Synopsys

Phase A — Understand the business

Company Overview

Synopsys is the global #1 in electronic design automation (EDA) — the software chip designers use to design, verify, test, and tape out integrated circuits — and, since July 2025, the new owner of Ansys, the gold standard in engineering simulation & analysis (S&A: structural, fluids, electronics, optics). Management's framing is "silicon to systems": own the entire flow from a transistor to a finished physical product.

How it makes money. Two reportable segments:

• Design Automation — FY2025 revenue $5,302.4M (+26% YoY). Holds EDA tools, Ansys S&A, and "Other." This is the engine.
• Design IP — FY2025 revenue $1,751.8M (−8% YoY). Pre-built, silicon-proven IP blocks (interface, memory, security, processor IP). This is the problem child (Lens 4/5).

Four revenue groups (Q2 FY2026 mix): EDA 51.0%, Ansys 28.7%, Design IP 20.0%, Other 0.3%. Ansys is already nearly a third of the company.

Contract structure — the quality tell. EDA sells mostly via multi-year Technology Subscription Licenses (TSL), 2–3 year terms, recognized ratably → recurring, sticky, visible. FY2025 revenue split: time-based 49%, upfront 29% (hardware/IP/perpetual S&A), maintenance & service 22%. ~71% is ratable/recurring; the 29% upfront (emulation hardware, IP, S&A perpetual) is what makes quarters lumpy. Backlog $11.0B at Apr-30-2026 (vs $8.1B a year before the deal, $11.4B at FY25 year-end) including $1.8B non-cancellable FSA; ~49% to convert within 12 months.

Customers/suppliers/competitors. Customers = essentially every chip and systems designer on earth (NVIDIA, AMD, Apple, Intel, Samsung, TSMC's ecosystem, the hyperscalers' custom-silicon teams). No customer concentration disclosed at ≥10% (none flagged in either filing). "Suppliers" are minimal — this is a software/IP business; the real input is ~21,000 engineers (75% of 28,000 staff). Competitors: Cadence (the near-mirror), Siemens EDA (#3), plus Keysight and Zuken at the edges, and customers' in-house tools.

Plain terms: Synopsys rents the indispensable software that makes modern chips possible, on multi-year contracts, to a customer base that cannot function without it — and just bought the company that simulates whether the resulting physical product will actually work.

Supply Chain

EDA/IP has no physical supply chain in the semiconductor sense — but it sits at the most leveraged choke point of the entire silicon value chain, so mapping the value chain matters more than inputs.

Upstream (inputs): human capital (≈21,000 engineers, >half with Master's/PhD ); compute (Synopsys Cloud runs on Microsoft Azure — first EDA SaaS ); and foundry process-design-kit co-development with TSMC, Samsung, Intel Foundry, GlobalFoundries — the tools must be certified for each new node (3nm, 2nm, GAA) before customers can design on it. That co-dependency with the foundries is the real "supply" relationship.

The node: Synopsys tools + IP → fabless designers and IDMs (NVIDIA, AMD, Apple, Qualcomm, Broadcom, MediaTek, the hyperscalers' silicon teams) → designs sent to foundries (TSMC, Samsung, Intel) → packaged → end systems. Synopsys' emulation/prototyping hardware (ZeBu, HAPS) is the one physically-built product line — contract-manufactured, carries inventory ($365M ) and is the lumpiest revenue line.

Chokepoints / single-source dependencies:

1. Advanced-node PDK certification — being the tool that's qualified first on each new node is a near-single-source position; designers won't risk a bleeding-edge tapeout on unproven tools.
2. China export-control transmission — the U.S. BIS can switch off ~11–16% of revenue with a letter (it did, May 2025; Lens 5/8/10). Synopsys is a chokepoint the U.S. government uses as a weapon.
3. Ansys channel — S&A goes partly through independent distributors; the Q2-FY26 gross-vs-net reclassification shows this channel is being pulled in-house.

Named stakeholders along the chain: inputs — Microsoft (Azure), TSMC/Samsung/Intel/GlobalFoundries (PDK partners), Arm (AMBA fabric licensed into IP). Buyers — NVIDIA, AMD, Apple, Qualcomm, Broadcom, MediaTek, Samsung LSI, Intel, and the hyperscaler custom-silicon teams (Google TPU, AWS, Microsoft, Meta). Regulators — U.S. BIS/Commerce as an on/off switch.

Competitive Advantages (moats)

This is one of the widest moats in software, narrowing at exactly one seam.

• Switching costs / workflow lock-in (very high). A chip design represents years of work encoded in a specific toolchain; methodology, scripts, IP, and verification environments are tool-specific. Migrating mid-program risks a multi-hundred-million-dollar tapeout. Customers negotiate total contract value, not unit price — the hallmark of a vendor that isn't price-shopped.
• Scale of R&D (structural). FY2025 R&D $2,479.3M = 35% of revenue. Only Cadence can match that absolute spend. EDA is a "you must run to stand still" business — each node needs new tools — and only two firms can fund the treadmill. This is the deepest moat: a new entrant would need ~$2B/yr and a decade.
• IP estate / standards. >3,800 patents (expiry through 2044); silicon-proven IP for the newest interfaces (UCIe, UALink, HBMHigh-bandwidth memoryMemory stacked in layers beside a processor so data reaches it faster. The usual bottleneck in AI chips., PCIe, DDR/LPDDR) — being first-to-silicon on a new standard is itself a moat.
• The new moat attempt — silicon-to-systems. Ansys lets Synopsys claim a flow no one else has end-to-end: chip → package → board → multiphysics (thermal/structural/EM/CFD) sign-off in one environment ("Multiphysics Fusion"). If it lands, it's a genuine new moat; if it doesn't, it's $35B of goodwill.
• Bargaining power. Over customers: very high (indispensable, switching-cost-locked). Over "suppliers": n/a. The one party with power over Synopsys is the U.S. government (export controls).

The narrowing seam: Design IP. The −8% FY2025 / −6% Q2-FY26 decline isn't all macro — management admits "roadmap and resource decisions that did not yield their intended results". A moat that's being actively pruned (selling the Processor IP business to GlobalFoundries, Lens 5) is a moat the company itself judges sub-scale in parts.

Segments

All `` and filings/10-q-2026-q2.md, Note 5/17 unless noted.

By segment (revenue):

Caveat: DA's +26% / +62% is inflated by Ansys ($756.6M in FY2025; $652.4M in Q2-FY26 alone). Strip Ansys and organic DA is healthy but not explosive; the headline growth rate is acquisition-fueled.

Segment (adjusted/non-GAAP) operating margin — the cleaner read on segment health:

• Design Automation: 6-mo FY26 adj. op income $1,736.6M (+101%) — strong, accelerating.
• Design IP: FY2025 adj. op income $419.3M, margin 24% (down from 38% in FY2024, −14pts); 6-mo FY26 $176.8M (−36%). This is the story — a high-margin IP business whose margin collapsed ~14 points in a year.

By product group (Q2-FY26): EDA 51.0%, Ansys 28.7%, Design IP 20.0%, Other 0.3%.

By geography (Q2-FY26): US $974M (43%), Europe $378M (17%), Korea $265M (12%), China $240M (11%), Other $418M (18%, incl. Taiwan/Japan). Ansys structurally shifted the mix toward US + Europe (Ansys is more Western/industrial). China fell from ~16% of revenue (FY2024) to ~11% — the export-control + design-start damage, partly recovering off the trough.

Trend & cause: DA accelerating (AI/HPC design demand + Ansys); Design IP decelerating then declining (China export controls disrupted design starts + a weak major foundry customer + self-inflicted roadmap misses). The company is reallocating IP resources to higher-growth pockets and divesting the laggard (Processor IP).

---

Phase B — Measure performance

Earnings Result (latest print: Q2 FY2026, ended 2026-04-30, reported 2026-05-27)

The headline GAAP number looks like a disaster; it isn't.

• Revenue $2,276.0M, +42% YoY ($1,604.3M PY). Beat — and guidance raised.
• GAAP operating income $120.4M (5.3% margin) vs $376.4M (23.5%) PY — −68%. GAAP net income $17.1M; GAAP diluted EPS ~$0.09 (vs $2.25).
• Non-GAAP net income $643.7M; non-GAAP diluted EPS $3.35 vs $3.67 PY.

Reconcile the gap (the entire thesis lives here): Q2-FY26 carried $403.6M of acquired-intangible amortization ($248.4M COGS + $155.3M opex), $115.9M restructuring, and $133.4M interest expense. Those three lines (~$653M) are the Ansys deal's GAAP footprint. They are real cash in the case of restructuring/interest, non-cash for amortization — and they crush GAAP while the underlying franchise compounds. Non-GAAP op income proves the point.

The genuinely uncomfortable fact: non-GAAP EPS fell YoY ($3.35 vs $3.67) despite +42% revenue. Why: ~32M new shares issued for Ansys (dilutionDilutionIssuing new shares, so each existing share owns a smaller slice of the same company.), higher non-GAAP share count, and $133M/qtr interest. Ansys is dilutive to per-share earnings right now and stays so until synergies + deleveraging arrive (management/Street: FY2027+). Anyone buying the "accretive mega-deal" story should sit with that.

• Lines that drove it: DA +62% (Ansys + organic AI/HPC strength); Design IP −6% (still bleeding).
• Guidance raised: FY2026 revenue midpoint $9.665B (from $9.610B), non-GAAP EPS $14.76. Ansys FY26 contribution ~$2.96B.
• Balance-sheet flags: cash down to $2.41B (from $2.89B); AR $1.5B (collections + Ansys); deferred revenue $2.81B (healthy, +); inventory flat. Long-term debt $13.5B.
• Market reaction: stock dipped on the print despite beat-and-raise — the market is fixated on the non-GAAP EPS decline + Design IP, not the revenue beat. Tells you expectations are high and the per-share dilution is the sore point.
• Unusual vs own history: GAAP op margin of 5% is unprecedented for this name (normally mid-20s GAAP, high-30s non-GAAP) — 100% explained by Ansys accounting, but it means GAAP screens (ROE/ROIC) are temporarily meaningless (Lens 9/7).

Earnings Calls (sentiment trend)

No transcripts on the research shelf (transcripts=0); this is ``.

Tone across the last ~4 quarters has arced from defensive (mid-2025) → cautiously confident (early-2026) → confident-with-caveats (Q2-FY26):

• Mid-2025 dominated by the China export-control shock (guidance suspended May 29, 2025) — crisis-management tone.
• Q1–Q2 FY2026: management pivoted hard to the synergy narrative — "accelerating synergies," "strong cost discipline," "expanded operating margin". CEO Ghazi's recurring frame: "the most transformative decade in modern engineering," "silicon-to-systems," "re-engineering engineering in the age of AI".
• What they now emphasize: Ansys cross-sell, Multiphysics Fusion, AI-driven EDA (Synopsys.ai, Copilot), deleveraging path, cost synergies ahead of plan.
• What they stopped saying: the bullish China commentary of 2023–24; Design IP is now discussed as a fix-it/reallocate story, not a growth story.

Net: management is selling the future (systems + AI) hard precisely because the present per-share math (dilution, Design IP, GAAP optics) is unflattering. Credible, but promotional-leaning — watch whether synergy claims convert to non-GAAP EPS in FY2027.

Comps

Peer set = the EDA/design-software complex.

5-yr-average ROE column (per battery spec): SNPS ~16%; CDNS not separately 5-yr-sourced but currently ~34%. Dividend yield: both ~0% (neither pays a meaningful dividend; SNPS has suspended buybacks).

Read: Synopsys trades at a ~10-turn forward-P/E discount to Cadence (~35x vs ~45x) — unusual, since SNPS is the #1 by share. The discount is the market pricing (a) the Ansys integration/dilution overhang, (b) the broken Design IP leg, and (c) the GAAP-depressed reported earnings. The bull reframes that discount as the opportunity; the bear says Cadence deserves the premium because its returns (ROIC ~28% vs SNPS's structurally-lower, goodwill-burdened ~12% normalized) are simply better and cleaner. Both at ~35–45x forward are priced for years of flawless execution — there is no valuation margin of safety in either name.

Stock-Price Catalysts (last ~5 years; mostly ``)

What actually moves SNPS >5%:

1. China export controls (the dominant 2025 catalyst). May 23–29, 2025: BIS "is-informed" letter → guidance suspended → stock −9.6% (Cadence −10.7%). July 2–3, 2025: controls rescinded → stock +~5%. Single biggest swings of the period. Lesson: this name trades on Washington as much as on bookings.
2. Earnings beats/misses vs the upfront-revenue line — because ~29% of revenue is lumpy (hardware/IP), quarters surprise; the Q2-FY26 dip-on-beat shows the market now reacts to non-GAAP EPS and Design IP, not the revenue line.
3. The Ansys deal arc — announced Jan 2024, regulatory wait (China SAMR approval was the long pole), closed Jul 17, 2025. Each milestone moved the stock; Elliott Management involvement added an activist overlay.
4. AI/semis beta — SNPS rides the broader AI-infrastructure trade (NVIDIA-adjacent design demand); a January 2026 "Nvidia AI deal" report was a notable up-catalyst.
5. Analyst re-rates — Morgan Stanley downgrade Feb-27-2026 ($550→$480) on the dip; Stifel ($600) and Citi ($610) reiterations Jun-11-2026 on "Q2 beat and IP recovery".

Pattern: the market reacts hardest to (a) U.S.-China policy and (b) signs the Design IP leg is healing — not to the EDA core, which it (correctly) treats as a compounding annuity.

---

Phase C — Judge people & books

Management

• Sassine Ghazi — President & CEO (since Jan 2024). 26-year Synopsys lifer; joined 1998 as an applications engineer, ran the largest (digital/custom) business group, then COO, now CEO. Engineer-operator, not a finance import. Track record: as COO/CEO he drove the silicon-to-systems strategy and executed the largest deal in company history ($35B Ansys) to close, through a hostile regulatory environment (China SAMR). Skin in the game: standard exec equity (no founder-scale stake; specific insider % n/a from a proxy here).
• Aart de Geus — Executive Chair. Co-founder (1986), CEO 1994–2024. Still on the board; founder DNA retained at the chair level. This is the rare large-cap where the founder handed the keys to a 26-year internal successor — strong continuity signal.
• Shelagh Glaser — CFO (since Dec 2022). Ex-Intel CFO (Data Platform Group) and ex-Zendesk CFO. Heavyweight operator-CFO who financed a $14B+ debt raise cleanly (Lens 10).
• Mike Ellow — CRO (since Nov 2025). The most revealing hire in the file: he was CEO of Siemens EDA (the #3 competitor) until Nov 2025. Synopsys decapitated a direct rival's sales leadership and installed him to run its own. Aggressive, and a tell about relative competitive momentum.
• Janet Lee — GC (since Jul 2025). Came over from Ansys (was Ansys GC) — integration-continuity hire.

Capital allocation: the defining act is the Ansys deal — ~$35B (cash + stock), funding the cash leg with $10B senior notes + $4.3B term loan. Bold, transformative, and the entire investment debate: brilliant platform expansion or value-destroying overpay (goodwill + intangibles = $39.6B, more than total pre-deal assets). They suspended the buyback to delever — disciplined, but it removes a support for the stock. ROE/ROIC: 5-yr avg ~16%/~12%; currently GAAP-crushed to ~2.5%/~2% by amortization + goodwill — temporarily uninterpretable. Cadence's ~28% ROIC is the cleaner benchmark and a fair stick to beat them with.

• Red flags (governance): none material. No related-party deals flagged; comp is equity-heavy but normal; the Ellow raid is aggressive not unethical. The only "flag" is the size of the bet relative to the balance sheet.
• Archetype: professional operator-engineer atop a founder-rooted board. For a maturing, M&A-driven scale play, that's the right archetype — execution and integration discipline matter more than founder vision here.

Forensic Red Flags

Acting as a forensic analyst. All figures `` from the two filings unless noted.

• Revenue recognition (medium-complexity, well-disclosed). Heavy use of multi-element TSL arrangements with judgment on combined-vs-separate performance obligations; ~29% upfront revenue recognized at a point in time (hardware/IP/perpetual S&A) → genuine quarter-to-quarter volatility, disclosed as a risk factor by the company itself. The Q2-FY26 channel-partner gross-vs-net reclassification (now reporting reseller revenue gross) added ~$12.5M and is the kind of presentation change worth watching, though immaterial here. Not a red flag, but the revenue model rewards close reading of bookings/backlog over any single quarter.
• Goodwill & intangibles (the real watch item). Goodwill $26.9B (from $3.4B) + intangibles, net $12.7B (from $0.2B) = $39.6B, on a $48.2B balance sheet. The Ansys allocation used royalty rates of 35–45% and a ~10% discount rate. If the deal underperforms, impairment risk is enormous — and there's precedent for write-downs ($53.5M core-tech impairment in FY2024). Amortization runs ~$1.5B/yr for years (2027 $1,545M, 2028 $1,385M…) — a structural GAAP drag that also flatters non-GAAP by exactly that amount. The single biggest accounting judgment in the company is whether $39.6B of Ansys carrying value holds.
• SBC flattering non-GAAP. Standard for the sector; non-GAAP excludes amortization, SBC, restructuring, and deal costs. The non-GAAP/GAAP wedge is now huge ($3.35 vs $0.09 in Q2) — legitimate but means non-GAAP is doing a lot of work; a skeptic should haircut it.
• Cash flow vs earnings. FY2025 OCF $1,518.6M vs GAAP net income from continuing ops $1,336.1M — OCF exceeds GAAP earnings (deferred-revenue/ratable model generates cash ahead of P&L), which is reassuring quality-of-earnings. AR jumped to $1.5B but that's Ansys consolidation + timing, not a receivables-outrunning-revenue red flag.
• Below-the-line earnings inflation (FY2025). GAAP net income from continuing ops ($1,336.1M) sat above GAAP operating income ($914.9M) only because of a $548.9M gain on divestitures (Optical Solutions Group → Keysight; Software Integrity tail) and $277.7M interest income — i.e. FY2025 GAAP EPS was helped by one-time gains while hurt by amortization. Messy year; trust the segment-level adjusted figures and OCF over headline EPS.
• Tax rate (low, sourced). FY2025 effective rate 4.0% (R&D credits, FDII, $148M valuation-allowance release); Q2-FY26 12.5% (6-mo 17.0%). Normalizing upward as the one-offs roll off + OBBB R&D-expensing reduces the FDII benefit — a modest forward EPS headwind.
• Leverage covenant. Term loan + revolver carry a max-consolidated-leverage covenant; in compliance at FYE. With ~$13.5B debt and OCF ~$1.5B, deleveraging is the multi-year project that gates the buyback's return.

Regulatory findings (required sub-section).

• SEC Litigation Releases / AAERs: None. Verified via SEC EDGAR EFTS (LR + AAER), 2021-06-23 → 2026-06-23. total_sec_findings: 0.
• 10-K Item 3 / Q2 Note 20 (Legal Proceedings): routine business litigation only; no material matters. One non-SEC tax matter: Ansys Korea withholding-tax appeals (Korea NTS, tax years 2017–2023), net impact recognized with offsetting foreign tax credit — not material.
• Non-SEC enforcement (web): No FTC/DOJ/FDA/CFPB enforcement actions or consent decrees found against Synopsys. The material government exposure is export-control compliance (BIS), which is a regulatory-risk channel (it can switch off China revenue) rather than an enforcement finding against the company. The Ansys deal cleared U.S. and China SAMR antitrust review (with remedies) to close.
• Verdict: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 / 10-QForm 10-QThe quarterly version of the annual report. Lighter, and not audited. Note 20 as of 2026-06-23. The genuine regulatory risk is geopolitical export control, not accounting/enforcement.

---

Phase D — Project & stress-test

Forward Projection (EPS, next three fiscal years — FY2026 → FY2028)

Built bottom-up from FY2026 guidance + segment trajectory. Output ``; inputs labeled. No our model logged (watchlist breadth mode — no committed forecast per skill rules).

Anchors: FY2026 guidance revenue $9.665B, non-GAAP EPS $14.76. Synergies: $400M cost by year 3 + $400M revenue by year 4. Street: ~16% EPS CAGR, ~11.6% revenue CAGR.

Scenarios (FY2028 non-GAAP EPS):

• Bear ~$16–17: Design IP keeps bleeding, synergies slip past FY2028, AI commoditizes high-end EDA pricing, China stays capped. EPS roughly flat-to-FY26.
• Base ~$19.5–20.5: as above — mid-teens EPS CAGR, the consensus path.
• Bull ~$22–24: ~$1B run-rate synergies pulled forward, margins to high-30s, Multiphysics cross-sell inflects, China normalizes.

Valuation cross-check: at ~$455 and base FY2027 EPS ~$17.25, forward P/E ~26x; on FY2028 ~$20 → ~23x. So the ~35x on FY2026 compresses to ~23–26x two years out if the base case holds — that's the bull's whole argument (you're paying 35x for a 23x business if synergies land). The bear's: if synergies don't land, you paid 35x for a mid-teens grower with a broken IP leg and $13.5B of debt.

No forecast logged (watchlist rule). If promoted to a thesis, the scoreable base call would be: "SNPS FY2028 non-GAAP EPS ≥ $19.50," p≈0.55.

Bull vs Bear

Bull case. Synopsys owns the widest moat in semiconductor software (switching costs + $2.5B R&D treadmill only two firms can run), at the exact moment AI makes chip and system complexity explode. The Ansys deal — temporarily ugly on GAAP and per-share — builds a silicon-to-systems platform no competitor can replicate, with $400M cost + $400M revenue synergies that drive operating margins back toward the high-30s and non-GAAP EPS to ~$20+ by FY2028. Deleveraging restores the buyback. Backlog $11B gives 1–2 years of revenue visibility. At ~35x forward — a 10-turn discount to Cadence — you're underpaying for the #1 franchise because the market is squinting at GAAP optics and a fixable Design IP leg. Earnings surprise vector: synergies ahead of plan + China normalization + an AI-design super-cycle.

Bear case (the 2–3 things that could permanently impair or de-rate it).

1. The Ansys deal is a value-destroying overpay. $39.6B of goodwill+intangibles on a $48.2B balance sheet; ~$1.5B/yr amortization; $13.5B debt; and per-share dilution today (non-GAAP EPS fell YoY in Q2 despite +42% revenue). If revenue synergies don't monetize (not expected before FY2027) and cost synergies require cutting into the 2,000-person reduction's institutional knowledge, the result is a slower-growing, more-levered, lower-ROIC company — and a goodwill impairment that crystallizes the overpay.
2. Design IP is structurally cracked, not just cyclically soft. −8% FY25, −6% Q2-FY26, margin 38%→24%, and management is divesting the Processor IP business — i.e., conceding parts are sub-scale. If IP is secularly losing to in-house/open-source (RISC-V ecosystems, customers building their own IP), a ~20%-of-revenue, high-margin leg keeps eroding.
3. AI commoditizes high-end EDA pricing. If gen-AI design tools (including customers' own, or a new entrant's) compress the value of the premium toolchain, the switching-cost moat erodes at the margin and pricing power — the whole investment case — weakens.

Pre-mortem (18 months out, thesis broke — what happened?): Synergies slipped, a goodwill impairment hit, China re-tightened, Design IP kept declining, and the stock de-rated from ~35x to ~25x on FY2026 numbers (≈$370) even as the business grew — a multiple-compression loss, not a business collapse.

Are multiples too high? Yes, on near-term GAAP and even non-GAAP — ~35x forward leaves no margin of safety; it requires the synergy story. The honest read: the business is A+; the price is a B−.

Contrarian view (what the market refuses to see): the consensus debate is "EDA duopoly = forever-compounder," and the market is rewarding Cadence's cleaner returns with a premium. The thing being missed both ways: Synopsys' Ansys bet is the only attempt by either duopolist to escape the EDA box into the far-larger systems-simulation TAM — if Multiphysics Fusion works, SNPS isn't the discounted #2-quality EDA name, it's the first systems-design platform and deserves a premium over Cadence, not a discount. That optionality is currently priced at roughly zero.

Devil's Advocate (short-seller)

Dismantling the bull case.

• What structurally breaks the model? Two things. (1) AI-native design — if a new flow (or a hyperscaler's in-house tooling) lets designers bypass the premium toolchain, the switching-cost moat — the entire thesis — degrades. (2) The foundries or mega-customers vertically integrate — NVIDIA, Apple, and the hyperscalers have the engineering depth to build more IP/tools in-house; every block they internalize is Synopsys revenue gone.
• Revenue concentration / what shifts? No single ≥10% customer disclosed, but revenue is concentrated in (a) the leading-edge fabless/IDM cohort and (b) geography exposed to U.S.-China policy (~11% China, switchable by a BIS letter — proven in 2025). And ~29% upfront revenue means a few large customers' timing can miss a quarter badly.
• Why is the moat weaker than bulls think? Because management itself is shrinking part of it (selling Processor IP), Design IP margins fell 14 points, and RISC-V/open IP is a real long-tail threat to the IP business. The EDA core is strong; the "comprehensive portfolio" moat has a soft flank.
• Most dangerous competitor bulls underestimate? Cadence — cleaner balance sheet (no $13.5B deal debt), higher ROIC (~28% vs SNPS's normalized ~12%), and not distracted by a $35B integration. While Synopsys spends 2026–27 integrating Ansys and cutting 2,000 people, Cadence can out-invest in core EDA + its own system-design push. The market already prefers it (45x vs 35x).
• Worst capital-allocation move? Betting the balance sheet on Ansys at the top of an AI-euphoria multiple — taking on $14B+ of debt and suspending the buyback to do a deal that's dilutive to per-share earnings for 2+ years and creates $39.6B of impairment-exposed carrying value.
• What must hold for today's price (~$455, ~35x fwd)? That synergies land on schedule, Design IP stabilizes, China doesn't re-tighten, AI doesn't commoditize the toolchain, and the multiple stays ≥25x. That's a lot of "ands."
• If growth disappoints by 20–30%? FY2028 EPS lands ~$16 instead of ~$20; at a de-rated 22–24x that's ≈$350–385 — ~15–25% downside from here, before any impairment headline.
• Single scenario that permanently impairs? A multi-billion-dollar Ansys goodwill impairment + sustained Design-IP erosion + a structural AI-driven pricing reset — the trifecta that turns "A+ franchise at a fair price" into "levered, slowing roll-up at a rich price." Plausibility: low-to-moderate, but non-trivial given the leverage and the multiple.

Management Questions (ordered by information value)

1. Non-GAAP EPS declined YoY in Q2-FY26 despite +42% revenue — at what specific quarter do you expect Ansys to turn per-share accretive, and what are the two biggest swing factors (interest, share count, synergy timing)?
2. Walk us through the revenue-synergy ($400M by year 4) bridge — which specific cross-sell motions (EDA→S&A, S&A→EDA) are already booked vs. still pipeline, and what's the attach-rate evidence so far?
3. Design IP fell −8% (FY25) / −6% (Q2-FY26) with margins from 38%→24%. How much is China/macro (cyclical) vs. "roadmap and resource decisions" (self-inflicted), and what's the realistic margin and growth target post-Processor-IP divestiture?
4. What is the goodwill-impairment sensitivity on the $39.6B Ansys carrying value — at what revenue/margin shortfall does a write-down trigger, and how are you stress-testing the 35–45% royalty / 10% discount-rate assumptions?
5. With $13.5B debt and ~$1.5B OCF, what's the deleveraging path and target leverage before the buyback resumes — and how do you weigh debt paydown vs. resuming repurchases at today's price?
6. How real is the AI-commoditization threat to high-end EDA pricing — are customers' in-house/gen-AI design efforts changing your pricing conversations, and how do you defend the premium?
7. What share of IP revenue is exposed to RISC-V / open-source IP substitution over five years, and what's your strategy if customers internalize more IP?
8. The 2,000-person reduction touches a just-doubled org — how are you protecting institutional knowledge in the cut, and what's voluntary attrition running post-deal?
9. Hiring Siemens EDA's CEO (Mike Ellow) as CRO — what specific go-to-market change does that signal, and what's the early read on win-rates vs. Cadence and Siemens?
10. China is ~11% of revenue (down from ~16%). What's the planning assumption — does it recover, stay capped, or could a re-tightening take it lower — and how much design-start damage is permanent?
11. What's the through-cycle operating-margin target for the combined company, and the path back to high-30s non-GAAP — synergy-led or revenue-led?
12. How should we read backlog ($11.0B, down from $11.4B at year-end) as a forward indicator given the upfront-revenue lumpiness — what's the organic bookings trend ex-Ansys?
13. Capital allocation post-deleveraging: more M&A (and in what — more S&A? AI?), or return-of-capital? What's the M&A appetite after a deal this size?
14. Where does Multiphysics Fusion stand — design wins, revenue, and the realistic timeline to it being a differentiated, monetized flow rather than a marketing concept?
15. What's the single biggest risk you see over 3 years that the Street is underweighting — and how would we know early if the Ansys thesis is breaking?

---