Synopsys, Inc. (SNPS) Earnings Call Transcript & Summary

Good morning, everyone, and thank you for joining us here again on day 2 of the Morgan Stanley Global Technology, Media & Telecom Conference. I'm Mark Edelstone. I'm the Chairman of our Global Semiconductor Investment Banking practice, and it's really a pleasure and honor here today to be hosting the session with Synopsys. So I'll do an introduction in a second, but I did want to just announce one disclosure here that for important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. With that, let me introduce Aart de Geus, who is the Founder and CEO of Synopsys, one of the legends of our industry and someone who has just incredible insight into what's going on kind of across the technology landscape. So Aart, welcome.

Thank you for having me.

Okay. Great. So let's start off. You guys are doing a lot of interesting things today as a company. Maybe you can just provide an overview of kind of the big fundamental trends and business dynamics that you're seeing in your 3 business segments today, which are EDA, IP and Software Integrity.

Sure. And of course, both EDA and IP are dominated by our market, which is semiconductors. And I don't have to tell you that semiconductors is absolutely on a roll. And this is very visible to many of your customers because their volumes are going up, there's even shortages, which, for once, is actually a positive sign. But at the same time, the part of the roll that you may not see is the fact that advanced design continues at rapid speed, right? And so where we sit, we help people develop the chips, be it with our EDA tools or with the IP. And then after some period of time manufacturing, they ultimately go to market. And the absorption is very high. And the reason it's on a roll is 2 reasons. One, of course, very visibly, so the whole world has gone virtual. And by the way, it's not going to go back. Even when COVID disappears, we will keep a duality between working in the office and from home because it's so effective. But the second reason is that after so many years of Moore's Law essentially being a driver of more transistors, now it's the end markets that have become the pull. So we've come from a push to a pull. And the pull is because with the combination of large data that is very portable and in a lot of AI computation, there's an enormous hunger to use that information to create a smart world. And so that's going to continue at a very rapid pace for many years. Now as you know, the other part of our business is Software Integrity. And one of the reasons we invested there is because we understood that there's a continuum from silicon to software. And at some point in time the 2 interacts, then we have quite a bit of business there. But above that, the notion of quality and security become super important. And recent events have shown again that with the power of these huge systemically complex systems comes the danger of hacking, of vulnerabilities and so on. And so the notion of security is actually going to be driven and through the entire stack of software and hardware and silicon underneath that. So all of these things are actually on a roll.

Great. Well, there are so many threads to pull on that. Let's first start with the pull that you talked about. Relative to kind of broader semiconductor industry and as Moore's Law has been kind of slowing down, what are you guys doing uniquely to drive more of that pull for your customers to help them continue to innovate?

Well, so that depends if the customers are in the manufacturing because there, we help them still go to smaller geometries. But there's a second phenomenon that I consider actually a booster on a Moore's Law classic, if I can call it that. And that is that a number of companies have now invested in going what we call 3DIC, meaning you actually stack chips and you really connect them very closely. And so that is an opportunity to dramatically still increase the number of transistors. They're just not all on the same chip. And so meanwhile, chips are still becoming more complex. So you need to get sort of a multiplying effect. And from where I sit, I've been looking at it for many years. And I think this is truly another decade booster of complexity. Now the second thing, of course, is we work with all the people that do super advanced chips. And immediately, you would focus on computation. And computation has gone from general computation to more and more specialized, such as GPUs, graphics processors, and now you have a zillion AI processors. Actually, we're serving over 200 companies that are all building the best AI processor ever, of course. And they're all right until they're not right. But it illustrates this push forward. And around these chips, of course, you need storage, you need transportation, you need things that collect data, so IoT devices. And we touch literally all of those. But invariably, the state of the art drives the most sophisticated consumption. And precisely because I see that continuing at a rapid clip, I think the rest of that market will continue substantially as well.

Yes. Maybe we could talk about the AI for a second because I think it's interesting. Obviously, AI is going to be everywhere. This is going to be one of the probably most profound trends we've seen in our lifetimes in technology. And you tie to NVIDIA, which is one of my favorite companies in the sector, that has really helped to get first-mover advantage on [ this AI ] both on the training and inference side of things. But as you mentioned, a couple of hundred players that you're supporting. I'm sure you don't want to name names. But maybe just kind of more conceptually, how do you think this plays out given so many companies that all think they are going to be the next dominant player?

Well, first, commenting on what you said. I think this phase is as profound as computation was literally 40-some years ago because it's transformative in terms of the value that it brings, right? Then the previous phase was really unbelievable human efficiency to do many tasks on a computer, now it is actually to add insight to the very notion of living and being in a world that's very complex. And so from that perspective, the fact that there's so many companies, it's just an indicator that we're in the early phases of development. To me, it's also an indicator that we are going to see very specialized computation because if you and I were an AI algorithm, we would have only one wish in life. Give me 1,000x more compute because then I can really do something, right? Well, that 1,000x is very difficult to achieve by just saying, well, make transistors smaller. You don't get there easily. So what has happened is that the notion of architecture is reinvented. And you mentioned NVIDIA, and what they did is they really moved from general computation to graphics. And that was their architectural change. And then they found that, that architectural change was actually very powerful for AI. And that gave them a head start. Now since then, of course, there's much more of that. And so this AI will gradually orient itself to different verticals to say, well, in this vertical, what aspect of learning is most important? What is the characteristics of the data here? And everybody will already optimize their architecture and their end direction on the end markets because that's why I call it a pull versus a push. The money is going to come top down. This is not you buy chips and then you put a computer together. No. You have a problem and you want to figure out how to buy AI or data manipulation for that. And that is why we also look at our world as being silicon to software as a continuum. And I think that opens for us doors that have still many opportunities above that.

Great. So clearly, you've set this company up in a great position, you're mission-critical, you're driving a lot of things from silicon through software. In the quarter you just reported a couple of weeks ago, I thought you did a pretty good job of encapsulating all that and basically showing the strength and positioning of the company. But despite those great results, the stock has sold off. So what am I missing?

Well, I guess the stock is your job. I did my job.

That was old my job. That was my old job.

Yes, I know, of course. No, no. The -- but the point is, I think we had a very strong first quarter. We didn't immediately change the outlook for the year. And actually, we very rarely do that after a first quarter just because it's not particularly good management and to have some degree of steadiness is good. But the fact is, with a strong first quarter, the overall year becomes immediately lower risk and in a better situation. And as a matter of fact, right now, we're on track to continue well against the plans and the outlook that we have set. So I think the market will take care of itself over time. But our growth rate is very good. The -- we made great progress, by the way, on our ops margin at the high end of the range that we had set for the year. And moreover, you know that late last year, we said that we're working this year on the Rule of 40, meaning growth plus ops margin, but that we are already laying in all the plans for Rule of 45. And so it's a very meticulous and steady improvement of financial profile while making sure that we invest enough in all these opportunities that are around us.

Yes. And I -- personally, I love the Rule of 40 and 45, 50, even better. So as I think about your business and the 3 components of it and I look at the things that are growing faster, namely IP and the software security, it would seem to suggest that your overall growth rate ought to be accelerating. So let me talk about what the potential could be when you think about the company in the next sort of 3 to 5 years. How could that really play out as the faster growth of those other businesses impact the total company?

Well, the nice thing about the Rule of 45 is that on purpose, you leave a certain degree of freedom to say, do you get there by growth or do you get there by ops margin improvement? And the answer is yes, right? The answer is yes. I am personally always partial to growth because I think that growth makes ops margin improvements easier. But the reason we like the rule and use it a lot in internal Synopsys communication is that you always have the question, well, investment versus growth, where do you go? And the answer is, well, to 45, that's the whole point. And leaving -- or giving I guess the richness of understanding that both interplay is key. But having said that, I do think that we are in a market situation that is very, very good for Synopsys because we are so central to literally almost every advanced chip in the world. Our portfolio now literally goes from deep silicon to increasingly higher in the software stack. And all of these things are less disconnected as they used to be in the past. They used to be very layered. But as the technology evolved, things become much more intertwined. Meaning, the design needs to understand how the manufacturing goes, for example, the software needs to be exercised on designs that are not finished. And these interceptions demand a systemic complexity that Synopsys is particularly well structured for. And by the way, that is a skill that we think differentiates us. And that's why I think we are de facto becoming more important to our customers than before.

Yes. And no doubt about that. So maybe we can dig into the software security business a little bit, which...

Sure.

To me, it just seems like a fantastic strategy and great opportunity that, in my mind, should only increase the importance basically of advancement for intelligent devices, especially if the edge continues to increase. So maybe first off, what's your view on that? Maybe talk about kind of your larger strategy for this part of the business.

Sure. And so this is a business that, literally, we built from scratch in the last 6 years or so. We grew it very fast to be about 10% of Synopsys. This year, we clearly are focusing on one thing, which is the scaling of our capabilities because we did see a number of growth challenges as we have grown very fast and our systems, our -- the balance of our approach between product development and go to market were somewhat out of balance. And we added management in the last 6 months. That is at a substantial different level of sophistication and experience. That's turning out to have been a very good decision. And so their focus this year is entirely all the things like the go-to-market strategy, the balancing between point tools and platform and the focus on the consulting teams that are necessary to deal with high-level sophisticated customers that have really complex issues. And 2 quarters into it, I can report that our own feel is that we're making very good progress. And so as we exit this fiscal year, I think that we will be at a different growth rate again. And back to sort of the next phase of growth, it's round numbers between -- around $350 million. Clearly, the $0.5 billion to the $1 billion is the next objective, and the systems are being laid -- put in place for that.

Okay. Great. So as you kind of go down this path and, obviously, you've been building it for half a dozen years and increasing the focus here, how do you see as kind of primary competitors? And how do you differentiate versus those?

Well, it's an extremely fractured market because it's a new market. So there are many, many companies. The names of some size that stand out would be a Checkmarx or a Veracode. And I have nothing negative to say about any of those. I think we're all racing forward in trying to provide truly powerful point tools because some of these problems are quite difficult. At the same time, we have already moved forward to providing more and more of a platform because we noticed quite quickly that maybe in contrast to the similar history 30 years ago in EDA, the security problems tend to bubble up very quickly all the way to the Board. And when a company has a problem, the Board says to the management team, do something. And that means the CTO, the CSO, the CIO are the people that look at this. And while there are specialties underneath, they want to have an overview of what to do in aggregate instead of having 15 different small companies all doing pieces. And that's an interesting challenge because that means you have to move forward all the point technologies as fast as the providing of a platform. And I think we're very well placed for that.

Okay. That's great. So it seems to me like from your comments in your -- at least in your recent strategy that -- or tactics, let's say, that you guys are focused on the growth part of the Rule of 45.

Absolutely. Yes.

And so that's great. Obviously, you want to build scale and drive it. Today, your Semiconductor & System Design business basically had a 32% operating margin last quarter and software had 9%. So obviously a lot of room to go there. So just talk about the dynamics there. What could this business be when you think about when it's $1 billion plus?

Well, it's obviously good news, right? There's a lot of room. And the reason I say there's a lot of room is because there's no reason to believe whatsoever that this would be any different than the rest of our software business. The software is a complex, high intellectual property business, and it should be in the Rule of 45. That's its objective, right? So as I said, this year, we -- you called it tactical. I would agree with that term. We're making a tactical set of investments for scaling up, meaning growing. The growth will drive the profitability with it. And I think it absolutely is aimed at the same horizon that we have for the rest of the company.

Yes. Okay. Great. So we talked a little bit about AI in terms of your customer base, but talk about it in terms as well of your tool set. So how are you basically implementing AI and machine learning into your deliverables for your customers? And how is it impacting what we're able to do?

I love that you bring it up because it's one of my favorite topics right now, and for a key reason. Obviously, we've invested in AI for a number of years and not only to support our customers, but for our own use. And I would say that the vast majority of all of our products today have embedded a variety of machine learnings. Some are relatively simple, they help the tools be better, others are more profound, they give insights to the tool that otherwise would be difficult. But in the last 9 months, I think we have now some really significant breakthroughs in not only being able to make individual tools be more performance, some more competitive, but actually pieces of design flow. And the results that we've gotten specifically in the last 4 months or so are amazing. And I'm careful with words like that because I see it somewhat as not dissimilar to literally with starting Synopsys where Synopsys was the company that made it possible to describe the chip by writing it in a language, and we would create the thing that was underneath, right? Well, we are now doing with AI some of these tasks at the level of entirely sub -- entire sub chips today. And what is so exciting is that in lesser time than a human, same as 30 years ago, we get better results, in most cases, than what a team of humans can do on chips. Now there's all kinds of caveats and so on and it's a bit early and so on, but it's not that early because our customers are adopting it very rapidly. And literally, we've seen in a time frame, I would say, of about 9 months, people going from some experiments to say, "Oh, no, no. This is now part of my production tape-out." And so I do think, for very good reasons, we should look at AI as a complementary computational method. The traditional computation is really deductive, meaning if then else, you do this, then that and so on, whereas think of AI as we recognize patterns. Well, recognizing patterns is really difficult for a human, in one hand, when there's an enormous amount of data. We're very good at looking at out the window and immediately spotting that's a bird and that's a tree. And you say, well, that's unbelievably difficult. And yes, making car drive autonomously is not easy. But looking at a chip is many, many, many more dimensions because everything is intertwined with everything else. And so great architects can see immediately that if you put this piece of the chip this far away from the other, it will never be fast enough. That's easy. Except it's not 2 pieces. It is a billion pieces. And so you can just get a feel that pattern matching and deductive reasoning when you can't combine them has a lot of opportunity. So ergo, I think we're at the beginning of that. And what I'm telling you is that while we're at the beginning, we are now getting better results in number situations than groups of humans can do in a reasonable amount of time. That is super promising. And by the way, I would have said exactly the same about the word synthesis, which is what Synopsys started 35 years ago. We got better results than humans were able to do in a reasonable amount of time.

Great. So one other major trends out there is cloud. And maybe you could just talk about how you think cloud impacts Synopsys' business in the future and maybe specifically your engagement with some of the cloud service providers.

Sure. And actually, that's -- that, too, is a stratified question because you can look at it from as a user versus as a supporter of them. Because as you all know, the cloud suppliers, also often referred to as the hyperscalers, they essentially have now pretty much all invested in developing their own technology, their own chips and so on sometimes in collaboration with others, sometimes they acquired some companies, sometimes they're building it from scratch. And what makes them different than other customers is that they are the one category of customers that is building chips not to sell the chips, but to use themselves. So they're very driven to exactly that specialization, call it in AI or in any other area that they may want to work in, and at the same time, they're still learning the trade. But it's a segment that's growing very rapidly. The second perspective, of course, is cloud as a compute environment for ourselves or for our customers. Well, ourselves and our large customers have enormous clouds themselves, right? And that turns to be economically super cheap compared to actually buying time on the cloud. However, buying time on the cloud has the benefit that you can get a lot of computation for a short amount of time. So if you want to accelerate something. And so this is where this intersects with our previous discussion. There are variety of verification and AI tasks, specifically, where you may want to have burst compute and quickly do a lot of experiments and then look at the results. And that is where I think that the cloud will have most impact in our field because we are super, super, super sophisticated users of computation, right? So it's not like verifying check account numbers or something like that, not to trivialize that, but -- and so we are very demanding of what a cloud can do, but we see the opportunity there that we will see growth forward. Presently, all of our tools are completely cloud capable, and we have many, many customers that use it with or without our knowledge.

Okay. Wonderful. If we go back to the beginning of Synopsys, you guys have used an M&A strategy over time to bring in new technologies, capabilities and so on to help build the business in addition to what you can do organically. Can you just talk about how your M&A strategy has evolved? And how you'd expect it to play out here going forward?

Well, in many ways, it has evolved because we've done, I don't know how many, but over 100 acquisitions. And so it has evolved by virtue of having a lot more experience, what works, what doesn't, how do you integrate, how do you look at M&A being additive or being multiplicative. Meaning, it changes who you are. And -- but at the same time, we've always looked at M&A through the lens of adjacency. Meaning, is it technically adjacent, is it customer adjacent or is it channel adjacent? And if you can say yes, yes and yes, it's significantly less risk than if it's completely not adjacent, right? And I think we have made -- had a good balance between making a few acquisitions that were further away in adjacency. And initially, the IP business was that. And in the last decade, the SIG business, the Software Integrity business, was like that. And then you complement it with things that are very adjacent because it's technologies that are really promising, that can be tuck-ins, that can round out an overall offering or sometimes it's also market positions that round out your position on the market. And I think that balance has served us well from a return of investment. And we actually have -- we have looked at return on capital quite carefully. We don't disclose the numbers. But we have a bar set, and we review that with our own Board every year. And you -- we can see that we do quite well against the expectations.

Right. And obviously, I'm sure you -- I mean my outside end view is that your M&A strategy has been successful over time. But when you're thinking about that versus return of capital to shareholders, how do you think about those 2 vectors of how you use your free cash flow basically?

Well, we fundamentally use free cash flow for 2 things, right? M&A or M&A of ourselves, which is by buying stock or buyback. And we've had actually I think a quite good track record of maintaining a share count that has essentially been pretty much flat by doing that. But I always look at M&A as what do we need to do to make sure that we keep growing in the future. And we need to understand that in super high-tech fields like ours, there is no monopoly to innovation. They are 3 guys in a garage and they have a great idea. It may not go anywhere because it's actually very difficult to build a channel around that and a successful company, but it's still a great idea. And so sometimes acquiring that makes a lot of sense. At the same time, sometimes investing in internal R&D, we have a lot of good ideas, too. But humility clearly shows that there's no boundaries to that.

Right. Okay. Great. So I've been doing this for over 30 years now, as you have. And I've never been more excited about what's going on in the semiconductor industry. And my observation is that the companies that are going to be the most successful going forward are going to be more vertically integrated. So for years and years, we basically provided this incredible platform that everybody else built on top of and captured the lion's share of the value. And companies like NVIDIA have completely flipped on its head. Many of these new AI start-ups are doing the exact same thing. What's your view of that as companies basically move up the stack, look to try to grab more value, is that a legitimate thesis? And then how can Synopsys sort of benefit from that as well?

Well, I'll give you the techno-nerd justification for your thesis, which is we have a big face shift from Moore's Law as a push up to AI potential as a pull from the top. And so the reality is very -- the value is being created in the end markets in the verticals, and that value is all by adding smarts to whatever products or problems one has. Now each vertical does have some peculiarities. And let me take one that Synopsys actually has invested a lot in and is very successful in is automotive. And the first thing that is often not spoken out but the thing that stands out for automotive, that thing better be safe. Well, safety is a top-down demand. You're not asking your PC to be safe. Yes. Sure. The battery shouldn't heat up. But the safety for a car is a very different attribute. And so that's a great example of a vertical saying, "Hey, I can provide you money if you make my car really, really, really smart. But be damn sure of that, that you make it safe at the same time." And by the way, safe and secure are words that are sitting right next to each other, right? They impact each other. And so we see these needs in the vertical coming up. And you'll see absolutely the same in the health departments in the biotech fields. They all have characteristics that if you can meet them, you now have essentially linked deep physics all the way to big functional impact in the field. And at Synopsys, we have focused on a number of verticals. Of course, our traditional vertical from day 1 was computation. And then really in the late '90s, early 2000s, mobility added a whole new vertical and actually have larger chips. And you say, what was different? There was one thing different between computation, mobility. Low power. And that sounds like so yesterday now, right, of course. But we all remember that phone that needed to be recharged every couple of hours. Nobody thinks about that today. And yet, you have an enormous computer in your hand. So that was a technical need from that field that then reflected on everything else. And I think that -- this face shift of value creation being explicitly in the verticals is what drives companies to say, therefore, we should orient ourselves also upward into different fields.

Wonderful. Well, Aart, I could do this all day long with you, I'm quite confident. But we are out of time. So thank you so much for joining us and sharing your insights on Synopsys and the overall tech ecosystem. No doubt that you guys will continue to be a huge player in driving success of the whole global economy. So thanks for being with us, and I look forward to doing this again with you next year but hopefully, in person at the Palace Hotel in San Francisco.

Well, thanks, Mark, for making this an interesting discussion. These are the best. So...

Have a great day.

You, too.