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

Let's do it. So thanks, everybody, for joining us. I'm Drew Guevara from Morgan Stanley. I will begin first with our research disclosure. 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. So Aart, thank you so much for joining us. We were just talking about how the muscle memory kind of kicks back in. You have missed a beat in terms of tying that tie.

And I don't think I've used it for 2 years. So here we are.

Yes, exactly. So you guys missed the conference 2 years ago because the COVID thing was just going to smack us, right? And here we are 2 years later, and we have a new set of challenges in front of us. But if we kind of step back and look at that 2-year horizon, okay? The S&P has been up and down, but it's up about 45%. [ SoC ] is up about 85% during that period, and Synopsys is up over 100%. So one definition of your served market, your customers has outperformed the market, which is generally up, recent volatility notwithstanding, and you've outperformed that customer base. So all that speaks to higher focus on design, but talk a little bit about what's been going on in the industry and the company that's driven that?

Well, starting with the industry, semiconductor markets passed $550-or-so billion this past year. I'd like to call it $0.5 trillion. And there is a reason for that because the expectations are that this decade, it will pass $1 trillion. You would say, so why is it growing so well? And why is that now? Now in all fairness, semiconductor industry has existed for 50 years, has delivered an unbelievable technology capability set and yet has entered a new phase of great, great opportunity. And I'll put it immediately in the context of the previous phases. The first one, clearly, 70s, 80s was computation, followed by networking and mobility as connectivity. And now we've entered this phase of literally making everything smart in some way or another because the intersection of big data, the storage, the transportation of all that with machine learning, turning into AI, opens up essentially impacts on all the vertical markets. And what's exciting about that is that while we have grown up in a technology that kept pushing things forward. Now suddenly the end markets have a pool, which they all realize as big data, whatever it is, it could improve their economic position somewhat. So there's a demand there and we're sitting in the middle of it. So that's half of the story of why things are up, right? The other half is that Synopsys, we have had the good fortune to invest in multiple areas that are all super relevant at this point in time. One is clearly all the design tools that make it possible to advance the state of the art. And in our history, we have delivered 10 million x in productivity. That's actually a pretty big number. And crazily enough, we have now said we're going to do another 1,000x this decade. And the reason we think we can do that is because the technology continues to advance. All the tools that we need, we have together, but better than to get -- they work together, and we've layered on top of that a breakthrough in AI that certainly makes it possible to automate entire design flows. We have now a plethora of examples that show you can do that in a fraction of the time and get better results. So EDA is key. The second thing that has gone very well for us is IP. These are building blocks that are predesigned. I would like to compare chip design to using LEGOs. You used to have very simple LEGOs and then the collection became bigger and bigger and bigger. And today, anybody starts a new company or who does new chip says, well, let me start by, I need that block, that block, that block and that block, and then I need the tools to bring them together. And we are the broadest provider of these IP blocks, and that's an area of the company that's growing very well. And last but not least, a little further out in this picture, is we have invested in software integrity, meaning the quality and most importantly, the security because there's 1 continuum between chips and software. And all of these things will be impacted by security. And you see that almost a daily basis. So all of these 3 areas for us have done well, but we also benefit from the fact that the hunger to do much well are, meaning that all the advanced customers say, "Oh, how come you cannot be 100x faster? Well, ask that question again because now we're engaging, right? New opportunity.

Right. Exactly. So you call it Smart Everything. You've also called it in the past, smart and secure everything.

Yes. And sometimes, smart, secure and safe. Because by the time you put in the car, it's not just secure, you want the thing to actually stop when the chips go bad, right?

Right. So big data AI is driving demand for more complexity everywhere in the stack, more complex chips, systems of chips, more security, more safety. Everything wants to go up by a 1000x. On a relative basis where do you see the weakest link or the biggest challenge to have everything scale and tandem?

The weakest link is that -- is actually in the word 'in tandem', meaning in the past, you often look at TASCO, you have to do this thing and that thing additive. They're not additive, they're multiplicative, single zero and everybody gets 0. If you want a recent example that broad chips front of mind is the $0.47 door knob controller in the car stops the shipping of a $47,000 car, if it's not available. Well, that's a multiplicative effect. Now that's the value chain had granted these roles set up for just in time, and they were not thought through in the present situation. But the same is true for technology. If the IP doesn't work really well with the tools or is insecure, you have problems that impact everything where it's built in.

So looking under the covers of the customer base, big picture, how is the composition of the customer base changing?

Good question because there has been change. And so you have all the usual suspects have actually continued to grow and do very well. And many companies that are many decades old are enormously valuable and enormously engaged in yet another wave of newer technology. But there are at least 2 or 3 cast of characters that are new. One was a whole set of start-ups, all doing AI. And by the way, they're all designing the best ever chip, so they think. The race is on. And frankly, if some don't make it, it doesn't matter so much to the industry because these people will be recycled in another company in no time, right? There's a high demand for talent and experience. And there's a huge demand for AI because every vertical market will want something that's just slightly different, just optimized for them. So that's one grouping. The second grouping that has come about and I'm looking at roughly half a decade to a decade, where we've seen all these things radically change is what's called the hyperscalers. So these are people that in some other parts of the business, they figure out the way how to print money and then they invest it in big computation or big offerings and they have started to design their own chips because: a, they use a lot of them; and b, they have some specific needs of accelerating it or bringing the power down in their compute centers. And because they are, in general, I would say, wealthy, they can afford to, from scratch, build new teams, hire some very good people. And by now, we have a number of those that are doing truly sophisticated chips, not so 4, 5 years ago. It was a little amateur hour at that time, but now it's really capable. And then you have emerging the verticals [ license ] where suddenly a vertical and automotive is a great example, suddenly discovers, "Oh, wow, you know this autonomous driving, that could change my whole business model." And initially, it was a little bit of overreaction at the same time, they're all watching Tesla and say, how do these guys do it that they have cars driving and they download changes. That's a radically different concept of the life cycle of a car, for example. And then on top of that, now is the wave of electrification of all of it. And so suddenly, systemic complexity now comes to host inside of your car. Where is the security and the safety to boot, right? And so these are people that are now mostly investing in the architectural thinking because they realize full well that they can -- they're not living at the same cycle, let's say, like cellular phone. You have a mobile phone. These things are designed to be dropped in the toilet every 3 years. You need a new one. A car is 20 years of life cycle. And moreover, car takes 5 to 7 years to develop, cell phones come out every year, right? So you have 2 completely different time constants that have to somehow get married. And it's going to be interesting to see how that gets done.

And then big picture, you kind of touched on this, are there significant and obvious differences with the approach that these different categories of customers take to design or how their needs are different?

Well, yes, because they're all in different paces of learning. And so the way to mitigate the learning part is to hire people that have done it before. And -- but in every industry that takes some time because initially, when an industry learns, they don't know who to trust, right? And that's true in all fields in life. When you hire somebody, you need to know enough about it to know well, who are the people that actually have the goods and who are just good talkers. But the learning is actually relatively fast. And the other thing is that the supply chains adapt themselves to the new necessity, which invariably is at a higher level of abstraction because that's where the money is made. So a lot of providers look not only at their customer with their customers' customers to assess what will be the key capabilities that are most important.

Right. And part of what you do that enables the moving up the abstraction layer to the higher value place is DSO.ai, basically applies AI to chip design.

Yes.

So that liberates the human because the AI does a better job of optimizing for power and space and performance. And so the human is now free to go focus on other things right...

Or do more things, right? And so there's doing other things or if things go faster, you can do more of it. So let me give an example of doing more. You have a chip that actually is working perfectly fine and with a good customer to sell it to. And now you get another customer that says, "Well, I like something like that, but could you make a few changes to it?" Well, is that our whole new design, well, less so because with the AI design capability, we have learned from first chip many things. And then there are a few changes, but all the many things are still good. And so that's so-called derivatives can be done much, much faster. And by the way, the same is also true from taking an existing chip and moving it to a newer, smaller technology, and that can be very much automated.

You talked about how people maybe who don't succeed at a start-up will immediately get rehired and repurposed. Is there a very narrow group of people who know how to do this? Or does the automation technology actually make it possible for more designers to be effective utilizing the technology?

That question is good for the last 35 years. you can have -- ask that every time. And it was always a smaller group than the market could sustain. And right now, there's absolutely some would say, a war for talent. There's a lot of change maybe also due to COVID. But in general, the opportunity space in semiconductors is very high. And so people are looking for more engineers that have experience or that have learned these newer techniques. But it's absolutely also true that certainly, the AI wave that we have just started now, has the same impact of many years ago, automation that certainly made it possible to do things that you would never have been able to do before. And that's what I'm saying, we have been, in our company, maybe 80% of our products have always been the state of the art. And state-of-the-art meant follow Moore's law at that time. And now I like to call this age SysMoore as in systemic complexity, but with a Moore's law ambition, that exponential is still driving like crazy. That's why we're talking not about 2x or 4x faster, but 1,000x faster, right? That is an ambition that is hard to actually conceive.

Right. You made a point about how the hyperscalers are so wealthy, and they can afford to make these investments. They can afford to pay these people, maybe more than the industrial...

You know that too well.

Yes, right. And so then is there this thing where maybe the privileged few who are so wealthy are going to hog the ball in terms of the talent?

Yes. But, somehow justice in life, too, right? Sometimes the environments that are less wealthy are culturally very interesting or have ambitions that are wonderful to work at. And so for top-notch people like top-notch challenges. And that's why there's many different places that are great to work at. And suddenly, we've had the benefit on being able to attract top-notch people because we do really exciting things that touch many others. But that doesn't mean that sometimes we feel a little [ joined ] if somebody gets attracted away with money.

Right. I mean your value proposition to an engineer is that they can have a leverage impact on the world, but maybe the customer in the end market is really challenged to hire somebody like that.

The word challenge is the middle name of the field, right? And that includes how do you not only attract the best people, but continue to grow them, challenge them, help them work as teams. And I think I mentioned earlier, so many things have moved from plus sign to a multiplier, the same is true for people. And so in really sophisticated technology, the ability to learn from each other and cooperate is absolutely central. And by the way, the same is true for our customers. Real relationships get measured on everything that goes wrong and how do you act then, because that determines if you will, still make it to the other side of the bridge. Whereas if something goes wrong and it's your fault, it's my fault, we are not going to make it. And so teamwork in technology has increased in importance substantially.

Okay. One of the things you mentioned was the automotive workload or use case being a great test bed for new AI solutions. What are some of the other vertical market solutions or workloads that you're really excited about?

Well, there are many of the traditional ones, all the things having to do with computation. They can all be improved. But if you look, for example, at the medical world, actually, there's a guy at Stanford, who is doing an experiment with only 1 patient, he himself. He's measuring everything, everything he can on his body, he has 4 Fitbits and does the DNA test, and you name it. He alone has 2 petabytes of data on himself. And now, in this case, he's a medical doctor, he's been able with that data to self-diagnose lyme disease, which is quite interesting. But just imagine that you use that data to start correlating over a 10-year period, how you body functions. And then you do that not with 1 person, but you do with 1,000. Now you have an exabyte, and you can see the riches of what we're going to discover because the internal world is equally complex and sophisticated as the external world. And by the way, if you looked at the grand scale, you would say, well, we need to completely understand climate change. And there are a lot of variables in that. And the more we understand, the faster we could take action. Now the taking action is actually our problem. We understand already so much. But all of these areas are superb. One of the examples I often site is you have some AI algorithm, look at that skin photographs and spots where the cancer beginning. So that's nothing compared to the fact that with new sensors, you can look at 10 layers and the temperature and the heart rate and what was eaten yesterday and so on. Once you start new inferences of all this data, you're going to find patterns. And the key with AI from my perspective, in contrast to the early computation is chess was won with deductive reasoning, if and else. Go was won by finding patterns that are more winning or more losing. There are so many patterns to be found that we don't even know there is a pattern. And that's where the impact on medicine is potentially enormous.

Cool. Well, when we were about to start, you talked about how there's things going on in the world bigger than any of us or bigger than any 1 company. Talk a little bit about geopolitics to the extent you can, and how you think it relates to Synopsys, maybe starting with the U.S. posture around more domestic investment in semiconductors?

Well, the domestic investment is somewhat of a reaction of discovering that there are international dependencies. Now I've always been very positive about international dependencies because it's one of the best way to avoid wars in general. And by the way, it is one of the most efficient ways for humanity to move forward. And lastly, most importantly, all the truly big human problems are global. And so thinking that we can solve those as little independent nations makes no sense to me. At the same time, the reality, I understand perfectly well and highly visible right now in Ukraine is some really bad stuff can happen that is the result of humanity, the way it is. And so one needs to find a balance. And so the very fact that United States is now looking at the semiconductor segment thinking, "Hey, that's really important." Feels a little late. But on the other hand, you can also say, no, it was not that late because the U.S. is a leading company -- leading country, having leading companies in these technologies for a long time, and entrepreneurial behavior can move these things on steroids pretty fast. So the bottom line is still, even the U.S. recognizes that semiconductors and chips are key and that there's a whole wave of new opportunities coming out of that. So even just economically independent of the nationalistic needs can be very positive. And I see this at many other places as well, of course. So we are fortunate that we are a very global company. We have an experience in many countries. And I can absolutely see that there will be big investments coming, and that these will all drive an acceleration of a space that can have a big positive impact on human kind.

Also on geopolitics, though, having the technology then can be used to create pressure or leverage in the geopolitical equation. So where does China fit into the equation for you in terms of both demand for product, but then also source of competition?

Well, China is a country that's growing very fast that has had a strategy of investing in relevant technology themselves for a long time, has gradually become more and more competent in the semiconductor arena from the perspective of designing things and is a rapidly growing customer for us. Due to the international tensions, there is an increased degree of wanting to have domestic capabilities. For us, that's called competition. We've lived with competition for our entire existence, which is 35 years. And we need to race faster because the Chinese customers want the best tools, and we have those today, and we want to continue to be the provider for that.

Great. I've got more, but if anybody in the audience wants to ask a question, we'll pass the mic around. Anybody?

Just on the IP business. I was curious in terms of the revenue, is that on a per unit basis that your customers ship? Or is it more of a licensing for you? Just curious about the economics there.

There's a combination of things. It's more of a licensing fee that is per project. And so it is not so much. There are a few cases, but mostly for us, not on a unit price. It's a business that has grown very well. And actually, you had a question earlier as something changed. And I think I omitted to say that the other thing that has changed with new entrants is that many of the new entrants, they don't buy tools and then IP. They buy IP and then tools, meaning they have already idea, I want to build the system, this block, that block, that block, that block, I will need those. And so they essentially start at a high architectural level of thinking. And that's why the IP business is quite relevant because for them to use their most skilled engineers to develop these blocks is certainly a feat because they're extremely complex, but they're not necessarily differentiated. And so they should use their engineers where they can provide differentiation and buy from the catalog from us if the price is right, of course, what they can. That's the fastest way to get to market. And so not everything that's difficult is differentiated and not everything has differentiated is difficult, right?

Here's another question about customer buying patterns. How about -- you said in the past that a given customer will maybe procure from a different place in the organization for security versus maybe IP or a DSO.ai solution. What are those dynamics like?

Well, security in the last decade or so has been not entirely, but mostly anchored in software. That's where all the vulnerabilities were and so on. And that is still the case. But there are situations, especially those that have super high security like banks or so where they also watch exactly what patterns can you read by putting some thermometer -- something on a chip, right? And so I think that over time, the distinction between software and hardware in the middle won't be that much of a distinction anymore. And so we see it already on the hardware that there are certain security modules that are now being built into the chips on the software, of course, there's a lot of diagnostics for that. And literally actually under our logo, it's a Silicon-to-Software because we view this now for a number of years as the destination is continuing.

Right. And so -- got a question?

Yes. I guess just following up on that, has the software integrity piece of the business, just thinking about that sort of distinction being blurred, it's 10% of the business today. It's small. Where do you expect that -- when do you expect that to be material? And like what portion of the business would that be in 5 or 10 years?

Well, 10% is material. We fight for every dollar. I think that it has the potential to have a slightly higher growth rate than the company. So it will gradually grow over time. We put that business together relatively fast. And about 1.5 years ago, came to the conclusion that we needed to sort of put a year in, maybe 2 years of really focusing on how we structure it for scaling to the next level. We completed that year last year, and that went really well because we can see that the growth rate is accelerating again, and that we have the foundation, both from the structure, from the people there to continue to build on it. So -- and it's certainly a TAM that's wide open, right? There's a lot of space here, the enormous number of software companies. everybody has some issues in some way, and we can be a good provider. So we will continue to invest in that and focus on it and actually, very positive of it. Now we're fortunate that every 1 of our engines right now is doing well. And sometimes you have a portfolio, and some do better than others. At a given time right now, everything is clicking.

So, over here.

How are your custom design products for analog? I'm just curious how that product is going versus cadence and what traction is emerging there, I think?

Sure. Well, we have invested over a number of years because we were in a position where essentially there was 1 company that had the entire market. At this point in time, we have some very competitive technologies. And it is on the basis of that, that we're seeing our market share slowly grow, which is very positive. It is also increasingly integrated well with our IP because part of the IP we provide has custom or analog design built into it. So there's coherence in multiple angles. And lastly, there's a piece of that called simulation, which is where you essentially mimic on a computer how well the chip will work or that portion will work, where we are absolutely in a leadership position.

And one other follow-on just in terms of expanding your addressable market, you made a lot of inroads in the software developer front. So I'm just curious, given rising complexity there, what kind of tools and onboarding is working best for that customer segment?

Are you talking now about the software segment? Or are you talking about for the...

Probably on the customer base side, moving aside from just the analog space to just broadly your overall business?

Well, literally, every angle, the design community, be it in digital design, software design, custom design, is fundamentally following the same challenge, which is a side of scale complexity, more of the same, systemic complexity is becoming the bigger issue. And systemic complexity is essentially the intersection of different domains where 1 person may be expert this in another area. And I've often used the term that in our field, we have all these 6 PhDD (sic) [ PhD ] people, and now we need a bunch of 6 PhD wide to actually make sure that things come together. And that is, I think, going to be for both our customers and ourselves a big challenge and a big opportunity. And so I preach a lot about it. And the -- we are actually making great progress, and that's why I'm happy about the AI capabilities because we've been able to bring many tools together, and now not just optimize what an individual tool does, but an entire design flow. And I think that is the breakthrough that for us over the next many years will have major impact on how we do things.

Well, kind of a related point is you've had to build muscles not only in what you've done historically, but get into hardware more, get into consulting more, all those things and bring them together.

Well, that's called learning, right? I mean, the benefit of being gradually growing company that a lot of people count on is you're always forced to learn. And yes, hardware had its own challenges. Of course, we have customers designing hardware. But when you find out that one of the screws is no longer available in the supply chain, we, too, have immediately a sense of what our customers feel like. We've managed these things as well. But the fact is that was not ever on our radar scope before.

All right. We're just about out of time. Absent the other questions, are there any parting thoughts on things you feel like you'd love the audience to know that they don't appreciate sufficiently about the company?

Yes. Maybe what is not understood enough is how many different parties Synopsys connect with. Because most people don't know that we're deeply involved in the simulation of individual transistors, how they're built, in the manufacturing of these transistors, in the building of the basic building blocks, in the building of the big IP blocks and all the tools above, in the system mimicking of hardware and software at the same time and in the architecture evaluation, that's an entire value chain. And that's why I think we can be a catalyst between all of these parties for -- to handle the systemic complexity.

Excellent. Excellent summary. Aart de Geus, thank you so much for joining us.

Thank you for having me. Thank you very much for participating.