Subscribe

Synopsys 2Q26 Earnings Review: When a Beat-and-Raise Stakes Everything on Margins

Last quarter, I titled my Synopsys review Proved the Margins — Now Prove the Growth. The margins had arrived two years ahead of schedule; growth still owed us an answer.

2Q26 handed back half of that answer.

The headline numbers were flawless. Revenue of $2,276mn, up 41.9% year-over-year, cleared both the top end of company guidance and the Street's consensus. Non-GAAP EPS of $3.35 came in 6.7% above consensus. Revenue, operating margin, and EPS all beat guidance — a clean sweep. Management duly raised its full-year outlook. This was a textbook beat-and-raise.

But take the raise apart, and something curious surfaces: almost none of it came from actual business growth.

What deserves an analyst's time this quarter isn't the size of the beat, but its shape — where the beat came from, what the raise rested on, and one phrase that analysts circled back to again and again on the call, until it nearly eclipsed the numbers themselves: value capture.

Where the Beauty Lies — and Why It Matters

Start with the anatomy of the beat.

How the core business delivered on profitability this quarter

Source: Synopsys 2Q26 earnings release and Financial Supplement; compiled by the author

Revenue beat by just 1.1%. But step down the income statement and the beat widens at every rung: operating income by 5.0%, net income by 6.1%, EPS by 6.7%, with operating margin a full 1.55 percentage points above consensus. This is the classic signature of a beat that grows systematically larger toward the bottom line — the company didn't win by selling more, it won by squeezing more profit out of the same revenue.

Very Synopsys. And very much the post-Ansys Synopsys.

And yet, in the same quarter, non-GAAP EPS still fell 8.7% year-over-year. A company that beats on profit but posts negative EPS growth sounds like a paradox — until you open it up. The Ansys acquisition swelled the diluted share count from 156 million a year ago to 192 million, a 23% increase (including roughly 4.8 million shares issued to NVIDIA in a private placement late last year). Layer on higher interest expense, and 12.4% growth in non-GAAP net income gets diluted all the way down into negative EPS.

In other words: the business is accelerating, while the accounting is still digesting the aftershock of a $35 billion acquisition. This isn't an illness. It's a bill coming due.

The 5.3% Hole on the GAAP Line

If negative EPS growth is merely "dilution," then the GAAP figures are where the true cost of this deal becomes something you can feel.

GAAP income flow: Ansys amortization and restructuring devoured most of operating income

Source: Synopsys 2Q26 earnings release; compiled by the author

Non-GAAP operating margin this quarter was 39.5%; GAAP operating margin, just 5.3% — a chasm of nearly 34 points between them. Follow the revenue down the Sankey diagram above: $2,276mn at the top, run through cost of revenue and operating expenses, narrowing to a trickle by the time it reaches the GAAP operating-income node — just $120.4mn left. GAAP earnings came to $0.09 per share, against $2.24 in the same quarter a year ago.

The gap traces almost entirely to two items: amortization of intangibles from the Ansys deal, which reached $403.6mn this quarter (versus just $11.7mn a year ago), and $115.9mn of restructuring charges tied to integration layoffs (zero last year). Together, those two lines swallowed more than half a billion dollars of GAAP profit.

These are non-cash or one-time items; they don't change the company's true operating profitability — a point I made at length in the 4Q25 piece, Ansys to the Rescue, and won't rehearse again here. The one thing worth holding onto: the accounting cost of this acquisition will weigh on the GAAP statements for several more quarters — but it hasn't obscured the real core margin of 39.5%.

The Wound Everyone Feared Most Is Starting to Scab Over

Now to the most important part of that answer sheet: Design IP.

If you've followed this series of reviews, you'll remember where this thread began. Back in 3Q25, I called the stall in Design IP's profitability the alarm bell on the whole "confidence curve"; management itself reached for an unusual phrase — "an unusually negative signal." By 1Q26, IP revenue was still running well below expectations, segment margin sat at just 16%, and a 31-point gulf yawned between it and the Design Automation segment.

This quarter, that line finally lifted its head.

Segment margin trajectories: Design IP has climbed for two straight quarters off its 13.8% trough

Source: Synopsys quarterly earnings releases and Financial Supplement; compiled by the author

Look at the dark-green line along the bottom of the chart. Design IP's adjusted operating margin climbed from its 13.8% trough in 4Q25 to 16.2% in 1Q26, then jumped to 24.4% this quarter — ahead of the 20.2% implied by consensus. Revenue rebounded 11.6% sequentially, from $407mn last quarter back to $454mn.

Management offered a judgment that was unusually crisp: IP bottomed in the first fiscal quarter, and the second half will improve sequentially, quarter by quarter. That picks up a thread left dangling in the 1Q26 piece — back then, management had described the IP recovery as "Q4-weighted," something requiring patience through the first half. Now, by the second fiscal quarter, it has already pulled margins back up — a touch faster than the sell-side was modeling at the time.

But restraint is warranted here. IP revenue is still down 5.8% year-over-year; in absolute terms it hasn't returned to where it stood a year ago, and management itself would only go so far as to call full-year IP growth "muted." One quarter's sequential jump is a scab, not a cure. Whether the trend holds will take another two quarters to confirm.

At a minimum, this quarter shifted the sharpest open question from "will it recover?" to "how fast?"

The Truth About the Raise: This Was a Margin Raise, Not a Growth Raise

Which brings us back to the thread from the opening: how much did business growth actually contribute to this raise?

Breaking down the full-year revenue guidance raise: the pure organic increment is just $35mn

Source: Synopsys 2Q26 earnings call; compiled by the author

The company lifted the midpoint of its full-year revenue guidance from $9,610mn to $9,665mn — a $55mn increase. That sounds like good news. But the waterfall chart splits that $55mn into three pieces: +$60mn from switching Ansys channel revenue to gross-basis recognition (offset by an equal rise in expense, so entirely EPS-neutral), −$40mn from divesting the Processor IP business, and a genuine, organic increment of just +$35mn — less than 0.4% of last quarter's revenue base.

Set against that, EPS guidance went up by $0.34, and the full-year operating-margin target rose from roughly 40.5% to 41.0%. The CFO put it plainly: the core driver of this raise is margin — the early arrival of Ansys synergies — not the top line.

This actually bears out a guess I floated in the 1Q26 piece. The company had already posted a 42.1% operating margin in a single quarter, yet full-year guidance called for only 40.5% — I suspected management was keeping some powder dry. This quarter, some of that powder was fired: they committed to completing roughly half of the contracted cost synergies by year-end (against the $400mn of annual cost savings promised at the time of the deal).

So here is a company with formidable execution, winning yet again on the battlefield it knows best: cost, integration, margin. The catch is that a margin story has a ceiling — you cannot raise on squeezed profit forever. The question the market will eventually ask hasn't vanished; it has merely been deferred: when does growth take the baton?

The Real New Storyline: From "Value Creation" to "Value Capture"

Management plainly knows this question is waiting for them. So on this quarter's call, they deliberately turned the narrative onto a new term.

Throughout the Q&A — analysts pressing on the staying power of the IP recovery, the durability of growth, when agentic AI and Multiphysics Fusion would monetize — CEO Sassine Ghazi answered nearly everyone with the same sentence: the company sits at multiple inflection points in its move from value creation to value capture.

That isn't PR boilerplate. It connects directly to the judgment I reached in March writing about Converge 2026 — that Converge had proven the "new Synopsys" was technically possible (a proof of concept), but had yet to prove it made economic sense (a proof of economics). At the time, Morgan Stanley named two unanswered questions: agentic AI had no pricing model, and the Ansys synergies had no financial figures attached to them.

This quarter, what the company offered wasn't numbers but a roadmap. The roadmap, though, was more concrete than last quarter's:

• A new business model for IP. Management said that by year-end, a handful of customers — chiefly hyperscalers — will sign agreements that depart from the traditional use-fee / NRE structure, letting the company "capture more dollars than a conventional license fee." These customers are willing to accept the change because their entire customer-owned-tooling (COT) strategy for in-house silicon already rests on the availability of Synopsys IP.
• The contract structure for agentic AI. Evolving from the subscription license an engineer holds toward "subscription plus consumption as agents use the product" — framed as upside for the EDA and simulation businesses as agents become commonplace in customer workflows.
• Revenue synergy from Multiphysics Fusion. Ghazi gave a firm date — FY2027 — tied to a $400mn semiconductor multiphysics opportunity, with an eventual target of $1bn. He kept returning to one principle: "one plus one has to be greater than two, or customers won't pay for the new technology."

Note what these three have in common: every one of them comes due in FY2027 or later.

On the durability of growth, Ghazi offered a candid line: "We don't expect customers to pay 20% or 30% more for us delivering the same thing; but when they inject agents into the workflow, inject human-machine collaboration, and need to add licenses heavily to train and run those agents, we absolutely expect the monetization model to change."

That sentence lays out the logic of value capture clearly — and its fragility just as clearly. The whole thing depends on customers actually pushing agents into their workflows at scale, and being willing to pay for it. Whether the technology works is one question (Converge has largely answered it). Whether customers will rewrite their procurement contracts for it is another (this quarter didn't).

An Activist Joins the Board

If anything makes the value-capture promise look more like a binding plan than a passage of call rhetoric, it's a move at the governance level.

The company announced a cooperation agreement with Elliott Management, appointing Jesse Cohn to the board. In the Q&A, Ghazi described the two sides as "immediately aligned, with no debate" from day one on two things: first, whether there was an opportunity to further monetize the value the company creates and the irreplaceability of its assets; and second, whether efficiency could be improved and converted into better operating margins.

Place that next to this quarter's margin discipline and the loop closes: a company already known for formidable execution now has, on its board, a shareholder famous for wringing out margins and value capture. This fiscal year, the company raised operating margin more than 300 basis points above where it ended last year — and on that point, Ghazi said, there was "no debate."

The next milestone is the Investor Day on September 30. Management has saved the full growth narrative, the value-capture framework, and the combined company's long-term financial targets all for that day. Five sell-side shops alike flag it as the next key catalyst. In other words: this quarter gave us the roadmap; the numbers wait until the end of September.

A Look at the Other Side: The Weak Spots in This Half-Finished Answer

With the bullish story told, several countervailing threads have to go on the table — otherwise this review would be incomplete.

On "how long can the margin story run": This is precisely the core reason Morgan Stanley is the lone holdout at a neutral rating. Its argument rests not on sentiment but on arithmetic — with core EDA growth still muted and R&D spending ongoing, it sees little chance of more than 200 basis points of margin leverage in FY27 on low-double-digit sales growth. The slope of margin expansion will flatten sooner or later; what it wants to see before changing its stance is further clarity on post-integration profitability and a return to double-digit growth in core EDA. This is the most direct stress test of the margin story's sustainability.

On "the quality of the IP inflection": China revenue grew 52.6% year-over-year this quarter — a flattering figure, but both management and BofA punctured it: almost all of that growth came from Ansys, which wasn't consolidated in the year-ago period, while the underlying design-start environment remains "challenged" from the cumulative effect of export controls. More to the point, Synopsys's historical IP footprint in China is larger than Cadence's, so the blow from export restrictions lands more directly on it. The durability of the IP recovery cannot get around the China obstacle.

On "growth is finally back": Don't forget the open question the 1Q26 piece left behind — strip out Ansys, and Synopsys's organic growth was only about 4.7%, against double-digit expectations for Cadence over the same span. This quarter's organic increment (+$35mn) didn't meaningfully narrow that gap. The rebound in IP margin is real, but the central proposition raised in 1Q26 — the leap from single-digit to double-digit growth — still went unanswered in 2Q26. Citi added a note of its own: Synopsys's backlog growth lags its peers.

On how much the valuation already prices in: The current share price corresponds to roughly 33x forward CY27 earnings. That the raise was conservative (a vanishingly small organic increment) admits a charitable reading — management is "leaving room," with upside to follow. But it can be read the other way too: if even in the quarter that should be most optimistic, a beat-and-raise, management would only move the business numbers this little, then the premium the market is paying for the FY2027 story itself needs that story to deliver on schedule to hold up.

(As usual, whether the valuation is high or low — and any buy-or-sell call — falls outside the scope of this review; that's the reader's own decision. What I can do is set out the facts and the logic of the debate.)

Closing Thoughts

Read the four chapters together, and 2Q26 is really the story of a company proving one thing in two steps.

Value capture has two layers. The first is squeezing more profit out of the same revenue — through synergies, cost discipline, integration execution. This layer Synopsys has now demonstrated to everyone for three quarters running, and there is now a shareholder seated on its board who will watch it keep proving the point. This is the part it can control itself.

The second layer is harder, and matters more: getting customers to pay more for the value it creates — to pay for agents, to pay for multiphysics, to pay a new price for the IP it claims is irreplaceable. This layer doesn't hinge on Synopsys's execution but on its customers' willingness. And so far, it remains a set of promises pointing toward FY2027.

A company that can lift its margin from 37% to 41% shows it knows how to run itself; whether it can translate "value creation" into "value capture" shows how much pricing power it truly holds across the entire AI supply chain. The former is its skill. The latter is its fate.

Until September 30, it remains a word with no number filled in yet.

Earnings Call Recap

Open this more visual friendly version in a new tab/点击跳转查看原文,左上角切换中文

I will now share a more detailed analysis on this earnings result behind the paywall:

Upgrade your subscribtion?

This post is for subscribers only

Become a member to get access to all content

Subscribe now