Cadence Design Systems, Inc. (CDNS) Earnings Call Transcript & Summary
April 27, 2026
What were the key takeaways from Cadence Design Systems, Inc.'s April 27, 2026 earnings call?
Cadence Design Systems, Inc. reported a strong first quarter for 2026, with revenue of $1.474 billion, representing a 19% year-over-year increase. The company raised its full-year revenue growth outlook to 17%, driven by robust demand for its AI-driven solutions and a record backlog of $8 billion. Non-GAAP EPS for the quarter was $1.96, while GAAP EPS was $1.23, reflecting solid operational performance despite some dilution from the recent Hexagon acquisition.
What topics did Cadence Design Systems, Inc. cover?
- Revenue Growth Acceleration: Cadence achieved a 19% year-over-year revenue growth in Q1 2026, driven by strong demand across its AI-driven portfolio. Management stated, "First quarter bookings were ahead of expectations, resulting in a record backlog of $8 billion."
- Increased Guidance: The company raised its 2026 revenue growth outlook to 17%, signaling confidence in its AI-driven solutions and market demand. CEO Anirudh Devgan noted, "Given the accelerating momentum of our business, we are raising our 2026 revenue growth outlook to 17%."
- Hexagon Acquisition Impact: The integration of Hexagon is expected to be dilutive to margins in 2026, with management indicating it will take time to realize synergies. CFO John Wall explained, "The strategic opportunity there is very large, but the 2026 P&L reflects the timing of integration... we expect it to be dilutive to the tune of about $0.28."
- AI-Driven Solutions Demand: Cadence's IP business saw a 22% year-over-year revenue growth, attributed to increasing demand for AI, HPC, and automotive workloads. Devgan highlighted, "Our IP business continued its strong momentum, with 22% year-over-year revenue growth driven by accelerating demand of AI, HPC and automotive workloads."
- Operating Margin Guidance: The company guided for a GAAP operating margin of 27.5% to 28.5% for 2026, slightly lower than previous guidance due to Hexagon integration costs. Wall stated, "We expect it to be dilutive... the margin impact on the $160 million is kind of in the 5% to 10% range."
What were Cadence Design Systems, Inc.'s April 27, 2026 results?
- GAAP EPS: $1.23
- Non-GAAP EPS: $1.96
- Operating Margin: 45%
- Backlog: $8 billion
- 2026 Revenue Guidance: $6.125 billion to $6.225 billion
- Operating Cash Flow: $1.875 billion to $1.975 billion
Cadence's strong Q1 results and raised guidance reflect robust demand for its AI-driven solutions and a healthy semiconductor environment. However, the integration of Hexagon poses a near-term risk to margins. Investors should monitor the execution of the Hexagon integration, the adoption of agentic AI solutions, and overall market conditions as potential catalysts or risks moving forward.
Earnings Call Speaker Segments
Richard Gu
executiveThanks, operator. I want to extend a welcome to all of you for our earnings conference call covering the first quarter of 2026. With me on the line are Anirudh Devgan, who serves as President and Chief Executive Officer, and John Wall, our Senior Vice President and Chief Financial Officer. You can find the webcast of this call, along with a copy of the prepared remarks we're delivering today, on our website, cadence.com. This discussion today includes forward-looking statements, among them our outlook regarding future business and operating results. Because of risks and uncertainties, actual results could differ materially from what is projected or implied in what we discuss today. To learn about factors that might cause actual results to differ, please see our SEC filings, which include our most recent Forms 10-K and 10-Q, the CFO commentary, and today's earnings release. All forward-looking statements made during this call rest on estimates and information available to us as of today, and we disclaim any obligation to update them. Additionally, unless otherwise specified, every financial measure discussed on this call is non-GAAP. The non-GAAP measures should not be considered in isolation from or as a substitute for GAAP results. Reconciliations of GAAP to non-GAAP measures are included in today's earnings release. <<operator_script>> Now I'll turn the call over to Anirudh.
Anirudh Devgan
executiveThanks, Richard. Good afternoon, everyone, and I appreciate you all being here today. I'm happy to share that Cadence got off to a robust beginning in 2026, with AI demand picking up speed and execution that stayed disciplined, giving us one of the strongest first quarters the company has ever seen. Our backlog reached a record $8 billion, which came in above what we had planned, showing how much confidence customers have in our AI-focused portfolio and the crucial part it plays in helping them deliver their chip and system design road maps, which keep getting more complex. With our business gaining momentum at an accelerating pace, we're lifting our 2026 revenue growth guidance to 17% and we expect to hit the Rule of 60 for the very first time. John will walk through more specifics shortly. The era of agentic AI has arrived, and Cadence is driving the transformation of semiconductor and system design. At CadenceLIVE Silicon Valley 2026, we made a major move toward completely autonomous chip design, breaking new ground with the industry's most sophisticated and broad-ranging agentic full flow platform. We rolled out AgentStack, the head agent framework for our AI Super Agent, which lets knowledge be shared throughout the design flow and stretches autonomous design from chips to 3D-IC to systems. Expanding on our groundbreaking ChipStack AI Super Agent for RTL design and verification, we rolled out 2 new breakthrough AI Super Agents, ViraStack for analog and custom design, and InnoStack for digital implementation and signoff. Taken together, these offerings reach across the whole chip design flow, forming a connected, continuously learning platform that moves the industry nearer to full automation. As the industry starts shifting toward agentic AI, the requirement for physically accurate and deeply mathematical EDA solutions grows even more essential. Our agentic AI offerings rest on decades of domain expertise, proprietary data and tightly integrated physically accurate engines, producing high fidelity results. We still think of our platform as a 3-layer cake, with accelerated compute and data forming the base layer, principled simulation and optimization as the crucial middle layer, and agentic AI as the top layer. As I've noted previously, we believe the biggest value stems from these layers being tightly coupled, strengthening one another to produce far better outcomes. As these super agents call upon our simulation, verification and implementation engines at scale, we anticipate they will significantly grow EDA consumption and push higher usage across our platforms; we announced a strategic collaboration with Google to optimize the ChipStack AI Super Agent with Gemini on Google Cloud. By pairing LLM reasoning with GCP scalable compute, this collaboration provides a cloud-native platform for next-generation chip development. In Q1, we deepened our long standard partnership with MediaTek through a wide-ranging expansion spanning our new agentic AI offerings and core EDA, 3D-IC and system analysis solutions. Physical AI is surfacing as the next big wave of intelligence as AI extends into autonomous systems, autos, drones, and robotics, and Cadence holds a unique position to lead this transition. Bringing Hexagon's D&E structural and multibody dynamics leadership tech into the fold moves our system analysis offerings into a leading spot for physical AI, which lets customers create and train entirely new kinds of AI world models by closing that crucial sim-to-real divide. During CadenceLIVE Silicon Valley, we unveiled a broader AI and robotics collaboration with NVIDIA. Marrying our agentic AI-powered products to NVIDIA's cutting-edge tech means we're speeding up engineering workstreams and lifting output in chip design, physical AI systems, and hyperscale AI factories. Let me now turn to updates on each of our businesses. Momentum stayed robust in our IP segment, posting 22% year-over-year revenue growth on the back of surging need for AI, HPC, and automotive workloads. Rising complexity in advanced node designs plus chiplet-based architectures keeps fueling heavy demand for our distinctive Star IP lineup spanning interface, memory, and foundation IP. We notched significant competitive victories and account growth at marquee customers, which speaks to how broad our portfolio is and, even more so, to how differentiated our solutions perform; we signed a record-setting agreement with a top global foundry, our biggest IP engagement with that customer so far, cementing our lead at the most advanced nodes; with favorable market forces, a sharpened strategy, and widening customer adoption, we're sitting in a very strong position to keep growing IP. Our foundational EDA segment turned in yet another solid quarter, with revenue up 18% year-over-year, fueled by broader uptake of our offerings among market-shaping customers. Our AI-powered products, and increasingly our agentic ones, are turning into a key element of customer renewals and expansions. Hardware demand surged in Q1, giving us our strongest quarter on record, powered by AI HPC clients plus rising need from automotive and robotics. Palladium Z3 remains the benchmark for emulation and fueled several competitive takeaways. Verification software momentum climbed, notably in Xcelium and Verisium SimAI, while ChipStack drew enormous customer curiosity, with many evaluations in progress. Propelled by our AI-powered Cadence Cerebrus offering, our digital platform keeps capturing share, particularly at the most advanced nodes. A worldwide semiconductor design leader sharply boosted their Innovus usage and brought in our digital sign-off products, and a marquee AI infrastructure firm broadened its use of our sign-off offerings for its leading-edge ASIC designs. In custom and analog, our AI-powered Virtuoso Studio maintained strong traction in design migration and layer automation as analog and mixed signal leaders adopt it more and more in pursuit of better productivity. Our System Design and Analysis unit posted 18% year-over-year revenue growth, since AI-powered multiphysics simulation and 3D-IC have become vital for tackling mounting system challenges. Momentum is strong in 3D-IC, where our unified multi-die, integrated design-to-analysis flow helps customers handle their escalating chiplet and advanced packaging complexity; we also enjoyed strong traction in Sigrity and Clarity as several memory and advanced IC packaging customers widened their deployments amid a shift to higher-speed interfaces. We're seeing more customers adopt this, since they want to tackle signal integrity, power integrity, and thermal issues sooner in the design flow by rolling out a complete Cadence sign-off flow. To wrap up, I'm happy with how well we executed and with the momentum we're seeing across our business. As the agentic AI era takes shape, Cadence is out in front, pushing toward far greater design productivity, and with design complexity rising, that growing demand for productivity is opening up a compelling long-term opportunity for Cadence. Given our differentiated solutions and our expanding agentic AI portfolio, I think we're in a very strong position to lead this shift and keep bringing meaningful innovation and value to our customers. At this point, I'll hand things over to John, who will share more details on the Q1 results and our updated 2026 outlook.
John Wall
executiveThank you, Anirudh, and a good afternoon to everybody. I'm happy to share that for Q1 of 2026, Cadence posted outstanding results, with momentum building and strength seen broadly throughout every one of our businesses. Strong design activity, along with the solid way we executed, produced revenue growth of 19% compared with a year earlier and an operating margin of 45% in Q1. Bookings during the first quarter came in above what we had expected, which gave us a record backlog totaling $8 billion. Let me walk through several financial highlights from Q1, beginning with the P&L. Revenue on a total basis came to $1.474 billion. On a GAAP basis, operating margin reached 29.3%. The non-GAAP operating margin figure was 44.7%. GAAP earnings per share came in at $1.23, while non-GAAP EPS was $1.96. Now moving on to the balance sheet along with cash flow. We held a cash balance of $1.407 billion, and the principal value of our outstanding debt stood at $2.925 billion. Cash flow from operations totaled $356 million. DSOs came to 67 days, and we spent $200 million buying back Cadence shares. Before I move to our updated outlook, I want to point out that it includes the customary assumption that export control regulations in place as of today stay substantially similar through the rest of the year; for our updated 2026 outlook, we anticipate revenue falling between $6.125 billion and $6.225 billion. GAAP operating margin should land between 27.5% and 28.5%. Non-GAAP operating margin should be between 43.5% and 44.5%. GAAP EPS should come in between $4.39 and $4.49. Non-GAAP EPS should be between $7.85 and $7.95. Operating cash flow should fall between $1.875 billion and $1.975 billion, and roughly 50% of our free cash flow is expected to go toward repurchasing Cadence shares during 2026. With all of that in mind, for Q2 we anticipate revenue between $1.555 billion and $1.595 billion. GAAP operating margin should be between 28.5% and 29.5%. Non-GAAP operating margin should land between 44.5% and 45.5%. GAAP EPS should come in between $1.07 and $1.13, and non-GAAP EPS should be between $2.02 and $2.08; and as we normally do, a CFO commentary document has been posted on our Investor Relations website, and it covers our outlook for additional items along with further analysis and GAAP to non-GAAP reconciliations. To wrap up, Cadence has gotten off to a solid start this year; we are lifting our 2026 revenue outlook to roughly 17% growth compared with a year earlier. As I always do, I want to express thanks to our customers, our partners and our employees for their ongoing support. And now, operator, let's begin taking questions.
Charles Shi
analystAnirudh, my question is rather big-picture, yet it's likely front and center for many investors. We've clearly come to understand that agentic AI probably benefits EDA and helps license consumption, among other things; however, concerns persist regarding how well AI can actually produce software, plus there are doubts about whether AI could genuinely build superior EDA base tools — by base tools, I mean things such as Virtuoso, Innovus, that sort of tool; and, of course, plenty of EDA start-ups keep appearing all the time. So my question becomes: does AI's capacity to write software make you uneasy about how defensible the EDA base tool business is? Again, we recognize that agentic AI helps consumption within the base tool business, but I'd like your perspective.
Anirudh Devgan
executiveYes. Charles, thank you for asking. So, look, this breaks into several pieces. Naturally, I'm extremely enthusiastic about agentic AI being applied to chip design and EDA. Your question targets the base tool more narrowly, and whether AI could author those base tools. So to begin with, I hold strong confidence in where we stand with the base tool and in our competitive edge, all right? And let me remind everybody, we currently have roughly 15,000 people at Cadence, with about 10,000 working in R&D; over — half of those hold advanced degrees. I believe over 1,000 among them earned PhDs from leading universities. So regardless, we'll apply AI inside the company, just as we're doing, to improve how we write our software. Yet I have no worry that another party could develop base tools any better. So — besides, our rival in the base tool is already best-in-class, and nothing suggests that will shift down the road, all right? Now, what truly excites me is what we introduced at CadenceLIVE — the agentic piece and how agentic tools interact with base tools, that AI orchestration paired with physically accurate base tools. That opens fresh opportunities for us, including TAM expansion. Because agentic AI lets us offer products in areas where we previously had none, such as RTL generation and verification plan generation; and those offerings, I believe, will be bought more through a subscription plus consumption model. So this represents a completely new category for Cadence. And then, as you noted, agentic AI will pull more usage of our base tools. So I feel quite positive about this 3-layer framework we've described and confident about what's ahead.
Jason Celino
analystPerhaps this is only a point of clarification. I saw that the guidance for operating margin has slipped somewhat. What's driving that primarily, John? I realize the Hexagon acquisition is being incorporated, yet in absolute terms, the OpEx contribution from it is fairly minor. Could you walk us through how the margin guidance works?
John Wall
executiveAbsolutely, Jason, and thanks for asking. What you're observing largely stems from bringing the Hexagon design and engineering business into the present forecast. The strategic upside in that area is enormous, yet the 2026 P&L captures the integration schedule we disclosed in the press release at the time the deal closed, namely that we anticipate $160 million in revenue during this year. That figure now sits in the guide. We anticipate dilution of roughly $0.28. Regarding margin, the effect from that $160 million falls somewhere between 5% and 10%. The dilution arises because 30% of the acquisition price was paid in shares while 70% was paid in cash, meaning the interest component on the — or rather the foregone interest income from the cash — accounts for much of the dilution in the near term; we anticipate it turning accretive in 2027. The — yes, so financially, I'd frame it as 2026 being a year of integration. Contained in the guide are the acquired cost base, the financing impact, the acquisition-related integration costs and near-term dilution of that kind. That explains why revenue goes up while EPS and operating margin come in below the February guide. So yes, $160 million. And in Q1, I believe the EPS effect was a bit smaller, since we had roughly $20 million of revenue from Hexagon in Q1; that meant only around $0.01 of dilution. So absent Hexagon, EPS would have come in about $0.01 higher.
Vivek Arya
analystAnirudh, over the past 12 months, the only thing we've heard repeatedly is various reports regarding chip shortages, rising chip prices, and the pricing power so many of your clients now hold. So my question is this: what impact do those shortages, plus the reality that your customers wield greater pricing power, have on the way they work with Cadence? Does it limit the number of chip starts? Does it push them toward products with higher ASPs? Simply put, what effect do semiconductor shortages have on your engagement trajectory and your growth? What's different now? And what are you seeing in how your customers behave?
Anirudh Devgan
executiveYeah. Vivek, thanks for asking; let me cover a few points. First, the overall environment is quite healthy, for system companies and semi companies alike. That's always a positive. As you're aware, some hyperscalers and AI semi companies were already performing well last year, but at this point the memory companies are performing well, and even analog companies are performing well. Naturally, we want our customers to succeed, and that fosters a favorable climate for engagement, particularly with the new solutions we offer. So that represents a fairly notable improvement across the last 3 to 6 months. That's point one. Second, the shortages don't directly — what I mean is, the customer remains committed to long-term R&D road maps. And at times, they might, as I've witnessed in a few instances, the customers, for instance, might use multiple foundries or nodes to ensure there's capacity at a given node or foundry. That would directly generate more design activity for us; broadly, when a customer is healthy because revenue is climbing, they'll not only do more on current designs to speed them up, but might also kick off new designs. That's the second thing, I'd say. And the third thing, which is more thrilling for us, is that since we have these agentic solutions, they can yield greater productivity for our customers, and we can provide more value ourselves. The more value we provide, the more chance we have to capture a portion of that value. And customers are quite receptive to those conversations given there's more automation. So we are in fact — as I noted, there's substantial engagement with ChipStack and also the new AgentStack, InnoStack, ViraStack. There's zero pushback whatsoever. If we can deliver productivity, the customer is more than ready to engage. So that, Vivek, I'd say, is — at minimum the 3 broad areas I observe in today's environment.
Jim Schneider
analystCould you perhaps break down what you said about agentic solutions, particularly regarding your suggestion that they'd lead to higher usage of base tools? Maybe you can discuss the pricing for those tools a bit? How exactly are agentic solutions priced? And then overall, could you give us a sense of how you might ultimately capture greater revenue value on a net basis when comparing agentic and traditional licenses?
Anirudh Devgan
executiveYes. Thank you for asking. So the opportunity here is considerable, in my view, and particularly with agentic, because what — and this took place across the past, shall we say, 6 to 12 months, from my perspective, and even more so within 6 months, is — not only have agentic tools progressed, but agentic tools can — we're able to embed skills within them, enabling far greater automation. As an example, we introduced ViraStack, an analog automation offering. Analog has long been a challenge to automate, correct? Extremely hard to automate. Yet today, with these agentic flows and skills, that automation is achievable. So what does that imply regarding pricing or the way these things get consumed? So first, as I mentioned, this type of automation simply wasn't achievable previously. So all of this work was previously performed by the customers on their own, correct? And even in that scenario, I spoke with one major customer. For instance, they told me that for analog or even digital, each new design demands 2x more engineers. And regardless, it's not — it's akin to unattainable head count growth, since they cannot bring in 2x more engineers each time. Let me walk through how we plan to generate revenue here, and the initial signals we're seeing are positive. The first piece is that we're going to introduce capabilities that didn't exist in our lineup before — things that customers used to handle manually, like doing analog design work or writing RTL. These will be sold through a subscription plus consumption structure, very similar to what you see with other top AI offerings. That represents a completely new segment for Cadence. It should also reduce the growth trajectory of customer head count, although the head count trajectory people projected was never realistic anyway. As you know, this pattern has played out throughout the history of automation in EDA. It's something we constantly have to do. However, on this occasion, we can accomplish that using an agentic AI workflow. So after the agent is running — for instance, when someone is designing a chip, which is pretty standard, right? Let's say, to keep things simple, that chip contains 100 blocks, and you have 100 engineers, with each engineer handling one block. That means a single engineer — he or she — will conduct one or two experiments to determine which settings or which design works better. Yet when those blocks are run by the agent, it might test 10 or 100 different variations of them. On top of that, AI investigates much more extensively than a human ever could. Thus an agent doesn't just boost output—it inherently operates a greater number of the base tools. That explains why, in this sort of setting, you'll see our consumption of base tools climbing quite sharply. So those are the 2 mechanisms. And within such settings, a conventional business model—but for the base tools—yet demand for those will grow. Then there's the fresh model, one centered on automating through agentic flows what used to be done by hand.
John Wall
executiveRight. Jim, one more thing I'd mention is that the way Q1 looked to us — well, the broader picture on pricing has gotten better. Clearly, our approach to pricing is still value-based. The value we deliver to customers is enormous, particularly through agentic flow, and as those customers do well in that space, we should share in the upside. On top of that, whenever you see customers move spending away from labor and toward automation, that kind of change tends to stick, and it tends to pick up speed as time goes on.
Siti Panigrahi
analystAlright, let me turn to the IP segment. Anirudh, you mentioned that IP is now in its third consecutive year of robust expansion. Can you share what happened during Q1? Also, do HBM, LPDDR6, and those items remain the primary growth engines? Or are newer foundries such as Rapidus and Intel Foundry now adding significantly to IP demand? And John, one clarification regarding your EPS outlook: you referenced $0.28 of dilution, yet you only reduced it by $0.20. I just want to confirm that on an organic basis, you increased EPS by $0.08.
John Wall
executiveI'll address that final point first. Yes, indeed we did. We raised it by $0.08.
Anirudh Devgan
executiveAnd Siti, this is a really strong way to begin the year, alright? And it's not only IP — it's everywhere. When I went through it with our team, I'd say this ranks among the biggest increases we've ever posted in Q1. We handed you guidance only in February. So here we are 2 months later, and I'd say this is among the biggest increases we've ever posted. Every one of the businesses is performing well, and IP in particular has gotten off to a terrific start, alright? And based on what I believe I'm seeing, it should keep doing well down the road; and in my view there are at least 3 major factors behind IP's growth; and as I mentioned, we're in year 3 now. So we're reluctant to speak about things prematurely. But following 3 years of robust growth, I'd call that a solid trend. So factor one is that the quality and performance of our IP has simply improved. There's a new team, and the performance is just — since these are standard-based IPs, correct, such as DDR or PCIe. So the specification doesn't change. But when our power numbers beat the competitor's or beat what the customer itself can achieve, they'll purchase our IP; so what encourages me most is that thanks to how strong our R&D team is, our PPA comes out ahead, and that's producing plenty of competitive wins at rather significant major customers. And I pointed to several of those during CadenceLIVE. So these are, you know, genuinely big marquee names. So that gives me confidence the team is executing well. So that's factor one. Factor two, our portfolio keeps growing, as we've pointed out with, you know, HBM. And part of that comes organically, part through acquisition. Take HBM — we bought it from Rambus and then made it better. But UCIe, a vital chip-to-chip technology, came entirely from organic development, alright? So the second factor is that our portfolio keeps growing. The third factor is these new foundries, alright? And watching that is quite encouraging. Naturally, we intend to be best-in-class at TSMC, the leading foundry. But today there are at least 3 additional major foundries — Samsung, Intel and Rapidus, as you're aware, at advanced nodes, plus Global and others at mainstream nodes. So the sheer volume of design activity involving AI and the growing count of foundries means more IP is needed. So that's why I'm genuinely happy to note today, as in the prepared remarks, that we landed a fairly significant IP deal — among the largest ones — at a leading global foundry, alright? And to be clear, that one is not Intel, alright? We're actually pleased with how our discussions are going with Intel, with Lip-Bu and team on 18A and particularly on 14A. From my perspective, Intel understands that a bigger commitment to 14A is necessary, and that this time around they must be better prepared, since having IP and EDA solutions ready when 14A arrives is essential when they sit down with customers. So things are moving along quite nicely with Intel, and before long we'll be able to share additional details about the work we're doing together. That said, I'm equally encouraged by what's happening with the other global foundry. Taking a step back, IP growth looks strong, and I'm quite happy with the position we're in; EDA has always been an area of real strength for us. In past years, though, our performance in IP wasn't as good. As things stand now, I believe we're sitting in a very strong position, and the same holds for SDA.
Joe Quatrochi
analystPerhaps I could just build on what was said regarding EDA a bit earlier. I suppose if you step back and consider what portion of R&D expense goes to EDA, and obviously, across quite a few companies we're witnessing R&D expense speed up. In what way should we view EDA's role in that, or its percentage of it? And considering the value you might be delivering through AI, where might that head? Since we're likewise seeing memory costs climb, and similar items, which also have to move through that R&D line.
Anirudh Devgan
executiveYeah, that's a fair question. And we need to watch it carefully, correct? You know how we are — we would prefer to, say, report actuals rather than forecast what's coming, since demonstrating beats speculating — yet as you're aware, in the past we've stated EDA was 7% of R&D and today it sits closer to 11% of R&D; thus it has risen and R&D spend on its own is going to rise considerably. Still, I believe there's genuine upside, particularly given agentic AI, for that 11% to increase. And every major CEO I've spoken with, they aren't merely open to it, they actually want to see it occur. They're eager to put money into greater automation and compute to bring it about; so I'm quite confident at present, I think it will rise; how far it rises, we'll find out, right? Yet I think there's a significant chance for automation to become a larger share of R&D, plus R&D itself to increase.
Ruben Roy
analystJohn, I'd like to return to what we were saying about operating margin. It's really encouraging to see that your team aims for Rule of 60 by the close of this year. When I think about it, though, that depends on revenue speeding up. Clearly, there are Hexagon integration costs involved. Still, once you're past $6 billion in revenue, how do you view the operating model in relation to operating margin? Compared with $5.3 billion, does that operating model look very different? Should we be thinking of operating margins in that 43% to 45% range? The reason I'm asking is that you're clearly putting money into agentic AI and other new product areas. I'd just like a sense of how you're viewing the operating margin structure longer term at this revenue run rate as Hexagon gets integrated.
John Wall
executiveYes. Of course, Ruben. I appreciate the question. Yes, when we examine our organic incremental margin, these days it's nearer to 60% than 50%. When it comes to integrating these acquisitions, our usual timeline runs somewhere between 12 and 18 months before profitability climbs to a level roughly in line with what we're targeting for Cadence. The pattern I'd point to as a comparison is what happened with BETA. During '24 and '25, the operating margin picture looked like this: the BETA deal weighed on margins in '24, but by '25 they had risen sharply once synergies kicked in and we captured the gains from running that business more profitably. For '26 and '27, I'd anticipate Hexagon following a comparable trajectory. There's a bit of a short-term drag we're dealing with, though we still see lots of room to boost how profitable that area can be. On top of that, given the gains we're witnessing when it comes to customers engaging more and more with agentic AI, I believe we can push that incremental operating margin even further down the road.
Harlan Sur
analystIf I take the guidance you provided for 2Q and look at what that suggests for the back half — the typical quarterly revenue pace there ends up slightly below where 2Q sits. Is it possible some unevenness is hitting the Hexagon business later in the year, maybe because customers are moving toward multiyear license deals? Or is the explanation unevenness in the core business, perhaps a shipment pattern for hardware or IP that leans more heavily toward the first half?
John Wall
executiveYeah. Appreciate that, Harlan. Yeah, absolutely. Look, the first half has been really strong; and as for the second half, I'd call it a case of sensible caution. Regarding Hexagon's D&E segment, you're right. Their business tends to skew toward the first half of the year. Looking back at Hexagon's revenue last year, Q3 and Q4 struck me as the two weakest quarters they had; a lot of their contracts tend to be dated early in the year. Still, on the whole, the second half — I mean, it's not — Hexagon's results don't swing the first-half/second-half split all that much. Honestly, as Anirudh mentioned, our Q1 guidance marks among the biggest increases we've ever made at this point in the year. Typically, our preference is to hold off on lifting the guide until two quarters are behind us. But with how strong Q1 bookings were, and the broad-based momentum we witnessed, we felt we had no choice but to raise it. That said, we decided to hold off until July before refreshing the second half.
Lee Simpson
analystI'd like to inquire about physical AI. You've completed a few quite strong acquisitions, I mean. Collaborations have now been announced, particularly one involving NVIDIA. So I'm simply attempting to gauge the momentum at this point, along with what genuinely remains the early stage of this breakout. And specifically, I'm thinking about how your emulation tools get adopted — above all when it comes to narrowing the sim-to-real gap for robotics and perhaps self-driving chips too — and whether that will actually produce disproportionate value capture for Cadence. Also, at what point does this show up in the results?
Anirudh Devgan
executiveYes. Appreciate that question, Lee. So, as I've mentioned endlessly by now, we view this as a cake with 3 layers, correct? And that cake has several slices, with data center AI, or infrastructure AI, being slice number one. Physical AI is the second large slice; naturally, I've been saying this across 5 years, since my belief is physical AI will exceed data center AI by a wide margin, given that we're discussing something like trillions of dollars in product opportunity. It will additionally reaffirm the data center layer — the data center slice — because to deploy an AI model in the car, for instance, you must train it on the data center regardless. Thus, my view is it will even benefit the data center slice. As for our own piece, indeed, we did this acquisition that we're tremendously enthusiastic about, and we now have this training flow for world models plus a more thorough simulation environment. What's thrilling regarding Hexagon is that, by combining our earlier technologies such as Millennium, Cascade, and BETA, we at last possess a full solution for physical AI within the middle layer — more or less the principled simulation and optimization layer. That can then be leveraged to create these world models, which will differ at the top layer. Yet another point I'd like to stress, beyond SD&A and the AI side, is that physical AI will itself propel a great deal of silicon design. Hence it benefits EDA and IP as well. And this is — you're beginning to observe it. Naturally, firms such as Tesla noting they lack sufficient silicon owing to physical AI. So physical AI isn't merely beneficial for SDA and AI; it's genuinely beneficial for silicon too. Plus it sits squarely in Cadence's sweet spot, since Cadence has always offered both analog and digital solutions, which explains why we've always done well with every major semiconductor company in automotive and now with all the system and OEM companies in automotive; and as that extends to drone and robots, it will likewise supercharge the silicon business. That's the reason I've consistently been enthusiastic about physical AI — not only for the AI and SDA portion, but for EDA and IP as well.
Gianmarco Conti
analystTurning to hardware — clearly another strong quarter. Looking toward the upcoming refresh for Palladium and Protium, though, you've typically kept to about a two-year rhythm. Does that mean Z4 and X4 should be landing within the coming 12 to 18 months? Or, since customers adopted that initial generation so recently, has the threshold for upgrading risen? And related to that: does any of your in-house agentic AI tooling materially cut internal hardware development schedules to the same extent customers report getting that identical 10x productivity on RTL?
Anirudh Devgan
executiveYes, absolutely. Great question. To start, as I mentioned, the bulk of our employees work in engineering—that covers both R&D and customer support. That's why we aim to have our own solutions in use across both hardware divisions, which make up a sizable design organization; within Palladium and Protium we handle software, hardware, and the complete system design. And let me also remind you about our IP team. Things are going really well—our IP group and EDA group collaborate closely, since when it comes to IP, demand is extremely high. Rather than just adding more staff — and we're constantly careful about how much we grow our workforce — we're bringing people on across every function, IP included, though agentic AI lets those teams accomplish far more. Turning to hardware: honestly, I'm quite happy. It's been an outstanding way to begin the year. Our standing against competitors is tremendous, since as you're aware, nobody else designs their own silicon; with Palladium we hold a lead of no less than a decade. Beyond that, Protium is now also in play, where the FPGA approach comes into use. Let me make sure that's clear — building next-generation systems is something we always do. Since the entire stack sits under our control, from how the system is architected to how the silicon is architected, it's worth keeping in mind that our pace will be considerably quicker than whatever cadence FPGA development follows. Next-generation FPGA designs will of course also be produced by the FPGA vendors. Yet the fact that the chip is ours and the design work is done in-house means our speed far outpaces FPGA. The upshot, then, is that palladium's advantage over FPGA-based systems is only going to grow wider with each new product we bring out, all right? That said, I won't go into specifics about the timing for those new products, since what we have on the market right now is performing incredibly well. Naturally, Z4 and Z5 are in the works. Still, keep in mind that today's Z3 platform can already handle designs with 1 trillion transistors, all right? At present, the largest systems anywhere sit at 100 billion to 200 billion transistors. That leaves us considerable room. By 2030, the industry is expected to hit 1 trillion transistors. What I can promise you is that a Z4 system will arrive ahead of 2030. So there's no question about whether Z3 meets the capacity and requirements. Working alongside our customers is something we're happy to do. And we also want to give investors and customers assurance: when it comes to hardware systems, our road map is very strong.
Jay Vleeschhouwer
analystAnirudh, since you've closed the Hexagon MSC deal, it looks like your company now ranks as the fourth-biggest player in non-EDA simulation — industrial simulation with multiphysics, as we might call it. Your slice is maybe one-tenth of that overall market, once again leaving EDA simulation aside. So here's my question: having pulled together all of these components and put over $5 billion in across the past 5 or 6 years, could you go into some depth about which main technical and/or go-to-market goals or deliverables you'll pursue over roughly the coming year? Synopsys described its plans with Ansys, so maybe you could do something similar regarding your own components. It also looks like you're turning somewhat more vertically integrated on the go-to-market side by buying a channel partner you've worked with for a long time; so perhaps address a few of those key elements for growing revenue and share in that business.
Anirudh Devgan
executiveYes, Jay, that's quite a bit there, right? Quite a bit there; so allow me to try to break some of it down. I'm sure we can discuss further if I don't reach every piece there. Well, to start, we feel good about how our SDA business now looks in terms of scope following this acquisition. So I mean, these are rough numbers. So I think it'll come to around $1 billion of run rate. And what excites me more is that it centers on the 2 significant areas of SDA. I've been a supporter of SDA for some time now, I don't know, perhaps 8 years now. But not every bit of SDA is equal, okay? For me, we want to handle the portion of SDA that either is expanding nicely or sits tightly connected to EDA. So the portion of SDA sitting tightly connected to EDA is, naturally, 3D-IC, okay? So we hold an inevitable position in 3D-IC with Allegro being the top packaging platform, and then we rounded that out with Clarity and Sigrity and Celsius, so all the thermal electromagnetics. So — and Integrity. So I'm quite pleased with the 3D-IC portion, which is like the nearest to chip design, the portion of SDA nearest to chip design and the portion expanding the most thanks to AI. Now the other portion now with Hexagon is all this physical AI and for designing cars and robots. So that, with this acquisition, is complete, and we can achieve a much stronger integration of that portion of SDA; and multiple things are happening there, okay? There are at least 2, 3 key things. So the first thing is we'll integrate the entire solution. I know you asked me this previously, when will you integrate? So I think now that we hold all the components of critical mass, this is the right moment to integrate since we have CFD now, we have structural, we have multibody dynamics, we have pre and post, okay? We have a significant amount of work ahead to create a complete flow solution and bring those pieces together. I gave a hint about this during CadenceLIVE. Beyond that, the method for bringing these solutions together — which holds for EDA but will also hold in this space — is agentic flow. So you're going to see us deliver an agentic flow for system design. That portion of the market hasn't experienced much of this — automation there is actually worse than in chip design, which already had considerable automation. Yet agentic flow will arrive, and it will bring all of these elements together more effectively. The next item on our list: there's substantial headroom to enhance these solvers, particularly given our track record of upgrading the underlying solvers, introducing GPU acceleration, and adding physics AI or AI surrogate models. For instance, there's potential for at least a 10x gain in these new solvers' performance. That's the second item we'll pursue on the R&D side. And the third item — something else about Hexagon that pleases me — is that we obtained what amounts to a solid go-to-market team. That's an area where we hadn't been as robust, since we were — most of the rest came largely through organic means. We did relocate some of our personnel into go-to-market. But through the Hexagon D&E business, we gain a considerably stronger go-to-market team. And as we said, we additionally bought some resellers to bolster go-to-market, all right? Right now, I have great confidence in our R&D solution, and agentic solutions will make it better. Speeding up the solvers will make it better, yet we also must put money into go-to-market, and Hexagon provides us a solid starting point; so you'll see that as well. Those, then, are the 3 areas of focus for improving SDA.
Kelsey Chia
analystAnirudh, you said AgentStack assists with closing the talent gap for chip designers. From what you're describing, it seems adoption of AgentStack is only speeding up going forward. Given what you've discussed, would you say that's true? Or do some customers instead choose to create or rely on their own agentic stack rather than take up Cadence's? And can Cadence charge for AgentStack, or for the additional base licenses, as an incremental add-on inside a fee or contract that already exists? Or does that monetization depend on renewals?
Anirudh Devgan
executiveYes. Thank you. Those are plenty of solid questions, alright? So let me be sure I — I'll begin, and John is welcome to chime in. Now first off, I think to put it plainly, customers are always going to author their own agents too, if I grasp the first portion of what you asked. Even back in our pre-agentic flow, we provided lots of flexibility to our customers. We offered a Tcl/Tk or a Python interface into our tools, and they would invariably maintain their own flows. I mean, that's simply natural for large customers. I mean, these are the who's who of tech companies. So they always seek some differentiation between one flow and another. So — and that's going to occur within the agent world as well. So I think the majority of our customers are writing certain agents of their own. But the crucial point is that the essential agents, okay, like, those big super agents we discussed, like RTL design and verification, analog design and physical design, these resemble super categories. And beyond that, the worth of the agentic flow isn't only in the agent itself, it's always the pairing of the agent with the base tools, since we run the agent at a far lower level of interaction, these API calls, which customers cannot do. So what has taken place, for instance, as we demonstrated InnoStack or ViraStack and ChipStack to our customers, they come to realize, oh, there's no purpose in writing these kinds of agents, okay? So they would prefer to use the super agents we offer, because we aren't just strong in agentic flow, we're strong in the coupling to the base tools. Now they'll still write certain agents to customize things that are particular to them, and we obviously welcome that. And then AgentStack lets the environment — for the customer to write its own agent, but also for the customer to write its own skills. We want customers to write their own skills in InnoStack, which might be particular to a part of design; so this has always been our strategy to be more open to customers kind of customizing their own environment, okay? And I think the second question concerns renewals versus new — I mean, it's a mix of that always, John, maybe you'd like to comment on that.
John Wall
executiveAbsolutely. Thank you, Anirudh, and thank you, Kelsey. For our customers, the subscription approach continues to serve as the foundational structure. Incremental add-on revenue then arrives via agentic workflow offerings that are somewhat usage- or consumption-driven for capacity, as well as through our token and card models. The distinguishing feature of agentic AI is that it does not substitute for the core EDA engines. Instead, it invokes them with greater frequency and does so intelligently. As a result, there are really two avenues for monetization: you have the newly introduced agentic workflow products, and you also have greater utilization of the foundational tools beneath them, driven by additional exploration, additional verification, additional optimization, and additional compute. That being said, we are clearly exercising discipline regarding our 2026 outlook. Within the guidance, we are not presuming an abrupt step function in AI monetization, yet we maintain the view that agentic AI widens the opportunity for Cadence's growth over the long term.
Andrew DeGasperi
analystI've got a question with two parts. First, on the marquee item — in your prepared remarks, I believe you noted that a marquee AI infrastructure company broadened its usage of sign-off solutions. I'd like some clarification: was that company a cloud provider? Second, at CadenceLIVE, you talked about physical AI, saying the adoption time line is roughly 2 years; however, you also pointed out that hardware has been adopted by automotive and robotics companies. So I'm curious: does that suggest the physical AI time line has moved up? Or is it simply the natural way these new markets will take up EDA? And if that's the case, when does that software start to see the benefit?
Anirudh Devgan
executiveYes. Regarding physical AI, and agentic AI more broadly, well, yes — for quite a while I've talked about 2 contract cycles, and by and large that holds. That said, given this new category of TAM expansion, which relates more to labor productivity together with the base tools, there's a possibility, I think, that agentic AI monetization arrives in fewer than 2 contract cycles, all right? I'd rather not forecast too much; and as John mentioned, we aren't including it in our guide. Still, I believe there are definitely more opportunities, driven by all the shortages, all the build-outs, and physical AI, so we are — and as with the earlier question, we can always include it in the renewal, but we always have the ability to do add-ons, which we've already observed, okay? That's what I'd like to say. As for sign-off, we're very pleased. Innovus has led as the solution for implementation, particularly at TSMC and now more and more with Samsung, Intel and Rapidus. But sign-off is very — it's gaining strength at TSMC and other customers. And we're engaged with all the top AI players; and the one we specifically referenced, I think, is a major AI infrastructure/ASIC company, and we're happy to see that adoption.
Gary Mobley
analystJohn, unless I've got this wrong, 2026 looks like it will be a light year for renewals. What I mean is that longtime customers who are due to renew during this year, somewhat similar to how 2022 played out. So did the robust bookings we saw in the first quarter reflect some add-on deals, with salespeople pushing to hit their quotas? And should we expect that sort of pattern to continue for the rest of the year?
John Wall
executiveAppreciate that, Gary. Yeah, 2026 is somewhat lighter than 2025 when you look at actual renewals measured on an annual value basis; however, it's often the case that those turn out to be among our best growth years, thanks to all of that add-on activity. Yeah, we felt really, really good about how strong Q1 bookings were, and it was broad-based across every line of business; so yes — so Gary, that's a good sign for the year. That said, it's only a single quarter. You know how we prefer to see a couple of quarters go by before raising the guide in the second half. And even though Q1 has come in strong over the past few years, and this particular one was very, very strong. So at the close of Q1, we felt we had to raise the guide.
Clarke Jeffries
analystI'd like to inquire about the biggest intellectual property deal announced today involving a worldwide foundry. What contributed most to making that the largest arrangement you've ever witnessed — was it truly expanding that contract into further nodes, having a wider range of content, or bringing in agent-ready AI flows?
Anirudh Devgan
executiveCorrect, this is an IP-specific contract. That particular one centers on intellectual property. Two factors pushed it forward: it involves a fresh node — a new advanced node, specifically 2-nanometer — plus additional IP content, since our portfolio is now considerably broader.
Josh Tilton
analystThis might come as two separate parts, and they're only loosely connected, so my apologies. Is there anything you'd highlight as the reason China had such a strong quarter? Then here's the second piece. Could you walk us through what's behind the really impressive organic increase you're guiding to for the full year versus the organic outperformance you posted this quarter? You did bring up the record backlog, but can you offer anything a layer deeper, particularly given that — it sounds as though you're signaling that despite a fairly meaningful raise, there's still some conservatism baked into the second-half guidance. Any clarity on that would be much appreciated.
John Wall
executiveSure, Josh. Thanks for asking. I'll handle this one. So Josh, yes, on China, it accounted for 13% of Q1 revenue, and that was pretty much in line with our expectations. Yes, we continue to anticipate China at roughly 13% for the year. It can be uneven on a quarter-to-quarter basis, I'd say. So I think the year-over-year comparisons likely appear flattering since Q1 in 2025 wasn't particularly strong in China; so with it at 13% of revenue in Q1, the growth rate probably looks robust; but it's simply — it's a hugely important region for us that — yes, and we're extremely pleased with that 13%. Regarding the guide, yes, I mean, we're — look, Q1 was a really strong beginning to the year. We beat every one of our metrics. And I suppose once we strip out Hexagon, the $160 million from Hexagon and the $0.28, we're essentially lifting the year by $65 million at the midpoint on revenue and around $0.08 on EPS. Also when it comes to cash flow, that operating cash, the manner in which we funded Hexagon, the reported guide contains roughly $180 million of preclose Hexagon tax liabilities that are economically part of the acquisition consideration, yet get classified within operating cash flow. I think it's just the geography and the accounting that require us to run it through operating cash. If you adjust our operating cash guide for that underlying — for that preclose Hexagon tax liability we're paying, the operating cash flow outlook comes to approximately $2.1 billion, which would be roughly $100 million higher than our original guide; so there's considerable strength we witnessed across the businesses; so the $65 million is the amount we raised revenue by, but we're seeing $100 million additional in cash where there's possible strength in the second half, but we felt it was too soon to raise the second half at this point.
Blair Abernethy
analystI'd like to focus on the Millennium platform. Anirudh, how's the uptake progressing? And broadly speaking, what's the state of health across some of your non-semi verticals — automotive, aerospace, industrial equipment, that sort of thing? Any thoughts you could share would be wonderful.
Anirudh Devgan
executiveYes, of course. Millennium is performing very well. I'm not sure whether you noticed, but Jensen showed up at CadenceLIVE and signed a Millennium box, which was a nice touch. So the partnership we have with NVIDIA makes us happy. There are 2 ways to think about it — 2 broad categories of application; we've been working on this CFD or SDA application for some time, and progress is good, particularly in auto and in drones as well, right? So there's plenty of — as you know, the Cascade acquisition we completed excels at very high accuracy CFD, which extends to aerospace and defense too; so autos are in there, but A&D is also a Millennium uptick, and we have multiple customers. A few we're able to discuss, a few we aren't, correct? That covers the traditional Millennium side. On the other piece, this year, as I mentioned at CadenceLIVE, we now have every kind of EDA application running on Millennium, which is really exciting. And the standout among EDA applications on Millennium is 3D-IC sign-off, because at present, the largest problem is how complex these 3D-IC systems are — not merely designing them, which Integrity and Innovus let us do, but signing them off. So you have this enormous system that requires thermal simulation, electromagnetic simulation, power delivery simulation. These behave more naturally like a matrix. To avoid getting too technical, they're closer to a matrix multiply numerical solver, which suits GPU acceleration nicely. So at this moment, I view Millennium as reaching more traditional segments such as autos, plus new segments such as aerospace and drones, and then reaching 3D-IC sign-offs; so the Millennium opportunity, together with our traditional hardware systems, has us very excited.
Anirudh Devgan
executiveI appreciate everyone being here this afternoon. We're at a thrilling moment for Cadence, kicking off 2026 with leading products and real momentum in our business. And along with our staff and our Board of Directors, we want to express gratitude to our customers, partners and investors for the ongoing trust and confidence they place in Cadence.
Operator
operatorTo all of you listening, we appreciate your joining us for the Cadence First Quarter 2026 Earnings Conference Call; that brings us to the end of today's session, so you're free to hang up now. Farewell.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Cadence Design Systems, Inc. transcript — plus 132,000+ transcripts from 7,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.
Get the API View API docs →This call discussed
For developers and AI pipelines
Programmatic access to Cadence Design Systems, Inc. earnings transcripts and 132,000+ others is available through the
EarningsCalls.dev REST API. Plans from $39.99/month — full transcripts, speaker segments,
full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.
More tools for stock research
OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in Cadence Design Systems, Inc. or any ticker on your watchlist.
StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.