Cadence Design Systems, Inc. (CDNS) Earnings Call Transcript & Summary

July 27, 2026

NASDAQ US Information Technology Software earnings 64 min

What were the key takeaways from Cadence Design Systems, Inc.'s July 27, 2026 earnings call?

In the second quarter of fiscal year 2026, Cadence Design Systems reported robust financial results, with total revenue reaching $1.584 billion, reflecting a 24% year-over-year growth. The company raised its full-year revenue guidance to a range of $6.260 billion to $6.340 billion, indicating a projected growth of 19%. Non-GAAP EPS for the quarter was $2.11, and the operating margin was 45.5%, showcasing strong profitability amid increasing demand for AI-driven solutions.

What topics did Cadence Design Systems, Inc. cover?

  • Revenue Growth Acceleration: Cadence achieved a remarkable 24% year-over-year revenue growth in Q2, driven by strong demand across all product groups. Management stated, "We are seeing growing demand for our AI-driven solutions across our expanding customer base," indicating a solid market position.
  • Increased Guidance: The company raised its revenue guidance for the fiscal year to $6.260 billion to $6.340 billion, reflecting a 19% growth expectation. This was described as "the highest raise we ever had" by management, signaling strong confidence in future performance.
  • Agentic AI Demand: Management highlighted the increasing traction of their Agentic AI solutions, with early customer results showing productivity improvements of up to 40x. Anirudh Devgan noted, "Agentic AI is a demand accelerator for Cadence," indicating its potential to drive future growth.
  • Record Backlog: Cadence reported a record backlog of $8.1 billion, which reflects strong future revenue visibility. John Wall stated, "Bookings resulted in a record backlog," emphasizing the solid demand pipeline.
  • Strong Performance Across Segments: All product groups delivered double-digit year-over-year growth, with the IP business growing over 40%. Management emphasized that "all product groups delivered double-digit year-over-year growth," showcasing broad-based strength.

What were Cadence Design Systems, Inc.'s July 27, 2026 results?

  • Non-GAAP EPS: $2.11
  • Operating Margin: 45.5%
  • Record Backlog: $8.1 billion
  • Revenue Guidance: $6.260 billion to $6.340 billion
  • GAAP Operating Margin: 28.4%
  • Operating Cash Flow: $635 million

Cadence's strong Q2 results and raised guidance indicate a robust growth trajectory, particularly driven by AI solutions and strategic partnerships. The record backlog and broad-based growth across product segments provide a solid foundation for future performance. Investors should monitor the ongoing adoption of Agentic AI and the integration of new partnerships as potential catalysts for sustained growth.

Earnings Call Speaker Segments

Richard Gu

executive
#1

Thanks, operator. I want to extend a welcome to all of you to our earnings conference call for the second quarter of 2026. With me on the call are Anirudh Devgan, our President and Chief Executive Officer, and John Wall, our Senior Vice President and Chief Financial Officer. A webcast of this call, along with a copy of the prepared remarks we're delivering today, can be found on our website, cadence.com. During today's discussion, we will make 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; for details on 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. Every forward-looking statement made during this call rests on estimates and information available to us as of today, and we disclaim any obligation to update them. Additionally, unless otherwise specified, all financial measures we discuss on this call are non-GAAP. These non-GAAP measures should not be viewed in isolation from, or as a replacement for, GAAP results. Reconciliations from GAAP to non-GAAP measures appear in today's earnings release. For the Q&A session today, we ask that you observe a limit of one question only. Should time permit, you're welcome to re-queue with additional questions. At this point, I'll hand the call over to Anirudh.

Anirudh Devgan

executive
#2

Thank you, Richard. Good afternoon, everyone, and thanks for being with us today. I'm delighted to share that Cadence posted excellent financial results for Q2 of 2026, with every key metric coming in ahead of the guidance we gave. We closed the quarter with record backlogs that surpassed what we had anticipated. Demand keeps rising for our Al-driven solutions among a customer base that continues to expand. The Al transformation is fueling robust, broad-based performance on both the design for Al and Al for design sides. Since our business is gaining steam and demand keeps accelerating, we're raising the outlook for the entire year to 19% revenue growth, plus better profitability, because we're becoming increasingly essential to our customers as a strategic, trusted partner. John is going to cover additional details regarding our Q2 results and the updated financial outlook. Let me start by talking about the broader environment. Design activity keeps growing as Al drives design complexity up exponentially, along with a new generation of system architectures that reach across hyperscaler infrastructure and physical Al. Customers are putting money to work aggressively in these areas, with Al and HPC leading the way, and we keep seeing indications of progress in the more traditional analog and consumer verticals. Developing chips and systems involves tough engineering problems. Solving them requires engines rooted in deterministic physics, data tied to proprietary silicon, and extensive design expertise. With our three-layer cake approach, we bring those pieces together in a way no one else does: accelerated computing and data sit at the bottom, physically accurate simulation and optimization solvers occupy the middle, and Al agents plus orchestration make up the top. Because autonomous agents expand the space of design possibilities and call on our physically accurate engines more often, agentic Al serves as a demand accelerator for Cadence, creating a lasting tailwind and a meaningful chance to grow our long-term TAM. We extended our lead in agentic AI through AuraStack AI Super Agent, which provides as much as 15x greater productivity along with a 2x quicker path to market for PCB and advanced packaging design. When it comes to agentic solutions spanning the entire electronic system design flow — from digital and analog design and verification through advanced packaging and PCB — Cadence is now the only provider. Early traction across our AI Super Agent portfolio is strong, with first customer results demonstrating substantial productivity gains and improved design outcomes. More than 20 customer engagements are underway for our ChipStack AI Super Agent, which drives higher verification productivity and shorter design cycles, and it is already in production deployment on several chip designs. During Computex 2026, we were there with NVIDIA and introduced what the industry has never had before: a completely autonomous virtual AI design engineer, which pushes ChipStack toward much greater autonomy. When it comes to early results from customers, RTL validation runs over 40 times quicker, shrinking a normal verification cycle of five weeks down to under one day for a design on a cutting-edge advanced node. Turning to analog and custom design, ViraStack is attracting considerable customer interest, with over 25 customer engagements and productivity gains ranging from 2x to 10x relative to conventional design flows. InnoStack is building momentum too, as customers embrace agentic AI for advanced-node SoC design. On this quarter's call, Rapidus outlined its intention to incorporate the Cadence InnoStack Al Super Agent within its Al agent design solution, pursuing design turnaround times potentially up to twice as fast. The strategic alliances we hold across the ecosystem continue to gain momentum. Together with Intel, we expanded our collaboration through a multi-year commitment focused on supporting its 14A process, leveraging our design IP and agentic Al-driven EDA to jointly optimize tool flows and methodologies for next-generation HPC and mobile design; across the next several years, this arrangement ought to act as a significant growth driver. Also, we strengthened our collaboration with Samsung Foundry around 2-nanometer and 3D IC technologies, combining our AI-powered flows with design IP so we can support next-generation AI, HPC, and mobile systems. Turning now to our businesses: we were happy to see every product group achieve double-digit growth compared with a year ago. Our IP business had an excellent quarter, growing more than 40% year over year. Data movement, memory bandwidth, and advanced packaging are increasingly becoming bottlenecks for AI performance, and adoption of our differentiated IP portfolio remained strong. This was evident in robust demand for our star IP portfolio in AI and HPC applications, such as PCIe, UCIe, HBM, and LPDDR6. We also saw broader engagement with major memory, semiconductor, and aerospace customers. In our work with STMicroelectronics, we landed our very first design win for Tensilica DSP, which strengthens what we offer in automotive and audio. Our Core EDA business expanded 18% compared with the same period last year, fueled by wider uptake of our Al offerings. We also kept seeing our digital full-flow solutions spread further. More leading-edge designs adopted the Tempus and Certus sign-off tools, and we picked up wins among hyperscalers, major semiconductor firms, and startups. At Frontier Al companies, plus a prominent ASIC silicon vendor, our implementation and sign-off presence also increased — a sign of the distinct value we bring in supporting the most sophisticated designs in the industry. On the analog side, we achieved a notable competitive victory involving Spectre at a top-tier semiconductor supplier. Our FastSPICE simulator, Spectre FX, achieved a number of production wins at top customers. Meanwhile, our hardware business turned in another record quarter, powered by sustained momentum in Palladium Z3 and Protium X3. With designs growing to unprecedented scale, hardware-assisted design and verification is turning into a strategic capacity layer supporting our customers' AI roadmaps. These customers are designing some of the most complex chips and systems anywhere in the world. Our scalable, high-performance hardware platforms are essential to bringing their designs to life. Demand is particularly robust among AI and HPC customers, which include hyperscalers and top semiconductor companies. We brought on 12 fresh accounts and experienced substantial growth among multiple prominent Al buyers, plus a key victory against rivals involving a leading Al infrastructure supplier. Revenue from system design and analysis rose 37% compared to the same period last year. With Al systems growing more intricate, buyers are more frequently choosing our sophisticated packaging and PCB offerings. Multiple customers picked up Allegro X Al, motivated by a major cut in layout design hours. Our 3D IC capabilities, combined with the advanced packaging work we're doing alongside TSMC's 3DFabric solutions, mean customers can confidently create state-of-the-art silicon for AI workloads whose demands keep rising. Turning to structural simulation, our BETA CAE unit won a number of competitive displacements, and integrating Hexagon's D&E business, which we acquired not long ago, is progressing well, with important agreements signed with leading customers. Interest is strong among customers in our integrated full flow, which unites our multi-physics offerings across electrical, CFD, and structural domains, allowing us to best address next-generation system design requirements, including the emerging field of physical AI. On the whole, Cadence enjoyed a fantastic Q2. I'm very pleased with how our business keeps building momentum. We're seeing design work pick up speed, and our execution remains solid. Our standing against competitors is the best it's ever been, because we're at the front of the shift toward agentic AI in chip and system design. With that, I'll turn it over to John, who'll walk you through our Q2 results and the revised outlook we have for 2026.

John Wall

executive
#3

Appreciate it, Anirudh. Hello, everyone, and good afternoon. For Q2 of 2026, Cadence posted outstanding results, with momentum in AI picking up and strength broadly spread across every one of our businesses. Solid design activity along with customer demand pushed revenue up 24% compared with the same quarter a year ago, and each of our product groups grew by double digits. Thanks to solid execution, our Q2 operating margin came in at 45.5%. Bookings during the second quarter produced a record backlog totaling $8.1 billion. Let me walk through several financial highlights for Q2, beginning with the P&L. Revenue totaled $1,584 million. On a GAAP basis, operating margin was 28.4%. Non-GAAP operating margin reached 45.5%. GAAP EPS came to $1.33. Non-GAAP EPS was $2.11. Now to the balance sheet and cash flow. Cash stood at $1,440 million, and the principal amount of debt still outstanding was $2,500 million. Cash flow from operations was $635 million. DSOs came in at 65 days. We spent $200 million buying back Cadence shares. Before sharing our revised outlook, I want to point out that it rests on the helpful assumption that export control rules as they stand at present stay largely the same through the rest of the year. Under our revised outlook for 2026, revenue should now land between $6,260 million and $6,340 million. GAAP operating margin should fall between 27.75% and 28.75%. Non-GAAP operating margin should be between 43.75% and 44.75%. GAAP EPS should range from $4.76 to $4.86. Non-GAAP EPS should range from $8.05 to $8.15. Operating cash flow should be roughly $2 billion. We anticipate spending about 50% of free cash flow on repurchasing Cadence shares during 2026. Looking at Q3, revenue should come in between $1,595 million and $1,625 million. 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 range from $1.11 to $1.17. Non-GAAP EPS should range from $2.01 to $2.07. As we normally do, we've posted a CFO commentary document on our investor relations website that covers our outlook for other items plus additional analysis and reconciliations between GAAP and non-GAAP figures; to wrap up, I'm happy with how strong our first half turned out and with the solid pipeline and momentum we carry into the second half. At the midpoint, we now anticipate 19% revenue growth, 44.25% operating margin, $8.10 in EPS, and $2 billion of operating cash flow for the year. As I always do, I want to finish by thanking our customers, our partners, and our employees for their ongoing support. With that, operator, let's open the line for questions.

Joe Quatrochi

analyst
#4

Thanks for taking my question. To start, I was hoping you might offer some assistance. You mentioned agentic AI represents a chance to grow TAM over the long run. Could you provide any figures for that TAM right now? And perhaps, how should we view that pushing EDA, as a share of R&D spending, potentially upward as time goes on?

Anirudh Devgan

executive
#5

Hi, Joe. I appreciate the question. As we've mentioned previously, what's excellent about agentic AI is that it creates a fresh TAM opportunity. Simultaneously, it demands more of our core physically accurate software. If we revisit that three-part framework, there's a fresh opening at the upper tier, and it strengthens what sits in the middle. Overall, the level of interest has been gratifying — genuinely remarkable, in fact. Nearly every major account, and really nearly all accounts, are eager to work with our agent stack, and we now offer four super agents. That strikes me as a significant opportunity. Turning to performance, we naturally posted strong results in Q2 and across the year to date, with considerable momentum across various segments of the business. What stands out to me with particular pride is how well our software operations have performed, especially when you look at recurring growth. That was largely fueled by momentum in add-on business. We're observing add-on growth on two fronts: one comes from design work for AI, since our clients are creating a greater number of chips, and the other comes from AI applied to design, meaning our agentic and AI offerings. That shows up in the numbers we posted. What stands out most — and I believe this is the largest increase we've ever announced — is how widely spread it is, with software and AI both fueling that expansion; we'll monitor how the remainder of the year unfolds.

John Wall

executive
#6

Joe, let me just jump in here — as Anirudh mentioned, customer engagement keeps picking up speed. We're noticing more evaluations, pilots, and early deployments, and our expectation remains that we'll monetize via the new workflow products along with greater use of the underlying engines. To put it plainly, our guidance still does not factor in any abrupt step function. Even so, the opportunity keeps shaping up nicely.

Joe Vruwink

analyst
#7

Alright, thank you. Continuing along these lines, I'd like to raise a question regarding open source models that work on chip design. Perhaps, taking Kimi at face value, it appears an agent went looking for EDA tools and then arranged the flow once given a chip design task. So I suppose what I'm asking is, what does all of this imply for Cadence — two ideas come to mind. First, if customers these days possess agents that can reach your EDA tools, does that ultimately result in greater usage and higher net consumption? And second, where would you say the differentiation sits when a customer purchases Cadence mental models for orchestration, as opposed to customers perhaps choosing to develop their own?

Anirudh Devgan

executive
#8

Yeah, thank you for that question. As I've mentioned previously, and I've been saying this for a number of years now — four or five years — the genuine orchestration and monetization of AI is going to come about via this three-layer cake. Let me just remind everyone: the uppermost layer consists of AI agents and orchestration; the layer in the middle is our conventional physically accurate tools, ground truth; and the layer at the bottom is compute along with data. What we've heard recently only serves to validate that framework; this is going to take place across every market. AI's value is going to become increasingly vertical rather than horizontal, and that's something I've been saying for quite a while. In chip design as well, the value resides in the agentic framework and the entire mental model, all the knowledge graphs, and then invoking the physically accurate tools across a rich hardware set. With regard to this latest news, in Kimi's case, they mentioned a chip, though my belief is it's a small block, and that's about technology that's 20 years old in terms of frequency, which is 20, 30 times slower than the frequency we have today. Even for designing a small bug at a node that old, they required EDA tools to accomplish it; this pattern is going to repeat itself over and over. Open source EDA tools have existed for some time, I'm not sure, for decades, and they get used in certain university or specialized contexts. When it comes to doing real designs, people turn to Cadence for that. Differentiation is going to be across all three. We aim to differentiate ourselves in each of the three parts of the cake. Our knowledge graph, our mental model, and the way we handle reinforcement loops at the agents are genuinely differentiated. The manner in which we then invoke the middle layer, via deep API access and the strength of our middle traditional tools, is differentiated. In the bottom layer too, as you're aware, we have Palladium, we have Millennium, we have special hardware for that purpose. Our differentiation is going to be in all three, and with all three combined, we are more differentiated than at any point before. I take great pride in the differentiation of the mode we possess, and in the fact that these three layers will reinforce one another. Customers may always maintain their own agents, exactly as they maintain their own flows at present. For truly mission-critical tasks, they are increasingly reliant on Cadence, as you can observe in our engagements.

John Wall

executive
#9

Right. Joe, Anirudh keeps making the point that agentic AI ends up driving more demand, since agents are calling EDA tools nonstop as they work through a wider set of design options — and Kimi really showed that well.

Vivek Arya

analyst
#10

Appreciate you fitting my question in. The IP segment, as I understand it, has sped up to a growth rate beyond 40%. What I'd like to know is what's behind that? How much of it comes from organic sources and how much from inorganic ones, and what growth rate can IP sustain over time? Stepping back, I also hoped John could make clear the current contribution from Hexagon as well as the EPS dilution. Thanks.

Anirudh Devgan

executive
#11

John, would you like to take that first?

John Wall

executive
#12

Absolutely. Regarding what Hexagon brings, it's performing exactly how we initially anticipated, and it keeps fueling SDA expansion — the vigor behind our SDA figures extends across a much wider base. Momentum is showing up in 3D IC, advanced packaging, PCB, multi-physics, and physical AI. Bringing Hexagon D&E together is moving along nicely, and over time we foresee a major chance to reinforce both the tech lineup and our route to market. Then, I suppose on the IP front, IP posted a fantastic quarter fueled by AI, HPC, activity at advanced nodes, memory bandwidth, chiplets, and advanced packaging. Customer engagement was robust and wins were significant, though IP revenue may fluctuate depending on timing from one quarter to the next. The momentum makes us happy, but I wouldn't take any single quarter and annualize it. Our standing against competitors keeps getting stronger in interface IP, memory IP, and foundation IP. With Intel, as Anirudh pointed out, it's yet another instance of customers opting for a wider strategic relationship with us. Anirudh, anything you'd like to add?

Anirudh Devgan

executive
#13

Sure thing, John. Vivek, we're really happy with how IP and SDA have done. Before diving into particular IP lines, the encouraging part is that both are expanding. IP, naturally, is expanding strongly. SDA is expanding strongly too, and by now they've got sufficient scale. In EDA, I think we remain the top provider, covering analog, digital, verification, packaging, and 3D IC. Broadly put, IP and SDA are each nearing a billion-dollar run rate. That gives our portfolio plenty of leverage when working with customers. On IP specifically, as I've said previously, I see three major mega trends. First, obviously, our IP is far stronger than it used to be. The caliber of our IP, the PPA — power, performance in the area — for TSMC and the top nodes is stronger. We're landing many competitive wins in IP where, 2 years back, we wouldn't have even taken part. That's one item. The second factor is that our IP approach is more concentrated. Our IP approach always has been concentrated and will stay concentrated on leading nodes, on star IP, on Al and HPC segments. I've discussed these 5 key IPs, which are interface IP and memory IP. We've broadened into foundation and other areas, but particularly chip-to-chip IP, memory IP, and interface IPs are extremely critical, and they're expanding nicely. All right. The third item is that foundries keep multiplying. We spoke about Intel. I take real pride in this new partnership with Intel, and it's obviously far wider than IP, though IP is one piece of it. Our work with Samsung, which we brought up last quarter, and Rapidus. The foundry ecosystem is far more varied than it used to be. Given these three factors, our IP business is performing phenomenally. Just to be clear, the bulk of it is organic growth. This outstanding growth we reported, the bulk of it is organic growth. Now, how it unfolds going forward? We'll see, but every signal right now is favorable.

Siti Panigrahi

analyst
#14

Thank you for fielding my question. Sorry about the noise behind me. Anirudh, I'm over at the DAC conference, and what everyone keeps talking about here is agentic AI, which rather confirms your point. My main question: you mentioned a few of these agents — super agent, ChipStack, ViraStack — that your customers have been working with. I'd like to know what sort of feedback you're receiving, plus the savings in cost and the value delivered to the customer. John, I recall you spoke earlier about monetization, which could require a contract renewal or cycle time. Given the usage you're observing, are you noticing any acceleration in adoption that might allow the timeline to be compressed?

Anirudh Devgan

executive
#15

Siti, as I said, demand for these agents is great. What's exciting is that the use cases — we've discussed many of them publicly — show 2x to 10x, and in certain cases 40x improvement; those are only the ones we're able to discuss publicly. That represents a very small portion of our engagement. The number of use cases and the benefit are real, okay? Interest is certainly real when it comes to the number of engagements and how many customers wish to work with us. The strength of our portfolio, with the three-layer cake, is very well differentiated. I'm very pleased. We're perhaps six months past launching these products; the launch happened in Q1, but I'd say we've been working with our customers for roughly six months. We'll watch how it develops. As I said, the early add-on business is encouraging, but we're still in the early days, so we'll see how it goes. So far, demand is tremendous.

John Wall

executive
#16

Siti, from our perspective, this is something that speeds up demand. Our customers aren't looking to cut back on design work at all. What they're dealing with is design complexity growing at a pace that engineering headcount simply can't match. That's a point we've made consistently. When agents widen the space for design exploration and invoke the Cadence engines underneath more frequently, it opens the door to new agentic workflow offerings and more usage of our core tools.

Jim Schneider

analyst
#17

Good evening, and thank you for taking my question. Picking up on the agentic AI topic, perhaps you could share a bit about the add-on engagements you're seeing for those tools, and how broadly they're showing up beyond the roughly 20 to 25 customers you've already mentioned? If you could put a number on what those add-ons contribute, whether in the guidance arrays or in what they might mean for core EDA software revenue over the coming year, that would be great. Thank you.

Anirudh Devgan

executive
#18

Well, as you'd expect, we're quite cautious when it comes to projecting future, next-year figures. Stepping back for a moment, there are three things I'm really enthusiastic about, and the first is that the overall environment has improved considerably. That helps us a great deal too. It isn't only the AI companies — the hyperscalers' commitment to silicon is far greater than it was 12 months ago, and you can observe that, since you track all the hyperscalers. The volume of designs, and how many designs every hyperscaler is pursuing, is impressive. The AI semi companies are expanding enormously. Analog, memory, and consumer semi companies are now performing well too. The overall environment is far better than it was one year ago, which naturally benefits us, correct? That's the first thing. Second, I'd like to stress that our competitive position, in my view, has never been stronger. We're gaining a great deal of share across various customers, achieving much deeper engagement, whether that's agentic AI, hardware, or IP, and the numbers reflect that. All right. The third element is this new TAM expansion opportunity — agentic — which we're obviously very enthusiastic about; we remain in the early stages. When you put those three elements together, that's what's driving such strong results and such strong guidance. Just as a reminder, this is the largest annual revenue raise we've made in a single quarter. All right. To roughly 19% revenue growth, with better profitability. I'd add that some of the benefit from this agentic and other areas is already present, and regarding next year and the year after, we'll stay as prudent as always and watch how things develop.

John Wall

executive
#19

Jim, we certainly receive plenty of inquiries regarding agentic AI. What I want to emphasize is that the upward revision we just made for Q2 and the remaining portion of the year is a reflection of widespread momentum throughout the business, not something tied to one particular customer or offering. Our Q2 performance was robust in core EDA, IP, hardware, and SDA. Naturally, all of that is gaining from ongoing vigor in demand fueled by AI too. This momentum spans every business and every region.

Harlan Sur

analyst
#20

Good afternoon, and thank you for taking my question. Anirudh, during the second half of last year, Al inferencing compute workloads surpassed training workloads — and as we know, inferencing requires significantly more memory, is that right? We've seen this diversification unfold, with a range of memory architectures emerging to address inferencing. Alongside HBM DRAM, there's been development around SRAM-based offload architectures. We've also observed CXL-based conventional DRAM offload, and even enterprise SSD or flash-based memory being put to use, correct? Given all this focus on memory architectures and memory controller architectures, does that create some tailwinds for your custom Virtuoso family of EDA tool solutions, or tailwinds for your CXL-based or memory compiler IP portfolios, or both?

Anirudh Devgan

executive
#21

Yes, Harlan, that's a great point. Like John mentioned, the momentum is widespread, and the analog business, part of EDA, is certainly seeing really powerful traction too. Whether it's memory or analog, all of it gets completed here, and Virtuoso stands as the top platform in the industry for analog, mixed-signal, and custom design. It makes me very happy to witness both the broader landscape and this particular circumstance — all the innovation fueled by inferencing is benefiting each of our segments: analog, digital, and verification. The part that thrills me, and you're already aware of this, is that inferencing brings about a much wider range of architectures, just as you pointed out, along with a much wider range of customers. At this point, every large customer is convinced they'll naturally rely on standard offerings from chipmakers — the giants like NVIDIA, who are performing really well — but they're equally convinced they'll develop their own bespoke silicon. Beyond that, they want multiple variants of that bespoke silicon, tailored for memory access and for networking too, correct? There's plenty happening on the networking front as well. I'd say that across the past half year, I'm noticing considerably more startup momentum. Previously, startups were largely inactive. Over these last six months, several extremely prominent startups have emerged — not only in Al, but also in networking and even CPU, right? Overall, I'd say the environment is healthy, and that's showing up across every one of our businesses. Analog certainly, hardware-assisted verification, the IP segment, digital implementation — and 3D IC is a major area where we hold a leadership position. Those are the pieces driving this broad-based momentum. The hyperscalers are determined to build their own silicon and to experiment with architectures that differ from the usual, as you mentioned, and that development is inevitable. Whenever a single choke point appears, customers respond by devising alternative memory architectures to get around it, or alternative networking architectures. My expectation is that this keeps going. You're aware that as the market expands, that happens. As the Al infrastructure market expands, innovation aimed at optimizing each segment within it will only increase. Every bit of that innovation is going to need Cadence products in order for it to become reality.

Charles Shi

analyst
#22

Hello, good afternoon. Hello, Anirudh. There are a hundred angles I could take on AI, but the thing sitting at the front of most folks' minds right now — or rather the situation, an especially drastic situation that scares people more than anything else — involves giving a prompt, and I'm not sure when this becomes possible, to a highly capable LLM down the road that includes your chip design requirements, after which that LLM could by itself produce GDS II code and have it delivered straight to a foundry for tape-out, skipping any use of commercial EDA tools whatsoever. That's a picture some folks have been painting; our view is firmly the opposite — do you consider this kind of end-to-end chip design driven by an LLM to be genuinely plausible? Or, given what you brought up about a three-layer cake. Thanks.

Anirudh Devgan

executive
#23

Yes, Charles, as I've mentioned previously, this is something I've stated for a number of years. The manner in which this progress will take place, and naturally AI will bring plenty of progress, is going to come via that three-layer cake; there will be agents, just as we have super agents. Our tools sit at the core, and they'll keep sitting at the core of it; they'll operate across a range of hardware. I don't foresee that shifting. Occasionally some folks might feel anxious about it. What's actually true will win out. Right; this three-layer framework is what will win out. When it comes to commoditization, my view is that what will probably occur isn't that EDA tools become commoditized. What will probably occur is that over at the agentic layer, LLMs will present plenty of options. If you observe what's genuinely taking place in the market at this moment, it's that buyers want options when it comes to their LLMs. To illustrate, take GLM-5.2 and Nemotron — obviously an excellent launch from NVIDIA — plus every commercial model. What customers keep telling me is, "Given how quickly LLMs are advancing, can the agent get smarter about picking the appropriate model for the appropriate task?" That, in my view, is what will most probably occur. Within the three-layer framework, the essential role of our tools isn't going anywhere. Yes.

Lee Simpson

analyst
#24

Great. Thanks for fitting me in. I believe the majority of my questions have already been covered. Perhaps I'll pose one that's a bit more general and competitive in nature; it really does appear that Cadence has broadened its DTCO partnerships, now including Intel, while continuing to develop its Samsung roadmap. You've also strengthened ties with TSMC. When it comes to stack die and multi-chip designs, your positioning is roughly on par with or ahead of your peers, you'd argue at this point. Its exposure to digital design and IP interface perhaps differs from Synopsys. I suppose what I'm really asking is, where do you feel the greatest competitive pressure is coming from among your peers, in disputed accounts, given some of the other agentic AI efforts at your competitors — are you gaining or losing share in that digital implementation and verification at the leading edge? Thanks.

Anirudh Devgan

executive
#25

Thanks for the question, Lee. First off, I simply want to say that I'm extremely proud of this new Intel collaboration, because Intel is a company we've attempted to work more closely with for a very long time; this isn't an issue that's one or two years old; it's more like something that's 10 or 20 years old. Okay. At last, we have a strong collaboration with Intel, with Lip-Bu and his new team, and I think we've been working on it for some time now, but it's nice to announce it in Q2; it's a fairly broad-based collaboration. Naturally, beginning with what we announced a month or two ago, 14A and DTCO, our agentic EDA solutions, our IP portfolio, which is much stronger. I think it extends beyond that. You'll see that we're engaging Intel across all parts of Intel with all parts of our product portfolio. I'm genuinely pleased to see our position improving at Intel and our collaboration being just like it is at all the other leading companies. The same thing occurred with Samsung over the last six to 12 months. Regarding what we were weak at previously, a few years ago we were doing great with the TSMC ecosystem; we have a strong partnership with TSMC. I've said for some time that we were weak at Intel and Samsung, and that has definitely changed. There's still further to go, but at least the trajectory has certainly changed in my view. That particularly applies to digital and verification businesses. Even in digital, we've always been very, very strong in implementation, place and route; now, as we mentioned in my prepared remarks, also strong in sign-off. The depth of our digital engagement is also improving at all customers. Overall, I'm quite pleased with our position, and we simply always believe in simple things, right? Team, technology, and customers. We have the best team, I believe, develop the best products, and listen to these demanding customers, and that's how we remain ahead. We're not looking at who else is doing that, but rather are we truly satisfying the demanding workload of our customers? My sense is that at this moment, we find ourselves sitting in a very strong spot.

Jason Celino

analyst
#26

Excellent, thanks. It's good to learn about yet one more record-setting quarter for hardware. John, I'm aware you said this is invariably a pipeline-type operation, and you hold off until midyear before gaining clearer sight of the back half. Could you perhaps describe the sort of demand activity you're observing on the hardware side? I couldn't help but see that inventory rose fairly well during Q2, on both a year-over-year and a quarter-over-quarter comparison. Thanks.

John Wall

executive
#27

Yes, that's a terrific question, Jason. Yes. We keep witnessing robust hardware demand, especially among AI and HPC clients. Hardware-assisted verification is turning into a strategic capacity layer for customers who design the most complex chips and systems. Quarterly timing effects may exist, demand stays solid, and we keep anticipating that 2026 will be yet another record hardware year. My characterization of hardware would be that it's still constrained by supply relative to customer demand instead of being constrained by demand. We're assembling the systems as fast as possible so we can deliver against the backlog. Yes, some of the rise over the course of the year came from hardware strength, yet we're observing strength across every area.

Gianmarco Conti

analyst
#28

Yeah. Thanks for fitting me in too. Yeah, IP results were fantastic. Perhaps you could say a bit more about the Intel win — what exactly does it involve? Which areas of the portfolio? Did it replace a competitor? Roughly how large is the contract expected to be drawn down over what period? Is it reflected in guidance? Just the outline of the details, if you can share any of that. Thank you.

Anirudh Devgan

executive
#29

Sure. I can add a little more; this is a multi-year arrangement, and naturally, part of the benefit lands this year, though the majority is still ahead. We'll also put more investment into Intel and Intel customers, which is to be expected. Regarding IP, it's far broader than IP since it covers EDA and DTCO. On the IP side, we have a fairly strong portfolio. We'll make that available on Intel process. This doesn't cover the fact that as Intel Foundry gains more customers, they purchase IP from us. This is simply our arrangement with Intel at present. As you know, we always stay conservative in those projections. Intel Foundry is, as you know, speaking with many customers, and it's possible those customers will obtain these IPs and any differentiated tools that emerge from this DTCO. We'll see how it goes. IP strength is, of course, Intel is part of it, but it's far more broad-based. Even our overall strength, I think I want to emphasize and John already noted, is not coming from one particular thing. There are four or five things driving this raised outlook. Intel is one of them for sure. IP is one of them. Hardware is a key focus, but hardware, and if you look at our recurring growth it's very good, right? Hardware is important growth, but so is EDA and agentic solutions and 3D IC and SDA. I feel right now there are four or five engines driving our growth, but we are definitely very proud of the Intel agreement and the new partnership.

John Wall

executive
#30

Gianmarco, I realize your main focus is revenue and similar items, but I'd like to point out that the second half carries certain costs too. That's because we're putting money into opportunities such as Intel, and also working to bring Hexagon's D&E business into the fold, since our attention is squarely on lifting margins in the coming year. You'll see that margins in the second half come in a bit under those of the first half, yet that simply shows we're making purposeful investments. These investments are intentional, and in no way do they signal any weakening of the core model. Our organic incremental margins are still quite appealing, and we anticipate that acquisition profitability along with the profitability of IP will keep getting better as we move toward 2027.

Ruben Roy

analyst
#31

Thank you, John. It seems like my question was already covered by what you said. Let me confirm I've got it right. So, I saw an implied operating margin of roughly 43%, and R&D costs likely rising about 19% year over year, going by the implied guidance for the entire year, compared with roughly 10% growth in the prior year. Naturally, Hexagon is responsible for a portion of that. So, how much of this comes from the core business, I wonder, and I was considering agentic Al and go-to-market — is that hiring something you've already locked in? Is that behind part of the expense increase, and what's your expectation for how it carries into 2027? Thanks.

John Wall

executive
#32

Yes, hiring isn't the only piece — there's also investment in systems and all of that; we're working hard to invest significantly to ensure a complete and proper integration of Hexagon's design and engineering business, plus certain other businesses, the smaller ones we've brought into systems design analysis; that's a major focus for us during the second half of this year. I believe I pointed this out last year — we had, I think, $20 to 25 million earmarked specifically for second-half investments; of course, there's always a degree of prudence in what we expect for expenses. Ruben, I always want to give the team sufficient room to invest and seize the greater profitability opportunities available to them. What we're really focused on, in this year's second half, is capturing those opportunities and positioning ourselves to achieve improved operating margins in the next year.

Kelsey Chia

analyst
#33

Hello there. Hi, Anirudh. When it comes to Intel, would that engagement involving 14A probably act as an extra boost to the roughly 20% to 25% expansion the team has been posting for the IP business? And also, might it serve as a significant lift for our EDA business across the upcoming quarters, or what timeframe ought we use when considering that path, given they're now involved in EDA?

Anirudh Devgan

executive
#34

Well, let me just confirm that I've got the question right. The Intel work we made public, as I see it, adds entirely to what we do — 100%. An existing Intel agreement was already in place; this represents a fresh deal layered on top, and it spans multiple years with several components of that operation. What I'd also note — and I believe Lip-Bu has stated this openly too — is that we put out news about 14A, and I think for Intel to win in the foundry space, they need to go beyond 14A; we're already in discussions with them about additional future roadmap items for Intel Foundry. Naturally, Intel has various segments, as you're aware — the product groups — and they're putting money into their server business along with their client business. Once again, we take pride in collaborating closely with Intel, the same way we do with other major customers. To me, it's really about our relationship with Intel becoming more normalized, much like how we operate with every other well-known company. This progress makes me very proud.

Jay Vleeschhouwer

analyst
#35

Thanks. Good evening. Anirudh, I want to ask about what's actually involved in practice — the practical side, or the requirements — when it comes to putting AI and agents into place, plus everything else you mentioned tonight. In other words, if you consider the support you provide before a sale and after a sale, along with customer support, how does that stack up against, say, what you previously had to deliver for classical EDA? Is there anything different now — whether measured quantitatively, qualitatively, or technically — that you're required to do which wasn't necessary before? What I'm thinking about, as an example, is that during the past few months there's been a quite obvious upward trend in your AE job openings. That's a classic leading indicator for customer adoption. You've also described GenAI as, quoting, "the critical path for agentic adoption." Perhaps you could address that too.

Anirudh Devgan

executive
#36

Yeah. Thanks, Jay, for asking that. That's a great question. I mean, broadly speaking, naturally, we are expanding, and we will put money in, right, in both R&D and application engineering. Agentic AI doesn't call for some enormous step up in investment. I just want that to be clear. It resembles more of our traditional business, since we are, naturally, very asset light, right? We aren't constructing compute farms. Our customers do all of that. Okay? Just so that's clear. Now, naturally, certain skills differ, and our team is already expert in computation software, as you're aware, so they're able to pick up agentic AI. Some hiring we will do. It's more business as usual, I would say. Also we can make AI — there are implications that we're applying plenty of AI internally, okay, to make things even more efficient. For instance, AEs — yes, we are hiring AEs — AI can dramatically cut AE workload and make them far more productive. More of the AEs can take part in pre-sales activity rather than post-sales support, right? Same thing in R&D. Naturally, we're deploying AI for software development, and naturally, AI deploying our agent stack and all our agents for IP development to make them more efficient. This is what John mentioned earlier, so we'll see how that goes. I think AI has the opportunity to even lower our cost in some cases. You shouldn't model in some massive investment. I think the investment we discussed is more for SDA, for the integration. You always talked about, Jay, how we need to have a full flow, and I feel that at last we have a full flow in SDA, so investing in that, investing in Intel. The agentic AI will go through our regular sales motion and regular AI and R&D support. Yeah.

Josh Tilton

analyst
#37

Hi, everyone. Appreciate you fitting me in. I've got what's maybe one point of clarification plus one broader question, both for Anirudh. On the clarification side, could you walk us through what's behind the momentum in other recurring revenue? It really caught our attention this quarter, and any color you could offer would help. Then on the broader side, Anirudh, unless I heard you wrong, during your prepared remarks at the start, you mentioned moving toward being a more strategic partner for your customers. This is aimed at you, but John, please chime in as well. Maybe help those of us who are financial analysts grasp what that looks like from a business standpoint. Is wallet share expanding? Do you have room to charge higher prices? In what way does your company capture value financially now that you're turning into a more strategic partner for your customers? Hopefully that's clear. Thank you.

John Wall

executive
#38

Should I kick things off, Anirudh? Josh, I'll handle the recurring revenue one. In Q2, recurring revenue was up roughly 24% compared to a year ago, and the main driver was robust core EDA growth. There's some Al-driven demand, share gains are happening, renewals are healthy, and expansions come through add-on business. Inside that figure, Hexagon likely added about four points. Strip that out, and on a normalized pro forma basis, recurring revenue sits in the high teens to 20%, which we see as a very strong result. Our expectation remains that the full-year mix lands near 80% recurring and 20% upfront; regarding the agentic Al side, basically, in terms of how we sell it, the first thing is that customers keep buying our underlying EDA software. What happens is they buy Cadence agent licenses that orchestrate engineering workflows. Broadly speaking, our economics scale alongside customer adoptions. Anirudh, do you want to add anything to cover the remainder of Josh's question?

Anirudh Devgan

executive
#39

Josh, here's the point we're making: with our semiconductor customers, we've always taken a strategic approach, correct? Naturally, engineering is where we live. R&D is where we live, right? We aren't comparable to other enterprise-type software players. What we do is engineering software, and we sit at the core of how they build their products and generate revenue. What's shifted recently, say across the past 12 months, is that even within semi companies, the level at which we engage has risen considerably, because EDA, chip design, and agentic AI possibilities matter a great deal to our customers. Roadmaps are getting more demanding, and Moore's law is slowing somewhat. Everybody knows this—it isn't delivering sufficient gains anymore. Those gains now have to come from design efficiency, improved optimization, smarter deployment of AI agents, plus that middle layer, correct? The PPA delivered by—we handle this alongside the foundries and alongside the customers, and DTCO factors in, so power and performance area can be optimized better than would have been achievable when Moore's law was truly advancing rapidly. Today, it has decelerated. That covers the semi side. Turning to the system side, the understanding that semiconductors are indispensable has now arrived over roughly the past 6 to 12 months. Every one of the Magnificent 7 companies, all those big, genuinely household names, treats silicon as a vital piece of their roadmap. With these customers, Cadence's involvement is absolutely essential. The method for turning that into money is straightforward: we deliver greater value to them, and in return we capture greater value for ourselves, exactly as our results demonstrate.

Gary Mobley

analyst
#40

I appreciate you letting me squeeze in here; this is probably something better directed at John. The thing that really jumps out to me is what looks like roughly a 55% rise in bookings during the first half of the year compared with the same stretch a year ago. My assumption is that you'll continue building on what tends to be a seasonally robust back half of the year, and my understanding was that this was a lighter renewal window for a few of the larger customers. So maybe you could talk about what's behind that booking momentum. Does it reflect how strong the chip cycle is? Is that strength coming from heavy chip design activity, or does it come from certain AI tools that are pushing greater use of additional copies of traditional EDA tools?

John Wall

executive
#41

Absolutely, Gary. That's a wonderful question. I believe Anirudh touched on this somewhat when responding to Josh's question. As we always put it, strategy comes first, correct? We've kept on carrying out our intelligent system design strategy. Anirudh has laid that out for us over roughly the last 10 years. What we observe is that every one of the fundamental structural demand drivers keeps getting stronger. Semiconductor complexity, AI infrastructure investment, engineering productivity, physical AI, agentic workflows — all of those trends appear to remain at an early stage, and I believe that's contributing to genuinely strong bookings for us; this year is likely among the lighter ones when you consider the 3-year renewal cycle. This year is likely one of the lower ones out of those 3 years. We're witnessing quite strong momentum in add-on opportunities, as Anirudh said earlier during the call. I'm extremely happy with the progress and how everything is going. Anirudh, is there anything you'd like to add?

Anirudh Devgan

executive
#42

John, that was an excellent recap. According to John, Gary, the current year represents the low point for bookings, and typically during the first half we pull from our backlog. This year, growth has been strong. We'll monitor how that unfolds, but we're quite happy. John noted that the environment is favorable. I want to highlight that in my view the three major factors are — as John also mentioned — the environment is terrific, encompassing both the new Al entrants, the traditional Al players, and the standard companies. Our offerings and competitive standing are outstanding. Then there's this fresh TAM opportunity involving agentic Al. When you put all three of those together, the first half has been excellent. It positions us well for the remainder of the year, and we'll watch how things develop.

Anirudh Devgan

executive
#43

We appreciate everyone being here this afternoon. Cadence is in a thrilling period as we move into the latter portion of 2026, powered by Al-driven product leadership and robust business momentum. Together with our board of directors and our employees, we want to express gratitude to our customers, partners, and investors for the trust and confidence they keep placing in Cadence.

Operator

operator
#44

Thank you, everyone, for joining us on Cadence's Q2 2026 earnings conference call. That wraps up today's session. You're free to hang up now.

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.