We're gonna get the next session started. Thank you. I'm Tim Arcuri. I'm the S emiconductor Analyst here at UBS, and very pleased to have Cadence with us. We have Anirudh Devgan, who's the President and CEO. Thank you, Anirudh. Before we start, I have to read a statement on behalf of the company. Today's discussion may 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. For information on factors that could cause actual results to differ, please refer to our SEC filings, including most recent Forms 10-K and 10-Q. All forward-looking statements during this meeting are based on estimates and information available as of today, and Cadence disclaims any obligation to update them. With that out of the way, thank you. So as a first question, maybe we can just start with having you give a brief overview of Cadence. What is EDA? I think most people in the room may or may not know what EDA is. You're not a semiconductor company, you are a software company for semiconductors. So just explain EDA. Where does it fit in the semiconductor and the broader universe? Yeah, thank you for the question, and thanks for joining today. Thank you for your interest. So Cadence, basically, we make software to design chips and electronic systems. So almost any chip design today uses some form of Cadence software, and these chips could be, you know, for any application, you know, any vertical and in any geography. And so this is kind of what I would call a computational software. This is very mathematical, numerical software, and these chips cannot be, as you may know already, cannot be designed by hand. They're all designed by software. And then recently, we have expanded in the last few years into not just making software to design chips, but also making software to design systems, because there's this merger of semi and systems. You know, a lot of system companies are semi companies, semi companies are system companies. EDA is for IC design, and then our new initiative, what we call SDA, System Design and Analysis, is for system design, and we can talk more about that. Great. Since you took the helm of Cadence two years ago, the company's grown by leaps and bounds. What's your management philosophy and secret ingredient that's driven that growth? Wow, that's a deep question. Well, I think it comes down to several things. I mean, first of all, you need to have the right strategy, and I can talk more about that. You know, you know, core, make sure you are number one in your core business, and then grow synergistically in areas that are synergistic from a customer R&D standpoint. Also, of course, execution. You know, any strategy is no good without the right execution and the right team culture, you know? We wanna make sure we have the best team, we are aligned with the winning customers, you know, the customers that are driving. And in the end, right, we are in the technology business, so we have to have the best products, right? So I think culture is a very important part, and it's something that we have worked on over the last 10 years to really have a true... You know, a lot of people talk about it, but to really have an innovative culture and then execute and align properly with the customers. Yeah. And so what's been the primary driver of the recent growth? Is it the proliferation of the customer base? You know, it used to be that the customer base was consolidating. Now it's actually expanding again because there are more different companies that are designing semiconductors. We're seeing a massive explosion in the, in the number of companies who are designing semiconductors. So is that the biggest reason for the growth the past couple of years, or is it something else? I mean, there I would say, like, probably three main reasons for growth. One of them is, like you mentioned, like, more companies doing silicon. So right now, about 45% of our business is coming from what we call system companies. And you know, and 55 is coming from traditional semi companies, and they're both growing, but this, the percentage of system companies has increased, and also these are a lot of the household names now doing silicon. So this whole move into, into system companies doing silicon and us selling more system products is, is a big reason for growth, and I think we are still in the early innings of that. I mean, we are still, in like baseball terms, probably in the second or third inning, and we can talk more about that. There will be more and more domain-specific silicon, and our job is to make the design easier and easier for both semi and system companies. The second thing, which is like a turbocharge to Moore's Law, I mean, we benefit from regular Moore's Law anyway, but one big trend, as you know, is chiplets and 3D -IC, which is accelerating Moore's Law, putting more things in a package. So Cadence has, you know, majority share in package design software, like 80% share, and then the whole 3D -IC solution is very, very well positioned, both from the IC side, the package side, and the analysis. So I can talk more about it. So 3D -IC is a big trend. And third, of course, is AI, you know. A lot of design has been happening and is happening, building the AI infrastructure and also applying AI to our own products. So I would say it's a combination of systems and 3D -IC and AI that's driving the growth for Cadence. And, speaking of AI, you have introduced a complete suite of AI-enabled tools across all your platforms. Can you describe those tools and maybe how your AI portfolio differentiates you from your competition? Yeah, absolutely. I mean, the good... AI can be quite transformative. Of course, one thing right now is everybody calls everything AI, so you can't-... But I think in terms of our thing, we have been doing this for some time. So for effect on Cadence, there are three big ways that AI can affect us. So first thing is, you know, a lot of the, you know, AI infrastructure, we are working with all the leading companies that are building the infrastructure. So these are the GPU companies or the big data center companies. They're all very, very deep Cadence partners, okay? And some of them we can talk publicly, like our partnership with NVIDIA, some of the things we, we call Tesla, but some of them we can't talk publicly. But in general, we are involved with all the major companies that are doing, you know, AI infrastructure, and that's a pretty significant portion of our business and growing. The second way the AI can benefit us is, applying it to our own products. So we have, like you mentioned, a whole suite of tools, and, you know, these are true GenAI, you know, reinforcement learning-based, a new way of doing optimization. I can talk more about it, which we launched about two years ago now, you know, well before the whole AI super hype cycle. But I think they can be quite transformative in optimizing the overall workflow, and I can give some examples later. You know, we are seeing like, you know, order of magnitude improvement in productivity, you know, 10%-15% improvement in performance, which is significant. Okay? Normally, when you go from one node to another, like from seven to five, you may get, like, 10%-20% improvement in PPA, you know, power, performance, and area. These AI tools can give half of that or almost that from better software. So the value is huge. So that's the second part of it. Then the third way we can benefit AI is, like, of course, AI is gonna have new applications that are gonna emerge, not just our current products. There could be a lot of them. You know, there's a lot of discussion on what will be the biggest new application of AI. So in my mind, the biggest new application has to be life sciences, okay? So we last year acquired a biosimulation company. You know, I'm super excited about this whole design of molecules and using AI, and a lot of the techniques are very similar to design of chips and then digital twins, okay? Whether it's molecules or physical systems, and we have a lot of partnerships. We also made acquisition and like digital twins, we have a digital twin for a data center. So that's the way we look at AI. One is the AI infrastructure, second is applying AI to our own chip design tools, and third is new markets in, in life sciences and digital twins. As you look at your outlook into next year, roughly 85% of revenues are recurring and the visibility there is very good. But, you know, macro is still a bit challenging too. So can you help us with kind of the puts and takes as you see them, as we look into 2024 and beyond? Yeah, we are not, of course, providing actual guidance for 2024. We will do that in our call in February. But in general, I think what you have to remember about Cadence is... And we talked about growth, but also we focus not just on growth, but also on margin, okay? So if you look at last six years, we have delivered, you know, double-digit revenue growth and more than 50% incremental margin. So right now, our—like this year, our growth is 15% revenue growth and for almost 42% operating margin. So people talk about Rule of 40 being best in class. So this is Rule of 57 right now, and we have done this consistently over several years, okay? The other thing, you know, so we are very disciplined, not just from the technology side and all the new trends, but also, of course, financially, we have to be very disciplined. And the other thing I watch carefully in terms of margin is not just operating margin, but operating margin after we take out SBC, stock-based compensation. So we put that in our disclosure. You can see what the SBC is, because some of the software companies, they may have margin, but then they also have high SBC. So we are pretty disciplined in SBC as well. So we have grown margin last several years. We have grown revenue last several years. We talk about double-digit revenue growth and at least 50% incremental margin. So we have done that for the last six years, and I don't, I don't see any reason why that won't continue going forward. So our goal is, of course, give revenue growth, but also more importantly, EPS growth, okay? And we also, you know, over half our cash flow, which is roughly this year $1.4 billion, half of that we purchase our shares. So the share count is also going down. So the revenue goes up, the margin goes up, and the share count goes down. So together you can have good EPS growth. So I think last two years, our EPS is up, like, 55%. Yeah. Right. So you know, you'll surpass 50% incremental margins this year, despite the fact that you had you know, picked up a bunch of expenses from acquisitions. So as those roll off, you should certainly be able to expand incremental margins. Yeah. So, I mean, last few years, our... You know, we promised 50% and up. I think last few years it has been more like 55%, okay? And our current margin is close to 42%, like I said. So, so there is still a lot of runway to keep improving margin and of course, keep improving revenue growth. So we are pretty disciplined that way. And the reason for that is, I mean, everybody would like to have this kind of position. The reason, first of all, is to have technology leadership and, and high barrier to entry. So it's very difficult to design this kind of software. So there's not many companies in the world who can do that, and we also have the most complete portfolio and also aligned with the future trends. So we will make sure we continue on that. Yeah. ... Can you talk, you know, one of the biggest boons for your business has, has been the slowdown in, Moore's Law, actually. And there's a growing interest in the, you know, 3D -IC and chiplet designs. Maybe you can provide us a status update in terms of that, that trend and, how, how you're positioned to, to serve that trend. Yes. So I mean, Moore's Law is like big debate, right? Some people say it's going strong, some people say it's already, you know, done. I remember, like even 20 years ago, we used to argue that it's kind of slowing down, right? So if you look at last, because the classical scaling, the Dennard scaling, you know, Bob Dennard is the one who wrote the original paper on transistor scaling. So according to the actual Bob Dennard scaling, you know, it's probably done 10 years ago from the traditional scaling standpoint. So what has happened in the last 10 years, which you may know this already, of course, all this is, it's not pure Moore's Law, it's more, area scaling. So the chip—the number—the CPU may not get faster, like a mobile CPU may be at 3 gigahertz, but instead of one, you have, like, four of them or eight of them. So when you go from 14 nanometer to 10 nanometer, you know, effectively the number of transistors doubles, and that is continuing. So the 10 to 7, to 5, to 3, to 2, to 1.4 to 1. Okay. So that's 30—that's 2x improvement. So that will go on for next 10 years, okay? So the area scaling is continuing. So if area scaling continues, then there's more design, and you can put more things on the chip, but the things are not necessarily getting faster. This is the problem. That's one reason that parallel computing, you know, I knew this from early 2000, so we invested so much in parallel computing for last 15 years. And the same reason why GPUs are much more efficient, because you can put more of them, and so you can do a lot more parallel computing for AI or other things, okay? So I think at one level that will continue, but the gains will slow in terms of performance, okay? So one way to extend that area scaling is, instead of just scaling the chip in one chip, you know, we can put multiple chips in a package. Okay, that's another way to do area scaling, and I think that will really take off, and it has taken off last few years. You know, we used to argue even in the late 1990s, that this- it should be not SoC, it should be SiP, system in a package. But it didn't happen because SoC had a lot of runway at that point, okay? Now that the runway reduces, you know, we have this emergence of SiP or 3D -IC, okay? And there are a lot of other benefits, as you know, is that if you have 6 chips in a package, you don't have to redesign all 6 of them. If you go to the next node, you know, you redesign only a few, two or three of them, or you can mix them in different nodes and all this. But all this is driven by basically package design complexity and, you know, aiding Moore's Law with 3D -IC s. The good thing about Cadence in this is that we are the leader in package design software. So about 80%, we have 80%, you know, share. Allegro is the franchise tool for package design used by most customers. And then the real thing to do 3D -IC properly is, you know, there are three layers to that. One, the bottom layer is, of course, the IC design. So we have franchise tools in analog and digital design. You know, Cadence is number one in, you know, analog and digital, so that's Virtuoso and Innovus. And then package design is Allegro, which is a franchise tool for package design. And then the third layer you need is analysis tools like thermal and electromagnetics, which we started about five years ago. So because thermal is a big thing. So if you have all these three things: analysis tools, package design software, and chip design software, that's the complete 3D -IC flow. So if you look at 3D -IC flow announced by all the major foundries, I think, TSMC announced last year, 3D blox and, you know, the other foundries. So most of them are based in collaboration with Cadence. So I feel that we are pretty well positioned as this 3D -IC becomes the dominant flow going forward for multiple reasons. You know, extension of Moore's Law, it's much more cost-effective to do that because you don't have to design everything at the advanced node. Just like, you know, 10 years ago, or you know, 11 years ago, when I joined Cadence, you know, Cadence was always strong in analog, but like 11 years ago, it was not strong in digital. Now, of course, we are very strong in digital. So I feel the same kind of position now. Because of Allegro and package design, we can be the lead developer for 3D -IC platform, and that's what I see playing out in the market. Can you talk a little bit about the IP business? With the recent Arm IPO, it's getting a little more focus from folks. Mm-hmm. It had a tough first half of the year, but it's been better in the back half of the year for you. What's driven that improvement this year in the back half of the year? And, and maybe can you, can you talk just about your IP business? Does AI drive more growth on IP? Yeah, that's a good question. So in general, you know, IP is good but not as profitable as software. Just because, you know, you can, it's a build versus buy decision, whereas for software, it's a buy decision. Even the biggest companies in the world don't write their own EDA software. So IP, you know, sometimes they can't develop IP, they buy it. Sometimes they can develop IP, but they buy it as they outsource it, right? So Arm is slightly different because Arm, you know, typically most companies don't do their own CPUs. Some of them do, okay? So we have a part of our IP business, which is Tensilica, which is DSPs and CPUs, which is more and similar to Spirit to Arm. And that's a very good business because that's a, it has royalty, it has, like, good margins. It is also soft IP, so you're not doing the back-end implementation. The customer is doing the back end. Now, the other part of IP business is physical IP, like DDR, PCIe, and all that. Now, that's also good, but it's not as profitable as EDA, but still can be very good in certain contexts. So we have focused on IP, on these kind of high-growth areas, and AI is a big part of that, you know, with this disaggregation of the chip into chiplets. So, you know, like, UCIe, which is the IP to connect the chiplets together, is becoming very important. SerDes, you know, DDR, also HBM. We recently acquired the physical IP assets of Rambus, and that came with HBM, which is critical for AI applications. And so IP is good business, but it's lumpy. I think the first few quarters were weak, and then Q4, we're expecting very good growth in IP. In general, I think it can grow well. We just have to be careful that we are investing in the more profitable parts of the IP. Yeah. You talk about EDA hardware, which is actually getting more attention from folks. Why has it become so important in chip design, and what types of designs does it benefit the most? And is this a sustained trend going forward? Yeah, so what—just to clarify, what we mean by hardware, EDA hardware is... Like, these are emulation systems. These are like supercomputers we build and sell, which will run EDA software much, much faster. So, for example, you know, we have simulators, like logic simulators, that will run on, like, Intel or AMD CPUs, and they do pretty well, and customers will run, like, millions of them or hundreds of thousands of licenses. But Palladium, which is our hardware box, this is a hardware-software combined solution, will run, like, maybe 1,000 times faster than a, you know, x86 CPU because this is special-purpose system designed to accelerate, you know, simulation. And what that allows is not just simulation, that allows what is called emulation. So you can actually, even before you have silicon, you can emulate the behavior of silicon using Palladium, and that's the leading platform in the industry and for all these AI chips. You know, by the way, NVIDIA is our development partner for Palladium for, I don't know, 2 decades, and all the other, you know, big companies use Palladium for these kind of... And the reason it became indispensable is that, you know, in the old days, you know, you would do silicon design, let's say this is silicon design, and then you would do software design, you know, bring up whatever application. So if this takes, like, two to three years, this take two years, so total, it's like five years to build that, which is too late anyway. You know, and so what customers want to do is they want to overlap hardware and software development. So they want to do software development before you have actual silicon, because otherwise it takes too long. And a perfect example is NVIDIA, right? They announce a chip, and then three months later, you can buy it from, you know, Supermicro or something, and, you know, everything, all the software works. So the reason it works is that during the development process, they are emulating everything on a Palladium, and they can boot software even before the chip is taped out. So we have this, you know, room full of Palladium boxes that will emulate the actual physical chip. And this is to, one, to make sure that the chip actually functions correctly. Two, is to do software development. And so, like we talked about in the beginning, all these system companies doing silicon, by nature, a system company has software, right? Whether it's YouTube or WhatsApp or some TensorFlow. So all this is emulated on Palladium before the silicon is designed. So that's the reason that there is more and more demand for these hardware systems or these emulation platforms. And we have, you know, of course, we work very hard to make sure we are best in class. As more and more companies do silicon and the silicon chips get more complicated, you need more and more of this hardware capacity. It, it's almost impossible now to design advanced chips without these emulation platforms. And we have done this for, I don't know, more than 25 years, and we do our own chip. So in that sense, we are a vertically integrated... So we make our own chip, this special, and these are like water-cooled, you know, super high-end servers that we will sell either as outright purchase for the big companies or for small companies, we also have a cloud model for them. Yeah. Can we talk about China? It seems to me that if the U.S. really wanted to shut down China's semiconductor efforts, Mm-hmm. EDA is a drop-dead way to do it. Can you just talk about where you fit and how the restrictions have affected you, and maybe how you see your business evolving in China, given the restrictions? China is a good area for us, and I think it has grown over the last several years because there's a lot of design activity in China, as you know. And, you know, that way, you know, we participate in all regions of the world wherever there is design activity: China, India, you know, Israel, of course, all these unfortunate events recently, but Israel is a great hub of design activity, Korea, Taiwan, of course, U.S. and Europe. And, you know, China has a lot of, of course, there are a lot of people and a lot of applications. So one thing about Cadence is that almost any, you know, any kind of application you need our software, whether it's like washing machines or TVs or phones or. So I think some of these restrictions are right now, and what I expect going forward, are targeted to a certain area, okay? So they're targeted either to some, you know, high-end AI applications, okay? Like you all know, or it is targeted to some manufacturing below a certain node. So I think it's more effect to the manufacturing companies. From a design standpoint, I mean, there's a lot of design happening in China. And then even if some AI companies or quantum companies are affected, you know, we are such a broad portfolio that, you know, all that is already included in our guidance, the effect on. The biggest effect for us on regulation was a few years ago when, you know, Huawei or HiSilicon used to be a big customer that was affected. I think now the U.S. policy is relatively stable, and it's more targeted, which is the right way to do it. So we'll see how that goes. But they, I guess the evolution of that question is that, if the—I mean, the goal is to basically add cost to China and keep them two, three, four years behind the U.S. I mean, the easiest way, sure, you know, you can restrict equipment, sure, but there are other parts of the world that produce equipment that are arguably more important than what you can buy in the U.S. So you don't have full control over equipment, but EDA, you do. If you're the U.S. No, I mean, it's easier said than done. First of all, you know, the software is very difficult to control. Right. There's piracy, okay? So you may not achieve the goal, and I think that's what... And also, that these are used for all kinds of applications. So you cannot say to 1.4 billion people that you don't have access to semiconductors, okay? Because there is all kinds of retaliation that can happen on the other side. So I think in a way, I don't want to get into policy debate, you know, that's not my—but I think what the U.S. is doing is a thoughtful way, and we support what the effort U.S. government is doing. We support China government, all the other governments and... But right now, I think the policy seems relatively stable for last few years. Yeah. Can you just talk about... I mean, I look at the stock, and the stock trades at an impressive multiple, and when you talk to investors and you hear the questions they're asking, what's the single biggest thing that, you know, not that you're selling the story to investors, but what's the single biggest thing that even at this multiple, even where the stock trades today, that you think people would be missing if they didn't buy the stock here? Well, one way to look at it is, you know, just look at our performance last, whatever, five, seven years, 10 years, right? And I can—I feel confident that our position now in the industry is better than it was 5 years ago. And so we are just accumulator of, of value, and we, you know, we want to, of course, share that growth with our investors. So like I mentioned earlier on, we will have revenue growth, we will have margin expansion, and we will, you know, continuously reduce share count, right? By reducing at least 50% of cash flow back to investors. So from an EPS growth standpoint, you know, I think last several years, we have grown, like, 20%+ EPS. Okay? And this is a high-quality operation going forward. So I think if... We are less volatile given the critical nature of our products, that we are very diversified geographically, and we are very diversified in end markets. You know, you know, it's not one particular, even though we can benefit for some market doing well, like AI, automotive, but we are very diversified. So it's, I think what you get from Cadence is consistent performance, compounder of value over the next five years. So you can, you can buy it and sleep well at night and say it will grow every year. I think that's what we deliver. We don't try to be, you know... That's, that's, that's what it is. Consistent performance is our goal. Yeah. Maybe one last quick question: so biosimulation, Mm-hmm. It's a $2 billion TAM, roughly. Should we—how should we consider that, the growth in that business- Mm-hmm. -going forward? Yeah. You see, bio, I'm super excited about it, you know, because, see, there are these big mega trends. And then, like, five years ago, I started all this effort into system design and analysis, you know, all this, and which is now you can see, you know, we are pretty. We are making more than $500 million in SDA. Also, all these system companies are doing silicon. Our portfolio has expanded. So if I look at next 5 to 10 years, you know, I think life sciences will emerge as one of the big beneficiaries because this has been our goal for the people in the industry is to... You know, what we do, you know. Sorry, what we do is computational software. What is computational software? Is computer science plus math at the basic level. And what we have applied that to, in if you look at semiconductors, is to physics, right? Last 40 years. So this combination of physics, computer science, and math has done marvels for the entire world population, right? With semiconductors, the design of all the software and AI. Okay. So like I mentioned in the past, the AI, the biggest application will be life sciences, okay? The biggest future application will be computer science and math, which is computational software applied to biology and chemistry. And the way to do that is design of molecules. So to do the design of molecules, the first thing you have to do is simulation of molecules. This is what we did with transistors. You know, first thing in EDA was simulation of transistors, then the design of transistors. So I feel that it is important to point out that there will be more and more computer science and math applied to biology. Life science is a very big area. What I choose to focus on is a pre-FDA approval. I mean, there will be a lot of things post-FDA approval. We don't want to get into that, okay? But we want to focus on more computational part of biology, which is drug discovery and simulation. This is a controlled bet, right? We'll see how it... The good thing is we are getting about $40 million of revenue, so it's a meaningful part. Not too small, not too big. I think it can, at the base case, it can grow faster than Cadence average because there will be more and more simulation needed, just like simulation is needed for thermal and power simulation. So that should be the base case. Now, the best case or better-than-base case is if really this thing happens, that AI and computer science and math revolutionize life sciences, then it can really take off. So it's like a controlled bet, but I believe that molecule, you know, design of molecules and life sciences will become a huge thing in the next five, 10 years, so. Great. Thank you. We are out of time, but thank you again. Yeah. Thank you. Yeah.