Onto Innovation Inc. (NYSE: ONTO) — The AI-Packaging Metrology Champion, Priced for a Flawless Two-Year Ramp

Independent equity research. The analytical body below carries no recommendation and no price target; the sole exception is the clearly-labeled author’s-opinion block.

As-of date: June 20, 2026 · Price referenced: $333.76 (close 2026-06-18) · Report type: Initiation (fresh coverage)

⚡ Claude’s Take

The author’s own independent opinion and general information only — not investment advice. The analytical body below carries no position and no price target.

Verdict: HOLD / own-for-the-quality / accumulate-on-weakness toward the 200-EMA (~$200–250) — NOT-A-SHORT. Conviction: medium. ONTO is a genuinely excellent business — the clear #2/#3 in semiconductor optical process-control behind KLA, with a real customer-captivity moat (50%+ gross margins, ~25%+ tangible ROIC), a net-cash legacy balance sheet, clean ~3%-of-sales stock comp, and ~$300M of real free cash flow. It also owns one of the most direct, in-hand exposures to the two fastest-growing pockets of the AI build-out: advanced packaging (HBM, 2.5D/CoWoS, panel-level) and advanced-node metrology (gate-all-around OCD). Management is guiding 2026 revenue >$1.3B (+30%), advanced packaging +50%, and an exit-Q4 operating margin >30% — and says 2027 looks stronger than 2026. The inflection is real and visible in the order book.

The problem is price. At $333.76 the stock sits at its richest-ever valuation on every lens — ~99th percentile of its own decade-long P/E, P/B and P/S history; ~48x forward non-GAAP EPS (~$7), ~12.7x forward EV/sales, ~37x forward EV/EBITDA, and ~155x trailing GAAP earnings — after doubling since December 2025 and rising ~270% off its August-2025 low of $90. The tape is a textbook high-beta (2.39) AI-semicap momentum melt-up: it loads on a “Chip Sector Superstars” basket (+1.51), Momentum (+0.55), Quality (+0.69) and Growth (+0.56), with a trailing-3-month annualized Sharpe near 9.5. The market is underwriting flawless execution across both 2026 and 2027 — a 25–30% grower compounding at peak-cycle multiples — while the quieter facts cut the other way: 2025 GAAP earnings actually fell 32%, the balance sheet just pivoted from net-cash to ~$1.7B of convertible debt to fund the Semilab + Rigaku deals, comp has no return-on-capital governor, and insiders own <1% and have made zero open-market purchases. This is a wonderful company at a wonderful-company-in-an-AI-melt-up price. Pay up only on a pullback. The framing is quality-compounder-at-the-wrong-entry, not falling knife and not short.

Catchy tag: “Best-in-class metrology, priced as if the cycle never turns again.”

Conviction: medium. Flips bullish on a meaningful de-rate (toward ~35x 2026 non-GAAP / the 200-EMA) with the advanced-packaging ramp and >30% exit margin confirmed — quality at a fair price. Flips bearish on a WFE/AI-capex digestion that stalls the 2H-2026 acceleration while the multiple sits at a record, or evidence the Semilab/Rigaku capital pivot is destroying returns.

📈 Stock Price Action — Five-Year Event Map

ONTO has round-tripped twice in five years. From ~$72 in mid-2021 it bottomed at $59.04 (Oct-14-2022) in the rate-shock semicap downturn, rallied to ~$219 (Jul-2024) on the first AI/HBM wave, then crashed −58% to a 52-week low of $90.00 (Aug-6-2025) as 2025 margins compressed and advanced-packaging orders digested — before exploding ~+270% to an all-time high of $335.96 (Jun-15-2026). Today’s $333.76 is essentially at the ATH, up ~+143% over twelve months, with a 52-week range of $90.00–$335.96. The stock is ~58% above its 200-day EMA (~$210) — a name in the steep, late stage of a momentum advance, not a value re-rating.

Cycle narrative. (1–2) ONTO is a high-beta semicap cyclical: it fell with the 2022 WFE downturn and rallied hard as the AI narrative took hold. (3) The first HBM wave drove it to $219 by mid-2024. (4) Then 2025 delivered an earnings air-pocket — GAAP net income fell 32% as the company absorbed its “extended-factory” manufacturing-offshoring transition and advanced-packaging orders paused — and the stock more than halved to $90. (5–6) From late-2025 the advanced-packaging cycle re-accelerated violently: the new Dragonfly G5 inspection platform won qualifications in HBM and at a 2.5D logic customer, the company raised 2026 guidance to >$1.3B (+30%), announced the Semilab and Rigaku deals, and attracted fresh sell-side Buy initiations (e.g., Deutsche Bank, Jun-5-2026) — carrying the stock to a new ATH. (Price moves are Fact; attributed drivers are Interpretation, cross-referenced to earnings dates, 8-Ks and the news feed.)

1. Executive Summary

Onto Innovation is a ~$1.0B-revenue (FY2025) semiconductor process-control company — optical metrology, macro-defect and advanced-packaging inspection, packaging lithography, and process-control software — formed in October 2019 by the merger of Rudolph Technologies and Nanometrics. It is one of a small handful of credible players (KLA, Nova/NVMI, Camtek, Onto) in a structurally attractive, oligopolistic niche of the WFE (wafer-fab-equipment) supply chain. Its franchise rests on genuine customer captivity: metrology and inspection tools are qualified into customer process recipes, where switching carries yield risk — a demand-side moat that shows up in 50%+ gross margins, high incremental margins, and very high returns on tangible capital.

The investment tension is entirely about price versus a real-but-cyclical inflection. The good: ONTO has direct, in-hand leverage to the two hottest pockets of AI hardware — advanced packaging (HBM, 2.5D/CoWoS, silicon photonics, panel-level) and advanced-node metrology (gate-all-around). Management guides FY2026 revenue >$1.3B (+30%+), advanced packaging +50%, advanced nodes +25%, and a Q4-exit non-GAAP operating margin >30%, with 2027 framed as stronger still. The balance sheet was net-cash, SBC is a clean ~3% of sales, free cash flow is real (~$300M FY2025), and the moat is verified by the financials.

The caution: at $333.76 ONTO trades at its richest valuation in its public history — ~99th-percentile P/E, P/B and P/S versus its own decade, ~48x forward non-GAAP EPS, ~12.7x forward EV/sales, ~37x forward EV/EBITDA, ~155x trailing GAAP EPS — after doubling in six months and rising ~270% off the August-2025 low. The factor signature is an unambiguous AI-semicap momentum melt-up (beta 2.39; loads on a chip-superstars basket, Momentum, Quality, Growth). Three quieter facts temper the enthusiasm: (i) 2025 GAAP earnings actually declined 32% on a margin air-pocket, so the record multiple sits atop trough-ish trailing profits; (ii) the company has just pivoted from a fortress net-cash balance sheet to ~$1.7B of 0% convertible debt to fund the Semilab (~$432M) and Rigaku (~$710M, 27% minority stake) deals — its largest-ever capital deployment, with no ROIC metric in management comp to govern it; and (iii) insiders own <1% and have made zero open-market purchases in 2.5 years. The business deserves a premium; the current price embeds flawless execution across both 2026 and 2027 with no margin of safety. The memo lays out the moat, the inflection, the cycle risk, and exactly what the price requires to be right.

2. Business Overview

What ONTO does. Onto Innovation designs, builds and services capital equipment and software that semiconductor and advanced-packaging manufacturers use to measure, inspect and control their production processes — the “process control” layer of the fab, distinct from the deposition/etch/litho “process” tools sold by Applied Materials, Lam and ASML. Process control does not add a layer to the chip; it measures whether the layers other tools created are correct, catching defects and process drift early enough to protect yield. As transistors shrink and chips move to 3D stacks and chiplets, the number of measurement and inspection steps per wafer rises structurally — process control grows faster than overall WFE over a cycle. ONTO is headquartered in Wilmington, Massachusetts, employs roughly 1,900–2,000 people, and reports on a calendar fiscal year.

Product portfolio (the franchise). ONTO’s lines map to the two demand engines:

• Advanced packaging & specialty. The Dragonfly macro-defect/2D inspection platform (the new G5 generation launched in 2026), 3DI 3D metrology for bumps/interconnects, JetStep packaging/panel-level lithography, and Firefly — used to inspect and measure HBM stacks, 2.5D/CoWoS logic, fan-out, silicon photonics/co-packaged optics, and emerging panel-level packaging. This is ONTO’s differentiated stronghold and the fastest-growing part of the company.
• Advanced nodes (front-end metrology). The Atlas OCD (optical critical-dimension) metrology platform (the G6 generation) and Iris thin-film metrology — used in leading-edge logic (gate-all-around) and DRAM process control. ONTO competes here against KLA and Nova.
• Software & services. Process-control software (standalone, group-of-tools, and factory-wide suites), spare parts, service, and software licensing — a higher-margin, more-recurring layer, now extended by the April-2026 Rigaku X-ray partnership (licensing ONTO’s Ai Diffract software onto Rigaku X-ray tools at near-100% margin).

How it makes money & revenue mix. The bulk of revenue is systems (capital equipment) sold to semiconductor device makers and OSATs (outsourced assembly & test), with a smaller, stickier stream of software, parts and services. In Q1-2026, of $292M revenue, management framed roughly $160M as “Specialty device & advanced packaging” (about two-thirds advanced packaging, ~$25M from the newly acquired Semilab, the remainder specialty/power), ~$80M advanced nodes (~60% memory/DRAM, ~40% logic), with software & services the remainder. Revenue is therefore majority non-recurring (tool sales) but underpinned by a meaningful recurring services/software annuity and — critically — by the recurring qualification position on customer process lines, which drives repeat tool orders as customers expand capacity.

End markets: AI/HPC compute and the memory (HBM/DRAM) and packaging that feed it; leading-edge logic/foundry; plus a long tail of specialty applications (power, CMOS image sensors, RF, MEMS, compound semis, silicon wafer). Verdict: a focused, high-quality capital-equipment-plus-software franchise levered to the highest-growth nodes of the semiconductor supply chain — but a tool-sale business model that is inherently cyclical and customer-concentrated.

3. Industry Dynamics

Structure: a good neighborhood of a cyclical industry. Wafer-fab equipment is an oligopoly, and process control is its most attractive room. Globally, process control / metrology / inspection is dominated by KLA Corporation (~55%+ share), with a competitive second tier — Onto, Nova (NVMI), Camtek (CAMT) — and the metrology arms of the big WFE primes (Applied Materials, Hitachi High-Tech). Barriers to entry are high: tools must be qualified into customer recipes over months/years; performance depends on decades of accumulated applications knowledge and proprietary algorithms; and the installed base creates switching costs and a service annuity. The profit pool is concentrated and the players earn high margins — the hallmark of a structurally good industry in Greenwald terms (genuine barriers to entry plus demand-side captivity).

Secular tailwind — content growth. Two structural forces lift process-control intensity faster than wafer starts: (1) node migration (gate-all-around, more complex 3D structures, exotic materials) multiplies the number of critical measurement steps; and (2) advanced packaging (HBM stacking, 2.5D/3D chiplet integration, panel-level) creates an entirely new, fast-growing inspection/metrology market where ONTO is differentiated. AI accelerators are the demand catalyst: every HBM stack and every CoWoS interposer needs inspection and metrology that did not exist at volume five years ago. ONTO’s stated thesis — advanced packaging +50% and advanced nodes +25% in 2026, both ahead of low-20s% WFE growth — is the company harvesting this content-growth dynamic.

Cyclicality and the Marathon lens. The flip side: WFE is violently cyclical, and the capital cycle is firmly in its attract-capital phase. ONTO’s own history proves the cyclicality — revenue fell ~19% in 2023 and the stock has twice halved in five years. High returns across the semicap complex are pulling in capacity and capital (every metrology player is expanding; new fabs are being built worldwide on subsidies). Marathon’s warning is apt: super-cycles over-build, and AI-capex enthusiasm could digest in 2027–2028, compressing both volumes and the multiple. The mitigant is that process control is a smaller, more concentrated, content-growing profit pool than bulk WFE — but it is not immune. Verdict: a structurally attractive, oligopolistic niche with a real secular content tailwind — wrapped inside one of the most cyclical industries in technology. Good industry, dangerous point in the cycle to pay a peak multiple.

4. Competitive Position

The moat is real, and the financials prove it. ONTO’s competitive advantage is principally demand-side customer captivity (switching costs) layered on technology/IP intangibles and niche economies of scale — Greenwald’s most durable combination. A metrology or inspection tool, once qualified into a customer’s process recipe and integrated into its yield-management workflow, is extremely costly and risky to replace: a swap means re-qualifying the process and risking yield on a multi-billion-dollar fab. That captivity is why ONTO sustains gross margins above 50% (55.7% non-GAAP in Q1-2026), high incremental margins, and very high returns on tangible capital (consolidated GAAP ROIC of ~6% is depressed by $643M goodwill + $278M of merger intangibles; on tangible invested capital the business earns comfortably above 25%). A “moat” that would vanish without the financial outcome it explains is not a moat — ONTO’s passes the test: strip the captivity and the margins compress to commodity levels.

Where ONTO is strong vs. exposed. ONTO is strongest in advanced-packaging inspection and metrology (Dragonfly/3DI/JetStep) — a market it helped create and where its packaging-specific engineering (handling die/wafer warp, rough CMP surfaces, sub-6-micron bumps) is differentiated. The Q1-2026 evidence is concrete: Dragonfly G5 qualified at a leading 2.5D logic customer and in HBM, with a pipeline of “over 15 distinct applications across over 10 customers,” and management characterizing G5’s addressable expansion as “nearly $1 billion in new markets.” It is more contested in front-end OCD/film metrology (Atlas/Iris), where it competes head-to-head with KLA (the 800-pound gorilla) and Nova — winning some gate-all-around evaluations but as a challenger, not the incumbent. The April-2026 Rigaku alliance (X-ray metrology, hybrid optical+X-ray solutions) is an attempt to extend the metrology moat into 3D/exotic-material measurement where pure optical reaches its physical limits.

Direct comparison. Versus KLA, ONTO is a fraction of the size and lacks KLA’s breadth and dominance, but is more concentrated in the fastest-growing advanced-packaging niche. Versus Nova (NVMI) — the closest pure-play comp, similar size and growth profile — ONTO is broader (inspection + lithography + metrology vs. Nova’s metrology focus). Versus Camtek (CAMT), ONTO competes directly in advanced-packaging inspection. Verdict: a durable, financially-verified moat — narrower and more contested than KLA’s, but genuine and concentrated in the highest-growth corner of process control. A strong #2/#3, not a fortress monopoly.

5. Growth History and Forward Opportunities

History — high-quality but cyclical compounding. Revenue has grown from $556M (2020, first full post-merger year) to $789M (2021), $1,005M (2022 peak), down to $816M (2023 downturn, −19%), back to $987M (2024), and $1,005M (2025) — a ~4-year CAGR of ~16% but with a deep mid-cycle trough. GAAP net income tracked the cycle and the margin structure: $142M (2021) → $223M (2022) → $121M (2023) → $202M (2024) → $137M (2025, −32%). The 2025 decline is the crux of the trailing-multiple optics: despite record revenue, net income fell because operating margin compressed from 19.9% to 15.1% (GAAP) — the company absorbing its “extended-factory” manufacturing-offshoring transition (start-up inefficiency, dual running costs) plus material-cost and tariff/fuel headwinds, with a particularly weak Q4-2025 (7.8% GAAP operating margin). This was a self-inflicted, transitional margin trough, not demand collapse — and it set up the 2026 inflection as those factory benefits are now harvested (Q1-2026 gross margin +110bps q/q to 55.7% non-GAAP, op margin 26.7%).

Forward — a visible, multi-driver inflection. Management’s 2026 framework is unusually specific and order-backed: total revenue >$1.3B (+30%+), with H2 at least 15% above H1, and a Q4-exit non-GAAP operating margin >30%. The drivers:

• Advanced packaging +50%: Dragonfly G5 ramping (doubling roughly each quarter off a small base), 3DI gaining OSAT customers on sub-6-micron bumps, HBM and 2.5D logic both growing, plus emerging silicon photonics/co-packaged optics and panel-level packaging (JetStep qualified at two AI-device packaging suppliers, 2027 ramp).
• Advanced nodes +25%: Atlas G6 OCD winning gate-all-around evaluations; DRAM next-gen development; early NAND recovery; a new TSV-metrology application win.
• New optionality: the Rigaku X-ray licensing/hybrid-metrology stream (small near-term, strategic long-term); software attach.

Management explicitly states 2027 “continues to look much stronger even than 2026” and that it expects to outgrow WFE again. Verdict: high-quality, content-driven, multi-product growth with genuine order/backlog visibility — but it is capex-cycle-dependent growth (it accelerates and decelerates with AI/WFE spending), not the steady annuity compounding of a software or consumables business. High quality, not low risk.

6. Financial Quality

Margins and the GAAP/non-GAAP gap. ONTO is a structurally high-margin business: gross margin has held ~50–54% through the cycle and is rising (55.7% non-GAAP in Q1-2026, guided toward 56.5%+). The critical analytical point is the large gap between GAAP and non-GAAP operating results. Q1-2026: GAAP operating margin 13.3% / GAAP diluted EPS $0.68 versus non-GAAP operating margin 26.7% / non-GAAP EPS $1.42. The ~$0.74/quarter wedge is driven almost entirely by amortization of acquired intangibles (the Rudolph/Nanometrics merger plus Semilab) and stock-based compensation — both real but non-cash and (for amortization) declining/finite. Trailing GAAP EPS (~$2.15 TTM) is therefore understated relative to cash earnings power, while non-GAAP EPS is the better run-rate anchor. One must value ONTO on non-GAAP/cash earnings and EV/EBITDA — not the headline 155x trailing GAAP P/E, which is an artifact. (Equally, one should not take non-GAAP uncritically: SBC is a real ~$28M/year cost.)

Quality of earnings — clean. SBC was $27.6M in 2025, only ~2.7% of revenue — strikingly clean for a semiconductor company (versus, e.g., 30%+ at smaller chip names) and a mark of conservative accounting. Cash conversion is strong: operating cash flow of $328M in 2025 exceeded GAAP net income 2.4x (working-capital release), and free cash flow was ~$300M (capex is light, ~$28M/year, ~3% of sales — a capital-light equipment model). FCF/share was ~$6.10 (2025). No serial impairments, no aggressive revenue recognition flagged. Earnings quality is high.

Returns on capital. ROE was 19.4% (2025), down from 35.2% (2024) and 84% (2022) as the cycle and margin trough bit and equity grew. Consolidated GAAP ROIC (~6%) is artificially depressed by the $921M of goodwill + intangibles from the 2019 merger — i.e., the denominator carries acquisition accounting, not cash deployed by current management. On tangible invested capital the franchise earns well above its cost of capital (consistent with 50%+ gross margins and light capex), confirming value creation. This is the same “GAAP ROIC understates a capability-rich, acquisition-built business” pattern seen across high-quality spin/merger names.

Balance sheet — mid-pivot. As of Q1-2026: $654M cash + short-term investments, only ~$17.5M debt (capital leases) → ~$637M net cash; positive tangible equity of ~$1.2B ($2.13B equity less ~$921M goodwill/intangibles); current ratio 6.2x. Inventory is elevated ($316M, mostly raw materials) as the company builds for the ramp. But the balance sheet is changing fast: in May-2026 ONTO issued $1.7B of 0% convertible senior notes due 2031 (conversion ~$381.80, ~14% above today’s price; with capped calls and a concurrent ~$205M share repurchase), to fund the Rigaku 27% stake (~$710M) that closes 2H-2026, on top of the Semilab acquisition (~$432M cash, closed Nov-2025). Pro forma, ONTO moves from net-cash to roughly net-debt-neutral / modestly net-debt — still investment-grade-strong, but no longer a fortress, and now carrying convertible dilution risk if the stock pushes through ~$382. Verdict: economics clearly improve with scale (operating leverage is steep and visible in the 2026 margin ramp), earnings quality is high, and cash generation is real — but the balance sheet has just absorbed a step-change in financial risk that the rest of the analysis must weigh.

7. Capital Allocation

The five-year record: disciplined and organic — until 2025. For its first five years post-merger ONTO behaved like a prudent, net-cash compounder: token, opportunistic buybacks ($65M in 2022 near the lows, $3M 2023, $25M 2024, $75M 2025 under a $200M authorization), no dividend, light capex, and only small bolt-on M&A. That profile fit the quality of the business.

The 2025–2026 pivot is the real capital-allocation story. In ~18 months management committed its two largest deals ever and changed the balance sheet:

• Semilab USA (~$432M cash + ~642K shares, closed Nov-2025): an operating acquisition broadening the specialty/metrology portfolio (contributed ~$25M revenue in Q1-2026).
• Rigaku 27% stake (~$710M, signed Apr-2026, closes 2H-2026): a non-controlling minority investment in a publicly-traded Japanese X-ray instrument maker (~$600M revenue, ~40% semiconductor), bought from Carlyle (a PE seller exiting). Accounted under the fair-value option — meaning Rigaku’s stock price will mark-to-market through ONTO’s “other income” each quarter, injecting earnings volatility unrelated to operations. The strategic logic (hybrid optical+X-ray metrology, Ai Diffract licensing) is coherent and the income (licensing + ~$7M dividend) is guided to offset foregone interest within a year — but it is a financial stake, harder to underwrite than an operating deal, and it consumes the cash cushion.
• Funded with $1.7B of 0% convertible debt (May-2026) — flipping a net-cash company toward net debt.

The governance gap. Two facts make this pivot worth watching closely. First, management compensation has no return-on-capital governor. Per the 2026 proxy, the annual bonus keys on Revenue (and non-GAAP Operating Income), and long-term PSUs key on relative TSR vs. the SOX index — there is zero reference to ROIC, ROCE, or ROE anywhere in the incentive plan. This is the classic “Marathon size-mis-incentive”: pay rewards growth and size with no explicit reason to ensure capital earns its cost — precisely as the company embarks on $1.1B+ of deals funded by leverage. Second, insiders own <1% (~252K shares across all directors/officers) and have made zero open-market purchases in 2.5 years — every Form 4 is a grant, a tax-withholding, or a 10b5-1-planned sale (the CEO has sold ~$25M, programmatically). There is no insider buy-side signal of conviction in the pivot. Say-on-pay support is high (~96%), and the deals may well prove value-accretive — but the alignment scaffolding that would discipline them is absent. Verdict: a strong, disciplined five-year record now giving way to an aggressive, leverage-funded M&A pivot that the incentive system does nothing to govern. Capital allocation has moved from a clear positive to “on watch.”

8. Changes and Headwinds — Last Two Years

• Advanced-packaging cycle: trough → boom (2025 → 2026). The dominant change. A 2025 earnings air-pocket (margin compression, AP digestion) gave way to a sharp 2026 re-acceleration (revenue guided +30%, AP +50%), driven by HBM/2.5D and the Dragonfly G5 launch. Strengthens the thesis operationally; is the source of the rich multiple.
• Dragonfly G5 launch and qualifications (2026). New flagship inspection platform; qualified at a 2.5D logic customer and in HBM; pipeline of 15+ applications across 10+ customers; framed as a ~$1B new-market expansion. Strengthens — the single biggest forward driver.
• “Extended factories” manufacturing transition (2024–2025). Offshoring/expansion of manufacturing footprint that depressed 2025 margins during the transition and is now delivering the 2026 margin expansion. Headwind turned tailwind; also a supply-chain resilience improvement.
• Semilab acquisition (closed Nov-2025, ~$432M). Portfolio broadening into specialty/metrology. Neutral-to-positive; integration risk.
• Rigaku 27% stake + $1.7B convertible debt (Apr–May 2026). The balance-sheet pivot. Strategic optionality, but adds leverage, MtM earnings volatility, and dilution risk above ~$382.
• CFO transition (June 2025): Mark Slicer out; Brian Roberts in as CFO. Neutral; new CFO executing the 2026 guide credibly so far.
• Macro/policy: tariffs, fuel/shipping and memory-input cost headwinds (management flagged these as 2026 margin pressures it expects to overcome); US-China export-control overhang on the broader semicap complex. Headwind, manageable to date.

Verdict: the operational changes (AP re-acceleration, G5, margin ramp) clearly strengthen the near-term thesis; the financial changes (leverage, minority stake, dilution) add risk. Net: a stronger business on a riskier balance sheet, at a far higher price.

9. Risk Analysis (Risk Matrix)

The dominant risks are valuation and cycle — they are correlated (a WFE digestion would hit both earnings and the multiple at once, on a 2.39-beta stock). Business-quality and solvency risks are low.

10. Valuation Discussion (Embedded Expectations)

Where the multiple sits. Anchor to non-GAAP/cash earnings and EV/EBITDA, not trailing GAAP. At $333.76, ~50M diluted shares → market cap ~$16.6B; pro forma EV (after the converts, Semilab and Rigaku flows roughly net to modest net debt) ~$16.5–17B. Against this:

• Forward non-GAAP P/E ~47–48x on FY2026E non-GAAP EPS of ~$7.00 (Q1 $1.42 + Q2 guide $1.69 + a H2 ramp consistent with >$1.3B revenue and >30% exit margin).
• EV/Sales ~12.7x on FY2026E revenue of ~$1.3B.
• EV/EBITDA ~37x on FY2026E non-GAAP EBITDA (~$440M).
• Trailing GAAP P/E ~155x — an amortization/SBC artifact, not the right lens.
• AZI own-history percentiles: composite 99.2nd; P/E 97.9th; P/B 99.9th; P/S 99.9th — i.e., the richest the stock has ever been on every metric.

Versus its own history and peers. ONTO’s EV/EBITDA has swung from ~9x (2022 trough valuation) to ~36x (2023) over the cycle; today’s ~37x forward is at the top of that range and well above the ~28x average. Versus peers, Nova (NVMI) — the closest pure-play comp — trades ~11x EV/sales / ~35x EV/EBITDA (also richest-ever); KLA trades ~10x EV/sales / ~22–25x EV/EBITDA at far larger scale and lower growth. ONTO at ~12.7x EV/sales / ~37x EV/EBITDA is therefore at a slight premium to Nova and a clear premium to KLA — defensible if its 30%+ growth and advanced-packaging leverage are durable, expensive if the cycle turns.

Embedded expectations — what the price requires. Back into it: at ~$16.6B equity value, the market is paying ~47x ~$7 of 2026 non-GAAP EPS and implicitly requiring continued strong growth in 2027 to make that multiple compress to a “normal” 30–35x on forward earnings. For the current price to be merely fair (not cheap), ONTO must deliver: (i) the full 2026 ramp (>$1.3B, >30% exit margin); (ii) a stronger 2027 (management’s claim) — say revenue toward ~$1.6B and non-GAAP EPS ~$9–10; and (iii) the premium multiple largely holding. That is a flawless, two-year, peak-cycle path. The market is underwriting the content-growth bull case in full and pricing essentially no cycle risk and no execution slippage. Conversely, the market is not crediting much for the Rigaku optionality or a 2027 above-consensus surprise — those would be the upside to an already-demanding bar.

Scenario sketch (3-year, directional — not a forecast):

• Bear (~30%): AI/WFE capex digests in 2027; AP orders air-pocket; revenue/EPS flatten or dip; the record multiple reverts toward ~20–25x EV/EBITDA. Outcome: a large drawdown (a 2.39-beta name in a semicap down-cycle), cushioned only by the franchise quality and net-cash legacy. Echoes the 2024→2025 −59% round-trip.
• Base (~45%): 2026 delivers ~as-guided; 2027 grows but decelerates; multiple drifts down to ~30x forward non-GAAP as growth normalizes. Outcome: roughly flat-to-modest returns — the business grows into a still-rich multiple, with the entry price doing the work against you.
• Bull (~25%): AP/AI super-cycle runs through 2027–2028; ONTO outgrows WFE again, 2027 EPS ~$9–10+, G5 captures the “~$1B new-market” expansion, Rigaku adds optionality; the premium multiple holds. Outcome: further upside despite the rich start — the momentum trade keeps working.

No price target, no recommendation (per framework). The valuation conclusion: a high-quality, high-growth franchise priced at the very top of its own and its peers’ historical ranges, embedding flawless execution across two years with no margin of safety.

11. Variant Perception

Consensus. ONTO is a high-conviction AI-beneficiary “must-own” among semicap growth investors — a smaller, faster-growing, advanced-packaging-levered alternative to KLA, with a clean balance sheet, a visible 30%+ 2026 ramp, and a stronger 2027. Sell-side is bullish (fresh Buy initiations; e.g., Deutsche Bank Jun-2026). The tape agrees: a high-beta momentum melt-up at all-time highs.

Strongest bull case. Process control is the best room in semicap; ONTO owns the fastest-growing corner (advanced packaging) at the exact moment AI makes it indispensable. The 2025 margin trough was self-inflicted and transitional; the 2026 ramp is order-backed and the G5/Atlas/3DI/JetStep portfolio opens ~$1B of new markets. Content growth + share gains let ONTO outgrow WFE for years; the Rigaku/X-ray extension deepens the moat. A 30%+ grower with a real moat deserves a premium, and in an AI capex super-cycle the multiple can stay high longer than skeptics expect. Quality + Growth + Momentum are all in favor.

Strongest bear case. The price already discounts all of that — and then some. You are paying ~48x forward earnings and ~37x EV/EBITDA, at the 99th percentile of the stock’s own history, for a cyclical tool-sale business whose GAAP earnings fell 32% last year and which has twice halved in five years. The factor signature (beta 2.39, momentum/chip-superstars loadings, ~9.5 trailing-3-month Sharpe) is exactly what a crowded late-stage melt-up looks like; specific volatility is ~41%. Meanwhile management has levered up the balance sheet with $1.7B of converts to buy a Japanese minority stake, has no ROIC governor in its pay, owns <1%, and isn’t buying a share in the open market. When the AI-capex cycle digests — and it will, at some point — a 2.39-beta semicap at a record multiple is the wrong place to be.

The 3–5 assumptions that matter most:

1. Durability of the AP/AI-capex cycle through 2027 (does the order book hold, or air-pocket?). Falsifies bear if 2027 accelerates as guided; falsifies bull if WFE/AP digests.
2. Margin ramp to >30% exit op margin (is the extended-factory leverage real and sustained?). Q2–Q4 2026 prints settle it.
3. Multiple persistence (can a 99th-percentile multiple hold as growth normalizes?). History says peak multiples mean-revert; bulls say content growth justifies a structural re-rating.
4. Capital-allocation returns (do Semilab + Rigaku earn their cost of capital, or destroy value under a no-ROIC comp regime?).
5. Competitive position in front-end OCD vs. KLA (does Atlas G6 keep winning, or does KLA defend?).

Factor-positioning read (Momentum agent). ONTO is, empirically, a crowded, high-beta AI-semicap momentum trade at an extreme, not an abandoned value name. Beta 2.39; loads on “Chip Sector Superstars” (+1.51), Semiconductors industry (+1.46), Quality (+0.69), Growth (+0.56), Momentum (+0.55), and negatively on LowVolatility (−1.28) and Value (−0.13). Risk-adjusted track record is spectacular but unsustainable-looking: y1 return ~+244% (Sharpe 4.2), m3 ~+67% raw (annualized Sharpe ~9.5). Factor twins are the entire process-control/semicap complex — KLAC (0.88), Nova/NVMI (0.88), LRCX (0.87), VECO, AMAT, FORM, CAMT. R² ~0.70, specific vol ~41% (meaningful idiosyncratic risk on top of the high beta). This is the empirical backbone of the Claude's Take framing: a genuinely great business riding a momentum wave to a record valuation — the kind of setup that rewards patience for a better entry rather than chasing the tape.

12. Fact vs. Interpretation Table

13. Open Questions

1. What is the actual 2027 revenue/EPS trajectory management is implying when it says “stronger than 2026” — and how much is order-covered vs. pipeline?
2. How cyclical is the advanced-packaging order book really — would a 2027 AI-capex pause produce a 2023-style −19% revenue down-year, or has content growth structurally dampened the cycle?
3. What return will the Rigaku stake earn? Beyond the guided interest-offset, what is the realistic licensing/hybrid-metrology revenue ramp, and how much MtM earnings volatility will the fair-value accounting inject?
4. Semilab integration: margin/revenue contribution trajectory and whether it earns its ~$432M price.
5. Customer concentration: what share of revenue is the top 1–3 customers, and how exposed is ONTO to a single memory/logic customer’s capex timing?
6. OCD competitive trajectory vs. KLA: is Atlas G6 sustainably gaining gate-all-around share, or winning evaluations that KLA later defends?
7. Convertible dilution path: if the stock pushes through ~$382, how do the capped calls and 2026 buyback net out on share count?

14. What Must Be True (Bull and Bear, with Falsification Tests)

Bull case — what must be true:

• The advanced-packaging/AI-capex cycle keeps accelerating through 2027 (not just 2026), and ONTO outgrows WFE on G5/Atlas/3DI/JetStep share gains and content growth.
• The margin ramp sticks: non-GAAP operating margin exits 2026 >30% and holds/expands in 2027; non-GAAP EPS reaches ~$9–10 in 2027.
• The premium multiple persists (the market continues to treat ONTO as a structural AI-content compounder, not a cyclical), and Semilab/Rigaku add value rather than destroy it.
• Falsification test: a quarter (or guide) showing advanced-packaging revenue flattening/declining sequentially, a cut to the FY2026 >$1.3B / >30%-exit-margin framework, or a 2027 outlook that decelerates rather than accelerates — any of these breaks the bull case and would likely trigger a sharp de-rate on a 2.39-beta name.

Bear case — what must be true:

• The current price (99th-percentile multiple, ~48x forward EPS) is unsustainable; a WFE/AI-capex digestion or simple multiple mean-reversion drives a large drawdown.
• The 2025 margin trough and 32% earnings decline are a preview of cyclicality the market is ignoring; the leverage/minority-stake pivot adds risk into the top of the cycle.
• Falsification test: ONTO delivers the full 2026 ramp and an accelerating, order-backed 2027 (>30% margins sustained, EPS toward ~$9–10), the multiple holds because the growth is genuinely structural, and Semilab/Rigaku visibly contribute — which would prove the premium was warranted and the bear simply wrong on durability.

The pivot: both cases hinge on the same two observables over the next 12–18 months — the durability of the advanced-packaging order book into 2027, and whether the >30% exit margin is real and sustained. That makes ONTO a genuine “show-me into 2027” name: the price has already paid for the bull case, so the asymmetry favors waiting for either a better entry (de-rate) or hard confirmation (2027 acceleration) before adding.

15. Source Appendix

See the separate Source Appendix (ONTO_source_appendix.md) and Diligence Questionnaire (ONTO_diligence_appendix.md) accompanying this memo. Primary sources: ONTO SEC filings (10-K FY2025; 10-Qs; 8-Ks dated 2025-11-17 [Semilab close], 2026-04-21 [Rigaku], 2026-05-21 [convertibles]; DEF 14A 2026-04-06; Form 3/4/5 corpus), CIK 0000704532; the Q1-2026 earnings call transcript (2026-05-05); ROIC.ai aggregated financials/ratios/EV (reconciled to filings); AZI price history and valuation-percentile data (2026-06-18); FactorsToday factor model; and peer EV data (NVMI, KLAC). All figures reconciled to filings where the filing is authoritative.

APPENDIX A — Standard Diligence Questionnaire

APPENDIX A — Standard Diligence Questionnaire — Onto Innovation Inc. (NYSE: ONTO)

Supplemental to the research memo; grounded in the same evidence base. Fact / Interpretation / Assumption labels applied where it matters. As-of June 20, 2026; price $333.76.

General

What thoughtful questions have other investors asked about this company? Whether the 2026 advanced-packaging ramp is durable into 2027 or a cyclical peak; how much of the guided growth is order-covered vs. pipeline; whether the >30% exit operating margin is sustainable post extended-factory transition; what return the $710M Rigaku minority stake will earn and how much MtM earnings noise it injects; how exposed ONTO is to a single memory/logic customer’s capex timing; and — the core debate — whether a 99th-percentile valuation can hold for a cyclical tool-sale business.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Interpretation: On a trailing basis, 2025 was a self-inflicted margin trough (net income −32% despite record revenue). On a forward basis, 2026–2027 is a sharp up-cycle (revenue +30%, margins ramping to >30%). So the stock is priced on rapidly rising earnings off a depressed base — neither classic peak nor trough, but accelerating into what could become a peak.

Driven by the external environment or internal actions? Both. External: AI/WFE capex super-cycle (advanced packaging, HBM, gate-all-around). Internal: the Dragonfly G5 launch, extended-factory margin harvest, and share gains. The 2025 trough was largely internal/transitional (factory move); the 2026 boom is both internal (G5/margins) and external (AI capex).

How stable are revenues? Fact: Cyclical and lumpy — tool sales to a concentrated leading-edge customer base. Revenue fell ~19% in 2023. A recurring services/software annuity and the qualified installed base provide a partial floor, but the model is inherently capex-cycle-driven.

Outlook for products/services? Strong near-term: AP +50%, advanced nodes +25% guided for 2026; G5 opening ~$1B of new markets; Rigaku/X-ray extension; silicon photonics and panel-level packaging as emerging drivers. Management says 2027 > 2026.

How big will this market be? Process control / metrology / inspection is a multi-tens-of-billions, growing segment of WFE, structurally outgrowing wafer starts via node migration + advanced packaging. Global, with leading-edge demand concentrated in Taiwan/Korea/US/China.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Interpretation: Structurally concentrated (oligopoly: KLA dominant; Onto/Nova/Camtek the second tier). Advanced packaging is a newer, fast-growing front where competition is increasing as the prize grows, but high qualification barriers keep the field small.

How profitable is the business (ROIC, ROE)? Fact: ROE 19.4% (2025), 35.2% (2024), 84% (2022). GAAP consolidated ROIC ~6% — artificially depressed by $921M of merger goodwill/intangibles; tangible-capital ROIC is comfortably >25%. Gross margin 50–56%.

How profitable is the industry; barriers to entry? High — process-control players earn rich margins; barriers are qualification/recipe lock-in, decades of applications knowledge, proprietary algorithms, installed base, and service annuities.

Can the business be easily understood? Moderately — the franchise logic (qualify-in, switching costs, content growth) is clear; the product taxonomy (Dragonfly/Atlas/3DI/Iris/JetStep/Firefly) and the GAAP-vs-non-GAAP accounting require work.

Can it be undermined by foreign low-cost labor? Interpretation: No — this is high-IP precision capital equipment; the competitive threat is technological (KLA, Nova, Camtek), not labor-cost. China-localization of metrology is a long-term watch item.

Do brands matter / nature of competition? Reputation and qualified position matter enormously (yield risk makes customers conservative); competition is on technology, cost-of-ownership, and applications support — not price alone.

Customers’ switching costs? High — re-qualifying a process and risking yield on a multi-billion-dollar fab. This captivity is the core of the moat.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The qualified installed base, applications IP, and customer relationships generate returns well above the book carrying value — intangible franchise value not fully on the balance sheet (the reverse of the goodwill drag on ROIC).

Off-balance-sheet liabilities? None material identified beyond ordinary operating leases. The new $1.7B convertibles are on-balance-sheet; the Rigaku fair-value accounting will run through the P&L (MtM).

How conservative is the accounting? Fact: Conservative — SBC only ~2.7% of revenue, OCF 2.4x net income, no serial impairments, light capex. The main complexity is the large GAAP/non-GAAP gap (intangible amortization + SBC) and the forthcoming Rigaku MtM volatility.

How CapEx-hungry is the business? Fact: Capital-light — capex ~$28M/year, ~3% of sales. Cash is consumed by working capital (inventory build for the ramp) and now by M&A, not by physical capex.

Capital Allocation & Management

How much FCF; how is it used; philosophy? Fact: ~$300M FCF (2025). Historically: opportunistic buybacks ($65M/2022 … $75M/2025 under a $200M authorization), no dividend, small bolt-ons. 2025–26 pivot: ~$432M Semilab + ~$710M Rigaku, funded by $1.7B of 0% converts — a step-change toward aggressive, leverage-funded M&A.

Significant acquisitions recently? Fact: Semilab USA (~$432M cash + stock, closed Nov-2025); Rigaku 27% stake (~$710M, signed Apr-2026, closes 2H-2026) — the two largest deals in company history.

Buying back shares? Modestly/opportunistically; plus a ~$205M convert-linked repurchase (May-2026) that offsets dilution rather than a valuation buy.

Issuing large amounts of stock to insiders? SBC ~2.7% of revenue (clean). $1.7B converts add potential dilution above ~$382 (capped calls mitigate).

Compensation policy / incentive alignment? Fact: Annual bonus = Revenue + non-GAAP Operating Income (70% corporate / 30% personal); PSUs = relative TSR vs. SOX. No ROIC/return-on-capital metric. CEO 2025 comp ~$7.8M; say-on-pay ~96%. Interpretation: a “size/growth + relative-TSR” plan with no capital-efficiency governor — a concern given the leverage-funded M&A pivot.

Motivations of management? Interpretation: Operationally credible and execution-focused; but insiders own <1%, have made zero open-market buys, and the CEO is a steady programmatic seller — limited personal-capital alignment.

Valuation & Market Data

ADR / MLP / K-1? No — US C-corp, NYSE-listed common stock; standard 1099 treatment.

Dividend policy? None — no dividend; returns of capital have been buybacks only.

How profitable is the business? Highly, at the gross/operating level (50–56% GM; non-GAAP op margin ramping toward >30%); GAAP net margin understated by amortization.

Net income diverging from cash from operations? Fact: Yes, favorably — OCF ($328M) ran 2.4x GAAP net income ($137M) in 2025 (working-capital release + non-cash add-backs). Cash earnings exceed GAAP earnings.

Risks & Downside

What would cause the stock to decline? A multiple de-rate from its record (highest-probability risk on a 2.39-beta name); an AI/WFE capex digestion or advanced-packaging order air-pocket; a margin-ramp miss; disappointing returns on Semilab/Rigaku; convertible dilution.

Risk of catastrophic loss? Interpretation: Low — net-cash legacy, real FCF, hard IP/installed base, investment-grade-strong balance sheet even post-leverage. The risk is a large drawdown (valuation/cycle), not impairment or insolvency.

Chance of total loss? Very low — diversified blue-chip customer base, durable franchise, strong liquidity.

Recent News & Events

Has the business environment changed recently? Fact: Yes, dramatically — advanced packaging re-accelerated from a 2025 trough to a 2026 boom; Q1-2026 beat and FY guide raised to >$1.3B (+30%); fresh Buy initiations (Deutsche Bank, Jun-5-2026).

Significant acquisitions? Semilab (Nov-2025) and Rigaku 27% stake (Apr-2026) — see above.

Change in accounting policies? The Rigaku stake will use the fair-value option (MtM through other income) — new earnings-volatility source from 2H-2026.

Recent changes — markets, facilities, management? Extended-factory manufacturing transition (2024–25); CFO change (Brian Roberts, June 2025); Dragonfly G5 platform launch (2026); $1.7B convertible issuance (May-2026).

APPENDIX B — Source Appendix

APPENDIX B — Source Appendix — Onto Innovation Inc. (NYSE: ONTO)

As-of June 20, 2026. Primary sources prioritized; third-party aggregated data reconciled to filings where the filing is authoritative. CIK 0000704532.

Primary — SEC Filings (EDGAR; trailing 5-year corpus)

• Form 10-K (FY2025) — annual report; revenue, segment commentary, risk factors, buyback authorization, share count. Filed ~Feb-2026.
• Form 10-Q (Q1-2026, period ended 2026-03-31) — quarterly financials; GAAP results; Semilab contribution.
• Form 8-K (2025-06-12 / 2025-06-16) — CFO transition: Mark Slicer separation; Brian K. Roberts appointed CFO effective 2025-06-16.
• Form 8-K (2025-06-30; amended 2025-10-10; closed 2025-11-17) — Semilab USA acquisition: final ~$432.3M cash + 641,771 ONTO shares.
• Form 8-K (2026-04-21) — Rigaku investment: Share Purchase Agreement with Atom Investments (Carlyle affiliate) to acquire 61,123,436 shares = 27% of Rigaku Holdings for ~$710M; $500M Goldman Sachs bridge backstop; close expected 2H-2026.
• Form 8-K (2026-05-18 / 2026-05-21) — $1.5B + $200M greenshoe = $1.7B of 0.00% Convertible Senior Notes due 2031 (conversion ~$381.80; capped calls); ~$205M concurrent share repurchase (805,325 shares).
• DEF 14A (2026-04-06) — proxy: compensation metrics (bonus = Revenue + non-GAAP Operating Income, 70/30; PSUs = relative TSR vs. SOX; no ROIC metric); NEO comp (CEO Plisinski $7.81M 2025); say-on-pay ~96%; insider ownership <1%.
• Forms 3/4/5 corpus (2021–2026) — insider transactions: zero code-P open-market purchases in 2.5 years; CEO ~$25M of 10b5-1-planned sales; balance are grants/tax-withholding.
• Earnings 8-Ks: 2026-02-19 (FY2025), 2026-04-16 (Q2 pre-announce), 2026-05-05 (Q1-2026).

Primary — Earnings Call Transcript

• Q1-2026 earnings call (2026-05-05) — CEO Michael Plisinski, CFO Brian Roberts. Source of: FY2026 guidance (revenue >$1.3B / +30%; AP +50%; advanced nodes +25%; Q4 exit op margin >30%; H2 ≥ +15% over H1); Q2 guide ($320–330M revenue, GM 56–56.5%, op margin 28–28.6%, EPS ~$1.69); Dragonfly G5 qualifications (2.5D logic, HBM; 15+ applications/10+ customers); Rigaku rationale and accounting; segment revenue split.

Quantitative Data Sources (reconciled to filings)

• ROIC.ai — multi-year income statement, balance sheet, cash flow, profitability ratios, enterprise value, valuation multiples (ONTO; peers NVMI, KLAC). Third-party aggregated; reconciled to filings.
• AZI price history & valuation-percentile data (as-of 2026-06-18) — daily OHLCV, EMAs, beta; valuation_index own-history percentiles (composite 99.2nd; P/E 97.9th; P/B 99.9th; P/S 99.9th; price $333.76, TTM EPS $2.15, BVPS $42.64).
• FactorsToday factor model — stock loadings (beta 2.39; Chip Sector Superstars +1.51, Semiconductors +1.46, Quality +0.69, Growth +0.56, Momentum +0.55, LowVolatility −1.28, Value −0.13); leaderboard (y1 +244% / Sharpe 4.2; m3 ~+67% raw); specific vol ~41%; related stocks (KLAC 0.88, NVMI 0.88, LRCX 0.87, VECO, AMAT, FORM).
• AZI news feed — recent items incl. Deutsche Bank Buy initiation (2026-06-05); Q1 earnings beat (2026-05-06); sector momentum.

Peer / Cross-Read

• Prior same-sector reports referenced for industry/valuation framing: KLAC, AMAT, LRCX (2026-06-10), ENTG (2026-06-19), Q/Qnity (2026-06-20).
• Peer EV/valuation data: Nova (NVMI) — closest pure-play metrology comp (~11x EV/sales, ~35x EV/EBITDA, also richest-ever); KLA (KLAC) — process-control leader (~10x EV/sales, ~22–25x EV/EBITDA at larger scale/lower growth).

Key Quantitative Facts (selected, all reconciled)

• FY2025: revenue $1,005.3M; GAAP net income $136.8M; diluted EPS $2.78; gross margin 49.7%; op margin 15.1%; OCF $328M; FCF ~$300M; SBC $27.6M (~2.7% of revenue).
• Q1-2026: revenue $292M; GAAP op margin 13.3% / EPS $0.68; non-GAAP op margin 26.7% / EPS $1.42.
• Balance sheet (Q1-2026): cash + ST investments $654M; debt ~$17.5M; net cash ~$637M; tangible equity ~$1.2B; goodwill $643M + intangibles $278M.
• Pro forma (post May-2026 converts + pending Rigaku): ~net-debt-neutral to modestly net-debt; $1.7B converts due 2031.
• Valuation at $333.76: market cap ~$16.6B; EV ~$16.5–17B; ~48x forward non-GAAP EPS (~$7); ~12.7x forward EV/sales; ~37x forward EV/EBITDA; ~155x trailing GAAP EPS.

Management commentary and guidance are treated as hypothesis, not evidence, and validated against filings and financials.