strength gives Morgan Stanley room to keep leaning into expanding market opportunities. The firm has accreted $18 billion of CET1 capital over the past ten quarters and ended Q2 with a standardized CET1 ratio of 14.8%, leaving at least 300 basis points of excess capital. That cushion supports continued investment in the core franchise, including AI-related advisory and financing capabilities, while also allowing MS to return capital through the $1.5 billion of share repurchases completed in the quarter.
The outlook remains constructive. Management said Investment Banking pipelines are healthy, global activity is broadening, and Q3 Wealth Management NII should increase modestly sequentially. Reflecting that confidence, Morgan Stanley raised its quarterly dividend by 15% to $1.15 per share.
1
❖❖ PNC Financial (PNC) delivered an excellent Q2 that affirmatively answered two key investor questions at once: whether core regional-bank fundamentals were improving, and whether the FirstBank integration would distract from execution.
PNC’s revenue climbed 21.6% year-over-year to $6.88 billion, beating estimates by roughly $450 million. Reported net income reached $2.1 billion, or $4.81 per share. FirstBank integration costs and other significant items reduced earnings by $0.04 per share, resulting in non-GAAP EPS of $4.85 – up 25% year-over-year and well ahead of the consensus.
The quarter showed clear balance-sheet momentum. Net interest income rose 4% sequentially to a record $4.11 billion, helped by commercial loan growth, higher noninterest-bearing deposits, and better deposit pricing. Average loans increased 4% to $363 billion, driven almost entirely by C&I growth, while average deposits remained stable at $457 billion and noninterest-bearing balances rose 4%.
Fee income added another leg to the story, rising 10% sequentially and 20% year-over-year to $2.3 billion. Capital markets and advisory revenue jumped 25% from Q1, helped by record M&A advisory fees and broad strength across capital-markets businesses, while asset management, card, cash management, lending, deposit services, and mortgage fees all improved.
Just as importantly, credit remained clean. Nonperforming loans fell 10%, delinquencies declined 8%, and net charge-offs were 0.25% of average loans. PNC’s CET1 ratio stood at 9.9%, near management’s operating target, while the bank returned $1.3 billion to shareholders and raised its quarterly dividend by 18% to $2.00 per share.
That performance gave management room to raise the outlook. PNC now expects 2026 average loan growth of about 12.5%, NII growth of 15-15.5%, total revenue growth of roughly 13%, and NIM above 3% by year-end.
The near-term watch item is normalization. Q3 fee income is expected to decline as capital markets cool from an elevated Q2, while management expects deposit pricing to drift back toward Q1 levels after Q2’s favorable repricing benefit. Still, PNC exits the quarter in a stronger position, with FirstBank converted, commercial demand broadening, credit improving, and NII moving higher.
1
❖❖ Interactive Brokers (IBKR) delivered another record quarter, once again highlighting its winning business model: the company scales like a fintech, monetizes like a broker, and compounds like a platform – while continuing to expand into new markets, products, AI interfaces, and potentially direct custody capabilities through a national trust bank charter.
Net revenues reached a record $1.90 billion, while adjusted EPS rose to $0.69 – both well above Wall Street expectations. Commissions rose 30% year-over-year to a new record, supported by higher trading volumes across stocks, options, and futures, while net interest income climbed 23% to $1.06 billion on higher customer margin loans and credit balances. Pretax margin held at 77%, marking the seventh straight quarter above 70% – a reminder of how efficiently IBKR’s automated infrastructure converts growth into profit. The company continues to operate with no long-term debt and about $10.3 billion of excess capital.
The operating metrics were even stronger than the headline numbers. Client accounts jumped 34% year-over-year to 5.19 million, client equity rose 40% to $930.3 billion, customer credit balances increased 27% to $182.4 billion, and margin loans surged 67% to $108.5 billion. Daily average revenue trades rose 36% to 4.82 million, showing that IBKR is benefiting not only from market activity, but also from a larger and more engaged client base.
The platform also continues to broaden. IBKR became the first e-broker to offer trading in Korea, expanded crypto access in Europe, added prediction-market connectivity through Kalshi and CME, and offered eligible U.K. and European retail clients access to the SpaceX IPO. Its new IBKR Connector, built with Anthropic, OpenAI, and xAI integrations, lets clients connect accounts to AI assistants while keeping trade execution under a human-in-the-loop model.
The most strategic development may be regulatory. IBKR received preliminary conditional OCC approval for a national trust bank charter, which management expects to make operational by year-end. If completed, the charter could deepen custody capabilities and strengthen IBKR’s position with larger, more sophisticated clients – adding another layer to one of the strongest financial platforms in the market.
1
Under Review
❖❖ We are placing Palantir Technologies (PLTR) under review. The business remains one of the clearest software beneficiaries of the AI cycle, with AIP adoption, U.S. commercial momentum, and government demand still supporting the long-term thesis. However, the stock’s recent behavior is concerning.
PLTR has remained under pressure even on days when the Nasdaq and most AI-linked names are rallying, suggesting that investors are still uncomfortable with the valuation and may be demanding another clean proof point before rewarding the stock again. That matters because Palantir trades at a premium that leaves little room for disappointment.
At the same time, earnings are expected on August 3, less than two weeks away. Selling before that update could mean exiting just ahead of the catalyst that might reset sentiment, especially if results confirm continued AIP monetization, strong U.S. commercial growth, expanding profitability, and raised guidance.
For now, we are holding the position, but the bar is high. If the report shows slowing growth, weaker guidance, softer commercial traction, margin pressure, or another poor stock reaction despite solid results, we may exit. If earnings restore confidence and the stock finally responds, PLTR can remain in the Smart Portfolio.
1
Portfolio Earnings and Dividend Calendar
❖ The Q2 2026 earnings season has officially begun, and several Smart Investor portfolio holdings are scheduled to reveal their quarterly results this week. Alphabet (GOOGL) will open the season for the U.S. Big Tech with its report today after hours. Teledyne (TDY) will also report today, while RTX (RTX) is scheduled for tomorrow, July 23. Vertiv (VRT), Generac (GNRC), Amphenol (APH), and Microsoft (MSFT) will reveal their quarterly results on July 29.
❖ The ex-dividend date for nVent Electric (NVT) is July 24, while for Bank of New York Mellon (BNY) it is July 27.
w
1
New Buy: Nova (NVMI)
Nova Ltd. operates at one of the most exacting points of the semiconductor value chain – the measurement and process-control layer that helps chipmakers turn increasingly complex designs into reliable, high-yield production. The company develops advanced metrology systems and software used across wafer fabrication and advanced packaging, giving manufacturers real-time insight into the dimensions, materials, composition, and uniformity of critical chip structures. Serving leading foundries, memory producers, and integrated device manufacturers, Nova plays an essential role in improving yield, controlling process variation, and accelerating the transition to next-generation nodes. As AI, high-performance computing, and advanced packaging push semiconductor manufacturing toward ever tighter tolerances, Nova has become part of the precision infrastructure behind the industry’s progress – helping chipmakers see, measure, and control what ordinary inspection can no longer capture.
1
Measure Up
Nova’s early history was shaped by a clear technical bet: as semiconductor manufacturing became more complex, chipmakers would need increasingly precise ways to measure and control production in real time. Founded in Israel in 1993, the company built its position in optical metrology, went public on the Nasdaq in 2000, and later listed in Tel Aviv. Over time, Nova shifted toward closer collaboration with chipmakers, selling more directly to fabs and gaining deeper insight into yield and process-control challenges.
The company’s platform widened in 2015 with the acquisition of ReVera, which added stand-alone materials metrology and thin-film composition capabilities. That deal marked the beginning of Nova’s expansion from a narrower optical-metrology specialist into a broader process-control company.
The transformation accelerated in the early 2020s. Nova launched Metrion, the first fully automated inline Secondary Ion Mass Spectrometry1 (SIMS) system validated for production process control, bringing lab-level materials analysis into the fab. As advanced logic and memory moved to 3D structures and new materials, Nova positioned itself where standard dimensional measurement was no longer enough.
In 2022, Nova acquired Germany’s ancosys, adding chemical metrology and extending its reach into process chemistry, backend manufacturing, and advanced packaging. This pushed the company closer to a fuller view of what happens inside the fab – measuring dimensions, materials, composition, and process chemistry across more stages of production.
The past two years added another layer. Nova enhanced its machine-learning software through Nova Fit for materials metrology, supporting higher-throughput and higher-precision measurements in advanced logic and memory, including Gate-All-Around2 (GAA) applications. In 2025, Nova’s standalone optical materials metrology platform ELIPSON was selected by a leading foundry for advanced GAA processes, while Metrion continued gaining adoption among leading manufacturers.
The company also moved deeper into advanced packaging. In 2025, it acquired Sentronics, adding modular wafer metrology tools for backend fabrication and packaging. That push became clearer with Nova WMC, a wafer-level optical metrology platform launched in 2025 for 2.5D and 3D advanced packaging process control. In 2026, Nova WMC was selected by a leading global foundry as tool of record for advanced packaging layers, strengthening Nova’s position in a fast-growing manufacturing transition.
Nova’s development through the years reflects a steady move up the value chain – from optical metrology specialist to a comprehensive process-control partner essential for the industry’s most demanding manufacturing transitions.
—
1 – Secondary Ion Mass Spectrometry (SIMS) is a technique that measures chemical composition and dopant concentration by depth inside semiconductor materials.
2 – Gate-All-Around (GAA) is a next-generation transistor design where the gate surrounds the channel, improving control as chips shrink beyond FinFET.
—
1
Fab the Lab
Nova operates in one of the least glamorous yet most essential parts of semiconductor manufacturing: measuring what is happening inside the fab with enough precision to keep increasingly complex chips manufacturable.
The company provides optical, materials, chemical, X-ray, SIMS, and software-based metrology systems used by leading logic, memory, and foundry customers to control advanced production processes. Nova helps chipmakers see whether the layers, films, materials, dimensions, chemical concentrations, and packaging structures inside a wafer are exactly where they need to be. As chip architecture becomes more three-dimensional, more material-intensive, and less forgiving, that measurement layer becomes harder to replace.
This is why Nova’s opportunity is expanding across several semiconductor transitions at once. In logic, Gate-All-Around is increasing process-control intensity as manufacturers move beyond FinFET architectures. Nova says GAA drives roughly 30% higher metrology intensity, and the company remains positioned across leading GAA manufacturers. In memory, advanced DRAM and HBM are becoming larger demand drivers as AI systems require faster, denser, and more power-efficient memory. DRAM exposure has already risen from about 25% in 2025 to more than 30% in Q1 2026, while Nova’s long-term mix still points to broad exposure across both logic and memory.
Nova’s strongest differentiation is its breadth. The company is no longer only an optical metrology specialist. Its “lab-to-fab” strategy moves techniques that were historically performed offline into automated inline production workflows. Metrion, the first fully automated inline SIMS platform validated for production process control, is a good example: it maps chemical composition across the wafer, monitors dopant concentration and contamination, and gives fabs faster feedback before defects become costly scrap. Management says inline SIMS remains in the early first-tool-per-fab stage, leaving room for multi-tool-per-fab adoption as advanced-node production scales.
Advanced packaging adds another major growth path. This business has grown from essentially nothing five years ago to more than 20% of Nova’s revenue, supported by logic and HBM demand. Nova WMC is the centerpiece of Nova’s advanced-packaging story, built for the increasingly difficult job of measuring warped, uneven, and highly complex package structures across large areas with nanometer-level precision. That matters as HBM and 2.5D/3D architectures push more value into packaging, where tiny measurement errors can damage yield, performance, and production economics.
In July 2026, WMC was selected by a leading global foundry as tool of record for multiple advanced-packaging layers after a competitive evaluation, while Nova also cited accelerating adoption across memory and foundry customers. Advanced-packaging metrology adoption remains far below front-end levels, giving Nova a clear runway as HBM, hybrid bonding, and 2.5D/3D architectures make backend manufacturing much harder to control without precise inline measurement.
Competition remains serious, especially from larger process-control players and potential challengers in newer materials tools. Still, Nova is gaining ground: its share in film and critical-dimension metrology rose from about 20.3% in 2023 to nearly 29% in 2025, including roughly 400 bps of share gain in 2025 alone, making it the second-largest vendor in that market.
From here, the next manufacturing inflections should work in Nova’s favor. Hybrid bonding3 and advanced packaging will demand tighter control over 3D structures and interconnects; CFET4 will raise the precision burden in logic; and next-generation memory architectures should make advanced metrology even harder to treat as optional.
—
3 – Hybrid Bonding is an advanced packaging method that directly connects chips or wafers without traditional bumps, enabling denser, faster, and more power-efficient 3D architectures.
4 – CFET (Complementary FET) is a future transistor architecture that vertically stacks n-type and p-type devices to keep increasing chip density beyond GAA.
—
1
The Yield Lens
Nova’s financial data shows a company already converting semiconductor complexity into scale, margins, and cash. Revenue rose 31% year-over-year to $881 million in 2025, while product revenue grew at a 27.5% CAGR from 2020-2025, far ahead of Gartner’s estimated 16% CAGR for process control market. That gap shows Nova’s growth is coming from sustained outperformance versus its industry, helping explain the sharp rise in its critical-dimension and thin-film metrology share.
Q1 2026 extended the pattern. Revenue reached a record $235.3 million, up 10% year-over-year and beating both the company guidance and the roughly $227 million analyst consensus. Non-GAAP EPS rose about 7% year-over-year to a record $2.33, ahead of the roughly $2.19 consensus. Nova has beaten non-GAAP EPS expectations in every quarter since analysts began issuing estimates in 2021, and has exceeded revenue estimates for at least the past eight quarters.
The quarter’s mix also showed why the business is broadening. Logic/foundry represented about 66% of revenue and memory 34%, with DRAM exposure rising from about 25% in 2025 to more than 30% in Q1. Product revenue reached $186.3 million, while service revenue hit a record $49.0 million, marking the 13th consecutive quarter of sequential service growth. That installed-base layer does not replace product sales, which still represent roughly 80% of sales, but it adds a steadier revenue stream as more Nova tools operate inside customer fabs.
Margins remain one of the cleanest signs of product value. GAAP gross margin expanded by roughly 40 bps year-over-year to 57.7%, while non-GAAP gross margin was 59.4%, near the upper end of Nova’s 57-60% target model. GAAP operating margin improved by roughly 50 bps to 30%, and non-GAAP operating margin remained exceptional at 34%, above the company’s 28-33% target model. Non-GAAP margins were slightly lower year-over-year, showing that further expansion may be harder from already high levels, but the profitability base remains unusually strong for a semiconductor equipment supplier still investing aggressively in R&D.
The balance sheet remains one of Nova’s strategic assets. As of Q1, total assets stood at $2.42 billion, supported by substantial cash, deposits, and marketable securities; by July, management was pointing to more than $1.7 billion in cash. Cash generation adds to that strength: Nova produced $218 million in free cash flow in 2025 and roughly $36 million in Q1 2026. Quarterly conversion can move around with shipments, collections, and working-capital timing, but the annual numbers show a business capable of supporting growth with internally generated cash.
There is one balance-sheet nuance: Nova’s $750 million of 0% convertible notes due 2030 create no meaningful cash-interest burden, but they do add potential dilution and capital-allocation complexity. Even so, that liquidity gives Nova room to fund R&D, capacity expansion, and disciplined acquisitions, with management focused on targets that are accretive, financially similar, and capable of adding technology or revenue synergies.
Guidance points to another step up. For Q2, Nova expects revenue of $245-255 million and non-GAAP EPS of $2.34-2.48, above consensus at the midpoint and implying roughly 14% revenue growth and about 10% non-GAAP EPS growth year-over-year. For 2026, management estimates that WFE spending will grow in the mid-teens and expects Nova to outperform, while remaining on track toward its $1 billion revenue target by 2027.
This growth trajectory is not risk free: export controls, supplier dependence, geopolitics, and semiconductor capex cyclicality can all disrupt even a stellar operator. Customer concentration is also material, with the five largest customers representing 52% of 2025 revenue, though this is structural in a market built around a limited number of advanced fabs. The key question is whether Nova can keep turning that concentrated demand into share gains, service revenue, and high-margin growth – and so far, the numbers say yes.
1
Measure Twice
Nova sits in one of the most specialized parts of semiconductor equipment: metrology and process control, where chip complexity, yield control, and advanced packaging are becoming increasingly important. Onto Innovation is the cleanest peer, with the closest overlap in optical metrology, inspection, process control, and advanced-node exposure. Camtek adds a sharper advanced-packaging angle, with inspection and metrology tools tied closely to chiplets, HBM, and AI-related manufacturing complexity. FormFactor is less direct technologically, since it is focused on wafer probe cards and test systems, but it provides a useful comparison for how investors value smaller semiconductor equipment companies exposed to production cycles, memory recovery, and advanced packaging demand. Together, these peers frame Nova as a high-margin, specialized process-control leader with both cyclical exposure and structural growth drivers.
The whole peer group has benefited from the AI manufacturing narrative throughout the past year, but the strongest gains went to the names with the most visible near-term leverage to HBM and advanced packaging. FormFactor led as probe cards became a direct AI memory bottleneck, while Onto was lifted by clearer advanced-packaging momentum. Camtek also rallied, though its move was more tied to expected second-half acceleration. Nova’s roughly 52% gain over the past year was the smallest in the group, reflecting its more balanced growth profile: a market reward for quality and consistency, with a smoother upward path and less thematic heat than the more HBM-centered peers. The more measured rally has left Nova’s stock with a considerable potential runway, as Wall Street’s average price target implies a potential upside of about 30%.
That potential is also supported by Nova’s valuation – which is not cheap in absolute terms, but looks more reasonable when measured against growth and profitability. At roughly 41x FY1 earnings, Nova trades close to Camtek and FormFactor and slightly above Onto, while its FY2 and FY3 multiples remain within the peer range. Its forward EV/Sales multiple is the group’s highest at 12.1x, while its forward EV/EBITDA is broadly in line with Onto and FormFactor and still below Camtek. The broader premium valuation is supported by stronger revenue growth than every peer, with forward revenue expected to rise 24.4%, and by one of the best long-term growth profiles in the group. Profitability strengthens the case further: Nova leads the peer set in gross margin, EBIT margin, EBITDA margin, net income margin, and return on equity. Nova appears priced like a premium asset – but it also delivers premium growth, margins, and operating quality, while its profitability, cash-rich balance sheet, and lower financial-risk profile deserve to be part of the valuation equation.
Nova has a long-standing buyback authorization, although share repurchases have been small and opportunistic, mostly intended to offset stock-based compensation. The dominant uses of capital remain R&D investment, balance sheet liquidity, and strategic bolt-on acquisitions. Nova repurchased approximately $35 million worth of shares during 2025, but did not repurchase shares in Q1 2026. However, as the company continues to steadily grow its revenue and free cash flow, buybacks could become a more meaningful capital-allocation lever over time.
1
Investing Takeaway
Nova is increasingly positioned as a quiet gatekeeper of the AI semiconductor cycle. As chips become more complex, manufacturers need tighter control over dimensions, materials, and defects, making metrology less of a support function and more of a yield-critical capability. The company does not offer the flashiest AI exposure, but its role becomes more important as advanced nodes, HBM, and packaging complexity scale. Risks remain around semiconductor capex timing, customer spending patterns, and export-policy uncertainty, yet Nova enters this phase with strong profitability, deep technical credibility, and a cash-rich balance sheet. That combination gives it room to keep investing, acquiring, and expanding alongside the industry’s most demanding customers, making the stock a compelling way to own the precision layer behind AI manufacturing.
1
New Sell: Synopsys (SNPS)
We are selling Synopsys from the Smart Portfolio. This is a tactical exit from a high-quality company whose investment case remains structurally attractive, but has not yet translated into stock performance.
Synopsys remains one of the most important infrastructure companies in the semiconductor ecosystem. Its design automation and IP tools sit at the center of modern chip development, making SNPS a natural beneficiary of rising AI complexity, custom silicon, advanced packaging, and system-level simulation needs. The Ansys acquisition also expands the company’s reach into broader engineering simulation, digital twins, and multi-physics design – all areas that could strengthen the long-term growth profile.
The problem is conversion. Since we added SNPS in April, the news flow has been largely positive, management raised guidance, and analysts have remained broadly bullish. Piper Sandler upgraded the stock, while BofA, Citi, Stifel, and Berenberg raised targets, pointing to Ansys synergies, stronger FY26 guidance, and AI-driven opportunities. Yet the stock has failed to respond meaningfully, even during a period when AI-linked names enjoyed strong rallies.
That gap matters. Investors appear to be waiting for proof that Ansys, design automation, and the IP recovery can translate into durable organic growth and cleaner earnings momentum. The acquisition also brought more than $10 billion of debt, while China softness, IP restructuring, and messy GAAP optics have kept the story more complicated than the headline AI thesis suggests.
Importantly, valuation is not the issue. SNPS now looks cheap versus its own history and trades at a notable discount to Cadence across several key multiples. But cheapness alone is not a catalyst, especially when the next major update, the fiscal Q3 earnings report, is still about a month away.
For Smart Investor, the opportunity cost is now too high. We still like the business and would revisit if earnings confirm stronger organic growth, faster debt reduction, clearer Ansys synergies, or a better stock reaction to positive news. For now, we are stepping aside and making room for a setup with clearer near-term momentum.
1
Smart Investor’s Winners Club
The Winners Club represents stocks from the Smart Investor Portfolio that have risen at least 30% since their purchase dates.
A sell-off was followed by a rebound, but the Club ranks have remained steady at 26 stocks: AVGO, GE, ANET, TSM, HWM, EME, APH, IBKR, VRT, ASX, CRWD, PH, MTZ, PANW, GOOGL, STRL, CSCO, ATI, BNY, KEYS, MS, JBL, RTX, NVT, JPM, and CRDO.
The first runner-up is still LLY with a 18.86% gain since purchase. Will it break into the Winners’ circle, or will another stock outrun it to the finish line?
1
|
