ASM International

Phase A — Understand the business

Lens 1 · Company Overview

ASM International makes the deposition tools that lay down the atom-thin films inside a leading-edge transistor. It is not a diversified equipment house like Applied Materials or a lithography monopolist like ASML — it is a focused deposition specialist, and within deposition it owns the two categories that matter most for the industry's move to 3D transistors: Atomic Layer Deposition (ALD) and single-wafer epitaxy (Epi).

• What it sells. 300mm ALD platforms (thermal ALD and plasma-enhanced ALD — the flagship is the Intrepid / XP generation), single-wafer epitaxy tools (Intrepid and Epsilon), plus PECVD, vertical furnaces, and — the quiet compounder — a Spares & Services annuity attached to the ~15,000+ tool installed base.
• How money splits. FY2025 revenue €3,173.2m: Equipment €2,457.5m (77%) + Spares & Services €715.7m (23%). Equipment is lumpy and tied to fab capex; Spares & Services is high-margin, recurring, and grew +14% (18% cc) in 2025 — faster than equipment — as ASM expands "outcome-based" service contracts.
• Who buys. The leading-edge logic/foundry names — TSMC, Samsung, Intel — plus memory makers (DRAM/HBM). FY2025 was "driven primarily by advanced logic/foundry as customers invested in 2nm GAA capacity". Memory fell to 16% of equipment revenue (from 25% in 2024) as China memory normalized.
• Contract structure. Capital-equipment sales (no take-or-pay; order-book driven — backlog €1,247m at end-2025, book-to-bill 0.9 for the year) plus a recurring services layer. As of Q1 2026 ASM stopped disclosing quarterly bookings and backlog — a transparency step-down worth flagging (see Lens 10).
• The odd asset. ASM holds a ~25% stake in ASMPT (Hong Kong-listed back-end/packaging equipment maker, 0522.HK), a legacy of the business it spun down from majority control in 2013. It is carried as an associate (~€25m/yr equity income) and is a recurring source of non-cash P&L noise and activist agitation (Lens 9 / Lens 13).

Plain-terms: ASM is the company you buy if you believe the transistor is going 3D (GAA nanosheets, then complementary-FET), because every 3D transistor needs more ALD and more epi than the planar/finFET one it replaces. It is a bet on process complexity, not on any single chip winning.

Lens 2 · Supply Chain

ASM sits in the equipment tier of the semiconductor stack — upstream of the foundries, downstream of precision-component and materials suppliers. Named map:

Upstream (ASM's suppliers):

• Precision mechanical / vacuum / gas-delivery components — the same specialist base that feeds all WFE (VAT Group for vacuum valves, MKS Instruments and Advanced Energy for RF/power & gas subsystems, Bruker/Edwards for pumps). ASM is an integrator of these subsystems into a tool.
• Precursor chemicals — the metal-organic and specialty gases consumed in ALD (e.g. molybdenum, hafnium, ruthenium precursors) supplied by Entegris, Merck/EMD, Air Liquide, Adeka. ASM co-develops precursor+process recipes; this is part of the moat (Lens 3).
• Contract manufacturing footprint — ASM manufactures in the Netherlands (Almere), Singapore, South Korea (new Korea manufacturing + innovation center completed 2025), and the US (Phoenix, Arizona — a "several hundred million euro" expansion announced).

The company → end customer:

Component & precursor suppliers → ASM (ALD/Epi tools + services) → Foundry/IDM fabs (TSMC, Samsung, Intel) & memory makers (SK Hynix, Samsung, Micron) → fabless designers (NVIDIA, AMD, Apple, Broadcom) → OEM/ODM → hyperscaler / device buyer

Chokepoints & single-source exposure:

• ASM is itself a near-single-source for high-volume single-wafer ALD at the leading edge — that is the bull case, not a vulnerability to ASM.
• Customer concentration is the real chain risk: the leading-edge buyer list is 3–4 names (TSMC/Samsung/Intel logic + a handful of memory makers). A capex pause at TSMC hits ASM directly.
• Geographic/geopolitical chokepoint: China was >30% of 2025 revenue and is governed by US/Dutch export-control policy ASM does not control (Lens 8/10).

This lens is names-complete for the tiers ASM actually touches; the component-supplier list is `` rather than ASM-disclosed, so treat those names as representative of the category, not confirmed ASM POs.

Lens 3 · Competitive Advantages (moats)

ASM's moat is process leadership fused to installed base — narrow but deep.

1. ALD share leadership. ASM holds the #1 position in single-wafer ALD — >55% share in the single-wafer ALD segments it competes in, and ~35% of the broader ALD equipment market by one framing. ALD is >50% of ASM's equipment revenue. Tokyo Electron is the #2 challenger at ~16% ALD share.
2. Recipe/precursor co-development (switching cost). An ALD process is a tool + precursor + recipe + qualification system. Once a fab qualifies ASM's tool for a specific film at a specific node, re-qualifying a competitor's tool costs months and yield risk. New penetrations — Mo ALD and Area-Selective Deposition (ASD) entering HVM at the 2nm node — are exactly this kind of sticky, node-specific win.
3. The 3D-transistor tailwind is a structural moat-widener. GAA nanosheets need conformal deposition of more, thinner layers than finFET; ASD lets fabs skip lithography steps. ASM management: "we are well positioned to gain further share in GAA at the 1.4nm node". Each node transition increases deposition intensity — ASM's served market grows faster than wafer volume.
4. Installed-base annuity. ~23% of revenue is Spares & Services, growing double-digits, with structurally higher margin — a moat against cyclicality and a compounding cash annuity.

Bargaining power. Vs. customers: moderate-to-strong at the leading edge (few alternatives for qualified single-wafer ALD), weaker in mature/trailing nodes where more vendors qualify. Vs. suppliers: strong (ASM integrates commodity-ish subsystems; precursor houses need ASM's tool as the delivery vehicle). The durable moats are switching costs + process IP; the eroding edge is that Lam and AMAT are attacking the newest ALD applications directly (Lens 13).

Lens 4 · Segments

ASM reports by product family (Equipment vs Spares & Services) and by end-market (logic/foundry, memory, power/analog) rather than a full segment P&L. No segments.csv on disk — all ``.

By product (FY, EUR m):

Read: Equipment growth is decelerating vs the prior up-cycle but positive, driven entirely by logic/foundry (2nm GAA capacity). Spares & Services is the standout — growing 2x equipment, mix-accretive, and now ~23% of the top line. This is the segment that de-risks the cyclicality bear case.

By end-market (FY2025 equipment mix):

• Logic/foundry — the dominant and accelerating driver (2nm GAA build-out at TSMC/Samsung/Intel).
• Memory — 16% of equipment revenue, DOWN from 25% in 2024. HBM-related DRAM demand was solid but offset by China memory normalization. This is a decelerating pocket with optionality: an HBM4 / DRAM capex leg would re-accelerate it.
• Power/analog — smaller, mentioned as the #2 equipment contributor in Q4.

By geography: China >30% of total 2025 revenue — the largest single geography and the biggest swing factor (Lens 8). Korea, Taiwan, US/Europe make up the balance (ASM does not publish a clean geo-revenue table in the release).

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 2026, reported 21 Apr 2026)

The most recent print was strong and beat, and it is the cleanest read on the current trajectory.

• Revenue €862.5m, +16% YoY at constant currency (+3% reported, FX headwind), at the high end of the €830m ±4% guide.
• Gross margin 53.3% — above the FY2025 record 51.8%, on favorable product/customer mix.
• Adjusted operating margin 33.1% — a quarterly record (reported operating margin 32.2%).
• Net earnings €238.5m; adjusted net earnings €246.0m.
• Guidance: Q2 2026 revenue €980m ±5% (a sharp sequential step-up from €862.5m), and H2 2026 > H1 2026 — an accelerating shape, not a plateau.
• Drivers: advanced logic/foundry strong (leading-edge 2nm), a sequential rebound in mature logic/foundry in China, and memory "solid" led by the most advanced DRAM for HBM.

Vs. the company's own recent history: revenue and margins are at/near records, and the Q2 guide implies a re-acceleration. The one thing to watch is that the YoY reported vs constant-currency gap is wide (+3% vs +16%) — a stronger euro is masking the underlying volume growth, and will keep pressuring reported figures if EUR/USD stays elevated.

Balance-sheet flags (from FY2025): clean. Cash €1,026.9m, total assets €5,337.0m, equity €4,005.8m. Accounts receivable fell to €562.1m from €789.0m — a positive working-capital swing (not a red flag). Inventories broadly flat at €552.1m. No net debt; the company is net cash and returns capital (dividend + buyback).

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on disk; sentiment read from the press-release narrative and management commentary across Q3'25 → Q4/FY'25 → Q1'26.

Tone arc:

• Q3 2025 (Oct): cautious-constructive. Orders −17% YoY (€637m), dragged by China export restrictions; margins strong (op margin ~31%). Management flagged China normalization and a projected double-digit YoY decrease in 2026 China revenue.
• Investor Day (Sep 2025): confident long-range — new 2030 target of >€5.7bn revenue (≥12% CAGR, outpacing WFE) — but simultaneously revised the H2 2025 outlook down. A "long-term bullish, near-term realistic" split.
• Q4/FY2025 (Mar 2026): more upbeat. "Ninth consecutive year of double-digit growth," record full-year margins, and — critically — an explicit reversal on China: "we now expect our sales in China to increase in 2026", a material improvement from the Q3 double-digit-decline guide.
• Q1 2026 (Apr): the most bullish of the sequence — record margins, "AI-led demand accelerated further," and an accelerating H2 shape.

Recurring phrases: "GAA," "2nm," "leading-edge logic/foundry," "outcome-based services," "positive mix." Stopped saying: the heavy China-caution framing of Q3 2025 softened materially by Q4. The tone has inflected positive over the last two prints — the swing factor being China moving from headwind to tailwind faster than the company itself guided one quarter earlier. (Surface this as a genuine guidance conflict: Q3'25 said China down double-digits in 2026; Q4'25 said China up in 2026. Both are ASM's own words, one quarter apart — the reversal is the signal.)

Lens 7 · Comps

ASM's key global peers are the other front-end WFE names. Multiples are `` with source/date; where a specific ASM figure isn't cleanly sourced I mark it rather than fabricate.

• Dividend yield / 5-yr avg ROE columns: n/a — not cleanly sourced for the full set. ASM's own ROE is very high (FY2025 net earnings €723.7m on ~€4.0bn equity ≈ ~18% reported ROE, flattered/distorted by the ASMPT impairment reversal; underlying operating ROE is higher).
• Read: ASM trades at the high end of the WFE peer band on P/E (~45x fwd), roughly in line to slightly below ASML, and above AMAT/LRCX/KLAC. The premium is earned on the highest margins-plus-growth combo and the cleanest secular story (deposition intensity), but there is no valuation cushion — ASM is priced as a quality-growth compounder, not a value name. Against consensus 2026 EPS of €21.85 the ~€959 price is ~44x 2026 EPS.

Lens 8 · Stock-Price Catalysts (moves >5% over ~5 years)

Pattern, mostly ``:

• AI/GAA capex-cycle repricings (2023–2026): the dominant up-driver. The stock ran from a 52-week low of €400 to a high of €1,092.50 — a ~2.7x range — as the market repriced the 2nm/GAA deposition-intensity thesis. Current ~€959 sits near the top of that range.
• China export-control announcements (Dec 2024; recurring 2025–2026): the dominant down-driver. New US/Dutch restrictions on tool exports to China repeatedly hit ASM and the whole WFE group; ASM's Dec 2024 statement that new controls were "largely aligned with previous assumptions" contained the damage, but China headlines remain the single biggest source of >5% down-days (see also the April 2026 Congress-driven WFE selloff ).
• Quarterly bookings/orders surprises: ASM's Q4 2025 bookings pre-announcement (€803m, +26% QoQ) moved the stock as the China order rebound landed. Note: ASM has now discontinued quarterly bookings disclosure — removing a historical catalyst and a leading indicator.
• Guidance revisions: the Sep 2025 Investor Day (2030 target up, H2 2025 down) is the template — long-range bullishness and near-term caution can move the stock in opposite directions on the same day.

What the market actually reacts to for ASM: (1) the China policy tape (fear), (2) leading-edge capex signals from TSMC/Samsung/Intel (greed), and (3) its own margin prints (quality confirmation). It is a high-beta expression of the "AI infrastructure → leading-edge capex" trade with a China-policy put/overhang.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO — Hichem M'Saad (Chair of the Management Board & CEO since 13 May 2024; reappointed to 2030 on 11 May 2026). A relatively new CEO — the record 2025 margins and the 2030 growth framing are being set on his watch, but his tenure is short enough that the capital-allocation track record is still being written. Background is deep semiconductor/technology operations. Archetype: professional manager / technologist, not founder-owner.
• CFO — Paul Verhagen (CFO since 1 Jun 2021; reappointed to 2027). The steadier hand — oversaw the margin-expansion and the shift to the Investor-Day financial framework (gross-margin target range raised to 47–51%, operating-margin 28–32%, FCF >€1bn by 2030).
• Track record / capital allocation: ASM's numbers speak well — record 51.8% gross margin and 30.2% adjusted operating margin in FY2025, disciplined SG&A (down to 9.2% of revenue), R&D held at ~12.5% of revenue. Capital return is shareholder-friendly and consistent: dividend raised to €3.25 (from €3.00) and a fresh €150m buyback. Net cash, no leverage.
• The ASMPT overhang: management continues to hold the ~25% ASMPT stake despite years of activist pressure (Elliott, historically) to divest. The stake generates non-cash P&L volatility (the €215m impairment/€215m reversal round-trip in 2025) and ties up capital that could fund buybacks. This is the clearest capital-allocation question mark — not value destruction, but value sub-optimization.
• Red flags: none of the classic kind (no related-party revenue, no promotional stock-hyping, no aggressive non-GAAP gymnastics beyond standard adjusted figures). The one governance step-down is the discontinuation of quarterly bookings/backlog disclosure from Q1 2026 — defensible (bookings are lumpy and were being over-read) but it reduces transparency into the leading indicator, and I'd note it against an otherwise clean scorecard.

Net: a high-quality, disciplined operating team executing a clean secular story; the debit is a new-ish CEO plus an unresolved ASMPT stake that activists have long wanted monetized.

Lens 10 · Forensic Red Flags

Forensic pass across the three statements — every figure or.

• Revenue recognition: capital-equipment + services; standard for WFE. No evidence of channel-stuffing or aggressive rev-rec. Accounts receivable fell to €562.1m from €789.0m while revenue grew — receivables are not outrunning revenue; this is a clean (arguably conservative) signal.
• Cash vs earnings: net cash from operating activities €1,060.9m in FY2025 vs net earnings €723.7m — operating cash exceeds accounting earnings, the opposite of the classic red flag. Free cash flow €433.9m reported, €615m adjusted for €181m of M&A-related cash payments.
• The ASMPT non-cash noise (the one thing to normalize): FY2025 net earnings €723.7m includes a €215m non-cash impairment reversal on the ASMPT associate stake (recognized as a €215m impairment in Q1 2025, then fully reversed — €181m of it in Q3 2025). This inflates reported net earnings and distorts the quarterly cadence (Q3 net earnings spiked to €384m; Q4 "dropped" to €166m purely because the gain was in Q3). Use adjusted net earnings (€740.8m) and operating metrics, not reported net earnings, to judge the business. This is disclosed and benign, but an analyst who anchors on reported net-earnings growth will be misled.
• Inventory: flat (€552.1m) against growing revenue — healthy, not building ahead of demand.
• SBC / non-GAAP: the gap between reported (30.2% op margin) and adjusted figures is modest — adjustments are principally PPA amortization on the ASMPT stake and normal items, not heroic SBC add-backs.
• Goodwill/intangibles: the associate carrying value is the item that moves; core goodwill is not a flagged risk.

Regulatory findings (required sub-section):

• SEC enforcement (EDGAR): None possible / none found. Per regulatory/regulatory-findings.md (generated 2026-07-06): "ASM International has no CIK — it is public and not required to file with the SEC. No EDGAR enforcement search is possible." Zero SEC LR/AAER by construction.
• Non-SEC enforcement (web search): No material enforcement action against ASM International N.V. found. The prominent 2024–2025 hit for "ASM" — a DOJ $3.5m HSR "gun-jumping" penalty — belongs to ASM Global (a venue-management company acquired by Legends Hospitality), a completely unrelated entity. Do not attribute it to the semiconductor company.
• 10-K Item 3 (Legal Proceedings): n/a — ASM files under IFRS-EU, not the SEC; there is no 10-K. Historical litigation of note is a long-settled patent dispute with Applied Materials (settled/dismissed years ago) and a 2019 patent settlement with Kokusai — no live material litigation surfaced.
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR/AAER — not applicable, no CIK), targeted web search (no ASM International N.V. enforcement; the DOJ/HSR hit is a different "ASM"), and the absence of any live material litigation disclosure, as of 2026-07-06. The one governance debit (not a legal one) is the discontinuation of bookings/backlog disclosure from Q1 2026.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (EPS, next three fiscal years — FY2026 / FY2027 / FY2028)

Built bottom-up from FY2025 actuals + Q1/Q2 2026 guidance + the 2030 framework. All outputs ``; inputs labeled. Note ASM reports EPS in EUR; use adjusted/operating earnings, not reported net earnings (which the ASMPT reversal distorts).

Anchors:

• FY2025 revenue €3,173m; adjusted operating margin 30.2%; adjusted net earnings €740.8m; diluted EPS €14.70 reported (€14.77 basic).
• Q1 2026 €862.5m + Q2 2026 guide €980m ±5%, H2 > H1.
• 2030 target: >€5.7bn revenue (≥12% CAGR 2024–2030), gross margin 47–51%, operating margin 28–32%.
• Consensus 2026: revenue €3.94bn, EPS €21.85. (Note the consensus €21.85 sits well above a "clean" operating-EPS build — it likely includes ASMPT equity income and possibly further associate revaluation; I flag the gap rather than reconcile away.)

Base-case arithmetic (illustrative): FY2026 rev €3.95bn × ~30.5% op margin ≈ €1.20bn op profit; less ~€200m tax at ~21% effective + ~€25m ASMPT income ≈ ~€0.96–1.0bn net ≈ ~€19.5 EPS on ~49m shares. This lands ~10% below the €21.85 street number — the delta is almost entirely how much ASMPT/associate income and any further revaluation you credit. My clean-operating base is more conservative than consensus; that gap is the number to interrogate before underwriting the stock.

Key swing variables (ranked): (1) China — the single biggest 2026 swing, and management itself flipped from "down double-digits" to "up" in one quarter; (2) GAA/2nm capex pace at TSMC/Samsung/Intel — the volume engine; (3) gross-margin mix — every point of mix is ~€30–40m; (4) EUR/USD — a stronger euro suppresses reported growth; (5) memory/HBM re-acceleration — currently only 16% of equipment, pure upside optionality.

Brier forecast: NOT logged (--watchlist unattended mode — the skill instructs skipping forecast.ts create in the loop). A base case worth logging later: "ASM.AS FY2026 adjusted diluted EPS ≥ €19.00" at p≈0.6.

Lens 12 · Bull vs Bear

Bull case. ASM is the highest-quality, purest expression of the deposition-intensity supercycle. The transistor is going 3D (GAA nanosheets at 2nm → 1.4nm → complementary-FET), and every node transition needs more ALD and more epi per wafer — ASM's served market compounds faster than wafer volume. It owns >55% of single-wafer ALD with a switching-cost moat (tool+precursor+recipe qualification), is winning new node-specific applications (Mo ALD, ASD at 2nm; positioned for 1.4nm GAA), and runs the best margins in the group (51.8% GM, 30.2% op margin, both records) with a growing 23% high-margin services annuity that dampens cyclicality. Net cash, disciplined capital return, and a credible >€5.7bn-by-2030 (≥12% CAGR) target that outpaces WFE. Earnings surprise potential: China flipping from headwind to tailwind in 2026 (already signaled), plus an HBM4/DRAM capex leg re-lighting the memory 16%.

Bear case (permanent-impairment risks). (1) The moat gets attacked at the newest applications. ALD leadership is category-wide, but the incremental GAA/backside-power ALD and Mo-deposition wins are exactly where Lam Research (ALTUS Halo — the first Mo ALD tool in production) and Applied Materials are pushing hardest. If ASM's share of the newest deposition steps erodes even while total ALD share looks stable, the growth premium deflates. (2) China is ~30%+ of revenue and hostage to export policy ASM cannot control — a policy tightening (US Congress is actively pushing ) could remove a large revenue block overnight, and management's own guidance on China whipsawed within two quarters (low visibility). (3) Expectations are fully priced — ~45x forward EPS, near the top of a €400→€1,092 range, with no valuation cushion; any capex air-pocket at TSMC/Samsung/Intel re-rates a 45x stock hard.

Pre-mortem (18 months out, thesis broke): most likely path — a 2027 leading-edge capex digestion (2nm capacity got built ahead of demand) coincides with a China export-control tightening, so revenue growth stalls to low-single-digits just as the market had underwritten mid-teens; the 45x multiple compresses to ~30x and the stock halves even though the business is fine. Second path: Lam/AMAT take a visible bite of the GAA/Mo deposition wins, and the "share leader" narrative cracks.

Are multiples too high? For the quality, no — a 45x multiple on a 30%-margin, ~15%-grower with a secular tailwind is defensible versus ASML's ~48x. For the entry, yes — buying at the top of the range with China visibility this low and consensus EPS running ~10% ahead of a clean operating build is a poor risk/reward at this price.

Contrarian view (what the market is refusing to see): the market treats ASM's ALD leadership as monolithic and permanent, and is under-pricing that the marginal, highest-value deposition steps of the GAA/backside-power era are a genuine three-way fight (ASM vs Lam vs AMAT). The share number can stay >50% while the mix of that share shifts toward lower-value films — a slow, hard-to-see margin/growth erosion that a 45x multiple has no room to absorb.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• Where revenue is concentrated / what breaks it: China >30% + a 3–4-name leading-edge customer list. Two independent single-points-of-failure. A China export escalation OR a TSMC/Samsung/Intel capex pause each independently craters the number. The bull "diversified secular demand" story is actually highly concentrated on the buyer side.
• Why the moat is weaker than bulls think: "ALD leader" conflates installed base (sticky, defensible) with new-application wins (contested). Lam already has the first molybdenum ALD tool in production (ALTUS Halo) and production Mo wins scaling through 2025; AMAT and TEL cover selective/plasma ALD for GAA and backside power. The most valuable incremental deposition dollars of this cycle are being fought over three ways — a share-erosion vector the >55% headline hides.
• Most dangerous competitor bulls underestimate: Lam Research. It has scale, an etch+deposition system-sell, and is explicitly targeting ASM's ALD turf at the exact nodes (GAA, Mo, backside power) where the growth is.
• Worst capital-allocation move: not monetizing the ~25% ASMPT stake after years of activist pressure — capital tied in a Hong Kong back-end name that generates non-cash P&L whiplash (the €215m impairment/reversal round-trip) instead of funding buybacks at a 45x stock. Also: cutting bookings/backlog disclosure right as China visibility is the whole debate — reduces the shorts' and the longs' leading indicator.
• Assumptions that must hold for €959: mid-teens revenue CAGR to 2030, op margins held at 30%+, GAA/1.4nm share maintained or gained against Lam/AMAT, AND China not de-rated by policy. That is four things all going right with a 45x multiple and zero cushion.
• If growth disappoints 20–30%: a 45x forward multiple on a decelerating equipment cyclical compresses fast. A 25% cut to the FY2027 growth path plus a re-rate to ~28–30x forward is a ~40–50% drawdown from €959.
• Single scenario that permanently impairs: a structural China decoupling (tools permanently barred, >30% of revenue gone and not recoverable elsewhere on the same timeline) combined with a visible GAA-deposition share loss to Lam. Plausibility: moderate — neither is a tail; both are live, actively-developing risks in 2026.

Lens 14 · Management Questions (ordered by information value)

1. Of your total ALD equipment revenue, what share comes from GAA/2nm-and-below node applications specifically, and how has that subset's share trended versus Lam and Applied over the last four quarters (not total ALD share)?
2. On molybdenum ALD and backside-power deposition — where Lam's ALTUS Halo is now in production — what is ASM's design-win position at the 2nm and 1.4nm nodes, and are you gaining or defending?
3. You guided China down double-digits in 2026 (Q3 2025), then up in 2026 (Q4 2025) — what specifically changed, and what is your current base/bear range for China revenue and its sensitivity to a US/Dutch export-control tightening?
4. What is the clean, ex-ASMPT operating EPS you would point analysts to for 2026, given the €215m associate revaluation distorted reported 2025 net earnings — and how much associate income is in the €21.85 consensus?
5. Why retain the ~25% ASMPT stake after years of activist pressure — what is the specific strategic or financial case versus monetizing it to fund buybacks, and under what conditions would you sell?
6. Why discontinue quarterly bookings and backlog disclosure now, and what forward indicator should investors use in its place to gauge the leading edge of demand?
7. What is the installed base of ASM tools, and what is the structural ceiling on Spares & Services as a % of revenue and its margin as the base grows?
8. On memory (16% of equipment revenue, down from 25%) — what does an HBM4 / advanced-DRAM capex leg do to that mix, and what is your served-market position in HBM-specific deposition?
9. The gross-margin target range was raised to 47–51%, yet you printed 51.8% (FY) and 53.3% (Q1'26) — is the current level sustainable through a downcycle, or is it a mix-driven peak?
10. What is your US manufacturing footprint plan (the "several hundred million euro" Arizona expansion), and how does onshoring change your cost structure and margin?
11. Where are you most capacity- or supply-constrained (precursors, precision components, skilled labor), and could that cap 2026–2027 upside?
12. What is your complementary-FET / post-nanosheet roadmap — does the deposition-intensity tailwind accelerate at the next architecture, and is ASM positioned to lead it?
13. How do you think about capital return priorities — dividend vs buyback vs the Arizona/Korea capex — at a ~45x earnings multiple?
14. What is the realistic downside scenario for revenue if leading-edge logic capex digests in 2027, and how far can margins fall before the cost base needs restructuring?
15. Which of your 2030 targets (>€5.7bn, 47–51% GM, 28–32% op margin) is most at risk, and what is the leading indicator you watch that would tell you first?

---