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

All right. Thank you all for joining us here on Day 2 of the Morgan Stanley TMT Conference. I'm Mark Edelstone with Morgan Stanley. And before we start, I want to read a couple of safe harbor disclosures here. So the discussion here with Cadence will contain forward-looking statements. They will make use of certain non-GAAP financial measures. Please see Cadence's most recent 10-K, 10-Q and website for a discussion of risk factors and their use of non-GAAP financial measures. Also, please see Morgan Stanley Research disclosures with the social website at www.morganstanley.com, research disclosures. If you have any questions on that, please reach out to your Morgan Stanley representative. So with that, I'm really thrilled to introduce Anirudh Devgan, the company's CEO. Joined Cadence in 2012, became CEO towards the end of 2021, following in the large footsteps of Lip-Bu Tan here in the audience.

Cadence is just an incredibly well-positioned company in my mind. But kind of walk us through, since you took over as CEO and just maybe your observations over the past decade or so of being at the company, your overall strategy and vision for the company?

Absolutely. Good to be here, and thanks for your interest in Cadence. So we basically make products, mostly software products to design chips and electronic systems. So almost any chip design in the world uses some form of Cadence software. And then more and more, our customers are also system companies. So about 45% of our revenue is coming from system companies, like car companies or phone companies and things like that. So in terms of our strategy, the first thing is to make sure we are good in our core business. So make sure our core EDA products are best-in-class. And we have the most a diverse portfolio for analog chips, digital chips, memory chips, all kinds of chips, we have a pretty broad portfolio in our core business. So that's always #1, right? Without being excellent in core, you can't really expand. And then the second part of that is the system companies, okay? So like I mentioned, a big portion of our customers are system companies, and they are designing more and more silicon. So that's a new business for Cadence. And along with them designing more silicon, we are also selling newer products to them for what we call not just EDA but SDA, System Design & Analysis. So this is simulation products for thermal simulation, power simulation, electromagnetics. So that's a huge market. That adds about $8 billion to $10 billion of TAM to our core EDA and IP business. So that system can provide a lot of growth for us. So that's the second part, apart from our core business. And then the third growth area is all data and AI. So there's a lot of opportunities to not only help customers design AI chips, but apply AI or AI and optimization to improve productivity of our own products. Like these days, the chips are about -- they have like 100 billion transistors in them, right? These chips 1 inch by 1 inch. But by 2030, they will have 1 trillion transistors, and then they are a lot more complicated to design. So the chip complexity and system complexity is going to grow up by 30x, 40x. So there's a lot of opportunity to apply AI and get more benefit to our customers. So at a high level, we want to make sure we are good in our core business, which is EDA and IP. Expand into system, both silicon design and system design and analysis. And then apply AI to our products.

Right. So fantastic growth drivers and obviously overlap well with a lot of the secular themes out there in the marketplace. When you just step back and look at those 3 opportunities that you have, where do you think you're best positioned today? And where do you think there's still more work to be done to have the position that you want in the market?

Well, I think we are well positioned in all those 3 areas. I mean the good thing is not to depend on 1 type of growth drivers. So these are 3 big kind of growth drivers. I mean EDA, we are very well positioned in our core business, EDA and IP. There are several reports saying we are the largest core EDA company. But in terms of EDA, we still have a lot of growth in a couple of companies, big companies that traditionally didn't use that much Cadence. So -- and now we are engaged with them. So there's a lot of opportunity even in core EDA to grow. And systems, I think, is still in the early innings. In terms of system companies doing silicon, I would still say this is still second inning of a baseball game analogy because typically, the customers do some chips. And if it works, they do more chips, and then they do more kind of in-house chips. So it's still very early on in the systems journey. And our kind of simulation portfolio, we generated about $400 million of revenue last year out of $3.5 billion. So that's about 12%. But the market is huge. The market is $8 billion to $10 billion. So still, we are in the very early stages on the system side. And AI, we're just getting started. I mean even though we launched our first products 2 years ago, I mean some of the deployment takes time and it's going to be. So what I like about our position is that there are multiple growth drivers. And at the same time, we are very financially disciplined. So not only our revenue growth has improved over the last several years. It used to be about 7% revenue growth a few years ago. And if you look at now, our 3-year CAGR is 15% revenue growth, okay? And our margin used to be about 26%, about 5, 6 years ago. Now our non-GAAP margin is 41%. So there are only a handful of companies that can grow 15% plus and have margin of 40%. And our incremental margin is 50%. So if we generate $100 million more of revenue, we want at least $50 million of profit. And I expect that to continue in going forward. So not only we have a lot of growth drivers and revenue can grow, but we'll make sure that the EPS grows even faster given our focus on financial discipline, right?

Fabulous business model for sure. Looks, we may come back to EDA, but let's focus probably more on systems and on the AI side of things. On the system side, it's been my observation, and you know I've been doing this for 4 decades, is that as we've seen the complexity of semiconductors increase, at the same time, we've gone through a lot of consolidation in the semiconductor sector by itself. And therefore, it's scarce resources. And therefore, I think systems companies definitionally have to have more semiconductor capabilities. How do you think about their problems and challenges to go design these next-generation really complex devices versus that of a traditional semiconductor company? And where is that -- how does that opportunity sort of fit in with Cadence and your growth and sort of modeling around?

Yes. Absolutely, yes. I mean, when we were -- we are fortunate to work with a lot of system companies that are designing silicon. And what I would say is that probably 3 main reasons they do it. And this one is, if you do some custom silicon, it's better for something, either it's better in power or performance because you can customize more data based on what the application is. And there are a lot of examples even like when you watch YouTube when it loads faster, there is some video chip that's making it load faster. But if you drive a car, there's AI engine, which is faster than. So that's always -- there's a lot of room to do a custom chip. The second reason is for schedule, right? If you have a certain release every year Cadence of products, you want to control your own schedule. And third reason, which I think often sometimes overlooked is if there is enough volume, it's actually cheaper to do it in-house. And that volume may be different for different industries, but it's definitely due for cars and phones and computers and data centers. So for those 3 reasons, this is almost an irreversible trend. I believe we have crossed activation barrier of that. So there will be more and more chips designed. And the good thing there is that in those companies, the system companies, they're starting from scratch. So they're always open to new ways of doing things. And they are also resource limited at least when it comes to silicon design because they haven't done that in the past. So sometimes we do strategic services to get them going. They often are more open to AI-based tools. They also do more hardware platforms like Palladium and Protium because if you're a system company, you naturally have software stack. Otherwise, you're not a system company. And these days to bring up the software stack, you need these ambulation and prototyping platforms. So it's a big tailwind for the entire industry. And Cadence is a unique position because we are also leadership in advanced packaging, not just core EDA and IP, but advanced packaging and then all the system simulation tools we have developed over the last 5 years. So naturally, if a system company is doing silicon, whether it's a car or a data center company or a phone company, they are also doing the mechanical design, thermal design and electromagnetics. So there are software components. There are mechanical -- electromechanical components, all this is good for our business.

Right. It seems like just the overall kind of content that you might have with a systems company would just definitionally be higher, I would think as well, just because the efficiency that they could have it's probably hard to be as efficient. So take an example of an Intel that is going to be designing CPUs for a lot of different users versus someone like Amazon or Microsoft or Google, is going to do it for their own usage. Is that a fair thought? Do you think just your ability to penetrate those companies and sort of get more content of Cadence products and capabilities into those systems companies would just actually be higher?

I mean, our semi customers are very good also. I mean, some of them have evolved and some of them have become almost like system companies. I think the -- I give you example of growing up in India, it could barely get like a phone line. Only a few people where I grew up had like real phone lines, and there was like a 1-year wait to get a phone. And then when the cell phone came, right, they basically skipped over land lines and just went to the cell phone. So to me, I think that's a better analogies to the system companies because they don't have a lot of legacy of design. So when they start, they can start with the latest what is available. But I think some of the semi companies have also done very well in moving to that latest. But since you're starting from scratch here, you're typically more open-minded. And this also happens in newer geographies like China or Asia because they're designing, or even in India, they're coming up from, so they just go to the latest technology. So we have a greater opportunity to provide a fuller portfolio for that reason.

Yes. Right. How about just there's so many secular demand drivers out here. We're still relatively early innings of cloud. Obviously, AI is still very early, autonomous, everything. When you look at these type of opportunities, do you think that your growth rate could actually accelerate just because your opportunity set would seemingly be larger, bigger TAM, bigger SAM and so on. So obviously had great growth here in the past handful of years. But when you think about the opportunities ahead of you, which to me are probably greater than ever before, how do you think that overlays against your type of opportunity set?

Yes, it's a good point. And of course, we guide like 1 year at a time. But what we say is, like I said, the revenue growth went from 7% to. And we always like to look at 3-year CAGRs because most of our contracts are 3-year in duration. So last year, we had a fabulous year of 19% revenue growth. But if you look at a 3-year CAGR, that's like 15%. And I think what we aim for is sustainable double-digit revenue growth, okay, and 50% incremental margin. And I think with both those things, you can provide a lot of profitability growth and EPS growth. And we are also returning about half of our cash. We generate, of course, a lot of profit. So half of that back in terms of share repurchases. So revenue can grow double digits. Profit can go more than that because the margin always improves every year. And then EPS can grow more than that because some of the shares we are buying back. And so I'm confident in that long-term model. So double-digit is already good. Of course, we try to deliver better than that. But I think for the investors, that's already -- that's our base plan of double-digit revenue growth, 50% incremental margin and at least 50% of cash flow is buyback.

Right. Maybe we can shift to AI. Obviously, it's been around for a long time as we well know. But it's clearly seen exponential type of just excitement, I think, as it's gotten more into the consumer market and things like ChatGPT have become buzzwords and so on. But just talk about what you're doing your joint enterprise data and AI solutions so JedAI. Talk about that, talk about just overall bringing AI into your product portfolio. And how you think it's going to -- how the opportunity of AI really impacts Cadence in the future?

Yes, AI is -- it can be transformational. Of course, there's a lot of hype on AI. Like everybody -- everybody calls everything AI now. That's one issue though. So because what we used to call statistics or computer science a few years ago is now AI, okay? So I took one Stanford course on AI, is great, by the way. I'm from CMU, but I love Stanford. But half the course was on regression, okay, if that's AI, then I was doing that AI. And that we joke around that like now -- if statement is AI to like it's artificial and it's somewhat intelligent. And then if you use the L statement, than its reinforcement learning because you learned from -- so this is a little bit of like AI washing going on. But part of it's our transformational, part of it does transformation. And what I always believe now, we didn't call it generative AI 5 years ago. But what -- if you look at fundamentals, right, there is like 3 kind of sciences, this is like classical stuff now. My dad is a mathematician. So they only talk in like 100-year increments, okay? So the classical science is like geometry, right? That's like 2,000 years. So that's science of place. Okay. Then the next science is signs of space, okay? That's derivative. That's Newton, right? That's Calculus. That's 400 years ago. And then there is science of pattern. And maybe AI is science of pattern we have been waiting for, right? So these are 3 fundamental sciences and they can last for hundreds of years. Look at Calculus. We are still doing. A lot of our computers now is differential equation like discrete differentials. So I believe that AI can be transformational, but that doesn't mean you don't need like geometry and calculus, okay? Sometimes people apply AI to. So there are certain things you need, which are inside-out, like from the physical sciences of physics and chemistry and biology. And then data science is AI from outside-in, okay? So what I always believed is you need the internals to be good, and then AI can be used for optimization. So that's what we have done for the last 5 years. To give you an example, right, we have, for example, optimization tool, Innovus, which is one of the flagship tools. But it runs like in the history of our software, it's 1 run. Like you give it to input, it gets you a output. You can see a very good output. But if you look at the customer, they are not running at one time, right? They're running it over and over again. So typically, what the customer does, they run something, then they change and they run it again, they change, they run it again, okay? And Cadence and our industry never provided any automation in that workflow because always single run environment. And the reason for that is there was no mathematical way to transfer knowledge from 1 run to the next run. But now with reinforcement learning, you can actually do that. So we actually build -- we have a JedAI, which is our data platform and then we have AI tools on top like Cerebrus and Verisium and Optimality, and I'll talk to you more about that. So what that does in this example is instead of the human searching the design space, you do it mathematically using reinforcement learning. So over 200 runs in this 1 particular case. Instead of doing it, I think you can do it exhaustively, run all combinations. That would be 4 million runs, it's almost impossible. But if you do it mathematically with reinforcement learning, over 200 runs, and that takes like about 1 or 2 weeks versus 1 or 2 days per run in about 3x, 4x more the effort, you can get much better answer, and also much shorter. Because the human would take like 6 to 12 months to do that. So it's a huge advantage. Now -- but you still need the base engine, it's not like, and then you add AI for optimization. So the real value of AI to me, is to get away with the mundane tasks and focus on higher-value tasks. And that's what is happening in generative AI. So you can't -- if you're Shakespeare, no problem. But if you're writing a regular poem or an e-mail, you can write that with ChatGPT, right? So the real creativity will still be there. So what we can enable with AI is for the designers to move up the stack. And instead of 1 designer doing 1 block, 1 designer can do 5 blocks or they can do architectural decisions rather than -- a lot of the customers are running these TDL tasks with our tools. So that's a huge opportunity. So that's a huge productivity improvement. So I always believe that AI for optimization is the real value, which is what we are seeing now. Yes. Right.

And where do you think you are on that curve for your products that are incorporating AI to make your customers more efficient, more productive?

So we have like more than 30 projects, okay, for over the last 5 years. Because 1 good thing about -- so what -- see -- so before I answer that question, okay, one thing that's very important to if you realize is what is your core competency, right? So from a Cadence standpoint, what is the core competency? Our core competency is what we call competition software, that CS plus Math, Computer Science plus Maths. That's what my background is, that's why we have out of 10,000 people there, 9,000 engineers, 6,500 people in R&D, that's their background. So when you apply that to silicon, that's EDA and IP. Then you can apply to system, that simulation. And then if you applied that to data, that's AI. So we have a lot of expertise we can bring to AI. So for example, AI, the classical training algorithm in AI is non-linear conjugate gradient, for example. We have used that in Innovus placement for more than 10 years. So some of the algorithms can be -- are well known in EDA can be applied to AI. Now in our journey, in terms of R&D, we have a lot of capabilities. In terms of products, the first products were launched about 2 years ago. And the response is phenomenal, because of the benefit they can do. And then over last year, we had this data platform because it's very important to have the data platform, which we call JedAI. So we have JedAI as the data platform because like I said, in the old days, EDA tools and all these tools were built for a single run, like open access, which Cadence pioneered. But JedAI is for multi-run, not just 1 instance of the design, you can have multiple instances. So you have the data platform and then you have apps on top. So the key apps are Cerebrus, which is for implementation. Verisium, which is for verification. Optimality is for system design. So we have launched 3 big products. And we will launch 2 more this year for PCB design -- PCB and package design and for analog design. So I would say that in terms of products, relatively soon we will be very complete. And the adoption is good, but I think it will take multiple years of adoption as people evolve their design flows. But that's good, right? I mean, so, right.

Okay. Before I open it up to the audience, I think we can talk a little bit about simulation, because that's, I think, a really interesting area for Cadence. It leverages so much of your core competencies and capabilities. Talk about your vision behind what you're doing there? Maybe some of the acquisitions you've done so far? And maybe what else is kind of the opportunities you see ahead for that simulation business?

Yes, thank you for that question. So like I said, I mean, we look at the world in 3 concentric circles. So the silicon circle, the system circle and the data circle, okay? And a perfect example is like the electric car, right? You have all the navigation data, then you actually have the physical car, hardware, software, electrical, mechanical, and then the chips that drive it, right? So when we do computational software for silicon, that's EDA, right? When we go to the system level, the next level up, there's a lot of software business. Actually, there's about $50 billion software kind of for systems. But for us, we want to do the computational part of it because that's our core strength, and that's what is going to grow. So the computational part of the system business is system simulation. Things like finite element, CFD, electromagnetics, designing cars and planes and thermal. So that's about $8 billion to $10 billion market. And that's why about 5 years ago, we invested significantly in that space, okay? And then there is a coupling of like 3D-IC, also couples the silicon and system together. And we are doing pretty well. That's growing like 27% last year and is profitable. And our advantage is that not only we can couple semi and system together, but the algorithms in EDA were always much more complicated because you're designing like chips with 100 billion variables and that Moore's Law is changing every 2, 3 years, where the system side has been more static and less kind of innovation. So we can build that R&D intensity we have on the chip side and apply it to the system side. And there's a lot of potential to couple these things. And it's going to happen. I mean, this merger of silicon and system you see happening now. I mean we saw that in the 5 years ago. So the question is who is going to do that? I think it's very difficult for the system companies to get into EDA, whereas we're assuming downstream when we go from EDA to system companies. And the customers want that, right? You remember like a few years ago, you go on the plane and you couldn't take your Samsung phone because it was melting or something like that. So that's the connection of silicon and system and still that problem exists. So I feel very good. And again, I think the other good thing about simulation of this $8 billion to $10 billion market out of the $50 billion market is simulation, always you can have multiple simulators. And our competitors in that space -- wish them well anyway. I think we are much better, but you can have multiple choices. There are some other parts of the system market like when you do like design of buildings or design of -- mechanical design, those are more UI-based products like PowerPoint. So then you don't want multiple choices. So we don't want to do things which are not computational. But the simulation is computational. The customers will have multiple choices. And the other thing I liked about system simulation is that -- so there is R&D synergy, because of the math that is similar. There is customer synergy, 45% of our customers are system companies, and the margin profile is even better than EDA because we want to grow in areas that have good margin rather than other way around. And -- so we started with simulation. So we have about $400 million in revenue. The other thing that is interesting in simulation is that EDA, what people don't realize, 1/3 of EDA is simulation because we do a lot of optimization like place and route, but you have to simulate first before you can optimize. And EDA simulation is also very profitable. That's the big -- the most profitable part of EDA is always simulation. But the other thing EDA or chip design is good at is optimization. This is the history of chip design, right? Whereas there is very little optimization in system design. Because they could barely stimulate like a wing or a car -- we have this partnership with McLaren F1. By the way, that makes me very popular at home. So we have this partnership with -- I visited McLaren, a beautiful car, and we are doing all the CFD simulation for them, right? And the range of the car and the speed of the car, it depends on how good the shape is. And if you look at those fins in the front, they have very sophisticated fins or a lot of them are designed manually. Because first of all, you can barely simulate the whole car, then you don't do enough optimization. Whereas with now optimality and AI, the EDA or chip design has a huge history of optimization. So we can automatically design those things. So I think that can also provide a huge benefit to the system simulation side is coupled that with optimization.

Yes. Yes. It's huge opportunities. It leverages our core competencies against these massive growth drivers in the marketplace. Let me see if there's any questions in the audience. Let's get there's a mic right there behind you.

Thanks for the opportunity. How do you think about China? How do you think about the way you kind of deliver product there? How do you think about it in some sort of continued battle with China, their ability to kind of compete and your ability to provide solutions?

China is a great market. And if you look at over the last 5 years, it has done pretty well for us, right? And we expect it to grow well. I think -- and they have a big investment in semiconductors. They have more than, I don't know, 1,500 or 2,000 companies. That's 1 other good thing now is that each country is investing in semiconductors. Even India, okay, and Israel and EU and U.S. So I think China is going to be a good opportunity. Now there is some -- we follow all the U.S. regulation. I think more of the issues in China is the U.S. has some regulations. But those are mostly targeted towards manufacturing, and we work with the government carefully. And so they're not as material to us on the design side because a lot of design in China -- even like there were some reports yesterday, I think the U.S. government is trying to restrict certain very advanced things for military use and other things, but a lot of the design is cell phones or TVs and electronics and washing machines. So even that's not even the intention of the U.S. government to effect. So all that requires design of chips and design of electronics. I talked about the same things that are true in U.S. and Europe are true in China with merger of system in semi. And then the other question we get asked is, is there local competition? So we are used to competition, right? So there is some local competition, but we are investing significantly in R&D, right, about 35% to 40%. And if a system company is designing chip, whether it's in U.S. or Europe or China, they want to use the best technology, right? So as long as we have good products, we will do well in China.

Can you just size the China revenue and not from global companies, but just the indigenous companies in China today for you?

So our China revenue is about 15%. And most of it is local. I mean, some of this is multinationals there, but yes.

Okay. Great. We're just about out of time, but you've covered a lot of great ground here in terms of the capabilities of the company and really how it overlays against the opportunity set in the marketplace. But -- when you -- and you've obviously spent a lot more time as CEO now in the past year and change. What's the one thing you think that investors just under-appreciate or just don't know about Cadence in terms of your capabilities and opportunity?

That's a great question. I think we -- what I would say is that investors and public at large don't realize how essential we are to the whole semiconductor and electronics ecosystem. What I joke around sometimes is that if we needn't to have our kind of software, we'll all be riding horses. It will be like Game of Thrones or something like that. It will be cool, but it won't be as productive. So I think what people don't realize we are essential to the design of chips and electronics. And we have the luxury to participate all over the world in all verticals. So it's having like -- I really love my job having like a front view seat of all the innovation happening globally, all the major players in the industry. So that's a great position. Yes.

And married against the business model that gives you phenomenal visibility versus almost anything else out there in tech land. Anirudh, thank you very much.

Yes. Thank you. Thank you for the opportunity.