Veeco Instruments

Phase A — Understand the business

Company Overview

Veeco is a niche advanced-semiconductor process-equipment maker headquartered in Plainview, NY (Delaware corp, organized 1989). It sells capital tools that solve specific materials-engineering steps rather than a broad platform: laser annealing, ion beam deposition/etch (IBD/IBE), MOCVD, CVD, advanced-packaging lithography, single-wafer wet processing, MBE, and ALD.

Four end-markets: Semiconductor (72% of FY25 sales), Compound Semiconductor (9%), Data Storage / HDD magnetic heads (6%), Scientific & Other (13%). The flagship franchises:

• Laser Spike Annealing (LSA) — "production tool of record at all three Tier-1 logic customers," with next-gen Nanosecond Annealing (NSA) in evals.
• Ion Beam Deposition (NEXUS / IBD300 / SPECTOR) — leading position in EUV mask-blank multilayer coating, expanding into EUV pellicles; SPECTOR IBD is the differentiated tool for InP laser facet coatings (the AI-optics story, below).
• MOCVD (TurboDisc / Lumina / Propel) for GaN/As-P power and photonics.

Business model: lumpy, project-based system sales plus a recurring aftermarket. Service, parts and upgrades were ~25% of FY25 net sales (up from 20% in FY24) — the only recurring ballast in an otherwise order-driven P&L. Contracts require deposits on many system orders; receivables run 30–90 days (up to 150 in some geos). Backlog (firm POs, customer-confirmed ship dates, deposits) was $554.9M at Dec 31 2025 vs $409.6M a year earlier — up 35%, a strong forward tell.

Small company: 1,265 employees (13 countries; 923 US), ~350 patents, ~7.6% voluntary turnover, >8-yr average tenure.

Supply Chain

Upstream: Veeco is an assembler/integrator — it outsources the manufacture of several of its systems to third-party partners while keeping some internal build capability, and sources critical components/sub-assemblies from single or limited-source suppliers (named as a risk). Supplier deposits were $9.8M (Dec'25) and purchase commitments $150.8M at Dec'25, rising to $205.8M at Mar'26 — mostly inventory, due within a year (a build-ahead for the InP ramp). Manufacturing/long-lived tangible assets are almost entirely US-based ($108.6M of $108.6M total long-lived assets in the US; China ~$0.2M) — a strategically clean footprint for a US-China-exposed toolmaker.

Veeco → customer: systems ship to semiconductor IDMs, foundries, OSATs, HDD makers, photonics manufacturers, plus research centers/universities. Named end-market pulls: Tier-1 logic (LSA/NSA), leading HBM/DRAM memory (LSA + IBD300 bitline metallization), EUV mask-blank makers (IBD-EUV), OSATs for 2.5D/3D AI-accelerator packaging (wet processing), and a leading 800G/1.6T optical-transceiver supplier for hyperscalers (SPECTOR IBD for InP lasers).

Chokepoints: Veeco itself is the single-source chokepoint in several steps (EUV mask-blank IBD, InP laser facet coating), which is the moat. Its own dependency risk is the reverse — reliance on outsourced system builders and single-source component suppliers, plus export-license gating on China shipments (an $8M revenue headwind hit Q1'26).

Competitive Advantages (moats)

Veeco's moat is narrow-but-deep process leadership in a handful of steps, not scale. "None of our competitors compete with us across all of our product lines." Principal competitors: Aixtron, Applied Materials, Canon, Grand Plastics Technology, Mattson, Screen Semiconductor Solutions, Shanghai Micro Electronics Equipment (SMEE), Suss MicroTec.

Durable edges:

• Tool-of-record lock-in. Being the qualified production tool at all three Tier-1 logic fabs on LSA is a multi-year switching-cost moat — requalifying an annealing step at an advanced node is expensive and risky for the fab.
• Differentiated physics in IBD. SPECTOR ion-beam deposition delivers low-loss optical films (tight thickness/uniformity/reflectivity) that e-beam evaporation, IAD and PVD can't match for InP laser facets — a genuine "why-Veeco" in silicon photonics.
• EUV mask-blank position. IBD is the enabler for the multilayer mirror on EUV mask blanks; a structurally scarce, high-barrier niche.

Bargaining power is mixed: strong versus customers where Veeco is single-source (InP facets, EUV mask blank, LSA), weak in the cyclical, price-competitive segments (Data Storage, some MOCVD) where Aixtron and Applied press. Customer concentration (below) caps pricing power at the top. The 2024 $28.1M Epiluvac SiC impairment is the counter-evidence — where Veeco tried to buy into a step it didn't lead, the market didn't come.

Segments

Veeco reports one operating segment (CODM = CEO, manages on consolidated net income), so "segments" = end-market + geography.

Sales by end-market ($000):

The FY25 −7% top-line is entirely a Data Storage collapse (−$59.6M) plus Compound Semi weakness (−$18.0M), masking a Semiconductor + Scientific advance. The mix shift toward Semiconductor (65%→72%) is the AI-leverage story.

Sales by geography ($000, FY25): US 101,387 (15%); EMEA 50,794 (8%); China 181,812 (27%, down from 36%); Rest of APAC 330,183 (50%, up from 32% — Taiwan $178.8M + Japan $69.0M); RoW 0.1M. The geographic story is de-risking away from China (export controls) toward Taiwan/Japan/Korea.

Q1'26 acceleration of that shift: China $20.0M vs $70.9M (−72%); Rest of APAC $90.4M (+51%); US $32.2M (+34%). China fell to ~13% of quarterly revenue. Data Storage +52% and Compound Semi +31% YoY off low bases — the early bookings inflection showing up.

---

Phase B — Measure performance

Earnings Result (latest print: Q1 FY2026, qtr ended Mar 31 2026)

The GAAP print was soft and cyclical-trough; the order commentary was the actual news.

Q1'26 P&L ($000):

Non-GAAP told a milder story: non-GAAP operating income ~$9M and non-GAAP diluted EPS $0.14; revenue landed within guidance. The GAAP-to-non-GAAP bridge is heavy — $8.5M SBC + $2.0M merger costs + amortization flip a positive non-GAAP quarter to a GAAP loss.

What drove it: gross margin compressed on unfavorable product mix + tariff/production costs; the China −72% collapse ($8M of it from a China export-license requirement) cut the highest-mix revenue. But management reiterated FY26 guidance: revenue $740–800M (+16% at midpoint vs $664M) and non-GAAP EPS $1.50–1.85, Q2 guide $170–190M, "growth accelerating in H2."

The order news (this is the print's real content): Veeco disclosed >$250M of aggregate InP-laser tooling orders (MOCVD + wet + SPECTOR IBD) for AI optical interconnects, deliveries starting Q3 2026 and ramping hard in Q1 2027, against a stated ~$2B market opportunity, and plans to 10x SPECTOR IBD capacity by early 2027. Data Storage is "fully booked in 2026 and extending into 2027" on the HAMR transition. The stock surged ~21% after-hours.

Balance-sheet flags: clean. Cash+equivalents $163.5M + short-term investments $226.8M = $390.2M, against $230M principal of 2029 convertible notes (2025/2027 notes retired) → ~$160M net cash. Undrawn $250M revolver. Equity $883.7M at Mar'26. Backlog $554.9M.

Market reaction / read-through: the tape rewards forward AI-optics order visibility over the trailing trough — exactly the setup that has pushed VECO to a premium over its own merger terms (Lens 7/8).

Earnings Calls (sentiment trend)

No transcripts on the research-layer shelf (transcripts=0); this lens is web-grounded.

Tone arc across the last ~4 quarters:

• Through 2025: defensive — China export-control headwinds, a −60% Data Storage air-pocket, cost discipline (R&D and headcount trimmed). Stock ranged $25–32.
• Q1'26 (May 2026): a clear tonal inflection — the vocabulary shifts to "multi-year demand tied to AI infrastructure," "silicon photonics inflection," "$2B opportunity," "fully booked into 2027," "10x SPECTOR capacity." Management is now selling visibility and ramp, not resilience. CFO Kiernan: first shipments against the $250M+ orders "in the third quarter," "most significant ramp" from Q1 2027.
• Cross-cutting constraint: calls are held inside a merger quiet period — forward strategy talk is muted because Veeco is 8–10 weeks from becoming an Axcelis subsidiary. Read management enthusiasm through that filter: they are talking up the combined asset's optionality.

Recurring phrases that appeared: tool-of-record, InP/co-packaged optics, HAMR, HBM/vertical DRAM, IBD300 bitline metallization. What they stopped emphasizing: China as a growth vector (now a managed risk).

Comps

The comp that matters is the deal itself. VECO is priced off ACLS via the 0.3575 ratio, so a P/E table on VECO's standalone earnings is largely moot — I give it, then the merger-math table that actually governs the price.

Peer multiples (as of Jul 9–10, 2026):

Sector context: semiconductor-equipment names trade ~34x forward on average (3-yr avg 33.7x); the group re-rated hard on AI-WFE. EV/EBITDA and 5-yr avg ROE columns: n/a at company level (I will not fabricate them). VECO standalone EVEnterprise valueWhat it would cost to buy the whole company: its market value plus its debt, minus the cash you would get with it. Often very different from market cap. ≈ $3.51B − ~$160M net cash ≈ **$3.35B ** → ~4.4x FY26 guided sales ($770M mid); that multiple is a deal artifact, not a fundamentals read.

The merger math that actually prices VECO:

• Exchange ratio: 0.3575 ACLS per VECO.
• Implied deal value/VECO = 0.3575 × ACLS $142.93 = $51.10.
• VECO spot = $57.45.
• VECO trades ~$6.35 / ~12.4% ABOVE deal value.
• Look-through: post-close VECO holders own 41.6% of the combined ~52.3M-share ACLS; that stake is worth ~$3.08B at spot ACLS vs VECO's $3.51B cap — VECO is ~14% rich to its own look-through.

Conclusion: this is an inverted merger-arb. Classic arbs buy the target at a discount to deal value; VECO trades at a premium, meaning the market is either (a) pricing standalone AI-optics upside above the deal, or (b) betting on a bump/break. Owning VECO to get the combined company is strictly worse than owning ACLS outright.

Stock-Price Catalysts (moves >5%, ~5yr)

• 2020 (COVID trough): VECO < $15.
• 2021–early 2024 (up-cycle): ran to ~$40 on the post-pandemic chip shortage + early AI/HBM optimism.
• 2025 (down-cycle): China export controls + Data Storage/HDD −60% air-pocket → sideways $25–32, 52-wk low $19.29. Insider selling + "China sags" headlines pressured it.
• Sep 30, 2025 — Axcelis merger announced: VECO becomes a deal-linked instrument; from here it tracks 0.3575×ACLS ± a deal-risk basis.
• Feb 6, 2026 — both shareholder votes pass: deal de-risked to a single China-SAMR gate.
• May 5–6, 2026 — Q1 print + $250M InP orders: +21% after-hours; the standalone AI-optics re-rating begins, pushing VECO to a premium over deal value. 52-wk high $86.63 was set in the subsequent ACLS/AI-optics euphoria (ACLS itself hit $193.78).
• Pattern the market reacts to: (1) ACLS's price (now the dominant driver), (2) AI-optics/InP order flow, (3) China SAMR headlines (deal risk), (4) cyclical WFE sentiment. Earnings matter less for their trailing numbers than for order visibility.

---

Phase C — Judge people & books

Management

CEO: Dr. William J. Miller — CEO + Director since Oct 2018; at Veeco since Nov 2002 (VP/GM Data Storage → SVP/GM MOCVD → EVP Compound Semi → EVP Process Equipment → President Jan 2016). Prior: Advanced Energy Industries. A deep insider-operator, not a hired-gun — 20+ years across every Veeco business unit. CFO: John Kiernan (fronting FY26 guidance). CTO: Peter Porshnev.

• Track record: navigated Veeco through the MOCVD/LED bust of the 2010s and repositioned the company toward semiconductor/AI leverage (Semiconductor 62%→72% of sales over three years). Delivered the LSA tool-of-record wins at all three Tier-1 logic fabs — the crown-jewel outcome. Against that: the Epiluvac SiC acquisition was a $28.1M write-off (2024), market penetration below plan — a capital-allocation miss.
• Skin in the game: Miller beneficially holds several hundred thousand shares (headline: 419,570 retained after a 100k sale). But he sold ~325,000 shares over the trailing year and bought none — tranches of 50k (Mar 3), 100k (May 7, right after the Q1 pop), 50k (Jun 2). Read it two ways: benign (rational de-risking when a fixed 0.3575 ratio caps your upside and your stock is about to become ACLS anyway) or mildly negative (no insider is adding into the "AI-optics inflection"). Net: a soft yellow flag, not a red one.
• Capital allocation: disciplined balance sheet (retired 2025/2027 converts, net-cash, undrawn revolver, no dividend, modest buybacks), R&D held ~18% of sales through the trough. The one blemish is Epiluvac.
• Governance note: one Veeco independent director also sits on the Axcelis board and recused from the merger vote — a pre-existing tie between the two boards worth flagging (it helped seed the deal but is a related-party-adjacent optic).
• Archetype: professional operator-CEO executing a sell-into-strength endgame — he is delivering the company to Axcelis at a moment of peak AI-optics narrative.

Forensic Red Flags

Accounting quality is clean; the risks are structural, not forensic.

• SBC > GAAP net income. FY25 SBC $37.0M vs net income $35.4M; Q1'26 SBC $8.5M on a $0.3M net loss. Non-GAAP EPS ($0.14 Q1) leans hard on adding SBC back — standard for semicap, but it means "profit" is a shareholder-dilution story. Shares grew 56.8M→61.0M over five quarters.
• Customer concentration. Customer A = 17% of FY25 net sales and 24% of receivables; the top-10 customers = 68% of net receivables (63% FY24). Single-customer receivable concentration is the balance-sheet watch item.
• Gross-margin compression + tariffs. 42.4%→40.0% (FY) → 35.3% (Q1'26) on mix + tariff costs; management warns of continued tariff pressure. A Feb 20, 2026 US Supreme Court ruling invalidated IEEPA tariffs; Veeco is filing for refunds but has booked no adjustment (refund amount/timing uncertain).
• Inventory/receivables vs revenue: FY25 operating-asset build (inventory + AR up) is consistent with a build-ahead for the InP ramp, not channel stuffing — corroborated by the $205.8M purchase-commitment step-up.
• Cash vs earnings: healthy — FY25 operating cash flow $69.5M > net income $35.4M (D&A + SBC bridge); FCFFree cash flowCash left after paying to run and maintain the business. Unlike profit, it is hard to flatter with accounting choices. ≈ $53.3M after $16.2M capexCapital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs.. No divergence red flag.

Regulatory findings (required sub-section):

• SEC Litigation Releases: none for Veeco, 2021–2026.
• SEC AAERs: none.
• 10-K Item 3 / Note 10 Legal Proceedings: "The Company is involved in various legal proceedings arising in the normal course of business… [not expected to] have a material adverse effect." No material litigation.
• Non-SEC (web): no FTC/DOJ/FDA enforcement actions found. The live "regulatory" story is export-control exposure, not enforcement — expanding US restrictions on China shipments (an $8M Q1'26 revenue headwind from a China export-license requirement), Chinese retaliation (2025 rare-earth ban), and tariffs. These are macro/policy risks to the business, not compliance findings against Veeco.
• The one regulatory item that can move the stock is the merger's China SAMR review (Lens 12/13) — an antitrust approval, not an enforcement action.
• Verdict: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3/Note 10 as of 2026-07-10.

---

Phase D — Project & stress-test

Forward Projection

Caveat up front: if the merger closes in H2 2026, Veeco will not report a standalone FY2027 — projections are for the break/standalone scenario and for valuing the asset Axcelis is buying. No our model logged (watchlist/unattended + event-driven). No fabricated consensus.

Anchor = FY26 guidance: revenue $740–800M, non-GAAP EPS $1.50–1.85.

Standalone base / bull / bear ($ revenue, non-GAAP EPS):

The base case implies FY27 non-GAAP EPS ~$2.30, roughly consistent with VECO's ~25.7x forward P/E on ~$2.2 NTM. But the price is set by ACLS, not this model — at 0.3575×ACLS, VECO's "fair value" today is ~$51 regardless of the standalone build. The projection matters only in the break scenario, where standalone VECO would likely trade $45–65 (a re-rated AI-optics small-cap, above its old $19–32 range but below the current deal-premium price).

Bull vs Bear

Bull case (for the underlying asset / the combined company). Veeco brings three genuine single-source franchises (LSA at all Tier-1 logic, IBD-EUV mask blanks, SPECTOR for InP facets) into an Axcelis platform, lifting combined TAM >$5B and pro-forma 2024 scale to $1.7B revenue / $387M adj. EBITDA / >$900M cash — the 4th-largest US WFE supplier. The secular levers are real and current: AI optical interconnects (copper→optics, $2B InP opportunity, $250M+ already booked), HBM/vertical-DRAM anneal, EUV/High-NA mask blanks + pellicles, 2.5D/3D advanced packaging, and SiC/GaN power (Axcelis implant + Veeco epi). Balance sheet is net-cash. Backlog +35% YoY.

Bear case (2–3 permanent-impairment risks). (1) China structural loss — 27% of FY25 revenue, already −72% in Q1'26; a permanent export-control wall removes a quarter of the TAM and hands share to SMEE/local Chinese toolmakers. (2) Cyclicality + concentration — Data Storage swung −60% then +52% in consecutive years; one customer is 17% of sales; a WFE down-leg with lumpy orders can halve earnings. (3) The merger caps standalone upside — holders can't fully monetize the AI-optics inflection because they're being paid in a fixed 0.3575 ACLS ratio struck when VECO was ~$30.

Pre-mortem (18 months out, thesis broke). Most likely failure: China SAMR slow-walks or conditions the deal, VECO drifts in limbo through the Sep-2026→Jun-2027 outside-date window; meanwhile the WFE cycle rolls, InP orders push right, ACLS de-rates from its ~34x forward, and VECO — having traded at a premium to the deal — mean-reverts down 20–30% as both the ACLS proxy value and the standalone premium compress.

Are multiples too high? For VECO specifically, yes on a relative basis — paying a 12% premium to deal value is negative carry into a likely close. For the combined ACLS, ~34x forward is full but defensible if AI-WFE holds.

Contrarian view (what the market refuses to see): the crowd treats VECO's post-InP surge as upside; mechanically it created downside — VECO now trades above its own merger terms, so absent a break the stock's most probable path is a grind down to ~$51 as the spread closes. The "good news" re-priced the stock into a headwind.

Devil's Advocate (short-seller)

Short the premium, not the company. The clean bear expression is long ACLS / short VECO (or simply avoid VECO): capture the ~12% convergence as the deal closes, with the combined company's fundamentals as your hedge. Structural break-points a short leans on: China SAMR is a real, non-trivial gate (US-China semicap consolidation, reviewed amid a rare-earth/export-control trade war — even under SAMR's simplified procedure, no unconditional clearance had posted as of mid-March 2026, and ION Analytics flagged "closer China antitrust scrutiny amid possible complaints"). Revenue concentration (Customer A 17%, top-10 68% of AR) and China (27%→13% and falling) mean a single customer or a policy step can crater a quarter. The moat is narrow — Aixtron and Applied contest MOCVD and adjacent steps; SMEE is the state-backed China substitute. Capital-allocation blemish: the Epiluvac SiC write-off. Insider behavior: CEO sold 325k shares, bought zero. What must hold for today's price: the deal closes near current ACLS and ACLS holds 34x forward — two bets, not one. If WFE growth disappoints 20–30%, standalone EPS drops toward the bear ($1.50–1.75) and ACLS de-rates, taking VECO down with it. The single scenario that permanently impairs value: a hard, durable US-China decoupling that walls off Chinese demand for both Veeco and Axcelis while a competing anneal/deposition roadmap displaces Veeco at a Tier-1 node — low probability, high severity.

Management Questions (ordered by information value)

1. What is the current, specific status of the China SAMR review — filing acceptance date, simplified vs. full procedure, and the exact conditions or remedies (if any) under discussion?
2. If SAMR conditions or blocks the deal, what is the standalone plan and does either party owe a break/expense payment (the $77.5M/$108.7M fees apply to Adverse Recommendation Changes, not a regulatory failure — walk us through the regulatory-failure waterfall)?
3. Given VECO trades above the 0.3575 implied value, is there any mechanism (collar, ratio adjustment, top-up) protecting Veeco holders if ACLS falls before close — or do they simply absorb ACLS downside?
4. Of the >$250M InP orders, how much is firm PO with deposits vs. LOI, what is the customer count and top-customer share, and what is the gross margin on SPECTOR IBD vs. corporate average?
5. What is the realistic 2027 revenue bridge from the InP ramp + Data Storage 2027 backlog + China recovery/loss — and the exit gross margin at that volume?
6. How defensible is LSA tool-of-record at the three Tier-1 logic fabs through the next two nodes — what would it take for a customer to requalify a competitor, and is NSA a defense or a new attack surface?
7. On IBD300 (DRAM bitline metallization) and IBD-EUV pellicles — where are these in the eval→POR→volume funnel, and what's the SAM each unlocks?
8. What share of revenue is now China-export-license-gated, and what is the downside case if the current license regime tightens another notch?
9. What is the combined-company integration plan for the overlapping wet-processing / deposition portfolios and the two sales channels — and the quantified revenue-synergy target with timing?
10. Post-close, how is Veeco's leadership and R&D footprint retained (San Jose, Plainview, etc.), and what is the employee-retention exposure given the pending change of control?
11. Capital allocation post-close: buybacks vs. reinvestment vs. the $230M 2029 convertible — how is the combined ~$900M cash deployed?
12. What did Veeco learn from the Epiluvac SiC write-off, and how does that change the diligence bar on the Axcelis SiC-implant + Veeco-epi wide-bandgap thesis?
13. Why has the CEO sold ~325k shares and bought none over the past year — how much of that is 10b5-1Rule 10b5-1 planA schedule an insider sets in advance saying when their shares will be sold, so the sales cannot be timed on private news. A sale made OUTSIDE such a plan is a live decision, which is why it carries more information. / deal-mechanical vs. discretionary?
14. What is the aftermarket/service growth path (25% of sales and rising) as the installed base of LSA/IBD/InP tools expands — can it become a recurring margin anchor?
15. If the deal does not close by the Sep-2026 outside date, what triggers the automatic extension to Jun 2027, and what is the operating plan for a 9-month regulatory-limbo period?

---