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

Great. Hello, everyone, and thank you for joining us today. I'm Joe Vruwink from the vertical software team here at Baird. Our next presentation comes from Cadence Design. Cadence is one of the leaders in the EDA industry. They provide software and, in some cases, hardware that really powers a lot of the innovation taking place in the semiconductor and electronics industry. Very happy to have with us today Anirudh Devgan, who's President; John Wall, Senior Vice President and CFO; and then from the Investor Relations team, Alan Lindstrom and Peter Jordan. I think maybe just to kick things off, a brief overview of Cadence would be good. And maybe in doing that, you can touch on some of the key pieces of what you define as the Intelligent System Design strategy of the company.

Thanks, Joe, and thanks for having us at the Baird conference that we always enjoyed every year. Before I begin, I need to mention the safe harbor statement. Today's discussion will contain forward-looking statements and make use of certain non-GAAP financial measures. So please see our most recent 10-K, 10-Q and website for a discussion of risk factors and our use of non-GAAP financial measures. With that out of the way, thanks for letting me say that. The -- yes, we're -- Cadence is a technical software company, largely in the EDA space or electronic design automation. Essentially, we're a company that was created by engineers for engineers. Our customers are chip designers spanning multiple verticals, including mobile, hyperscale, aerospace, data center, automotive and AI with a rough breakout of maybe around 55% of our revenue coming from traditional semiconductor companies and maybe around 45% from systems companies. Perhaps the best way to think about Cadence is we're a technical or vertical software company, largely in the EDA space, that provides software tools, intellectual property, hardware verification platforms and services to semi and electronic system companies. We have an IP business and a hardware business that -- where we get revenue predominantly on an upfront basis. That's just about 10% to 15% of our total revenue. Most of our revenue is recurring in nature. We have about 85% to 90% of revenue that's recurring. And that means that we typically know 90% of our next quarter's revenue before we even start the quarter and 75% of the next year's revenue before we start the year. Average contract length is about 2 to 3 years, and we execute against our Intelligent System Design strategy, the creator of which is here with us, Anirudh. So maybe best to hand over to him and let him talk you through that.

Thank you, John. So just a brief overview. Like John said, we are a technical software company or what we call computational software, which is like numerical mathematical software. And we have a lot of experience in that over the last 30 years in like chip design. So that's the core business of ours is the software for chip and system design. And then we try to extend from there and excited to grow from there. But our core strength is what we would call computational software. So we believe we are best in the world for this kind of stuff like mathematical solvers, matrix inversion, Boolean algebra, all these kind of good things. So we supply software for chip design. So the next adjacency around chip design is, of course, system design, and there's a lot of activity. And even John mentioned, we have 45% of our revenue coming from system companies. But they want more than chip software. They want software to design their systems. And that's about a $50 billion market. EDA is about $10 billion. System is about $50 billion. But part of the system market, which is more computational, is simulation -- system simulation. And we think that's a good adjacency for EDA to expand into, and that's about $7 billion, $8 billion market. So over the last couple of years, we have expanded into system simulation, and we can talk more about that. It's going pretty well. I believe we have good growth rates and a lot of new products, and we did several acquisitions. And that area is also growing. And it's more horizontal in the sense that it has more vertical applications. So definitely in the semi space, but also other verticals, all the way to medical. So we are excited about system simulation. So that's the second part of our strategy. The first part is the chip design silicon part. Second part is system simulation. And then the third part is the whole AI data analytics, which is also computational by nature. So that's the exciting thing. Even the general purpose software market is moving more towards AI and data analytics, which is inherently computational. So that's intelligence, which is AI; system, which is system simulation; and design, which is silicon design. So that's how we get Intelligent System Design, the 3 circles. The core circle being silicon design; next being system simulation, system analysis; and third, being AI and data analytics.

That's great. I want to mention if anyone in the audience has questions, you can use the prompt in the webcast, and those will be e-mailed to me. So Cadence today is not the Cadence of even 5 years ago, certainly not 10 years ago. There's been an incredible amount of new investment change and ultimately improvement in market share. Let's leave the system analysis strategy to the end because that's going to be a multipronged set of questions. But maybe in core digital design and also, I think, what's happening in verification, maybe you can touch on some of the strategies that have helped really rebuild and get Cadence to the point they're at today, where you're involved in a lot of advanced development work again?

Yes, definitely. And I've been here, what, 7, 8 years now, and my first job was digital verification and now this. So I can walk you through what we tried to do over the last few years. And so I think if you go back like 10 years, we were not really present in the top companies in digital. Digital -- analog, we've been doing analog for 30 years. Cadence is a leader in analog. But -- and we have a lot of customers. So one great thing is that Cadence always had a very, very wide customer base because of analog. There's a lot more companies. But the digital companies are a few, but they are, of course, dominant in their field. There are these household names. So going back like 5, 10 years, we had to do well in digital. Otherwise, it's not a good thing. And I think we have kept our strength in analog, and we really added our strength in digital. So we are now in 19 of the top 20 companies and a lot of the household names we're doing pretty well in. And I think that is also evident in our revenue numbers in digital. They have grown well over the last few years. And the other good thing in digital now is that a lot of new entrants, a lot of system companies, whether it's data center or automotive, they're doing digital and mixed signal chips. So I feel pretty good about our position, especially at advanced nodes. So I think below 16, we started doing pretty well. And some of the lower notes like 7 and 5, we have a lot of market share. And most of our R&D effort right now is at 2-nanometer. The other good thing that people keep asking is going back 10 years, is it going to be end? Is it going to end at 65? Is it going to end at 28? Is it going to end at 7? But fortunately, all the smart people working in this industry keep extending it. So right now, we definitely -- a lot of production is at 5, some customers are at 3, lot of R&Ds at 2. And there will be at least 1 more node after 2, at least. Okay. So that's already like 5 to 10 years of like pure silicon. And then beyond one, maybe there may be other structures like some of the foundries talk about. So I think Moore's Law in that sense is continuing. And then the other thing I want to mention here, which a lot of people already know is the other exciting thing, is all this disaggregation at the package -- this chiplet-based design, what people call 3D-IC or a 2.5D-IC, which is interposer and chiplet. So one thing I do want to mention is Cadence is in a very good position because we are the leader in package design for a long time. So Allegro is the most widely used package design tool in the industry. And this whole thing of IC and package coming together, and we launched a new tool actually yesterday, Allegro X, which is next generation of Allegro. So that also positions Cadence well with now analog, digital and packaging. So that's on the digital side, and I can tell you more about that. In digital, what we need to do, though, better going forward is there are a couple of customers that historically, going back 10, 20 years, Cadence hasn't done as well. So we have a lot of effort to improve our position in these 2 companies, which are non-TSMC company, okay. Now on verification, actually, we had a good year last year. And we expect a good year this year. And a lot of it is driven by hardware, okay. We have a lot of investment in hardware. I can talk for a long time. And we have Palladium. We are market leaders. And we have a new version of Palladium that was just released in April, Palladium Z2. And we have Protium, which is for software bring up. That market is exploiting with Protium. And we never had a good solution for software bring-up. But now over the last few years, we have Protium -- now Protium X2. So that is driving good growth. We had record hardware year last year. We had a record quarter in Q1. And so I feel good about our verification business. And I think you will see us take more -- I mean, one, the market is growing and we expect to take more share in verification going forward. So I'll stop here, Joe, if there are any questions I can answer. Yes.

Well, maybe I was going to ask about this earlier, but you brought up the new Allegro product that was introduced yesterday. And great timing because there's a specific capability of MCAD-ECAD integration within Allegro X, which is the topic of the week given Autodesk making a bid for Altium. Maybe just address strategically why this convergence between what typically have been totally separate undertakings -- why is this coming to a head now? And how is Cadence positioned to capitalize on that?

No, it's a great question. And I saw all these reports. And by the way, we are -- like I said, Allegro is very widely used. So if that companies were X billion, you can assume how much Allegro is. But -- and I think we've been doing this for a while. I mean there is like technology synergy I talked about, which is computational software. And then there is customer synergy, which is driving this convergence of electrical and mechanical design. And I mean, you just have to look -- I think the best example of these 3 circles I talked about is a car like Tesla, I think. Of course, this, as we all know, Tesla is the best example, okay, because they design their own chips, not a secret. Then they a lot of silicon teams. Then they have system design, so they have to do all the system simulation, and then they have a lot of data and AI for self-driving. And this electrification and convergence of electrical and mechanical things is definitely going to happen -- is happening already in the car space with Tesla, is going to happen to all the other car players. Same thing has happened in aero and defense. There's a lot of investment there. And of course, on the system side with phones and all. So I think this is -- and we saw this a few years ago. One is because our expertise, but, two, also because of our analog business, we also talk to a lot of people, and we have a much broader portfolio with analog and digital. And I think this just makes a lot of sense. I mean is -- the mechanical part has to be -- has to have integration with electronics. And then the third circle is important to the whole data analytics that goes with these systems, just like we're talking about Tesla. So the 3 pieces are critical there. And we have built the company to go to this new word that we believe is happening. So that's why this move into systems. And in the beginning, people said, like, what, like it doesn't make sense. But you can see now that this natural convergence is happening. And the other example, even this Allegro X and 3D-IC. See, there are some of these other companies that Autodesk is talking to. They're only at the PCB level. See the -- another thing with Allegro is not just PCB but also packaging. Because the real convergence is not just mechanical and electrical. But within electrical, this board and chip is going to merge. So the packaging part of it becomes much more critical. Even if you look at like Intel Foundry Services, there's this big presentation by Pat, the new CEO, and he talks about packaging so much in that. If you look at TSMC, they talk about packaging. So I think the electrical mechanical is there, but also this packaging part will become super critical. So there is this diffusion between these 2 words. And I think Cadence is in the best position. One, because of historical reasons, Virtuoso, analog, Allegro. And two, last few years, we have consciously seen this coming and made all these investments to complete the portfolio with system analysis, thermal analysis, Allegro X, data analytics. So we feel pretty good about where the industry is going and how we are positioned.

Great. You mentioned that your forward road map at this point, it's 2 nanometer, a node beyond that. So all of this is conducive to having really strong multiyear visibility. I wanted to ask the question in the context of financial model for the company. And Cadence has started putting 3-year CAGRs in its quarterly earnings, just to underscore that there's actually started to be fairly good consistency in delivering double-digit revenue growth and the margins are improving as well, even though, as you said, you're making a lot of investment. I'll just maybe pose the question. How do you think about the financial model for this company?

Yes. Sure. Joe, great question. I think -- I mean, while growing the top line is a priority, we're very focused on scalable and profitable growth. Generally, what we've been trying to do with the financial model is drive double-digit revenue growth across the whole portfolio of business at Cadence, and not just revenue growth for the sake of revenue growth, but sustainable and profitable revenue growth. When we build our models for our annual operating plans, we do that with a view to driving incremental margins of over 50%. Right now, I feel very confident in our ability to do that. It's easy to achieve that over a 1-year period. But what we've achieved in terms of incremental margins has been very consistent over the longer time horizon. If you look at our outlook for '21 and compare that to 2019, 2018, '17 or even '16, you'll find that the growth in non-GAAP operating profits over that period of time is more than 50% of the growth in our revenue. And what that means, of course, is that more than $0.50 of every dollar of revenue growth that we create is dropping through to operating profit. And we also see it showing up in our operating cash flow, which proves it's high quality, profitable revenue growth. And then we use that cash to continue investing in R&D to keep that virtuous cycle going, as Anirudh just described.

And then maybe can we just address quickly because I know this is guiding questions. Everything the company talks about, you read in the headlines, it would suggest things are accelerating, not decelerating. But the financial guidance for just this particular calendar year does entail decelerating growth into the second half. Why is that?

Sure. Good question. I mean design activity is really strong, and we haven't seen any slowdown in design activity. But last year was an unusual year for many of us, I guess, with the pandemic. But it wasn't just a pandemic year. We saw unusually high revenue in China for the second half last year, which was great, but mostly due to onetime things that triggered immediate revenue. And when we were doing our outlook, I derisked our guidance for China. My current guidance assumes that China's strength doesn't repeat in the second half, but reverts back to our usual average experience. That when you consider the revenue profile, we had more upfront revenue -- a more upfront revenue profile in the second half of last year in China than we would normally have. And I assume that, that kind of mean reverts back to the average experience we have. And I thought that was the right thing to do to derisk China because the strength last year in Q3 was with hardware. In Q4, it was more to do with license compliance and a catch-up of cash collections on debt that we previously thought was uncollectible. Those 2 things really benefited China, but they didn't feel like they were sustainable things that would recur the following year. So you're kind of comparing the second half of this year, again, to the second half of last year, which had those onetime things in them. Also, I think you have to consider the impact of the extra week. Our fiscal 2020 was a 53-week year for us. Our year-end is technically the Saturday that's closes to December 31, which means that most of our fiscal years, like this year in '21, is a 52-week year. But every 5 or 6 years, we have a 53-week year. I think your question refers to like we're comparing our second half guide to the second half of last year until the first half of this year. I mean if you compare second half of this year to the second half of last year, you're comparing a 26-week period against the 27-week period. And with so much of our revenue being recurring, comparing half year over half year, you're basically comparing a 26-week period to a 27-week period. And what that means, that's like a 4% headwind. But the -- yes, so I guess, I mean, last year was an unusual year by anyone standards. But comparing to last year, the midpoint of our updated '21 outlook implies around 10% revenue growth If you adjust for that extra week and the non-GAAP operating margin is continuing to accelerate. We believe it will land in the range of 35% to 36% now.

That's good. Now -- and I guess, to extend that out to next year, I mean, some of these comparisons will flip and be in your favor. But ultimately, next year should be probably more representative of what's happening in the industry organically. Do you think that's fair as opposed to all these considerations?

Yes, I think so. And then you pointed out in your earlier question that I started showing the 3-year CAGRs in our CFO commentary because that's the way I look at it. But typically, a lot of our customers purchase our technology over a 3-year cycle. Like on the software side, it always starts with a baseline typically 3-year contract. And our average duration of contract is typically about 2.4 to 2.6 years. And the reason for that is that everyone tends to start with a 3-year baseline contract. And then throughout that contract over 12 quarters, maybe somewhere between 4 and 7x, those customers come back and purchase add-on technology that we will have terminate with the underlying baseline contracts. So when the baseline contract comes up for renewal, typically, you have a big exercise of taking maybe 7 add-ons and the original baseline rolling it all into 1 new baseline contract. And of course, I mean, an add-on for the last 3 or 4 months of the contract, there's only 3 and 4 months in duration. So that's why we say that like our average duration is typically about 2.4 to 2.6. It's very, very consistent.

Okay. Okay. Great. Maybe let's finish up on just the system analysis strategy, which we started on earlier. And then maybe just to ground those conversations. So this was a bigger push into simulation solvers, 1 example. Can you talk about what drove the areas you started in? So starting in EM and thermal, why there? And then we can extend that into why have we seen what has come next, which is a bigger move then to CFD?

Yes. Let me take that. That's a good point. So first of all, I think when you look at EDA, I mean you may know already, of course, but about 25% of our revenue is simulation within EDA. Now this simulation is applied to chips design. So chip design has a lot of simulation, probably the most advanced simulation. To give you example, when we -- all these chips are simulated to a great extent before they are manufactured because you can't have first -- you have to have first time right. Because the cost of doing another re-spin on silicon is like tens of millions of dollars, okay. So for that reason, very advanced simulation has been developed in the EDA industry. And also, it is at very, very large scale. So we solved, for example, $40 billion by $40 billion system because it's 3-nanometer or 2-nanometer. So you have billions and billions of transistors that need to be simulated. And it is so expensive to do another re-spin that they are very accurate and very large simulations, okay. So that's why about 25%, 30% of our business is simulation already in EDA. And the other interesting thing of that part of the business is that it is always more profitable than the regular EDA business. And there are multiple reasons for that again. So that's the foundation of our -- and so EDA is -- of course, right now, things are great, and we expect that to continue in silicon. But if you go back 5 years ago or 10 years ago, the question of EDA was, okay, it's great, but your customers are buying each other. Is it going to be great for the foreseeable future? Okay. That was always the question, if you remember. So I think EDA is a good business and there's a lot of things improving, but we always wanted to diversify beyond EDA. And this is like -- I think this is not just for Cadence. This is true for EDA industry as a whole. And I think some of our competitors have chosen a different route. But I do believe the most synergistic route for EDA is system simulation because of the trends we talked about and because of the mathematical expertise you have. You have to be good at some -- whenever you wander out your core area, you have to have some core expertise that you can win with. Just because it's a big market, it's not a good thing. So -- but because we have this core expertise and analysis, mathematical software and the customers are asking for it, I think it's a very, very good fit. Actually, I am surprised why it was not done before. It should have been done 20 years ago, in my opinion. I was not in Cadence or commercial EDA, but it makes more sense to go into system simulation from EDA. Because mathematically, we are good at it. The customers will want it. Now it has accelerated in the last few years, but I believe -- so then the system simulation is, whatever, $7 billion to $8 billion and growing, right. I think that's the good thing about simulation is you always want to do more and more simulation. And if you look at those companies or those areas, they are also more profitable just like we know within EDA simulation is profitable. So other thing we like about simulation is not only it's an area that is very diverse, you can apply to multiple verticals, but also it's more profitable. So like we have our IP business. We like our IP business. But if you do too much IP, IP by its nature is less profitable than EDA. And this is some of the issues of I think some of the other companies in the EDA space. If you do too much IP, then your profitability can be affected. So we want to do IP at the right level and the IP is growing, but we want to go into areas which are more profitable, so simulation. So for all those reasons, we want to go into simulation. Now the question like you asked, Joe, is what to start with? And this is like this famous book, right. I have this I think the business strategy book, talk about like bowling pin analogy, right. That when you go into a new space, you have to go to the bowling pin that is closest to you and then the next bowling pin and the next bowling pin. So simulation has a lot of areas. The closest area was the electromagnetics part because it is naturally with the IC part. So that's why we entered electromagnetics. Now out of $7 billion, it's about $700 million, $800 million market. So it's smaller than the whole market, but it's very synergistic. Because all these electromagnetics or this FEA solvers for this finite element-based FEA solver. So that was the first area. That's Clarity. Again, we did pretty well for a lot of reasons there. Now the next agency is -- so in system simulation, I believe there are 2 main methods to do things. One is finite element, one is CFD, and these are the dominant. And finite element, we want to start with electromagnetic. And the next bowling pin is the overlap of finite element and CFD, which is thermal. Because when you do thermal, you not only need the power, which is consumed from the chip, but you need to model the airflow and fan and all that, and that's CFD. So you have finite element plus CFD has to be there in thermal. So that's why we did thermal next. And then full-blown CFD with NUMECA and these acquisition, Pointwise, because that's a huge market. And to really realize the breadth of this simulation into other verticals like medical and aerospace, you need CFD. So this is our journey. Finite element first is electromagnetic, then thermal, than CFD. I think it makes sense. Yes.

I wish we could keep going, but unfortunately, we're out of time. So Cadence is going to have a breakout session after this. So hope everyone can join for that. Otherwise, thanks to the Cadence team. Thanks, everyone, that listened in and hope everyone has a great rest of your conference.

Excellent. Thanks, Joe.

Thank you. Yes.