Built by Grok 4.6 High

SNPS | Earnings Preview

FY2026Q3 (ended ~July 31) has not printed. Synopsys reports Wednesday, August 26, 2026, after the close; the call is 2:00 p.m. PT / 5:00 p.m. ET. Last-reported baseline is FY2026Q2 (May 27): revenue $2,276.0 million (+41.9% YoY vs $1,604.3 million), Design Automation $1,821.8 million (+62.3%), Design IP $454.2 million (−5.8%), non-GAAP operating margin 39.5% (+150 bps), non-GAAP EPS $3.35 (−8.7%). That print cleared the then-live FQ2 guide at the high end on revenue and +6.7% vs the EPS midpoint, and raised FY26 on revenue, nGAAP OM, EPS, and FCF. The raise is still in force. Do not treat this quarter as reported.

Growth trajectory — Ansys mix is decelerating; organic is not accelerating. Headline YoY is a staircase down, and the live Q3 guide extends that fade:

Four-quarter headline run: +14.0% → +37.8% → +65.5% → +41.9% → guide +40.0%. That is Ansys mix and Q1 seasonality, not demand collapse. The organic tell is EDA ex-Ansys: +17.0% → +5.8% → +12.3% → +8.3%. Q2 organic ex-Ansys was ~3–4% (Glaser). Cadence already printed Core EDA +18% and IP +40% in the same season.

Watch items into Wednesday

• Guidance / the real bar. Live Q3 is $2,410 million–$2,460 million / $3.63–$3.69. FMP $2,437.7 million / $3.67 sits on the midpoint. A “meet the mid” print is in-line vs street and a miss vs the company’s own three-quarter beat pattern (~+0.8% rev / ~+5.7% EPS vs mid). The stock-moving number is whether FY26 ($9.625–$9.705 billion / $14.72–$14.80) is held or raised again, and whether the implied Q4 residual (~$2.545B / ~$3.98) is guided without borrowing from hardware.
• Tariffs / China. Not a goods-cost story. The analog is export-control and China design starts. Q2 China $240.4 million looked +52.6% YoY only because of Ansys plus an easy compare; Glaser has not changed the year forecast. Q3 laps the FY2025Q3 restriction-miss China print of $247.3 million. No new BIS letter in the window. A “design starts inflecting” claim needs a number.
• Design IP is the live P&L debate. Three straight negative-YoY prints through Q2 (−21.4% / −6.5% / −5.8%). Management promised a second sequential step-up off $454.2 million (“no doubt”). Cadence IP +40% is the uncomfortable compare. Processor IP sold to GlobalFoundries closed June 2 (−$40M already in the year).
• Investor Day, September 30. Elliott’s Jesse Cohn has been on the board since June 1. Value-capture (agentic consumption + IP business-model) and remaining Ansys synergies are parked for that day. This print is a bridge, not the algorithm event.

Peer tape into 8/26 is uniformly raised: TSMC capex $60–64B and FY growth 40%+, AMAT Semiconductor Systems +29.7% YoY on the same July quarter, every hyperscaler that reported held or lifted AI capex. There is no demand-side excuse. “Demand is strong” is no longer differentiating after Cadence printed +24% revenue and +40% IP.

Last beat-and-raise still sold off 8.6% the next session ($525.92 → $480.64 on 2026-05-28, Daloopa company_id 176). Street is already at the midpoint this time. A clean in-line is not a free option.

Issued May 27 on the FY2026Q2 call. Never revised. Synopsys guides one quarter of revenue, GAAP / nGAAP costs, and GAAP / nGAAP EPS, plus a full-year revenue / nGAAP OM / EPS / FCF package. Classification: CONSERVATIVE on EPS and Ansys; IN-LINE (not a gift) on revenue. Street sits inside both boxes.

Verbatim (CFO Shelagh Glaser, FY2026Q2 call, May 27): “Now to targets for the third quarter. Total revenue between $2.41 billion and $2.46 billion. … non-GAAP earnings of $3.63 to $3.69 per share.” FY raise bridge: +$35 million underlying business, +$60 million Ansys channel gross-up (EPS/FCF-neutral), −$40 million Processor IP sale.

What the +40% headline hides. FQ3 FY25 included only ~two weeks of Ansys (closed 17 Jul 2025). FQ3 FY26 is a full Ansys quarter against a ~$2.96B year. Strip ~$0.7B of Ansys run-rate out of the $2,435M mid and core Synopsys is low-single-digit YoY — consistent with muted IP and Glaser’s “organic ~8%” framing at FY26 initiation. EPS growth of +8% against +40% revenue is the mix tell.

Street is not asking them to stretch. The $50M / $0.06 printed ranges are tight (2.1% / 1.6% of mid). A “normal” cost-driven EPS beat of the size seen in FQ1–FQ2 FY26 (+6–7% vs mid) would land ~$3.85–$3.90 — well above the $3.69 high. So either the Q3 EPS box is more honest than usual, or the beat, if it comes, will again be opex / synergy, not revenue.

H2 / Q4 residual (not company guidance). H1 actuals $2,408.8 million + $2,276.0 million = $4,684.8 million; nGAAP EPS $3.77 + $3.35 = $7.12. FY mid $9.665B / $14.76 minus H1 minus Q3 mid leaves Q4 ~$2,545 million / ~$3.98. Holding the FY midpoint requires that sequential step-up. A Q4 guide below ~$2.53B without a FY cut would be internally inconsistent; a Q4 guide above ~$2.56B is a stealth FY raise.

FY26 walk (the real signal).

YoY only. Internal estimates are n/a. Synopsys is an EDA / semiconductor IP / multiphysics simulation franchise — the driver stack sits above consolidated revenue (Design Automation, then the EDA vs Ansys split, then Design IP).

(a) Current quarter — FY2026Q3 (upcoming)

FMP /analyst-estimates has no FY2026Q3 quarterly row; the $2,437.7M / $3.67 pair is print-level /earnings (epsActual / revenueActual still null as of 2026-08-17). Segment-level street is unavailable (no VA / Bloomberg).

Same-quarter-last-year anchors (FQ3 FY25) — Design Automation $1,312.1 million · EDA $1,193.8 million · Ansys $0 in the SNPS product-group series · Design IP $427.6 million · recurring 75% · nGAAP OM 38.5%.

Last-reported run-rate (FY2026Q2) — the baseline this print will be read against

Ansys H1 = $885.6 million + $652.4 million = $1,538 million, leaving ~$1,422 million for H2 (~$711 million per quarter if even). That is a residual, not a Q3 guide. Q2 Ansys included $12.5 million of channel gross-up (EPS/FCF-neutral); $60 million is guided for the year.

How to read the print

• Bullish: revenue at/above the $2,460 million high with a second sequential IP step (ex-Processor-IP) and EDA holding ≥8%; nGAAP costs below $1,440 million; FY26 nudged on EPS / OM / FCF; a named Fusion paid conversion or a signed IP new-model count.
• In-line / soft: $2.435–$2.450B / $3.66–$3.69 that is all Ansys / channel-accounting / hardware timing. Clears street; misses the company’s own beat pattern and does nothing for the Cadence gap.
• Bearish: revenue below $2,410 million, IP sequential stall, Ansys ≤$650 million without a Fusion-eval update, or a FY cut after the May four-metric raise.

(b) FQ+1 — FY2026Q4 (expected; will be issued Aug 26)

Q4 FY26 has not been guided. Synopsys issues one-quarter-ahead targets on each call. Italic rows are implied residuals from the standing FY26 guide after backing out H1 actuals and the Q3 guide — arithmetic, not management guidance. FMP has no FY26Q4 quarterly analyst-estimate row.

This is the number that matters. Holding the FY26 midpoint requires Q4 revenue of ~$2.545 billion against the Q3 mid of $2.435 billion, and EPS of ~$3.98. That H2 residual (not a same-quarter YoY rate) is consistent with Glaser’s “even stronger second half” and with hardware timing slipping Q3 → Q4. If management holds FY26 and guides Q4 to ~$2.53–$2.56B, the print is a bridge, not a raise.

(c) FY+1 — FY2027 (none issued) and the standing FY2026 guide

Synopsys has issued no FY2027 revenue, EPS, or margin guidance. The next planned framework event is Investor Day, 2026-09-30. Multiphysics Fusion commercial ramp and the first $400 million of Ansys revenue synergies were explicitly dated to FY27.

Street FY27 vs FY26 consensus: revenue +10.9%, EPS +17.2%. That EPS-over-revenue gap is the street underwriting the rest of the Ansys cost-synergy stack and some IP recovery. Management has not endorsed those numbers. FMP longer-dated (context, not guided): FY2028 $12.013 billion / $20.541 (14 / 9); FY2029 $13.195 billion / $23.317 (7 / 7).

Tone into the quarter

Management walks into August 26 having raised all four FY26 metrics on May 27 and having not updated those pins since. No 8-K, no pre-announce. The 2026-07-22 earnings-date release is silent on targets. Conviction on execution, margins, and “essentialness” is the highest of the last four calls; conviction on growth acceleration is deliberately parked at Investor Day.

Conviction read: The Q3 box is a floor they want treated as the forecast, not a sandbag they want the Street to fade into another +1% / +6% residual. The $35 million “business” raise is the only organic confidence signal. Growth acceleration is not this call’s job.

Assumptions still live

• The $35 million raise is “strength across the business,” not a single-segment call. Sassine declined to split it EDA vs IP.
• Ansys FY is held despite mid-teens H1. Fusion is still in eval — no FY26 monetization. Mechanical seasonality: Ansys’s old December Q4 now lands in SNPS Q1.
• IP: Q1 was the bottom; sequential H2 is committed; FY is still muted. New royalty / COT model is “a few customers with signed agreements by year-end” and is an FY27+ story. Processor IP is gone from the remainder.
• Intel Foundry / 18A / 14A is explicitly not in the FY26 guide. “We’re not accounting or taking into account in our guidance any upside.”
• Agentic EDA is upside, not a 2026 P&L assumption. 20 customers / 25+ agents in eval. Working thesis is subscription plus consumption.
• China is held flat-to-pragmatic. Design-start environment “remains challenged.” Ansys industrial is the offset.
• Hardware is lumpy across Q3/Q4. A light FQ3 hardware quarter is pre-excused if FQ4 catches it.
• Tale of two markets. AI / HPC / hyperscaler COT design starts are strong. Analog / industrial / automotive “still fairly muted.”
• Cost synergies: ~50% of the committed Ansys run-rate by YE26. Remainder and the path to mid-40s OM are reserved for September 30.

Tone arc — adjectives up on execution, parked on acceleration

Net shift Q1 → Q2: first four-metric raise of FY26. No defensive shift on the guided P&L. The pocket of caution that is more visible, not less, is the organic-growth / Investor Day gap. FQ3 is unlikely to be the venue where they raise the long-term growth rate.

Post-guidance updates (May 27 → Aug 17)

No formal guidance revision. No pre-announce. Daloopa has no calendar-2026Q2 / 2026Q3 SNPS filings indexed beyond the May 27 package.

No Bloomberg / sell-side conference transcript with a Ghazi or Glaser guide revision was found between May 27 and Aug 17.

What Wednesday has to do

1. Clear the tight FQ3 band. Street is already at the midpoint. An in-line print is not a miss vs street; a mix miss (IP sequential stall, Ansys below the “normal” H2 run-rate, hardware slipping and a soft Q4 comment) is.
2. Show the second sequential IP uptick off $454.2 million, ex-Processor-IP. A down sequential IP quarter would be the first broken intra-year promise of FY26.
3. Keep FY26 intact — and decide whether to raise again. After a four-metric raise in May, a second raise would tell that the $35M “business” raise was conservative. A hold is consistent with parking the algorithm at Investor Day. A cut would be a serious tone break.
4. Give a clean Ansys H2 number that makes the ~$2.96B year look like seasonality, not a 2H air-pocket.
5. Preview Investor Day without pre-empting it. Watch for any quantification of remaining cost synergies, subscription-plus-consumption, IP new-model dollars, or organic growth exiting FY26. Absence of those numbers is the base case.

Ranked by information value on August 26. “Consensus” = FMP / last primary management comment. Bloomberg and Visible Alpha are not connected.

Scorecard on the 26th

Where management and the Street actually disagree (investing-principles: the only contrarian setup is a credible team repeating something bullish the Street will not underwrite):

1. IP is a value-capture story the street still models as a cyclical repair. Management has now said, for several quarters, that Q1 was the bottom, H2 is sequential, and a new COT / royalty model will be signed with “a few” hyperscalers by FY26 year-end. Cadence’s +40% IP is the number the street is now using as the comparable. SNPS IP at −5.8% YoY can be “recovering” and still look like share loss.
2. Investor Day is an algorithm change, and FY27 street (+10.9% rev / +17.2% EPS) does not price one. Sassine spent the last Q&A insisting the inflection is monetization (agents as incremental consumption; IP as COT infrastructure), not “pay 20–30% more for the same.” Cohn is now in the boardroom for the same two levers. The 26 August call will not reset FY27. The 30 September Day will, or it will confirm the street is already right.

Not a real disagreement — China, Fusion dollars, 18A. Both sides have China as pragmatic, Fusion monetization in FY27, and Intel 18A/14A at zero in the guide.

Window: 27 May 2026 (Q2 print) through 17 Aug 2026. Company and sector news excluding the Q2 print itself, the raised FY26 guide, and the 22 Jul call notice. The Elliott cooperation agreement was announced the same afternoon as the print — it is not an earnings item and is the single most material governance event of the window.

What this does / does not change

• Does not change Q3 FY26 guide math. Company already guided $2.41–$2.46 billion / $3.63–$3.69. Nothing above is a Q3 shipment beat in hand. Fusion is GA; agentic is still eval.
• Does confirm the franchise is executing the Ansys-synergy and foundry-enablement script. Fusion shipped with named logos; Samsung 2 nm and Intel 14A certified; ARC closed early; DAC put NVIDIA / AMD / Microsoft on stage.
• Does start the Elliott value-capture clock. Cohn is inside the board. Numbers are promised 09-30, not 08-26. FQ3 Q&A will be read as a teaser.
• Does not invert sentiment. Last print was a beat-and-raise that still sold off 8.6% the next session. Newsflow since is constructive on the franchise and noisy on the multiple. This is confirmatory, not a worse→better inversion.

SNPS is a Consistent beater. Versus FMP Street it is 11 / 12 on non-GAAP EPS and 10 / 12 on revenue over the last 12 printed quarters — 11 / 12 on revenue once one stale FMP estimate is isolated. Versus its own quarterly box (midpoint, +1-quarter offset) it is 11 / 12 on both lines after discarding the one apples-to-oranges Software Integrity restatement. The only live, real miss in the window is FY2025Q3: IP deals that did not close (China export-control shock + a major foundry customer + roadmap/resource misses). Every print since has been a double beat, and beat magnitude has re-widened in the last two Ansys-era quarters. Classification is Consistent beater; magnitude is improving, not deteriorating, after the one discrete miss.

Versus own guide (the more informative bar). Offset: guide issued in Q(N) applies to Q(N+1). Last four reported:

nGAAP EPS last four reported vs the prior-quarter guide midpoint: Q4'25 $2.90 vs ~$2.78; Q1'26 $3.77 vs ~$3.55; Q2'26 $3.35 vs ~$3.14; Q3'25 $3.39 vs ~$3.85 (the miss). Those EPS beats were operational (cost + mix), which is why the FY EPS midpoint moved $14.38–$14.46 → $14.72–$14.80.

Verdict: improving, not deteriorating. Three facts, not the 75% last-4 rate:

1. The pre-Ansys revenue box was a $30M-wide, 20–40 bp beat machine. That is guidance-as-forecast, not a sandbag.
2. FY2025Q3 is a named, discrete miss — not the start of a miss streak. They cut the year, recast IP as “transitional,” and have beaten every print since.
3. The last two Ansys-era prints are the widest revenue surprises in ten quarters. Part of FQ2'26 is accounting ($12.5M gross-up, EPS-neutral). The rest is Ansys deal timing (FQ1) and cost/synergy flow-through (EPS).

Bar into Wednesday

Two implications. Directionally, another beat is still the higher-probability outcome — eleven of the last twelve scored quarters cleared Street on EPS; the three prints since the IP miss have all cleared both the mid and the high end of the revenue box. The magnitude the last two prints trained the tape on (+1% rev / +6% EPS) is the live analog, not the old 0.3% metronome and not a 5–7% revenue crush. A print that merely clears $2,438 million / $3.67 by a few million and a few cents is a beat on the scorecard and in-line with the new, slightly wider Ansys-era residual. The bear case on this record is not a miss; it is an in-line-to-mid print that confirms the box is now honest.

Last-print +1-day reaction (AMC convention, Daloopa OHLCV company_id 176): −8.6% ($525.92 on 2026-05-27 → $480.64 on 2026-05-28) on roughly 2× print-day volume (4.75M vs 2.44M). OptionsLam weekly implied move into this print is ±9.26% (expires Aug 28); monthly ±12.59% (expires Sep 18). The options market is pricing a move about the size of last quarter’s realized gap.

Do not treat FY2026Q3 as printed.