Well, good morning everybody, and welcome. My name is Marco Lagos. I run the semiconductor investment banking business for Morgan Stanley in the U.S. Before I introduce our guest and dive into today's conversation, I have to read this preamble. I'm gonna have to read the preamble. Today's discussion may contain forward-looking statements related to Synopsys' current outlook, expectations, and belief, which are subject to certain risks and uncertainties that could cause actual results to differ. Please refer to the company's most recent SEC filings for a discussion of risk factors that may materially affect these statements. With that, I'd love to introduce our speaker today. I'm sitting here with Aart de Geus. Aart, I'll start by saying, in my estimation, is one of the godfathers of EDA, so I'm pretty honored to be up here with him.
He is also an avid musician, has a band called Legally Blue, I believe.
Mm-hmm.
That he plays with. We're gonna try to riff for you guys up here today, if that's okay with you. All right?
Okay.
Aside from those two things, he's co-founded Synopsys in 1986. Quite some time in the industry. He's also the CEO and Chairman of the company today. With that, Aart, would love for you to tell us a little bit about the company before we dive on in.
Sure. Well, thank you for having me, and thank you for you being here. Many of you probably know Synopsys pretty well, but just to quickly set the basis, it's actually useful to look at Synopsys as really three businesses. EDA, you mentioned it already. It's round numbers, not exact numbers. 65% of the company. We have 25% of the company is IP blocks. Think of IP blocks as pre-designed pieces of chips and, you know, the USB on the back of your computer, for example, is for sure designed by Synopsys and provided. That is 25% of Synopsys, and then 10% is Software Integrity, which is looking at the security of software.
Of course, these three things, actually a theme that is particularly relevant, today, because if there's one big destination, and you can see it literally every day in so many things, which is Smart Everything. Everything is gonna get some smarts. With that comes an interesting, opportunity space, because on one hand, smarts and the electronics behind it have been pushed for many years, the, technology development under the label of Moore's Law, which is scale increase of more and more transistors that simultaneously become smaller and smaller. At the same time, what has fundamentally changed is now every vertical market that you can think of has figured out we have big data, we need to do something with that, and we want to use it, make it smart. That, of course, now brings a pull to the market.
You have this technology push up and the pull from the end markets. The end markets, of course, the minute they see that something can work, you see, for example, ChatGPT giving a whole new wave of awareness of the potential, they all say, "Well, you know, give me chips that are faster, that are more powerful, that can handle more data, 1,000x more." In order to do that, there's a lot of challenges in our field, but these are the precisely the challenges that Synopsys is extremely well-equipped for and well-webbed into the rest of the ecosystem, the people that are manufacturing, the people that are designing, the people that are testing the chips, and so on.
Moreover, there's an additional new set of markets that open up, such as automotive, that suddenly has figured out, well, smart needs to be not just smart, but needs to be also secure and really safe because errors could have disastrous impact. That is why the combination of all of these requirements are so well suited for providing a massive amount of technology, and we're in the midst of that. We ourselves are big consumers or developers of smarts, and our own field is making fantastic advances by using AI to automate and get better chip design going. This is one of the highlights, certainly, of Synopsys from a technology leadership again.
With that general preamble, I'd like to conclude with the fact that we have given strong guidance for this year, 14%-15% growth rate in a market that is somewhat wobbly, and every other day you can see is it over or is it the recession still coming? Well, you know, read The Wall Street Journal every morning, then you will for sure not know. On that basis, meanwhile, we keep moving forward very, very rapidly with the technology.
That's fantastic. Thanks for that, Synopsys. No pun intended. I guess let's kind of bring it down to investor level, just high level here. The iShares software index has been down 20% since the start of 2021. SOX has been up monthly, up about 9% over that same period. You guys sit at this beautiful intersection between software and semis. Synopsys has actually achieved a 43% return over the same period, stock-wise. What do you think has enabled investors to embrace the company at a time when other companies in software and the sectors overall in software and semis have struggled?
Well, it's always a combination of opportunity space and then execution with that opportunity space. you know, it's always better to be lucky than smart, maybe. The reality is that we positioned Synopsys now for many years to be at the fulcrum of all of these demands coming from the end market and the technology developments that are going through some very substantial discontinuities. At the same time, you have to execute really well. Our execution really falls fundamentally into two or three categories. One is a very high attention to the technologies that are going through discontinuities. Just to give you an example, you know, for many, many years, the focus has always been how can you put more on a chip and more on a chip and smaller and faster and so on.
While that is continuing, it's insufficient. There's a breakthrough that is heading now to let's do multiple chips at the same time that are very close together, and out of that, there's a 100 to 1,000x opportunity right there. Technology leadership is key. The second thing is relationship with the customers. The more there's complexity growing in what we have to do together, the more the word trust actually really matters because trust is built entirely on everything that can go wrong and how do you act at that moment. That is actually where the team play, or just like in a band...
Yeah.
-if somebody plays the wrong notes, the rest of the band has to make up for that in a second, right? Then the third one is ultimately execution by the company. In the last few years, I think we have put a high emphasis on driving the products and the capabilities, but also enhancing the financials fairly systematically by improving the ops margin and also having a high degree of focus on the growth. I think that's been very visible as we have relatively quickly passed from $3 billion to $4 billion to $5 billion, and now comes whatever the next one is.
Yeah. Look, I think the only thing better than growth with profitability is predictable growth with profitability. Can you talk a little bit about visibility and predictability of your model, your financial model?
Actually, one of the things that Synopsys pioneered in the software world was this notion of a model where you recognize revenue over time. Instead of selling software and maybe having a support fee later, what we moved into in the early 2000s is a model where we think of it as renting the software for typically three years or so, and recognizing the revenue as it gets consumed. That brought a very high degree of stability and also a different negotiation position with the customers because we're both interested in having a stable relationship over a long period of time. Not all of our business falls in that category, but certainly most of the EDA part, all the software part of EDA, the part of the Software Integrity.
Some other parts, such as the IP is a little bit more lumpy over time. In aggregate, it has given us a very high degree of predictability. Of course, the other part of the predictability is to make sure that the relationship with the customers are such that there's an interdependence. We dependent on them and them dependent on us in a fashion that is constructive.
Got it. Kind of taking a step back now to shifting to strategy, what is your sort of long-term vision for the company? As a second part to that question, I promised your team we'd talk about AI because you get.
Oh, okay.
About that. Let's talk about your long-term strategy and your vision first.
Well, the long-term strategy is look at the end market and, you know, we actually coined the term, Smart Everything already, I think in 2011 or so, because we saw all of these forces gradually coming together that would make it possible. And, you know, with many big concepts, you know, it's understood that you wish for something, and it's understood that there's a lot of things to actually do to get it right, and that took a good decade to really materialize. And then suddenly it's like, well, it's obvious. Everything is gonna be smart. How come you're not moving faster? How come you cannot provide another 100x? And actually, for this year, for this decade, sorry, we have set ourselves the goal to deliver another 1,000x in productivity.
Just for the record, in our previous history, Synopsys delivered 10 millionx , right, in productivity. We've been somewhat foundational to this entire field, and so we will continue in that direction. You mentioned AI. AI from my perspective is just a completely different computational way to deal with data. In, in somewhat simplistic terms is, you know, traditionally we have essentially had deductive reasoning. If this, then do that, or if that, do this, right. It choices. That has led to tremendous advances in automation for, you know, 50 years. The AI capability is really the ability of that sort of looks like this thing, with other words, the pattern matching. Pattern matching is something the human is unbelievably good at.
You know, when you drove here this morning, most of that you did completely unconscious, because if you were conscious, you would be stymied, you couldn't do it. Yet, by the time you apply AI, it has two characteristics. It is not as good as the human in many practical things that we do today, but it learns to optimize gradually, and it can see more dimensions than a human. Just take maybe a little wacky example, you're driving here. If you could see the temperature of the tires of all the other cars, you'd see some are pretty red. Those are dangerous drivers, for starters. Well, it's a little wacky example, but this is exactly what we do in domains. We see relationships that before were not visible, therefore we can optimize around them.
This is where the computational technology is now sufficient to do this. In order to be sufficient, it needs to handle a lot of data. When we look at the data of a chip, we are quickly talking about 1 trillion or more things to move. I always ask our own team, is our AI so successful because it found one thing that changed the whole picture, or no, it found 3 billion little things that all improved a little bit? I think the answer is yes.
Mm.
We do both.
That's good. Look, I think that addresses sort of AI from the solution for the customer, right? Like the.
Yeah.
-are solving for. How have you employed it internally to improve your own sort of, development?
Well, when I said, you know, you quickly have 1 trillion things to deal with, that's what one chip is. Yeah. It is unbelievable how much is on a chip because, you know, the transistors are now moving in size below nanometers, which is 10 to the minus 9, to angstrom. Right? Angstrom is, well, it contains the word angst. I always like to highlight that. With other words, it's not so easy to do. When you have that many details, a single detail can stop things. This is maybe one of the other characteristics that I would like to highlight for this whole next wave of technology, but also the whole next wave of industrial interaction. We're no longer dealing with additive problems, we're dealing with multiplicative problems. So a big set of interactions.
You know, I'll always like to highlight that in English we often say success, it's the sum of our efforts. It's not. It's the product of our efforts. A single zero, and we all have zero, right?
Mm-hmm.
If you take, you know, from the minute parts of the chip all the way to driving a car autonomously without killing somebody, you can see how that multiplication better be pretty good.
Yeah. No, that's great. That's fascinating. As we think about sort of pivoting a little bit to sort of the business composition and business mix, right? You talked a little bit when you opened about what the business is split between EDA, IP, and application security.
Yeah
... the SIG business, if you will. How do you see that evolving over time, given where you're trying to take the company to mix itself?
I think the mix will evolve a little bit. You know, the smaller businesses tend to always grow a little faster initially, that's where the we call it the Software Integrity.
Mm-hmm
... security testing is a good term for it. Has the potential to touch many, many things that are actually fairly distant from what the traditional home of Synopsys is. Although that distance is becoming smaller because the way many new products are conceived, and I'll use the car again as a great example, people are now talking about a car as a software-defined product, right? Well, that software-defined product itself, the software is defined by the workloads, and the workloads may have a lot to do with, you know, understanding the traffic and all that. There's also workloads that wanna know, "Oh, you were looking for McDonald's, you know, here it is." Right? There's so many opportunities there. The word that ties this together for me is really security.
Mm-hmm.
Making sure that things are safe from a being hacked perspective, and that obviously is extremely relevant. Then for some markets, you add the word safe because if it, if it can harm human life or impact it, be it medical automation or robots in general or automotive, then they have additional constraints. I think that that is over time a continuum. By the way, at the very bottom, we also have security devices in our IP collection that ultimately have to connect to the safety at top level. I'll add one more thing, which is a new area of growth, which is what's called Silicon Lifecycle Management.
That is to say, well, what can you put in chips so that they can self-diagnose sufficiently, so that when you're driving on the highway and one of the chips inside of the car feels, "I'm not feeling so good," how does that chip communicate, "I'm not feeling so good," to the system, to the dashboard that then says, "Okay, park the car right now," or, "Here's a short prayer," because otherwise it's gonna be too late. You know, it's that goes from the most minute physical details all the way to system decisions that have to be very smart to not endanger even the stopping of the car.
Got it. As I think about that answer, right? One thing comes to mind when you talk about the chips not feeling great, right? Data and the explosion of data obviously has a lot of different requirements. You folks have optimized designing chips and helping folks produce chips in a way that maximizes the ability to, you know, retain, transmit, process data. You know, among many things that you have to manage is heat, right?
Yes.
The power sink and whether or not that is dangerous to the broader electronics or the chip itself. Can you talk a little bit about your innovations in that regard around power and power management?
Sure. You know, power really became a topic in the early 2000s. Not for the heat reason, but for the battery life cycle reason. Remember, early 2000s really where the mobile phone came about, and while initially it was just a phone, within a matter of a few years, it was a complete data machine. That is where the initial emphasis went on to power. Before that, chips were optimized for performance, how fast, and area, how small can you make them, because that impacted cost. Power became the third leg on that. You're absolutely correct that as we pushed more speed into the chips, with the speed comes a high degree of power utilization. The power is used to switch something, but then has to get out in the form of heat.
Heat, dealing with heat has been a challenge for now quite a number of years, initially on a single chip, now between multiple chips. There's also a whole different dimension to that, which is, well, is your software gonna misuse computation and produce too much heat, or can you actually write software that is smarter in the computation? This is where it's super interesting for us because we have super high-speed simulators, so things that actually can mock up a design and make or have the software run on it, the ultimate software, and tell you, "Well, this thing really switches a lot here. Can you reduce that?" Because all of these switches, by the time you turn it into hardware, are actually gonna produce heat.
The power problem is not just a horizontal technology issue, it's also a vertical stack all the way to the end applications. You know, Synopsys is a leader in that area, but it's an area where we also have very strong partners, and one of them is Ansys, for example, as we have some of their source code in our product and a very active participation on driving the state-of-the-art.
Got it. Okay, well, on the topic of power, you know, I know you love to talk about sort of obviously what's next. You know, an exciting thing in semiconductors always is sort of the advent of new materials, right?
Oh, yes.
As we think about where things are headed now, we've been talking about lasers and photonics for some time. You know, right now, GaN and silicon carbide are super top of mind for a lot of folks, and rightfully so. How are you positioned to sort of exploit that pivot and that trend towards those new materials?
Okay. One of the things that is not that well-known because it's relatively deep technology is called TCAD, Technology Computer-Aided Design. It is very much predicated on the question of can you simulate, so mock up very small physical phenomena such as a single transistor, and understand what happens when you change the materials a little bit, what happens when you change the thickness, the orientation, and so on, and actually predict the behavior of these. You're absolutely correct that there is still a wave of new changes, initially very much driven by lithography, so how small can you make things?
Also now driven by materials. Can materials learn to align themselves instead of having so many manufacturing steps? Also new materials that are better suited for power devices, for example, actually, gallium nitride company was just sold to Infineon.
That's right. Yeah.
literally last week or so because they're focusing for automotive on massive power devices. These are things that we can simulate and that we can help optimize.
Right.
That is at, you know, as close to deep physics as you can get.
Absolutely. Put in, you know, dumb layman terms, it's enabling the things that enable the things to happen, which I think is.
Yeah. Well, you're not that much of a layman. Actually, it's enable to enable to enable to enable to enable and then happiness or whatever.
Whatever it is, the result.
Right.
Yeah. Well, look, switching gears a little bit again, you know, talk about your application security business, the SIG business. Sorry.
Software Integrity
...Software Integrity business. It's becoming increasingly important, right? Unfortunately, bad actors, you know, hacks, all sorts of problems with the integrity itself of the software. The company's been pretty aggressive over a pretty long period of time in building that business out. What does the future hold for that business? You know, that market, some would say, remains fragmented in many different ways. Is there a plan to sort of be the company in that regard, or what's the strategy there?
Well, I think in our area, we're already the largest and, you know, we're imminently close to passing half a billion. It actually starts to matter in the overall P&L, of course. It also matters in the sense that it's another puzzle piece in a puzzle that looks at security as having many different layers. Just that enablement joke we just made, it's actually real, right? We see that a lot of opportunities. It's still a fairly young market. You're absolutely right. It's very fragmented. You're also right to observe that in contrast to physics, which you can view as the enemy or your friend, but in fact, physics is actually a pretty neutral thing. Hackers are not.
You know, now you're dealing with human ingenuity that will constantly try to find some other breach opportunity. I think that is a field that will continue to grow in importance and also in sophistication.
Mm-hmm.
The very fact that so many things are becoming interdependent and it's gonna be difficult to verify a complete system, the push on having sub-subsystems to be verified and checked out at a much higher degree will continue. There's an opportunity for growth and execution there. You know, Synopsys has at least the experience of doing that in a couple of fields now over the last few decades. There are many similarities, although it's a much broader ultimate TAM because every software company in the world is potentially part of that, the technology is not dissimilar.
Yeah. Aside from the technology similarities and sort of the development core competency you can apply to that, there's a misconception that it's not aligned with your semiconductor business. That's a pretty broad misconception. Can you address that a little bit? It is linked.
Sure. Well, you know, let's start with the around 2000 misconception that we shouldn't be in the IP business. You know, I always looked at IP business as a shortcut. Instead of building everything yourself, you have a catalog of pieces that you can reuse. Initially, people said, "Yeah, but that's not your business, that's not your job," and so on. You know, the reality is they get tied over time. Obviously, we have to show over time that there's more synergy between these areas. We've always invested in technologies that we felt were adjacent either by technology or by channel or by end customers, and that would become important over time in the overall equation.
In that sense, I think that Synopsys has acted for the last 30 years in the, in the notion of multiplications rather than additions.
Mm-hmm.
While it's being viewed as an addition, that doesn't mean you continue to invest. It's just that over time, you understand how the interdependencies will grow.
You touched on your IP business. You know, it's overlooked sometimes that you with all the noise around Arm, particularly the Arm-Nvidia deal not that long ago. That you folks are actually the number two overall IP player, and you actually have a number one position in some very important IP categories. Can you talk about some of those around, you know, again, some of the similar themes around IP and data center, and AI?
Sure, sure. Actually, I'm glad you mentioned Arm because they're the number one provider of IP from a revenue point of view, as everybody knows, looking at going public again. Also a very good partner of Synopsys. The way to think about Synopsys is we're everything else, everything around Arm. Of course, when you say around, right there you see connectivity in so many different ways, right? I mentioned simple connectivity. Everybody knows USB. It's really old, except no, it's not old at all. There's new versions all the time. Secondly, there's new versions in terms of new silicon technologies. Designing these interfaces in 2 nm is very different than 10, which was very different than 135, right? Over years, this keeps evolving.
Now on top of that, you have a massive increase of new needs as the whole configuration of chips is changing from being chips on a board to now being chips literally on other chips. This is called multi-die. With multi-die comes the whole question of, well, how fast are the signals between the chips? Because that's always been the challenge. There's a reason why we like to have as much stuff on a chip as possible. It's really fast inside, right? Whereas if you need to go somewhere else, you have to take the elevator down to the next chip, or you have to go around, you know, the neighborhood corner. Can we speed that up? These interfaces are now a whole next wave.
There's a little bit of an illustration how all of these things hang together, how you do design, what are the tools to do design with, and what are the building blocks. I like to think of them as super sophisticated LEGO blocks.
Mm-hmm.
LEGO blocks, you know, people that know LEGO blocks, they know, A, they have grown in sophistication over the years. Secondly, they have been the masters of connectivity. The fact that these plastic things 40 years later still work is completely unbelievable from a chemistry point of view.
Absolutely.
Right? Well, we do the same, but electronically speaking.
Fantastic. That's a great analogy. Now that we've kind of done the walkthrough of your overall business and the three sort of business units, I'd like to open it up for any questions from the audience. One or two.
Sure.
Yeah. For everybody.
I'll repeat.
Yeah. Yeah, please go ahead. Yeah, sorry.
Yeah. Two questions. You know, RISC-V, what about it? Secondly, quantum computing, do you have it in your kitchen in a week?
No. So the RISC-V obviously is yet another processor. There are multiple processors on the market. Arm is the best-known one. Actually, the x86 is the best-known one, which has been used massively in general compute, I would say. RISC-V is an open standard with all the positives and challenges of open standards because making things work is both a technology challenge, but also ultimately a responsibility challenge of who do you go to. There has been certainly an increase of interest in RISC-V, and a number of people are trying it out or investing in it.
Some countries that want to be less dependent on the U.S. are looking at that as well. We have supported RISC-V for quite a while. We have many tools that are optimized for it, and so we're engaged with that. Quantum computing is a bit more distant from the silicon world today. The interfaces to quantum are going to be important. Synopsys had for a number of years an effort in a cryogenic computation, which is essentially the circuits that work at very low temperatures. Now, quantum is a step below that, but there too, think of it as puzzle pieces that we look at putting in there in time to be of interest.
The other thing I would say is there's already a lot of effort, specifically in the security side, for quote, "post-quantum," necessity of protections. It's based on the assumption that, well, if quantum suddenly works, you know, all secrecy is gone. Well, I don't think it will be. Instead, what's already happening right now is that people are in the traditional mechanisms trying to increase substantially the degree of protection. There's a lot of reactions around it. It's a field of high interest, and that's sort of my assessment of where we are today.
Good. One more question.
Sorry. You've been talking for a long time about AI. Congratulations.
Thank you.
I'm just wondering from your perspective, how much of an inflection is generative AI? Is it big? If so, how can you facilitate it?
Well, you know, there are two inflection points, what you believe and what you do. You know, what I find so fascinating about ChatGPT right now is that instantaneously it became a visualization of, wow, this is cool stuff. Let's see what it can do. Once you start putting the let's see on it, you'll find that there are many things that there's a lot of learning going. By everybody asking, you learn faster, right? This is not any different, I would say, than when the Go game was cracked or when many years earlier, the chess game was cracked with very different approaches. Suddenly it's that there's stuff here that could probably be used in many different directions.
Instantaneously, there's a battle royale of all the big guys that say, "Well, you know, we better up, you know, our search capabilities and our representational capabilities." There's a whole set of openings on can this be applied to different industrial applications or learning or what have you. Of course, you know, every college in the U.S. now is looking at, so those essays, where did they really come from, right? So change is a spark. The spark of the perception actually drives the spark of the technology because it accelerates things, and there's attention. This is one of the additional reason why I'm saying, you know, Smart Everything is upon us, and it, you know, it certainly feels smarter than previous searches.
There's a lot of questions, you know, intellectual property protection. What if you find a piece of your own code suddenly somewhere else? This is a story that has multiple chapters coming.
Sounds like many opportunities too. Aart, with that, we're out of time. We could talk to you for another week, but appreciate it.
Thank you, Marco.
Thanks very much. Yeah. Thank you.