Cadence Design Systems Inc (CDNS) September 2026 Conference Transcript

Hello, good morning everybody. Welcome to the Goldman Sachs Community and Technology Conference. I'm Jim Schneider, the semiconductor analyst here at Goldman Sachs. It's my pleasure to welcome Cadence and CEO Anirudh Devgan to the stage today. Welcome Anirudh. Thanks for having me. Thank you, thank you. Great to be here. I've been asked to read a safe harbor to begin. Today's discussion will contain forward-looking statements, including Cadence's outlook on future business and operating results. Due to risks and uncertainties, actual results may differ materially from those projected or implied in today's discussion. With that out of the way, let's get rolling. So first question for you, maybe high-level. I mean, I think Cadence has had a very strong first half of the year with double-jitted growth across pretty much every product group, record backlog, two increases to full-year guidance, Before we get into individual businesses, how do you characterize what's changed in customer behavior over, say, the last four months?

Yeah, thank you for the question. I mean, the customer environment is probably the strongest I have seen it because, you know, last few years, of course, some companies were doing phenomenally well, you know, the big AI companies or the hyperscalers, but some of them were not. But if you look at in 26, you know, universally, the industry is doing great. The semi-companies are doing great, and then all the system companies, hyperscalers, the commitment to silicon is the strongest that I have seen. Because sometimes you used to get questions like a few years ago, well, will all these hyperscalers really do chips or not? But you can see now the success. So I think that what I would say at the highest level is the environment is good, and we always try to check how long is the party going to last. but it looks like the party is only getting started. They talk to all the people. They're very confident next few years. That's number one. Number two, I think our products are performing great. Of course, we are in the tech business, so best product always wins. And our comparative position is very strong. That's second. And third, we have this new TAM opportunity, new expansion of, you know, agentic AI on top of our kind of traditional offering. So that's all new time for us. So if you put it all together, these things are what is driving this growth that you are seeing. Great.

Now you framed Keynes' differentiation as a three-layer cake, especially tempting as we get closer to lunch here. But anyway, a base that's accelerated computing data, middle layer that's physics-based simulation, and optimization top layer of AI agents. Why is that particularly relevant for EDA versus other kinds of software that's in the market today?

Yeah, and I've been only saying this, like, for five years now, I think, the cake. And people say, like, first of all, all things have to have three things. You know, the answer to life is E, right, the universal constant. That's what my advisor used to say is 2.7, you know. Because if it's less than three, it's, like, too little. And if it's more than three, nobody remembers anything. A pie is more than three. Yeah, slightly more, a 3.1. So whether it's 2.7 or 3.1, you can choose your favorite universal constant. So I think three, and the reason I call it a cake, you can call it a stack if you want. The reason I call it a cake is because if you eat a cake, unless you are two years old, you eat all the layers together. And you have to bake all of them together, means they they interact with each other so that's the reason to call it a cake you know and and the reason I put AI at the top and computer at the bottom you can put it in because so first of all the middle layer is super critical and this is going to happen in all by the way it's going to happen in all markets not just EDA or not just chip design you have to ground the AI with with physics especially in these kind of complicated engineering software or engineering workflows Now, in some cases, the middle layer may not exist or is maybe simple, but definitely in our business, you need to ground the AI, the physics, and then, of course, run it on compute and data. So all three are critical. And the real value will accrue to the vertical application, not the horizontal. Because in the beginning, it's always horizontal. In the end, it is always vertical. Like Waymo is a vertical application, for example. So if they have an AI model, Do you know what model it is? It doesn't matter, right? Can you get from point A to point B? They, of course, have control theory navigation, and then they have the silicon, just to give an example. And same thing will happen in chip design. And the reason I put agents in the top is because agents are very good at, you know, directional kind of orchestration. Like if you want to go from here to Palo Alto, you know, that's a directional thing. But actual navigation and detail, They are not as good, but they're great for orchestration, planning, optimization. So that's why the top layer calls the middle layer that sits on the computer.

Now, the bear case that investors often raise with me relative to the EDA industry is that if you have a sufficiently capable frontier model, you could basically automate chip design from a prompt and basically bypass the commercial EDA software flows. Why do you believe that's wrong? Specifically, why do you think deterministic physics-based engines and proprietary data are kind of essential, especially for leading-edge designs?

I think they're all going to be important. One thing with AI is people who graduated a few years ago, they think, well, I will make a model of everything. What do I need to know? And then people who graduated 30 years ago, they say, oh, it's all curve-fitting. What do you need to know is reality? how things work, whether it's physics or mathematics or economics or whatever it is. The reality is you need both. There's no need to take a side in that. You need both. And to have a successful thing, you know, AI itself, it's a nonlinear curve fit, right? That's what these LLMs do, is you give it input, output, it fits a nonlinear model to it. It used to be a transformer architecture. But fundamentally they cannot do nonlinear differential equations, state varying. This is mathematically not possible to do it. But together, they can provide a good combo. I do believe AI, like we have seen, can provide more scenarios to optimize, and then that can be optimized in the physics-based layer. So mathematically, it's not possible to do the middle layer. But we want to innovate in all three layers. We just don't want to innovate in the middle layer, which is classical physics-based. It's the combination of the three layers that will win. And you'll see that more and more in all industries. And then why, I mean, I guess the other question, follow-on is... And nobody's trying to do that, by the way. All the LLM companies, all the hyperscalers, they're all using our tools to design chips, just to be clear. There are no chips being designed without using our tools, yeah.

Just to push that back for a second, like why do you need all three layers together? why can't we have somebody else's solution for the top or a bottom layer and yours for the middle?

Yeah, that could happen. Yeah, you could have an agent, and some customers are writing some agents that call our tools and not use our agents. That could happen. What you have to remember is the top layer is a brand new TAM opportunity for us because what the top layer used to happen, and these AI agents, was basically done by humans in the past. So what agents are doing is they're not replacing the middle layer. They're replacing what humans used to do. Now, in some scenarios, agent could call our tools, but it is not that efficient. So because we wrote the middle layer, we wrote the top layer, so a lot of times we have access to the internals that is not exposed to the user. Yeah, but it'll be natural for some users, especially in the beginning, to write their own agent. But in the end, they realize, okay, you know, it's more efficient for Cadence to do it. And we have like these four super agents, which are more aligned with functions. You know, so like tools, we will have the middle layer, we'll have like 30, 40 products. The top layer, we have four super agents, like front-end design, physical design, analog design, design, and PCB and packaging. So they are integrated closely with our middle layer. And we have unique advantages. We have like 10,000 people in R&D. So they are writing both the top and middle. But even in the top layer, we don't need to get 100% of that market. Even if some of it is written by our users, or they could have like 10 agents, but the four big ones are by ours, and six could be their more domain specific. That's all fine. Even in the traditional flows, a lot of customers do customization on top of R2.

So then on Agentic, how do you think about the monetization of Agentic? Specifically, where do you expect to sort of drive incremental revenue above and beyond what you're already doing? Is that the new Agentic workflow products themselves? And how do you think about the opportunity for higher consumption of your existing calls?

It will be a combination of, like, we have new business model for the top layer, which is consumption plus, you know, subscription. And then, of course, our existing business model for the middle layer. And a good example of that is, because one very always is, if something is like, let's say, 5x more efficient, then you will use one-fifth of the middle layer. You know, this is also some perception in the market. And this is not new. Actually, in 2006, I launched a simulator, and it was like 10 times faster. And then my marketing team was worried that, oh, people will buy like 10 times less. But that never happens. You know, that's the history of EDA. And the reason for that, you know, there's a fundamental reason, which is different than almost all other software markets. That's why EDA is so exciting. And sometimes we get lumped in general software. You know, those guys never thought we were software. and we never thought they were software. Because our software is so mathematically complex that accessing a website or database, we don't consider that. That's just one small part of what we do. And they thought, oh, we are semiconductors or something like that. But it doesn't matter. I think what happens in this kind of application, EDA or chip design, the workload is exponential. Workload is exponential. So if you look at TSMC roadmap, You know, next five years, they think that they said that chip's complexity or size will go up by 48x. This is not happening in any other software market. So I talked to some customers, our big hyperscalers. They think every year they want to, if they continue like this, they need to hire 2x more engineers. It's not sustainable. So if the workload is exponential, the requirements of headcount is exponential, You need this 5 to 10x automation. You know, if the chip size is going to be 50 times bigger, there's no way they're going to hire 50 times more engineers. So you need this 5 to 10x improvement to even sustain the growth. I think the customer's headcount will grow, but with automation, you know, with AI, it will be less than... And this is the history, you know. Like, if you look at late 90s, early 2000s, we would design... Our customers would design a CPU. it would take them five years and 500 people this is not uncommon in in all these IBM Intel deck all these companies now you can design a CPU with 30 40 people within six months so that's 100 times faster than 20 years ago and the amount of silicon is only going up an amount of design activity only going up because exponentially the size is exponential, also the applications are... So this is going to continue. If you look at the roadmap from iMac and all that, this kind of exponential is still projected to go until 2042, which is still how many? 16 years at least. And by then they will have some other technology. So this is not going to slow down, which is very unique to any other software market. So we are always looking. We are always looking at improving the efficiency of our solution, and it gets absorbed even faster. You know, you get all the roadmaps from NVIDIA or Google or Apple, and they are doing even more and more with that. So this is something not to be afraid of. It's something to embrace, that the productivity will actually help us sell more, right?

And can you say something about sort of like what is, So if you think about the monetization of it in terms of revenue terms, you know, what is different about the Argentic flows that's actually driving, accelerating, recurring revenue growth today versus the past things like Cerebris and other AI features, which were maybe in your core offering where we didn't see that kind of like acceleration revenue?

Yeah, that's a very good question. And, of course, you know, we always did a lot of, you know, good work. But what is new with this agentic AI? And we always wanted to do it. This is going back, you know, decades. We wanted to automate, more automate the running of our tools. You know, our tools are fairly complex. And typically what happens is, you know, they run for a few days. This is not like, it doesn't run for five minutes, right? If you're doing some blog, you know, it will run for a few days, do all kinds of optimization. But what the customers are doing is they run it one time and design is naturally iterative, so they have an RTL, they would change it and then they would run it again and they change it and they run it again. And typically a user would do like three or four experiments at a time because that's what typically humans would do. But if an agent is running it, agenting is much more meaningful to us than Gen AI because some people said, well, Gen AI has been around for four years. Why did it not have a big impact on chip design? Because Gen.AI helps improve the I.O. of the tool, you know, you can talk to the, you know, look up documentation or whatever, but that's not, okay, that's useful, but that's not, that's not earth shattering, okay. What is, what is interesting in Gen.AI is that you can define a workflow or a graph or like you do A, you do B, you do C, if you get stuck, you do, and this is all relatively new with Claude code and all like about a year ago or a little more than a year ago. So this kind of workflow combined with our base tools can give a lot more productivity. And then when the agent runs it, it runs like 100 experiments. It's not running three or four experiments. But this kind of workflow is the new thing. So that's why I'm so confident that our agenting solutions will have a real impact versus Gen AI a few years ago. And Cerebris and all were good, but now with agenting and the base, but it calls more of the base than less of the base. and then this kind of productivity, this 5, 10x productivity or at least several x is possible will help meet the exponential demand of our customers. And the demand for all these, we have engaged with all the top companies with all our agentic solutions and of course the usage of the base tools is also going up like you see in our results.

Yeah, so if you think about that acceleration, where do you think Cainz is getting most competitive traction today and sort of what are the product areas that represent like the most remaining market share opportunity for the company over the next few years?

I mean, right now we are doing well in almost all of our products, which is great. You know, normally you always want to see that, but it doesn't happen that often. But right now I think we are hitting in all cylinders. And we are not dependent on one critical area, but right now all of them are firing, you know, EDA. Anyway, we have the broadest portfolio for chip design. I don't know how familiar you are. We not only do digital design, we do analog, you know, memory, mixed signal, packaging, PCB. So Cadence has always had the most complete portfolio, and then we work closely with TSMC for a long time, with ARM for a long time, and now with Intel and Samsung. So Core EDA is as strong as it has ever been, and then we put all the agenting on top of that, right? And I think we are definitely leading in agenting. and then hardware which is like hardware acceleration which can run things like thousand times faster we are the only company that designs our own chip actually at TSMC if you look at our hardware system these are as complex as you know the latest GPU or XPU system so these are liquid cool fully optically connected rack and then we have like a 10 15 year lead in designing our own so that's And the demand for hardware is going up because, first of all, more people are designing chips, but hardware is used in proportion to the size of the chip. So if the size is going to go up by 48x in the next five years, so that's a systematic improvement. And then IP was the weak point of cadence historically, and I didn't invest as much in IP because it's not as profitable as EDA. But now, I think, especially with AI and 3DIC, there is more opportunities than IP. So if you look at IP, our business is up 30% this year. It was up, I think, 30% last year, probably. So last three years, it has grown much, much higher than the market. And I feel that IP can still continue to grow well with all this Intel and Samsung and, of course, TSMC. So I feel all these three major areas, three or four, and system business is growing pretty well. So we are in a good position. And the main thing is, you know, our customers are growing. So if the customers are growing, they want to do more and more innovation.

Yeah. I want to get back to IP, but first to close out the loop on hardware for a second. I mean, you've talked about demand being supply constrained, I think. You know, what's structurally driving that demand for hardware? I mean, is it the scale, the designs, which you mentioned, or is it also kind of like, you know, your customers shifting towards emulation as more of a strategic capability rather than sort of a project-level thing.

Yeah, I mean, one thing, I don't know how familiar with this is, like, give me a few minutes to explain what these hardware systems do. I mean, we call it hardware, but it's hardware plus software. You know, people would call it, like, full stack, basically. But basically what happens is at this point, you cannot design any complicated chip without these systems. It's not possible. And there are multiple reasons for it. What these systems will do is even before, let's say you're designing a chip for like nine months or twelve months whatever it is six to twelve months typically is the design time you you want a more you want to verify the chip in your environment whether it's a software environment is like Windows or CUDA or iOS or whatever so we can have a chip behave like a chip you know RTL we can make it behave like a chip even before it comes back from TSMC or any foundry and that is used to not only develop software but also verify the functionality of the chip because if you can you know boot some OS on top of your chip and run your you know application correctly then of course you know the chip is correct and that's only possible with this kind of palladium kind of systems so then they become like irreplaceable but otherwise what will happen is you would do the design and then you would check and then you will redo the design and take few iterations, which is the old way of doing it. And only a few companies are doing it. Most of them have moved to hardware-assisted design And the second reason they are popular is not only you can verify the chip, you can write your software. Because if you emulate the chip, and these are custom chips that emulate the chip like 1,000 times faster than CPUs. I mean, they're still slower than real life, but much, much faster than anything else. So you can develop all your software. So all these system companies, hyperscalers, they're developing chips. Of course, they have software to develop. So for those two reasons, it became irreplaceable. And then the amount of hardware you buy is proportional to the size of the chip, which is going up. So one, it became irreplaceable. Two, there are more chip designs. Three, there are the size of the chip going up. So it has been record year for, I don't know, the last six years. I don't think that's going to slow down. Very good.

IP, let's come back to that one for a second. In terms of your market position there, you've got a very wide product breadth across a bunch of areas, including DDR, CERTES, PCI, even processor course to some extent. Maybe talk about the diversity of the IP offerings and what are the specific areas where you feel like you have the most competitive advantage?

I think IP the interesting part is is of course we focus on lower nodes and HPC IP which is exactly what is what is of course growing the most because we didn't want to do all parts of IP because it's not as as profitable and also we want to do of course where the fuck is going and we focus on like five or six critical pieces like some of it we developed some of them we acquired So this is like the 30s IP, the PCI, UCI, which is chip-to-chip, HBM, connection to memory, DDR. So these are, in terms of design IP, these are the critical IPs that a lot of customers want. And then the other key thing that happened is our team is much better than before. I mean, in the end, these are standard-based IP. So the customer will buy if the PPA is good. In the end, it's not just having the IP, just like in anything, it's how good is your IP. so our team is you know we anyway I personally believe all the leaders should be highly you know technical and engineering background so that's true for all my my GM's and I think we have it this is one thing that has changed in the last few years our IP our EDA teams are always world class okay hardware teams world class now our IP team is world class in terms of design capability and they can also use AI to further accelerate their own so then the output of the IPs are very competitive at TSMC and other foundries. And then the third thing that happened is these other foundries also want to get in, so we need to, you know, develop IPs for them. So whether it's Samsung, Intel, Rapidus, along with TSMC. So I think these three things, you know, our focus is correct in terms of the market segment. Our team is much better, and, you know, PPA is much better. PPA is power performance area of IPs. And then the market is naturally growing with newer foundries. Got it. Okay.

I want to move on to your last segment, system design analysis. My personal interest is, I think it's the most interesting segment you have in terms of the evolution. You've talked about SCA enabling companies like Aerospace and Defense OEMs to simulate a whole system. You know, how different is a product strategy when you're selling to somebody like a Boeing relative to somebody like NVIDIA or AMD? I mean, do they want the same physics models, or do you have to take a fundamentally different approach to R&D for that?

No, it's similar. That's why I did it. By the way, I don't know if you know this. I'm the one who started in 2017, and people thought this was preposterous. Like, why would EDA and SDA be together? Because they were not together, okay? And there were multiple reasons for it. I don't know how much time I have to explain the reasons, but at the highest level, first of all, the math is very simple r&d is very similar and sda is easier than eda of course the sda guys don't like it when i say that but but eda algorithms are much more complex than sda algorithms you know electromagnetics is much simpler than circuit simulation but they're in the same direction you know they're also mathematical software okay that's so you know all companies want to expand but you want to expand in your core strength so what is because they asked me like okay and you're going to be CEO you know I became president you're going to be CEO so what is your strategy okay so the strategy is that go amplify your core strength what is our core strength in cadence or my background or all the EDF is numerical analysis computational software this is not like I said in the beginning this is not like you know some database software or you know look up a website you know this is mathematically deep like as deep as you can get of course everybody thinks what they what they do is hard but you can look at up what we do it's like the most difficult you know CS plus math plus physics so that could be applied to two as two systems and then the question is why do you apply to systems you know because if you look at the market so that's our core strength you know mathematical software if you look at the market I always thought the market will evolve into these three concentric circles again this is obvious now but the silicon is in the middle then system and then data okay a perfect example is like a car right or a you know self-driving you have all the navigation data then you have the car which is mechanical plus electrical hardware plus software and silicon that drives the car and this could happen in all all markets so if you take those three concentric circles and overlay the strength of ours, which is computational software. Of course, computational software applied to silicon is EDA, chip design, EDA and IP, and that was always our core. Always wanted to make sure that we are number one in EDA because the other mistake people move is they expand into other markets but lose focus on the core market. So our always focus from the beginning is EDA should be number one. That's why OREM invested in EDA versus IP, even though IP is interesting now. But in EDA, we have the broadest portfolio. We are clearly the company to work with. But then if you apply computational software to systems, that's SDA. And we want to do things which are synergistic to chip design, so which is like thermal analytics, electromagnetic analytics, things which are 3DIC, which are closer. And then computational software applied to data is, of course, AI. By the way, the AI is even simpler than SDA. Okay, the AI people don't like that either. That's just linear algebra. That's like, you know, I took seven courses in linear algebra in undergrad. So don't forget even grad school. I think I forgot about that in kindergarten. So AI is just, you know, the algorithms AI are even simpler than, but it's a good, I mean, it has a lot of applications, but it is same kind of computational software applied to chip design, which is the most complex, and, of course, growing exponentially. then systems, and then data.

Just a minute or two left, but I wanted to quickly ask you about physical AI. So how should we think about sort of physical AI being a long-term opportunity for Cadence and sort of where are companies seeing practical value in those applications today, your capabilities today, and how should we kind of think about physical AI being, you know, in terms of magnitude of revenue contribution over time for Cadence?

Yeah, I'm super excited about physical AI. and have been for some time, and not that, you know, we are, of course, excited about the current trends of data center. I mean, those are huge. But also, physical AI will be a very big application. I mean, you know, if you talk about the cake in the beginning, you know, the three-layer cake, also for five years, talk about three slices of the cake. These are vertical slices. Because in the end, of course, the value will be vertical, right, not horizontal. So the big slice right now is data center and infrastructure, and I think we are very well positioned. We are working with all the Mag7, you know, like we discussed. You can see it in our results. But the other thing in strategic direction is you want to make sure you don't miss any of the other big things. See, one thing is you have to grow in your core strength, number one. So I explained, like, computational software. Number two, you have to grow with the market, So then chip companies are becoming system companies and AI companies, which is obvious now. Look what NVIDIA is doing or Broadcom and Google and Apple. And number three, you don't want to miss any big trend. So we always over-invest ahead of it. Not too much, but always ahead of the big trend. So then what are the big trends? If the three layers are horizontal, the three vertical slices are data center first.

We are very well positioned.

Then I believe physical AI will be huge. because these are all trillion-dollar markets. See, if AI is good enough to reason and talk and see, imagine what could happen in cars and robots and drones. And these are trillions, trillions of dollars of markets. And then the third slice, I always believed, is science is AI, which is life sciences and other deep sciences. I think what happens is people confuse that all these three are happening at the same time. And to some extent they are, but they have a peak of each cycle, So I think data center is in peak. I think physical AI may peak in the next three to seven years. And then life sciences and all may be five to ten years from now because that's another important thing. It's very difficult. So we want to invest in all these three slices. So we, of course, do life sciences, as you know. And physical AI, we did acquisition in Hexagon to get the best kind of middle layer for that. It's the best robotic simulator. So the opportunity for the physical AI is not just... The AI model will be different. It'll be a word model, right? If you go back to the three layers of the cake and put the physical AI slice, the top of the slice is different because it's a word model, not LLM. And there is no data for the word model. You have to do a lot more simulation. So therefore, we invested in hexagon D&E business for simulation. But also, it will drive a lot of silicon. So our traditional business, and the silicon in physical AI will be more mixed signal silicon, you know, for cars and drones. which is anyway, Cadence's traditional strength. And you can see that in Tesla or Rivian or BYD or Xiaomi. I mean, I just came back from China. It's amazing what's happening in Xiaomi and BYD and NIO. And they're all designing chips. They're all our customers. Same thing with this U.S. Some of these U.S. companies like Tesla, what they're doing is remarkable. Rivian and some of the traditional companies are. Because the criticism has been, oh, this is a very slow-moving market. But I think this self-driving is completely going to change that. And then drones and all. So it doesn't mean that we don't love data center. Of course, we love data center. But we just want to make sure we are ready for physical AI, ready for science AI.

It's a great place to end it. Unfortunately, we're out of time. Anurudh, thanks for being here with us here. Thank you.