Entegris, Inc. (NASDAQ: ENTG) — The Earnings Caught Up; the Valuation Only Halfway Did

Update report | 20 August 2026 | Reference price: $144.28 (19 August 2026 close)

⚡ Claude’s Take

This is the author’s own independent opinion and general information, not investment advice. The analysis in sections 1–15 below carries no position and no price target; the single directional view is confined to this clearly labeled block.

Verdict: HOLD / wait for a better entry at $144.28. The prior AVOID call is upgraded because the operating evidence improved and the stock de-rated, but the proof is not complete. A more attractive accumulate-on-weakness zone is roughly $115–$130, equivalent to about 21–23x a plausible $1.0 billion FY2026 adjusted-EBITDA base. Conviction: medium.

Q2 was the quarter the June report needed to see. Revenue accelerated 11.5%, APS grew 17%, adjusted EBITDA margin reached 28.4%, first-half free-cash-flow margin moved into the mid-teens, and net leverage fell to 3.4x. The stock also retreated from June’s extraordinary valuation: at $144.28, spot EV is about $25.4 billion, or 27.4x trailing adjusted EBITDA and 38.5x trailing free cash flow, versus roughly 35x and 69x in the prior report. That changes the asymmetry from plainly hostile to merely demanding.

The improvement is not clean enough to call victory. The 47.6% reported gross margin received roughly 150 basis points from longer depreciable lives; Materials Solutions grew only 5% against management’s 7–8% MSI estimate; Q2 free-cash-flow margin was 13.2% on a simple filing definition; leverage remains above 3x; and there is still no buyback. APS volume and fixed-cost absorption, rather than demonstrated broad pricing power or materials share gain, drove the best part of the print. The framing is therefore quality franchise plus balance-sheet repair, still priced for successful execution. Factor evidence says the stock is positive Quality but overwhelmingly a high-beta semiconductor/equipment exposure, with beta near 2.5 and 36% annualized specific volatility. June’s 42.8% spike and 41.8% reversal were largely cohort moves; the 15.5% Q2-results jump was company-specific. At today’s price, that makes ENTG neither a value setup nor an obvious momentum chase. It is a prove-it story with real assets and less valuation risk than two months ago.

📈 Stock Price Action — Five-Year Event Map

Entegris has completed several violent cycle turns in five years. The stock rose with the 2021 semiconductor super-cycle, fell about 60% through the 2022 rates-and-cycle shock as the debt-funded CMC acquisition closed, recovered with the AI semiconductor complex in 2023–24, sank to $68.64 in November 2025, and then reached $183.84 on June 22, 2026. The latest close of $144.28 is 21.5% below that high but still 110% above the November low. Price moves below are facts; driver attributions are interpretations based on filings, company events and identical-window peer moves.

The quantitative tape is unusually high dispersion: the latest one-year return is about 79.9%, annualized volatility 67.7%, and one-year maximum drawdown 41.8%. The 21-day EMA sits slightly above the latest price, the 50-day EMA slightly below it, and the 200-day EMA well below it. Those are descriptive trend measures, not valuation evidence.

Changes since June 19, 2026

• Operating evidence improved. Q2 sales rose 11.5% to $883.2 million; adjusted EBITDA was $250.7 million at 28.4%; APS grew 17% and its GAAP segment margin reached 29.3%. Management guided Q3 sales to $905–$935 million and adjusted EBITDA margin to 28–29%.
• The margin headline needs normalization. Longer useful lives reduced Q2 depreciation by $18.7 million and first-half depreciation by $37.7 million. The change supplied roughly half of first-half GAAP operating-income growth. Adjusted EBITDA, which adds depreciation back, remains the cleaner evidence of genuine improvement.
• Cash and deleveraging progressed. First-half operating cash flow was $339.2 million and simple free cash flow was $258.4 million, or 15.2% of sales. Entegris repaid $250 million of term debt in the half and another $25 million after quarter-end; filing-derived net leverage is approximately 3.37x pro forma for the July payment.
• The outgrowth test is only partially tracking. Management’s unit-driven revenue grew about 10% against MSI of 7–8%, while external wafer shipments grew 7.4%. That is only 2–3 points of unit outgrowth, and MS itself grew 5%. One borderline quarter does not establish the required several-quarter content/share trend.
• The tape reset, then rewarded Q2. ENTG gained 42.8% from June 10–22 and fell 41.8% from June 22–July 29 alongside its semiconductor peers. On August 4 it rose 15.5% versus 3.5–7.9% for MKSI, AMAT, LRCX and KLAC, a genuine company-specific response. The latest price is 21.5% below the June 22 high.
• The prior call changes, but the bear case is not falsified. Free cash flow is the only one of the four core thresholds presently met. Adjusted EBITDA margin remains below 30%, leverage remains above 3x, and buybacks have not resumed. No new acquisition was announced, so the capital-allocation kill switch was not triggered.

1. Executive Summary

Entegris is a $3.3 billion trailing-revenue supplier of yield-critical materials and contamination-control products consumed in semiconductor fabrication. Its two segments sell CMP slurries and pads, deposition and ion-implant precursors, specialty chemistries, liquid and gas filters, purifiers, wafer carriers, FOUPs and fluid-handling systems. Approximately 75% of revenue is tied to wafer starts and recurring consumption; about 10% is tied to wafer-fabrication equipment and 15% to fab construction. That mix makes Entegris less cyclical than a pure equipment supplier, though not immune to capital cycles or customer utilization.

The economic attraction is customer captivity. A filter, chemistry or slurry qualified into a process-of-record is cheap relative to the wafer value it protects. Requalification consumes engineering time and risks yield. Customers therefore do not switch merely to save a small percentage on input cost. Entegris reinforces those switching costs with process IP, clean manufacturing, applications engineering and a broad portfolio. The evidence is strongest in liquid filtration, advanced purity and selected materials niches; it is not a semiconductor-wide monopoly. The company names formidable competitors including Pall/Danaher, Merck KGaA, Qnity, Air Liquide, Linde, Shin-Etsu Polymer, Gudeng, Aicello and Chinese suppliers. A defensible label is a narrow-to-moderate moat with strong product-level captivity and uneven enterprise-level proof.

Q2 2026 was a meaningful operating step. Revenue rose 11.5% to $883.2 million, adjusted EBITDA increased to $250.7 million, and adjusted EBITDA margin improved 110 basis points to 28.4%. APS supplied nearly all the operating leverage: revenue rose 17% and GAAP segment profit rose 57%, taking segment margin to 29.3%. MS revenue rose 4.6%, segment margin reached 20.9%, and growth lagged both management’s 7–8% MSI assumption and external wafer-shipment growth of 7.4%. Total unit-driven sales grew 10%, only 2–3 percentage points ahead of the market. The quarter supports a recovery but does not yet prove the 3–6 points of durable content-per-wafer outgrowth on which the long-term thesis depends.

Reported profitability also contains a material accounting tailwind. Entegris lengthened useful lives for buildings, manufacturing equipment, cylinders, molds and laboratory equipment beginning in January. Depreciation fell $18.7 million in Q2 and $37.7 million in the first half; the change contributed about 150 basis points to Q2 gross margin and 48.5% of the first-half GAAP operating-income increase. A rough comparable Q2 gross margin is 46.1%, still better than 44.6% a year ago. Adjusted EBITDA is unaffected by the depreciation change, so its 110-basis-point improvement is economically meaningful. The correct conclusion is real operating improvement, not the full 320-basis-point headline gross-margin expansion.

Cash conversion and balance-sheet repair are the strongest parts of the update. First-half operating cash flow rose to $339.2 million; simple free cash flow was $258.4 million, or 15.2% of revenue, versus $79.4 million a year earlier. Trailing free cash flow is approximately $575 million, a 17.3% margin. But first-half capex was only $80.8 million against a roughly $250 million full-year plan, leaving back-half spending. Receivables rose $101 million and inventory rose $83 million, partly financed by a $115 million increase in payables and accruals. Cash quality is improving, but quarterly cadence and working-capital investment matter.

The company repaid $250 million of term debt in the first half and $25 million after quarter-end. Net debt was about $3.14 billion at June 27, or 3.39x trailing adjusted EBITDA; pro forma July leverage is about 3.37x. Management expects the ratio to begin with a two by year-end. The balance sheet has $1.10 billion of gross liquidity and no 2026–27 maturity, but $2.175 billion remains due in 2029 after the July payment. Debt-first allocation is appropriate. The absence of buybacks, continuing share-plan dilution and openness to future M&A mean the eventual use of post-deleveraging cash is still unresolved.

At $144.28, diluted market capitalization is approximately $22.16 billion and spot enterprise value about $25.37 billion. Against trailing sales of $3.326 billion, adjusted EBITDA of $925.6 million, non-GAAP EPS of $3.21 and free cash flow of $575.2 million, the stock trades at approximately 7.6x EV/sales, 27.4x adjusted EBITDA, 45.0x non-GAAP EPS and 38.5x free cash flow. Its own-history P/S rank has fallen from the 97.5th percentile in June to the 84.8th percentile. A reverse DCF still requires roughly 16.3% annual free-cash-flow growth for ten years at a 10% discount rate and 4% terminal growth. The valuation is less extreme, but it still assumes successful margin, outgrowth and deleveraging execution.

The institutional conclusion is balanced. The recurring consumables engine, qualification captivity, APS capacity absorption, AI/HBM exposure and falling interest burden are real. Against them sit the CMC acquisition’s still-poor returns, accounting-aided margin comparisons, unproven MS share/outgrowth, China and customer concentration, a late-cycle investment boom, and an equity valuation that offers limited forgiveness. The evidence that matters next is not another total-sales beat; it is several quarters of unit-driven growth at least 3–6 points above a consistent MSI benchmark, MS growth above market, adjusted EBITDA margin above 30%, full-year free-cash-flow margin at or above 15% after planned capex, and leverage below 3x without a new large transaction.

That evidence set is observable, time-bounded and capable of disproving either side of the debate.

2. Business Overview

Portfolio and segments

Entegris sits between raw-material suppliers, equipment OEMs and chip manufacturers. It formulates, purifies, filters, contains and delivers materials inside extremely clean process environments. The input cost is usually small compared with the value of a finished wafer, yet a contaminant or chemistry deviation can ruin yield. This asymmetry is the basis of both the business model and the moat.

Materials Solutions generated $1.407 billion of FY2025 revenue and $276.6 million of segment profit. It includes CMP slurry, pads and cleans; advanced deposition precursors; ion-implant gases; formulated etch and cleaning products; and other specialty materials. CMC Materials contributed much of this portfolio. MS is closest to the pure content-per-wafer thesis: products are consumed as wafers move through more steps and more complex structures. Q2 2026 sales were $371.3 million, up 4.6%, while segment profit was $77.7 million and margin 20.9%. Management is investing in molybdenum, selective etch and other programs, so near-term margin is not expected to expand rapidly.

Advanced Purity Solutions generated $1.799 billion of FY2025 revenue and $426.4 million of segment profit. It includes liquid filtration, gas purification, fluid management, wafer and reticle handling, FOUPs, sensing and control. The mix spans recurring consumables and capex-sensitive handling products. Q2 sales were $514.6 million, up 17%; GAAP segment profit was $150.9 million, up 57%, and margin reached 29.3%. On an adjusted basis, segment profit was $155.8 million and margin about 30.3%. Management cited the strongest FOUP demand in more than three years and a fourth consecutive record quarter in liquid filtration.

The segment contrast is essential. APS currently demonstrates the best growth and operating leverage, helped by the recovery in fab construction, filtration volumes and absorption of recently built capacity. MS owns many of the most strategically interesting materials positions, but its Q2 growth did not keep pace with market measures. Treating consolidated 11.5% growth as proof that every product gained content or share would obscure the actual mechanism.

Demand model and end markets

Management divides revenue into roughly 75% unit-driven, 10% WFE-driven and 15% fab-construction-driven. Unit-driven products are consumed with wafer starts; the relationship is recurring but not contractual in the software sense. Wafer starts can fall, customers manage inventory, qualifications can shift, and prices can move. Nevertheless, production consumption is structurally steadier than equipment orders.

Advanced logic represents about 40% of revenue, memory about 30%, and mainstream logic/other applications the balance. AI affects several layers. Leading-edge logic requires more deposition, etch, filtration and purity. HBM adds DRAM complexity, TSV and CMP intensity. Advanced packaging creates additional filtration, cleaning, bonding and handling steps. Management says advanced-packaging revenue runs around $100 million, material enough to monitor but still small relative to consolidated sales. Molybdenum precursor demand is reportedly growing rapidly as device architectures change, but product-level dollars and margins are not disclosed.

Customer and geographic concentration remain high. TSMC accounted for 16% of FY2025 sales and the top ten customers 50%. International markets supplied 82% of revenue. Q2 Taiwan revenue was $225.4 million, up 29%, and represented 26% of total. China was $168.8 million, up 2%, or 19%; within China, MS declined while APS grew. The different product trajectories make a single “China growth rate” analytically misleading.

Revenue visibility and capacity

Visibility comes from process positions, customer roadmaps and the order sequence from construction to tool installation to production consumption. Management tracks more than 20 leading-edge expansions and more than 50 engineering projects across its top 35 customers. It describes eight to ten advanced-logic projects, seven to eight advanced-memory projects and six to eight packaging projects. These counts are management disclosures, not audited backlog. They show opportunity breadth but not conversion probability or margin.

Recent capital spending created what management calls significantly more than $1 billion of incremental revenue capacity with limited additional investment. The KSP filtration facility in Taiwan is moving through qualifications and was expected near breakeven in the second half. Colorado and other sites should ramp later. That capacity can create powerful incremental margins if demand arrives; it can also depress returns if qualifications or utilization lag. The missing data are site-level utilization, product mix, incremental margin and timing.

Business-overview verdict

The core business is understandable and attractive: recurring consumption of yield-critical inputs, secular process complexity and costly qualification. Its complexity lies in a portfolio where product families have different cycle, share and margin characteristics. APS currently provides the strongest financial evidence; MS carries much of the content-per-wafer promise but needs better proof. The portfolio should be analyzed as a collection of qualified niches, not as one undifferentiated wide moat.

How the economics accrue

The recurring character should not be confused with contractual recurrence. A fab typically buys qualified products repeatedly because the recipe, tool configuration and quality system make continuity valuable, not because it signed a long software subscription. Revenue therefore depends on wafer volume, customer inventory and yield roadmaps. This distinction explains why Entegris can retain process positions and attractive gross margin while consolidated sales still decline during a semiconductor correction.

The profit pool also differs by product. High-purity filters and proprietary formulations can earn attractive contribution margins because the avoided cost of contamination is enormous. FOUPs, fluid-handling infrastructure and some construction products carry more competitive and volume sensitivity. MS has substantial acquisition amortization and R&D demands. Consolidated segment disclosure cannot isolate each niche, so analysts should resist applying APS’s current 30%-ish segment margin to the whole portfolio.

Working capital is another economic cost. Customers require assured supply, exact lots and regional redundancy; Entegris carries inventory and manufacturing capacity to provide it. Inventory of $703.9 million is more than one-fifth of trailing sales. Some of that inventory supports captivity and continuity, but it also reduces capital efficiency. The moat is therefore partly built with balance-sheet assets, not only intangible know-how.

The company claims more than 30,000 products and deep customer engagement. Breadth reduces dependence on any one SKU and creates multiple routes into a fab, yet it complicates pruning and disclosure. No single product is material enough to explain company performance, and a handful of large customers set demanding qualification and commercial terms. That combination tends to create stable product positions but also makes total-company pricing power less obvious than the yield-critical narrative implies.

3. Industry Dynamics

Market structure and size

The 2025 semiconductor-materials market reached $73.2 billion, up 6.8%, according to SEMI. Wafer-fabrication materials were $45.8 billion and packaging materials $27.4 billion. Taiwan was the largest regional materials market at $21.7 billion and China the second largest at $15.6 billion. Those figures are broad context, not Entegris’s serviceable market: they include silicon wafers, photoresist and packaging inputs Entegris does not broadly supply, while some APS equipment and handling products sit outside a materials-only measure.

Observable 2026 category anchors include approximately $3.9 billion for CMP slurry and pads, $4.1 billion for wet chemicals and $6.8 billion for electronic gases. They overlap with MS but cannot simply be added into a precise company TAM. A sound serviceable-market analysis requires product-by-product participation, qualification and geography filters, plus separate filtration, purification and handling categories. Public data do not support a reliable unified SAM or category-share series, so this report avoids false precision.

Industry economics are fragmented at the top level and oligopolistic in individual qualified niches. The relevant competitive market may be one chemistry at one process step for one node, not “semiconductor materials.” Scale helps through R&D, clean production, customer support and redundant local supply, but it does not erase specialist competitors. Customers frequently dual-source critical inputs to manage risk, and domestic Chinese suppliers are broadening capability.

Capital-cycle position

The industry is in a strong investment phase. SEMI forecasts 2026 semiconductor-equipment sales of $165.9 billion, up 23.2%, with WFE at $143.9 billion, up 23.1%, followed by 21.8% growth in 2027 and 14.1% in 2028. Installed fab capacity, however, is forecast to grow around 5% in both 2026 and 2027. The gap reflects high technology intensity and expensive leading-edge builds, but it also matters in Marathon’s capital-cycle framework: rapid investment today creates future supply and can eventually produce utilization or inventory digestion.

Entegris benefits in three waves. Fab construction pulls fluid handling, FOUP and purity infrastructure; tool installation pulls WFE-linked products; subsequent wafer starts consume filters and materials. This staggered sequence can extend the revenue cycle beyond equipment peers. It also creates analytical traps: today’s 20–30% WFE order-rate claim cannot be capitalized as permanent consumables growth, and fab counts do not guarantee loaded utilization.

The 75% unit-driven mix dampens the capital-cycle swing, but unit demand itself is cyclical. A smartphone, PC or memory inventory correction reduces wafer starts. New fabs may cannibalize old capacity. Subsidized Chinese capacity can create structurally low returns while still buying local materials. The appropriate conclusion is lower cyclicality than equipment, not acyclicality.

Technology and regulation

Gate-all-around transistors, more 3D NAND layers, HBM, backside power, EUV and advanced packaging generally increase steps, purity requirements or novel materials. Those trends are favorable to content per wafer. Entegris reported a cross-license with JSR/Inpria around metal-oxide-resist IP in May and favorable company-reported decisions defending elements of its CMP slurry patent portfolio in August. Both demonstrate relevant IP positions, but neither discloses near-term revenue economics.

China is a two-sided issue. It represented 18.6% of first-half sales, down from 20.3%, while dollars fell less than 1%. Local trailing-node expansion supports APS and selected materials; advanced-node controls, entity restrictions, origin rules and foreign-owned-fab licensing create risk. BIS controls cover specified semiconductor-manufacturing-equipment end uses, HBM and advanced capabilities, and the 2025 removal of VEU treatment tightened supply to selected foreign-owned China fabs. A January 2026 case-by-case policy for certain advanced computing chips is not a broad relaxation for manufacturing materials or equipment.

Entegris mitigates policy risk through local-for-local manufacturing, regional redundancy and product-specific licensing. The 10-Q nevertheless warns that tariffs, alternative suppliers, origin rules and foreign export controls on raw materials can add cost or delay. Local competitors such as Cobetter increasingly offer filtration, PFA tubing and valves, high-purity packaging and related systems. Their existence does not prove leading-edge equivalence, but it weakens any assumption that localization preserves price and share automatically.

Industry verdict

The industry seat remains structurally attractive: high yield sensitivity, recurring consumption, rising process complexity and niche oligopolies. Near-term demand is strong and supports Entegris’s underused capacity. The counterweight is classic late-cycle supply behavior—investment and WFE growth far above installed-capacity growth—plus policy-driven duplication and Chinese localization. The industry supports earnings recovery today while increasing the need for discipline in long-duration forecasts.

4. Competitive Position

Greenwald moat map

Demand-side captivity is the primary advantage. Semiconductor customers qualify materials and contamination-control products into process recipes. Reformulation and requalification take time, absorb engineering resources and can introduce yield risk. Because the product cost is small relative to wafer value, modest price savings rarely compensate for that risk. Once a supplier has a process-of-record position and performs reliably, inertia is economically rational.

Process IP and application know-how reinforce captivity. Entegris employs roughly 1,600 people in engineering and R&D and maintains hundreds of commercial/application personnel. Formulations, contamination science, patents, manufacturing recipes and accumulated failure data are category-specific intangibles. The August patent decisions involving colloidal-silica CMP slurry support enforceability in one niche; they do not establish a portfolio-wide monopoly.

Niche supply scale is meaningful but not universal. Clean manufacturing, regional applications labs, redundant capacity and customer support create barriers where volumes are too small for commodity scale but quality requirements are high. Recent capacity can improve service and reduce sole-source risk. Yet management does not disclose utilization or replacement-cost economics, so a broad cost advantage cannot be quantified.

Breadth helps co-optimization but is not a standalone moat. Entegris can combine slurry, pad, cleans, precursors, filters, containers and delivery systems. That may reduce customer qualification burden and create earlier access to roadmaps. Management’s own call also showed that different products can have separate buying centers, limiting simplistic cross-sell assumptions. No network effect exists, and brand matters mainly as a signal of validated reliability.

Tests against operating evidence

Gross margin stayed between roughly 42% and 46% through the 2022–25 downturn and reached a normalized roughly 46.1% in Q2. That resilience is consistent with switching costs and specialized value. APS’s record filtration quarters and 29.3% GAAP segment margin are favorable evidence. However, post-CMC enterprise returns remain poor: normalized trailing GAAP ROIC is approximately 5.6%, or about 8.1% after acquisition-amortization adjustments, far below estimated pre-CMC returns of 14–18%.

The distinction is crucial. A company can own good assets and still destroy shareholder value by paying too much. Greenwald’s market-share and return tests must be applied after acquisition price, not merely at product gross margin. Public category-share data are sparse, MS grew only 5% in Q2, and broader Qnity Semiconductor Technologies and Merck Semiconductor Solutions portfolios grew 16% and 17.3%, respectively. Those are not clean apples-to-apples comparisons, but they prevent an unsupported share-gain conclusion.

Pricing power is also unproven. Management embedded no material pricing in Q3 guidance and emphasized volume, absorption and execution. Longer-term supply agreements can deepen visibility and captivity, but there is not yet evidence of price realization consistently above inflation. The current margin recovery should be attributed primarily to volume, mix, fixed-cost absorption, productivity and accounting—not blanket price increases.

Competitive threats

APS competes with Pall/Danaher, Shin-Etsu Polymer, Cobetter, Gudeng, Aicello, Mersen and others. MS competes with Merck KGaA, Qnity, Air Liquide, Linde, Anji Microelectronics and specialist suppliers. Customers can sponsor second sources, local governments can subsidize entrants, and technology transitions can displace a process position. The danger is not usually an overnight commoditization of the whole company. It is a collection of incremental share losses or missed new-node qualifications that compound while consolidated reporting obscures them.

Customer concentration can strengthen and weaken the moat simultaneously. Deep TSMC engagement supports co-development and qualification captivity, but a lost leading-edge position at a 16%-of-sales customer would be material. Taiwan’s 29% Q2 growth is encouraging; it does not isolate TSMC or prove content per wafer. The most useful disclosure would be product-family organic growth versus the same MSI benchmark for eight quarters, with share and price/mix bridges.

Competitive-position verdict

Entegris has genuine product-level captivity and selected intellectual-property advantages. The moat is strongest in liquid filtration, high-purity handling and qualified specialty materials. It is not yet proven wide across the consolidated portfolio because share data are sparse, MS undergrew the market in Q2 and acquisition-adjusted returns remain modest. The correct judgment is narrow-to-moderate, durable in niches, and requiring better market-share evidence.

EPV and asset-value perspective

Greenwald’s earnings-power-value framework provides a useful cross-check. A protected franchise should earn sustainably above the cost of reproducing its assets. Pre-CMC returns and persistent gross margins suggest Entegris’s legacy niches did so. Post-CMC, normalized ROIC of 5.6% on GAAP economics or about 8.1% after amortization adjustments is near or below a reasonable cost of capital. That does not mean the moat vanished; it means the company paid away much of the franchise value to the seller and financed the purchase expensively.

Asset value is difficult to estimate because a replacement entrant would need more than PP&E. It would need clean manufacturing, qualification history, process recipes, customer trust, applications engineers and years of trial data. Book value understates those internally created assets. Conversely, the $3.949 billion goodwill balance records what Entegris paid, not what a new entrant must spend efficiently. The gap between real intangible barriers and inflated acquisition goodwill is why neither book value nor a simple asset multiple captures the economics.

For the moat to translate back into enterprise-level value, earnings must grow into the acquired capital base. That can happen through CMC product growth, co-optimized wins, utilization of new facilities, debt-funded interest savings and restrained future investment. It cannot happen merely by extending useful lives or reporting adjusted EBITDA before amortization. A rising normalized ROIC series is the decisive validation metric.

Share-stability evidence hierarchy

The strongest evidence would be independent market share by product and node over time. That is unavailable. The next best evidence is organic product-family growth against a consistent market denominator, customer qualification disclosures and stable normalized margins. Q2 offers positive liquid-filtration records and Taiwan growth, but negative relative MS growth. Management’s defense of CMP positions is informative but remains testimony. The evidence hierarchy therefore supports confidence in existing captivity, not confidence in broad incremental share gains.

5. Growth History and Forward Opportunities

Revenue rose from $2.299 billion in FY2021 to $3.282 billion in FY2022 largely because CMC closed in July, then peaked at $3.524 billion in FY2023 before declining to $3.241 billion in FY2024 and $3.197 billion in FY2025. This is not a clean organic compounding record. Portfolio sales and the semiconductor downturn complicate comparison, and FY2025 revenue remained below FY2023 despite years of added capital.

The 2026 recovery is broader. First-half revenue reached $1.695 billion, up 8.3%. Q1 unit-driven revenue grew 7%; Q2 grew 10%, WFE-linked sales about 10% and fab-construction sales 15%. Management increased its MSI assumption to 7–8%. Q3 guidance midpoint of $920 million and commentary for approximately 4% sequential Q4 growth imply illustrative FY2026 sales around $3.57 billion, about 12% growth. That is a calculation based on guidance and commentary, not formal full-year guidance.

The highest-quality growth vectors are those tied to process complexity rather than industry spend alone:

• Advanced logic and gate-all-around add deposition, etch, clean and filtration requirements.
• HBM and TSV increase CMP, cleaning and handling intensity; advanced-memory projects provide a multiyear qualification runway.
• Molybdenum and other new deposition materials can create high-value positions, though management’s product-growth claims are not separately filed.
• Advanced packaging adds process steps and purity requirements, but the current roughly $100 million run rate is only about 3% of sales.
• Liquid filtration benefits from both new capacity and recurring replacement consumption.
• Recently built sites provide operating leverage if qualifications convert, potentially allowing more than $1 billion of incremental revenue without another capex super-cycle.

The lower-quality growth is simple exposure to construction and WFE at a time when equipment spending is growing above 20%. It is welcome earnings, but it should not receive a perpetual content-growth multiple. The current quarter blends both forms: consolidated growth exceeded wafer shipments by four points, while unit-driven growth exceeded them by only 2.6 points and MS lagged. The proper score is early/borderline, not passed.

Growth risks include customer inventory, fab delays, qualification failures, export controls, local Chinese competition and utilization digestion after the build cycle. Management added direct factory labor by a double-digit percentage from Q4 to unlock capacity. If demand arrives, this improves absorption; if it does not, the labor and fixed base will pressure margins. The November 9 Investor Day should clarify capacity utilization, segment targets and capital allocation.

Growth verdict: the trajectory has improved and the pipeline is credible, but the mix of outgrowth is not yet strong enough to establish a durable 3–6-point content advantage. Several quarters of MS and unit-driven performance above a consistent MSI baseline remain the essential proof.

6. Financial Quality

Five-year record

The record shows three phases: a high-return pre-CMC franchise, a debt-and-capex shock in 2022–23, and cash repair in 2024–26. FY2022 free cash flow turned negative because capex reached $466 million as acquisition debt arrived. Revenue and GAAP operating profit have not yet surpassed FY2023, but capex normalization has lifted free cash flow materially.

Q2 earnings quality

Q2 sales rose $90.8 million to $883.2 million. Gross profit rose to $420.0 million and reported gross margin to 47.6%. GAAP operating income was $164.6 million at 18.6%; net income was $93.6 million. Adjusted operating income was $216.3 million at 24.5%; adjusted EBITDA was $250.7 million at 28.4%; non-GAAP EPS was $0.93.

The useful-life change is the main accounting caveat. Manufacturing-equipment life increased from 5–10 years to 14 years, buildings from 5–35 to 12–40, and several other asset classes also lengthened. It reduced first-half depreciation by $37.7 million, increased first-half net income by $23.3 million and EPS by $0.15. Removing it takes first-half operating margin from 18.1% to about 15.8% and Q2 from 18.6% to about 16.5%. Normalized adjusted operating margin is roughly 21.9% for the half and 22.4% for Q2. Adjusted EBITDA is unaffected, making its increase from 27.3% to 28.4% the strongest clean margin evidence.

Operating expenses contain a second nuance. First-half SG&A rose to 14.4% of sales from 13.9%, largely from variable compensation and professional fees. R&D fell to 9.3% from 10.8%, owing to lower project spending and depreciation. Some operating leverage therefore reflects lower R&D intensity. That may be sensible project discipline, but sustained content growth depends on reinvestment; a durable model should not assume R&D keeps shrinking as a percentage of sales.

Cash conversion and working capital

First-half OCF increased $85.3 million and simple free cash flow increased $179.0 million. The improvement is real but partly capex timing: only $80.8 million was spent against a roughly $250 million full-year plan, implying about $169 million in the second half if guidance holds. Company-defined free cash flow includes $5.4 million of government incentives and is approximately $264 million; the consistent OCF-less-PP&E definition used here is $258.4 million.

Working capital consumed cash. Receivables rose $101.2 million and inventory $83.3 million, while payables and accruals increased $114.5 million. The build is compatible with growth and supply preparation, but it is not free. Q2-only OCF was $156.2 million, capex $39.3 million and free cash flow $116.9 million, or 13.2% of sales. The first-half 15.2% margin passes the mechanical threshold, while the quarterly and capex-cadence evidence argues against declaring it permanently achieved.

Balance sheet and returns

June cash was $353.6 million, debt principal $3.495 billion, PP&E $1.625 billion, goodwill $3.949 billion and identifiable intangibles $814.7 million. Goodwill plus intangibles equal 56.6% of assets and 114.6% of equity. The company has $1.104 billion of gross liquidity, an undrawn $750 million revolver extended to 2031, and no near-term maturity. Principal maturities are $400 million in 2028, $2.2 billion in 2029 and $895 million in 2030 before the post-quarter $25 million repayment.

Using average invested capital and normalizing the depreciation benefit, trailing GAAP ROIC is approximately 5.6%; an acquisition-amortization-adjusted measure is about 8.1%. Both remain well below pre-CMC economics. The assets may earn attractive product-level margins, but the enterprise has not earned an adequate return on the acquisition price.

Financial-quality verdict: improving, not fully repaired. Adjusted EBITDA, cash flow and debt reduction validate real progress; reported operating-margin expansion is materially accounting-aided, capex is back-half weighted, working capital is building and ROIC remains modest. Cash evidence deserves more weight than GAAP optical improvement.

Sensitivity of the repair

The repair has several reinforcing levers. Every $100 million of debt retired lowers annual interest by roughly $5–7 million at current borrowing costs, before tax. Filling underused factories spreads fixed labor and overhead. Exiting subscale operations improves mix. Capex falling from the FY2022–23 range above $450 million toward $250 million releases about $200 million of annual cash even before earnings growth. These are tangible and largely controllable.

They also have natural limits. Debt can only be repaid once, capex cannot remain below maintenance needs indefinitely, and facility absorption depends on external demand and qualification timing. Once leverage approaches the target and capex normalizes, incremental free-cash-flow growth must come from revenue, margin or working capital rather than another step-down in spending. Valuation should distinguish temporary repair slope from steady-state compounding.

Interest and tax also complicate GAAP-to-cash comparisons. Acquisition financing increased interest materially after 2022; debt reduction now allows operating improvement to reach equity. The first-half tax rate was only 8.6%, while a normalized cash rate closer to 18% is more appropriate for return analysis. A model that annualizes first-half GAAP net income without normalizing tax and depreciation will overstate durable earnings.

Goodwill risk is less about an immediate cash outflow than what impairment would reveal. MS goodwill was tested quantitatively in FY2025 and its valuation was a critical audit matter. A future impairment would not itself consume cash, but it could confirm that projected cash flows or the acquisition return were overstated. Investors should monitor forecast headroom, discount rates and segment performance rather than dismissing impairment as “noncash.”

7. Capital Allocation

CMC Materials remains the defining decision. Entegris paid $6.021 billion of gross consideration in July 2022—$3.837 billion cash to shareholders, $1.266 billion of stock and $918.6 million of debt retired—creating $3.629 billion of goodwill and issuing 12.9 million shares. The strategic rationale was coherent: CMP materials, deposition and purification broaden the recurring portfolio and enable co-optimization. The timing and financing were poor. The deal closed near the cycle peak, used heavy debt and equity, and was followed by a downturn and export-control shock.

The subsequent cleanup produced about $1.257 billion of proceeds before contingent value and taxes from QED, the MacDermid termination, Electronic Chemicals and PIM. It also produced a $115.2 million goodwill impairment in FY2023. Portfolio pruning and deleveraging have been competently executed, but selling pieces soon after purchase and still earning mid-single-digit ROIC four years later leaves the transaction economically unproven. Strategic usefulness does not repair acquisition math by itself.

Current allocation is appropriately debt-first. Entegris repaid $315 million in the first half including revolver activity, with $250 million of term principal reduction, then $25 million in July. The quarterly dividend remains $0.10 and consumed $30.8 million in the half. No board-authorized repurchase occurred. Weighted-average diluted shares rose 0.9% year over year, while SBC was $35.1 million, or 2.1% of revenue. Once leverage is below 3x, management will face a test: continue debt reduction, repurchase shares, fund organic projects or return to M&A.

Compensation improved but still avoids the core acquisition lesson. The 2025 annual incentive weighted adjusted EBITDA margin 50%, revenue growth above market 25%, FCF margin 15% and customer experience 10%; the revenue-outgrowth component paid zero and total payout was 69% of target. For 2026, options were eliminated and CEO long-term awards shifted to 60% performance shares, split between three-year FCF margin and relative TSR. There remains no ROIC or absolute-value hurdle, and adjusted EBITDA adds back acquisition amortization and restructuring.

Insider ownership is only 0.53% including exercisable awards. A complete 24-month Form 4 sweep found zero code-P open-market purchases. Excluding amendments, code-S sales totaled 432,460 shares and about $53.7 million, though most of the former CEO’s sales were plan-tagged and many other transactions were exercises or tax withholding. Post-June events reinforce the need for code integrity: Robert Bruggeworth’s RSUs were a grant, and CEO David Reeder’s August disposition was automatic tax withholding, not discretionary selling. Susan Rice’s sale and Joseph Colella’s exercise-and-sale were discretionary and untagged; Olivier Blachier’s exercise-and-sale used his May 10b5-1 plan.

Governance is incrementally cleaner. Bertrand Loy retired from the board, James Gentilcore became independent chair and Qorvo CEO Robert Bruggeworth joined. Stockholders eliminated supermajority requirements, and say-on-pay support recovered to 95.5% after 77.2% in 2025. These steps help oversight but do not replace owner-like capital allocation.

Capital-allocation verdict: poor historical record, sound current repair, unresolved future discipline. CMC remains return-dilutive; debt reduction is correct; the compensation redesign is constructive. The decisive future signal is whether management reaches sub-3x leverage and chooses disciplined repurchases or high-return organic investment before another large acquisition.

8. Changes and Headwinds — Last Two Years

The business was reorganized into MS and APS in 2024, simplifying the customer-facing structure but breaking older segment comparability. CMC-related divestitures narrowed the portfolio. David Reeder became CEO in August 2025, Sukhi Nagesh became CFO in May 2026, and Bertrand Loy completed the executive-chair transition in July. The independent-chair structure is cleaner, though leadership turnover during a balance-sheet repair remains an execution consideration.

Footprint rationalization continues. Management is exiting a U.S. Life Sciences fluid-management activity described as below $20 million of annual revenue and margin-dilutive, and it is closing Logan, Utah. Q2 restructuring expense was $5.6 million and first-half expense $9.7 million. This is the third dilutive-facility action since late 2025. The moves should improve mix, but repeated “one-time” restructuring deserves monitoring.

China exposure fell as a share but not catastrophically in dollars. Q2 China sales rose 2%; first-half sales declined less than 1%. Local-for-local production helps, while controls, tariffs, rules of origin and domestic competitors remain risks. Taiwan growth and leading-edge projects have become more important, increasing both upside from TSMC-related demand and geopolitical concentration.

The most consequential accounting change is longer useful lives, expected to lower FY2026 depreciation about $73 million. Investors should rebuild comparable margins and avoid confusing noncash accounting relief with productivity. The most consequential operating change is APS absorption: a segment that compressed from a 28.8% margin in FY2023 to 23.7% in FY2025 reached 29.3% in Q2, albeit with some depreciation aid. The most consequential capital change is faster debt reduction toward a high-2x year-end goal.

The external environment strengthened. Wafer shipments grew 7.4% in Q2 and WFE spending forecasts rose above 20%. That supports orders but also makes current growth more cyclical than a pure secular-content narrative. The November 9 Investor Day becomes a key evidence event for through-cycle margins, site utilization, capital returns and the balance between organic capacity and M&A.

Changes/headwinds verdict: net positive on operations and governance, still mixed on quality. The recovery is real; comparability, cycle exposure and capital-allocation history prevent a clean upgrade of the enterprise economics.

9. Risk Analysis

The most plausible severe loss is not insolvency. Liquidity is adequate, maturities are manageable and recurring products generate cash. The realistic permanent-loss path combines three events: unit-driven growth fails, the industry enters digestion, and a premium multiple compresses before deleveraging creates equity value. A new leveraged acquisition would worsen that path. Total loss probability is low absent an extreme geopolitical or fraud event; a 30–50% equity drawdown is demonstrably possible given the one-year 41.8% maximum drawdown and five-year 59.3% maximum drawdown.

Risk interactions matter. A Taiwan disruption would hit customer demand, supply and valuation simultaneously. China localization could preserve revenue but reduce margin. A WFE slowdown could initially hurt APS construction/handling revenue and later reduce the installed base that drives consumables. Conversely, the 75% unit mix, broad product portfolio and no near-term maturity limit the probability that one ordinary downturn permanently impairs the franchise.

Risk verdict: business continuity is resilient, equity outcomes are not. High factor beta, premium valuation, leverage and concentrated customers make the stock substantially more volatile than the underlying consumables model suggests.

Downside pathways and mitigants

The ordinary-cycle downside begins with customers slowing tool installations or wafer starts. APS construction and handling revenue would react first; consumables would follow with utilization. Lower volume would reverse fixed-cost absorption at the same time that receivables and inventory release or become excess. Adjusted EBITDA could therefore move more than revenue. The mitigants are product breadth, the recurring unit base and management’s ability to slow discretionary capex, but the recent labor additions would make an abrupt turn painful.

A technology downside is more gradual. Entegris could retain legacy positions but fail to win molybdenum, HBM, advanced packaging or next-node processes. Consolidated revenue might still grow with the market, masking falling content opportunity. This is why growth versus MSI and product-level positions matter more than absolute sales. The mitigants are early co-development, more than 50 engineering programs and existing customer access; none guarantees conversion.

The balance-sheet downside is manageable in isolation but dangerous when combined with a cycle turn. There are no 2026–27 maturities and the revolver is undrawn, so near-term liquidity is strong. The 2029 wall creates refinancing exposure if cash generation disappoints or credit spreads widen. Continued debt reduction over the next two years is the best mitigant and also reduces interest sensitivity.

Finally, governance risk is capital rather than operational. Management could interpret a repaired balance sheet as permission for another transformative deal. The CMC record shows that a strategically compatible asset can still dilute returns when price, financing and timing are wrong. A published ROIC hurdle, leverage ceiling and repurchase framework would reduce this risk more effectively than another adjusted-EBITDA synergy target.

10. Valuation Discussion — Embedded Expectations

This section quantifies expectations and scenarios; it carries no recommendation or price target.

Spot capitalization and trailing multiples

At $144.28 and 153.6 million diluted shares, market capitalization is approximately $22.16 billion. Adding about $3.206 billion of net debt produces spot enterprise value of $25.37 billion. Filing-derived trailing figures are $3.326 billion revenue, $925.6 million adjusted EBITDA, $3.21 non-GAAP EPS and $575.2 million simple free cash flow.

These are live-price multiples, not period-end aggregator values. The June-quarter closing price was $179.86, which makes unadjusted third-party enterprise values stale for today’s decision. Likewise, GAAP P/E is distorted by acquisition amortization and should not anchor valuation.

Own-history and peer context

The August 19 own-history valuation index ranks composite valuation at the 78.0th percentile, P/S at 84.8th, P/B at 62.9th and GAAP P/E at 86.3rd. P/S has fallen materially from 97.5th in June. Entegris is no longer at the most extreme point of its history, but it remains in the expensive quintile on sales.

Closest peers are imperfect. Equipment leaders AMAT, LRCX and KLAC have different service mixes, scale and returns; MKSI carries more subsystem cyclicality; Qnity has a broader materials portfolio and PFAS liabilities. A current-price reconstruction places Qnity around 18.7x its FY2026 adjusted-EBITDA guide midpoint versus roughly 25x a plausible $1.0 billion Entegris FY2026 outcome. Definitions and subsequent guidance differ, so the comparison is directional. It suggests Entegris still receives a substantial premium despite higher leverage and lower post-deal ROIC.

Reverse DCF

Starting with $575.2 million of trailing free cash flow, a 10% cost of capital and 4% terminal growth, the current enterprise value requires approximately 16.3% annual FCF growth for ten years. The calculation is sensitive to the unusually favorable current capex run rate. If normalized free cash flow is below trailing output, the required operating growth is higher. The June version required roughly 24%; price decline plus higher cash flow has reduced, not eliminated, the burden.

What could produce 16%? Revenue outgrowth in the high single digits, EBITDA margin moving above 30%, lower cash interest, stable working capital and capex below depreciation could combine for several years. What makes it difficult? The base industry does not compound at that rate indefinitely, pricing is not proven, taxes normalize, back-half capex rises, and debt reduction transfers cash to creditors before it reaches equity holders.

Scenario analysis

The following FY2028 scenarios are analytical stress tests, not forecasts or targets. They start with illustrative FY2026 sales of $3.555 billion and value the enterprise at the end of 2028.

The bear case assumes cycle normalization and no durable outgrowth. The base assumes solid growth, partial margin recovery and continued premium valuation; it offers only low-single-digit annualized return because much of the success is already capitalized. The bull requires sustained content gains, APS/MS execution, low capital intensity, deleveraging and a 25x terminal multiple. The distribution is no longer uniformly negative, but the base remains demanding.

Valuation verdict

The valuation reset materially improved asymmetry relative to June. It did not create a conventional margin of safety. The current price still embeds sustained double-digit-ish profit/cash growth, successful deleveraging and a premium exit multiple. The operating evidence is moving toward those expectations, while only one post-prior-report quarter supports them.

Multiple quality and duration

The appropriate multiple depends on which cash flow is durable. A company with recurring qualified consumables, secular content growth and a low-capex installed network can merit more than a cyclical equipment supplier. Entegris also has leverage, acquisition dilution, customer concentration and a recent history of weak organic outgrowth. Assigning a premium is defensible; assigning a premium as if all repair objectives are complete is not.

Duration magnifies small assumption differences. At a high-20s EBITDA multiple, a one-point change in long-run growth or discount rate can move value materially. The reverse DCF’s 16.3% decade-long cash-growth requirement is especially sensitive to the terminal assumption. A 4% terminal rate is already generous relative to mature nominal economic growth; using a lower terminal rate raises the required near-term performance. Conversely, successful debt reduction transfers enterprise value to equity even if the enterprise multiple stays constant.

The scenario table deliberately varies both operating results and exit multiple. Keeping a 25x multiple in the bull case recognizes franchise quality; reducing it to 18x in the bear recognizes cycle and proof risk. A valuation that changes only earnings while holding today’s premium constant would understate downside. A valuation that forces the company to an industrial commodity multiple would understate the switching-cost franchise.

The most useful near-term bridge is FY2026 adjusted EBITDA. Q3 guidance midpoint implies about $262 million. If Q4 sales rise 4% sequentially and margin holds near 29%, full-year EBITDA can approach or exceed $1.0 billion. At current EV that is approximately 25x. The market therefore prices not just the 2026 recovery but further 2027–28 growth and deleveraging. The November framework and next two quarters will determine whether that premium reflects visible earnings or anticipation.

Equity versus enterprise-value path

At current leverage, equity value can grow faster than enterprise value if cash is directed to debt. Reducing net debt from roughly $3.1 billion to $1.8 billion by 2028 would add about $1.3 billion, or roughly $8 per diluted share, holding enterprise value constant. Most of the scenario upside, however, still requires EBITDA growth or multiple support. Deleveraging alone cannot bridge a very high purchase multiple.

Share count matters as well. Diluted shares increased 0.9% year over year and no repurchase offsets stock plans. Even modest annual dilution reduces per-share benefits from debt reduction. A disciplined repurchase program after leverage falls below 3x could improve per-share conversion, but repurchasing at an excessive multiple would repeat the capital-allocation problem in smaller form. The right sequencing is balance-sheet safety, high-return organic capacity, and price-sensitive capital return.

11. Variant Perception

The consensus bull case likely emphasizes three correct facts: Entegris sells recurring consumables, AI/HBM raises materials intensity, and completed capacity plus lower capex should drive exceptional free cash flow. Q2 strengthened all three. The variant question is not whether these forces exist; it is how much of consolidated growth is true content/share, how clean the margin improvement is, and what price already discounts.

The skeptical variant is that APS is currently riding a construction/WFE and absorption wave while MS—the cleaner materials-content proxy—lags the market. Reported gross and operating margins are helped by a depreciation-life change. TTM free cash flow benefits from a capex trough after a major build. A premium multiple may therefore be capitalizing cyclical, accounting and self-help benefits as if they were permanent organic economics.

The counter-variant to that skepticism is credible. Construction revenue today can seed WFE revenue tomorrow and recurring filters/materials later. The company may have intentionally built capacity before demand, so current absorption is the beginning rather than the end of earnings leverage. Debt reduction lowers interest and equity risk. If MS outgrowth returns in Q3/Q4, the market may be right to look beyond one segment’s comparison issue.

Positioning adds nuance. ENTG is not statistically a simple positive-Momentum factor exposure despite an 80% one-year return. The all-factor model loads strongly on Market, Semiconductors and Semiconductor Equipment, positively on Quality, negatively on Low Volatility, and zeroes Momentum. The recent equipment factor return is sharply negative over 21 days but positive over 126 days. That explains why cohort moves dominate the tape and why Q2’s company-specific outperformance matters more than the June spike.

The most decision-useful variant is therefore timing and composition: the market prices a broad secular compounder, while current evidence shows a levered high-quality portfolio whose earnings recovery is concentrated in APS and whose durable MS outgrowth is not yet established. The next two quarters can resolve that gap.

12. Fact vs. Interpretation

13. Open Questions

1. What is organic unit-driven growth versus one consistent MSI definition for each of the last eight quarters, split by MS and APS?
2. What are current market shares and process-of-record positions in CMP slurry, pads, cleans, liquid filtration, FOUPs and deposition precursors?
3. Why did MS grow 5% when management’s MSI estimate was 7–8%, and what evidence supports above-market Q3/Q4 growth beyond comparison timing?
4. How much of FY2026 gross and segment-margin improvement remains after the $73 million expected depreciation reduction?
5. What are utilization, qualification timing and incremental margins at KSP, Colorado and other recently built sites?
6. How does management bridge “more than $1 billion” of latent capacity by site, product and expected year of conversion?
7. What is normalized annual capex after the current $250 million plan, and how much maintenance capex is economically required despite longer accounting lives?
8. How much of China revenue is local-for-local by product, and what margin/cost penalty accompanies localization?
9. What customer or third-party evidence supports content per wafer at TSMC, HBM and advanced packaging?
10. What ROIC hurdle will govern M&A after leverage falls below 3x, and where do repurchases rank versus acquisitions?
11. Will the November Investor Day establish through-cycle segment margins and a quantified capital-return framework?
12. What portion of receivable and inventory growth is demand preparation versus slower conversion or deliberate supply redundancy?

14. What Must Be True

Bull thesis requirements

Bear thesis requirements

The skeptical case requires that recent strength is mostly cycle, absorption and accounting; that MS content/share remains weak; and that the premium multiple compresses before free cash flow retires enough debt. Evidence would be unit growth at or below MSI, normalized EBITDA margin below 29%, capex/working-capital pressure, APS order deceleration or renewed M&A.

The bear thesis is falsified only by the full bundle: two to three quarters of clear MSI outperformance, adjusted EBITDA margin above 30%, free-cash-flow margin in the mid-teens after full capex, leverage below 3x and a buyback restart without new leverage. Through August 20, only the free-cash-flow criterion is mechanically near/pass. The bundle is therefore not falsified.

Monitoring cadence

Q3 and Q4 should be scored on unit/MS growth versus the same MSI baseline, normalized margins, working-capital conversion, capex, net leverage and capital allocation. The November 9 Investor Day is the opportunity for management to quantify through-cycle segment targets, site utilization, cash conversion and post-deleveraging priorities. Forecast language should remain a hypothesis until filed outcomes confirm it.

15. Source Appendix

The full source register, filing corpus notes, transcript integrity checks, market-data methodology and primary links appear in Appendix B. All financial statements and insider conclusions were reconciled to SEC filings; third-party aggregators were used only where specified and were rejected when their definitions did not reconcile.

APPENDIX A — Standard Diligence Questionnaire — Entegris, Inc. (NASDAQ: ENTG)

General

What thoughtful questions have other investors asked?

The Q2 call centered on the questions that matter most: why MS grew only 5% when MSI was 7–8%; whether competitor CMP share claims are credible; how much Q3/Q4 growth is market versus content; when KSP becomes profitable; whether more than $1 billion of latent capacity can be filled without another investment cycle; and what happens to capital after leverage falls below 3x. Investors also pressed management on the three-wave construction/WFE/wafer-start opportunity, molybdenum, mainstream foundry utilization, China and segment margins.

Interpretation: these questions correctly reject consolidated growth as sufficient proof. The underwriting hinge is the composition of growth, normalized margins and the eventual return on the post-CMC capital base.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low?

Fact: Q2 wafer shipments grew 7.4%; SEMI forecasts 2026 WFE spending up 23.1%. Entegris’s Q2 revenue rose 11.5%, APS 17% and MS 5%. Adjusted EBITDA margin was 28.4%, below the prior proof point of 30% and below what recent capacity could plausibly earn at fuller utilization.

Interpretation: earnings are in an early-to-mid recovery, not a trough and not obviously at peak. The capex-linked quarter is strong enough to introduce late-cycle risk, while unit-driven consumables may have a longer runway as new capacity loads.

Are results driven by the external environment or company actions?

Both. External MSI, WFE and fab construction improved. Internal actions—facility absorption, productivity, footprint exits, lower capex and debt reduction—amplified cash and earnings. Roughly 150 basis points of Q2 gross margin came from longer useful lives, so accounting also contributed.

How stable are revenues?

Approximately 75% is tied to wafer starts and recurring consumption, 10% to WFE and 15% to fab construction. Qualification captivity stabilizes share, but unit revenue still moves with utilization and inventories. Historical sales fell from $3.524 billion in FY2023 to $3.197 billion in FY2025, demonstrating that “recurring” is not “noncyclical.” TSMC is 16% of sales and the top ten customers 50%, increasing event concentration.

What is the product and market outlook?

The strongest product outlook is liquid filtration, advanced purity, HBM/TSV-related CMP, advanced deposition and selected advanced-packaging steps. Management guides Q3 revenue to $905–$935 million and expects Q4 roughly 4% above Q3 sequentially; these are hypotheses until filed. The broad materials market was $73.2 billion in 2025, up 6.8%. Public data support secular growth but not a precise Entegris SAM.

Domestic or international?

International revenue is 82%. Taiwan represented 26% of Q2 and China 19%. The opportunity is inherently international; so are trade, currency, customer and geopolitical risks.

Business Quality & Competitive Moat

Is the industry becoming more or less competitive?

Interpretation: leading-edge qualified niches remain oligopolistic, while trailing-node and China-local categories are becoming more competitive. Qnity, Merck, Pall/Danaher, Air Liquide, Linde, Shin-Etsu, Gudeng, Aicello, Cobetter and Anji are meaningful competitors. Chinese capability is broadening, but public evidence does not establish leading-edge equivalence.

How profitable is the business?

Q2 adjusted EBITDA margin was 28.4% and normalized gross margin roughly 46.1%. Filing-derived normalized trailing GAAP ROIC is approximately 5.6%; acquisition-amortization-adjusted ROIC is about 8.1%. Product-level economics appear attractive, but enterprise returns remain well below estimated pre-CMC 14–18% because the acquisition price and financing expanded invested capital.

How profitable is the industry, and what are the barriers?

The value chain rewards yield-critical equipment and materials niches more than generic subsystems. Barriers include long qualification, copy-exact process control, customer yield risk, applications know-how, clean manufacturing, IP, and local/redundant supply. The moat is Greenwald demand-side captivity reinforced by niche scale and intangibles. There is no network effect.

Is the business understandable?

Yes at the economic level: Entegris sells consumable and handling products that protect semiconductor yield. It is harder at the product level because 30,000-plus SKUs, two mixed segments and customer-specific qualifications obscure share and contribution margins. An investor must model MS and APS separately and distinguish unit from capital-driven growth.

Can foreign low-cost labor undermine it?

Not by labor cost alone. Purity, repeatability, qualification and process support matter more than labor. Foreign competitors supported by local policy can still compete through patient qualification, bundled supply and local sourcing. Localization can erode share or price over time, particularly in China.

Do brands matter?

Brand is secondary. Reputation signals validated quality and supply assurance; the economic mechanism is switching cost and yield risk, not consumer preference.

What is the nature of competition and switching cost?

Competition occurs product by product on performance, purity, qualification time, total cost, service and supply resilience. Switching can require reformulation, testing and requalification and may endanger yield. Because inputs are inexpensive relative to wafer value, customers accept supplier continuity unless performance, availability or strategic dual sourcing justifies a change.

What evidence could disprove the moat?

Two or more quarters of MS or unit growth below the same MSI benchmark; independent evidence of lost process-of-record positions; normalized gross margin returning to 44–45% despite volume; or category-share loss to local competitors would weaken the moat assessment. Three consecutive quarters of 3–6-point unit outgrowth plus stable normalized margins would strengthen it.

Financial Condition & Balance Sheet

Are assets understated?

Company-developed recipes, qualification history, customer relationships and applications knowledge are not fully capitalized. They are economically valuable. That does not justify the entire goodwill balance, which reflects acquisition price rather than replacement cost.

Are there material off-balance-sheet liabilities?

The reviewed filings do not reveal a hidden liability comparable to funded debt. Ordinary leases, purchase commitments, environmental/product obligations and tax exposures exist. The larger practical commitments are customer supply reliability and maintenance of a global qualified footprint. The source corpus should be refreshed after each 10-Q/10-K for changes.

How conservative is accounting?

Mixed. SEC cash flow and debt disclosures reconcile well. The January 2026 useful-life extension materially reduced depreciation: $37.7 million in the first half and about $73 million expected for the year. This is disclosed but makes operating and segment margins less comparable. Acquisition amortization and recurring restructuring make adjusted results useful, while also risking normalization of genuine acquisition costs. Goodwill plus intangibles is 56.6% of assets.

How capex-hungry is the business?

Historically variable. Capex was $466 million in FY2022 and $457 million in FY2023, then $316 million in FY2024 and $299 million in FY2025. First-half 2026 capex was only $80.8 million against roughly $250 million planned. Management says recent investment created more than $1 billion of capacity, but maintenance capex and site utilization are not separately disclosed.

What is liquidity and refinancing risk?

Cash was $353.6 million and the $750 million revolver was undrawn, giving about $1.104 billion gross liquidity. No principal is due in 2026–27; $400 million is due in 2028, roughly $2.175 billion in 2029 after the July payment, and $895 million in 2030. Filing-derived net leverage is about 3.37x pro forma July. Refinancing risk is manageable if cash generation persists, material if a downturn and new M&A interrupt deleveraging.

Capital Allocation & Management

How much free cash flow is generated and how is it used?

Simple FCF was $396.2 million in FY2025, $258.4 million in first-half 2026 and approximately $575.2 million trailing. Current priority is debt reduction: $250 million of term principal was repaid in the first half and $25 million after quarter-end. The dividend is $0.10 quarterly. No Q2 repurchase occurred.

What is management’s philosophy?

Management prioritizes organic investment and deleveraging, then leaves shareholder returns and external M&A available after the balance sheet reaches its target. The open question is sequencing and price discipline. A formal ROIC hurdle is not disclosed.

Significant acquisition record?

CMC Materials closed in July 2022 for $6.021 billion of gross consideration, using debt and 12.9 million shares, and created $3.629 billion of goodwill. About $1.257 billion of assets/rights were subsequently divested before contingent value and tax, and FY2023 recorded a $115.2 million goodwill impairment. The strategic fit is credible; current 5.6–8.1% ROIC, leverage above 3x and lower FY2025 revenue than FY2023 leave value creation unproven.

Is the company buying back shares or issuing stock?

No active Q2 repurchase occurred. Diluted weighted shares rose 0.9% year over year. First-half SBC was $35.1 million, 2.1% of sales; stock-plan issuance is not a follow-on financing but creates dilution without an offset.

How is management compensated?

The 2025 annual plan weighted adjusted EBITDA margin 50%, revenue growth above market 25%, FCF margin 15% and customer experience 10%; total payout was 69% and the outgrowth metric paid zero. In 2026 options were eliminated and performance shares gained weight, split between FCF margin and relative TSR. Alignment improved, but there is no ROIC or absolute-value hurdle.

What do insiders’ actions show?

Directors and executives own only 0.53% including acquirable shares. The complete 24-month original Form 4 set had zero code-P open-market purchases and 432,460 code-S sale shares, many plan-tagged or associated with exercises. Grants, tax withholding and option exercises were separated from discretionary sales. The evidence shows weak open-market sponsorship, not a claim that every disposition is bearish.

Valuation & Market Data

Security structure and dividend

ENTG is a NASDAQ-listed U.S. common stock, not an ADR, MLP or K-1 issuer. The annualized dividend is $0.40 per share, about a 0.28% yield at $144.28.

How profitable is the business and does cash diverge from income?

Trailing adjusted EBITDA is approximately $925.6 million. First-half FCF of $258.4 million exceeded net income of $185.6 million because depreciation, working-capital components and low capex affected conversion. The divergence is currently favorable, but capex is back-half weighted and the useful-life change raises accounting income without cash.

What does the price embed?

Spot EV is approximately $25.37 billion, or 7.6x sales and 27.4x trailing adjusted EBITDA. Equity trades at 38.5x trailing FCF and 45x non-GAAP EPS. Own-history P/S is at the 84.8th percentile. A reverse DCF using 10% discount rate and 4% terminal growth requires about 16.3% annual FCF growth for ten years.

What are the valuation sensitivities?

Growth composition, normalized EBITDA margin, terminal multiple, steady-state capex and debt reduction matter most. An illustrative FY2028 bear/base/bull set produces about $102/$159/$206 per share under 2%/7%/10% sales growth, 27%/29.5%/31% margins and 18x/22x/25x exit multiples. These are analytical scenarios, not price targets.

Risks & Downside

What could cause a material decline?

MS or unit growth at/below MSI; a WFE/wafer-start digestion; APS under-absorption; China controls or localization pressure; loss of a major customer qualification; Taiwan disruption; normalized margin weakness; working-capital or capex reacceleration; a new leveraged acquisition; or premium-multiple compression. Factor beta and negative Low Volatility exposure can amplify sector moves.

Catastrophic and total-loss risk

Total loss probability is low because recurring products generate cash, liquidity is above $1 billion and there are no near-term maturities. Catastrophic equity impairment could arise from a Taiwan geopolitical event, severe customer/process losses plus recession, fraud, or leveraged M&A ahead of a downturn. A large but nonterminal drawdown is more plausible: the stock’s one-year maximum drawdown was 41.8% and five-year maximum 59.3%.

Recent News & Events

Has the business environment changed?

Yes. Q2 wafer shipments and WFE forecasts improved; Q2 operations and guidance beat expectations; the stock’s August 4 reaction materially exceeded peers. China controls remain restrictive, while local demand is product-specific rather than collapsing.

Significant acquisition?

No acquisition was identified in the IR/SEC record from May 20 through August 20. The company instead reduced debt and wound down a small life-sciences fluid-management operation. Management leaves future external M&A open.

Accounting-policy changes?

Effective January 2026, Entegris lengthened several PP&E useful lives, lowering first-half depreciation $37.7 million and expected FY2026 depreciation about $73 million. This is the main comparability issue.

New facilities, markets and management?

KSP filtration qualifications continue and recently built capacity is being absorbed. Management highlighted more than 20 leading-edge expansions, HBM, advanced packaging and molybdenum. David Reeder is CEO, Sukhi Nagesh CFO; Bertrand Loy retired from the board, James Gentilcore became independent chair and Robert Bruggeworth joined. The company is exiting U.S. Life Sciences fluid management and closing Logan, Utah.

APPENDIX B — Source Appendix — Entegris, Inc. (NASDAQ: ENTG)

Evidence standard and corpus integrity

Research was refreshed through 20 August 2026. Financial claims use SEC filings as the source of record. Ownership filings were reviewed with amendments excluded from transaction totals. The filed 8-K and company event page establish that the Q2 release and call occurred on 4 August 2026.

Primary — SEC filings and company materials

• Entegris FY2025 Form 10-K, filed 11 February 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000012/entg-20251231.htm
• Entegris FY2023 Form 10-K, filed 15 February 2024: https://www.sec.gov/Archives/edgar/data/1101302/000110130224000009/entg-20231231.htm
• Entegris FY2022 Form 10-K, filed 23 February 2023, including CMC purchase accounting: https://www.sec.gov/Archives/edgar/data/1101302/000110130223000021/entg-20221231.htm
• Entegris Q1 2026 Form 10-Q, filed 30 April 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000102/entg-20260328.htm
• Entegris Q1 2026 earnings release, published 30 April 2026: https://s205.q4cdn.com/144974603/files/doc_financials/2026/q1/ENTG_Q126_Earnings_Release.pdf
• Entegris Q2 2026 Form 10-Q, filed 4 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000150/entg-20260627.htm
• Entegris Q2 2026 Form 8-K, filed 4 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000146/entg-20260804.htm
• Entegris Q2 earnings release, Exhibit 99.1, published 4 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000146/entgq22026ex991.htm
• Entegris Q2 webcast page, event held 4 August 2026: https://investor.entegris.com/events-and-presentations/events/event-details/2026/Entegris-Earnings-Webcast-2026-Second-Quarter-Results/default.aspx
• Entegris 2026 proxy, filed 23 March 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000054/entg-20260323.htm
• Entegris annual-meeting Form 8-K, filed 8 May 2026: https://www.sec.gov/Archives/edgar/data/1101302/000114036126020041/ef20072733_8k.htm
• Entegris leadership/governance Form 8-K, filed 29 July 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000141/entg-20260728.htm
• Entegris issuer and ownership filing history: https://www.sec.gov/edgar/browse/?CIK=1101302&owner=include
• SEC company submissions file: https://data.sec.gov/submissions/CIK0001101302.json
• Entegris IR news feed, searched 20 August 2026: https://investor.entegris.com/news/default.aspx
• Entegris/JSR/Inpria EUV cross-license release, 26 May 2026: https://investor.entegris.com/news/news-details/2026/Entegris-and-JSR-CorporationInpria-Corporation-Announce-Non-Exclusive-Cross-Licensing-to-EUV-Lithography/default.aspx
• Entegris dividend release, 15 July 2026: https://investor.entegris.com/news/news-details/2026/Entegris-Declares-Quarterly-Cash-Dividend-02cc5095a/default.aspx
• Entegris CMP patent release, 18 August 2026: https://www.businesswire.com/news/home/20260818175989/en/Entegris-Achieves-Further-Success-in-Defending-Its-CMP-Slurry-Patent-Portfolio/

Primary — insider transaction examples

• Robert Bruggeworth Form 4, RSU grant, filed 3 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000114084226000004/wk-form4_1785861254.xml
• Susan Rice Form 4, sale, filed 11 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000110130226000152/wk-form4_1786555778.xml
• Joseph Colella Form 4, option exercise and sale, filed 13 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000181106126000002/wk-form4_1786730610.xml
• Olivier Blachier Form 4, plan exercise and sale, filed 14 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000189166126000002/wk-form4_1786982437.xml
• David Reeder Form 4, automatic tax withholding, filed 18 August 2026: https://www.sec.gov/Archives/edgar/data/1101302/000159713326000002/wk-form4_1787160280.xml

Management commentary

Management-provided MSI, WFE order rates, fab-project counts, utilization, content claims, molybdenum outlook, KSP timing, latent-capacity estimate and future Q4 growth are identified as management hypotheses in the report. Filed financial outcomes are cited separately.

Industry, competitors and regulation

• SEMI, 2025 materials-market revenue, published 12 May 2026: https://www.semi.org/en/semi-press-release/global-semiconductor-materials-market-revenue-reaches-record-73.2-billion-dollars-in-2025-semi-reports
• SEMI, Q2 2026 silicon wafer shipments, published 29 July 2026: https://www.semi.org/en/semi-press-release/semi-reports-worldwide-silicon-wafer-shipments-increase-7.4-percent-year-on-year-in-q2-2026
• SEMI equipment forecast, published 14 July 2026: https://www.semi.org/en/taxonomy/term/45726
• SEMI World Fab Forecast, Q2 2026 edition: https://www.semi.org/en/products-services/market-data/world-fab-forecast
• SEMI materials-market outlook presentation, September 2025: https://www.semi.org/sites/semi.org/files/2025-09/5 Clark Tseng_Building the Future-AI Investment%2C Equipment %26 Materials Market Outlook.pdf
• Qnity Q2 2026 results, published 4 August 2026: https://ir.qnityelectronics.com/_assets/_0439e220595f055b47a6bf1f5d556814/qnityelectronics/news/2026-08-04_Qnity_Reports_Second_Quarter_2026_Results_Raises_69.pdf
• Merck KGaA Q2 2026 results, published 6 August 2026: https://www.emdgroup.com/en/news/q2-2026-06-08-2026.html?global_redirect=1
• Cobetter SEMICON China 2026 exhibitor profile: https://expo.semi.org/china2026/public/eBooth.aspx?BoothID=639351&IndexInList=893&ListByBooth=true&Nav=False
• U.S. BIS EAR Part 744 / Section 744.23: https://www.bis.gov/regulations/ear/744
• U.S. BIS, closure of VEU treatment for foreign-owned semiconductor fabs in China, 29 August 2025: https://www.bis.gov/press-release/department-commerce-closes-export-controls-loophole-foreign-owned-semiconductor-fabs-china
• U.S. BIS, December 2024 China semiconductor controls: https://www.bis.gov/press-release/commerce-strengthens-export-controls-restrict-chinas-capability-produce-advanced-semiconductors-military
• U.S. BIS, advanced-computing-chip license-review policy, 13 January 2026: https://www.bis.gov/sites/default/files/documents/DoC Revises License Review Policy for Semiconductors Exports.pdf

Market price and quantitative evidence

• Nasdaq ENTG historical data, including 19 August 2026 close: https://www.nasdaq.com/market-activity/stocks/entg/historical
• MarketBeat FINRA-derived short-interest history, accessed 20 August 2026: https://www.marketbeat.com/stocks/NASDAQ/ENTG/short-interest/

Price windows use identical adjusted-close dates for all peers. Third-party beta and factor outputs are descriptive, not forecasts.

Analytical frameworks

• Bruce Greenwald and Judd Kahn, Competition Demystified: demand captivity, scale, market-share stability, ROIC and earnings-power-value tests.
• Marathon Asset Management / Edward Chancellor, Capital Returns: supply-side capital-cycle analysis, investment booms, capacity response and return mean reversion.

Calculation notes

• Simple free cash flow = GAAP operating cash flow less cash purchases of PP&E. Company-defined FCF also includes government capital incentives where noted.
• Spot market capitalization uses $144.28 and 153.6 million diluted shares; enterprise value adds approximately $3.206 billion net debt.
• Trailing adjusted EBITDA = FY2025 $886.2 million + first-half 2026 $476.8 million − first-half 2025 $437.4 million = $925.6 million.
• Trailing free cash flow = FY2025 $396.2 million + first-half 2026 $258.4 million − first-half 2025 $79.4 million = $575.2 million.
• Reverse DCF uses $575.2 million starting FCF, 10% cost of capital and 4% terminal growth. Scenario outputs are assumption-driven tests, not price targets.
• Normalized margin and ROIC calculations remove the disclosed $37.7 million first-half useful-life benefit and use an 18% normalized cash tax rate where specified.