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

All right. Good afternoon, and welcome to the third day of JPMorgan's 51st Annual Technology, Media and Communications Conference. My name is Harlan Sur. I'm the semiconductor capital equipment analyst for the firm. Very pleased to have Aart de Geus, Chairman, Chief Executive Officer; and Shelagh Glaser, Chief Financial Officer at Synopsys here with us today. Synopsys is a leader in electronic design automation or what we call chip design software and hardware solutions and also a leader in software integrity solutions to the software development community. It's been a busy earnings season. So I've asked Aart to maybe start off with some opening comments, and then we'll kick it off with some interesting Q&A. Aart, Shelagh, thank you for joining us this afternoon.

Thanks for having us, and good afternoon. Thank you for joining us. As you may know, Synopsys is actually doing quite well because we are in a market that is reflecting enormous changes in the world. By now, be it high [per] actually, a lot of it is not high but just great promise. We see a horizon of smart everything. And of course, the advances that you've seen recently in the AI software are just a teaser for what is to come. But underneath all of that is an enormous amount of computation. And the computation, of course, is prepared by the people that develop the systems and the chips that do this, and these are all our customers because what we provide is: a, the design tools to design the chips, and that's about 65% of our business; b the building -- the big building blocks that you can reuse such as the USB on the back of your computer, it comes from Synopsys. That's about 25% of our business; and then lastly, we also have a part of the company that's focusing on security in the software that increasingly becomes more and more important as it also intersects with all of this, and that's about 10% of our business. One of the reasons for the positive and optimistic outlook is because there are a number of breakthroughs that are very fundamental in terms of what will be possible in years to come. You're all familiar, of course, with Moore's Law, which has been for 60 years, this notion of smaller transistors and more of them, but you have also heard maybe it's slowing down. While there's an alternative that is just now starting to open up, which is the notion of taking big chips, breaking them into smaller ones and actually bringing them very closely together. This is sometimes called multi-die, multiple chips or chiplets. Be it as it may, the challenge of that is when you take multiple chips and try to bring them very closely together, it's all about the speed and the amount of energy of communication. And these are difficult problems. We love difficult problems because Synopsys for 37 years, I would say, for 75% of our products has been the state of the art of providing design tools. And here we are, we have a whole additional decade or two of opportunity to have many, many, many more transistors becoming available. And so we see rapid progress in that. The second is, of course, that all of the vertical markets have discovered this notion of smart everything as a great opportunity at the intersection of big data and computation. And because of that, a lot of companies are investing in the technology that is necessary to run this because it will take an enormous amount of computation. If you just take a Chat GPT search, for example, it's roughly 5x the energy that you get out of a normal Google search. So you can see that as this spreads there's going to be a lot of challenges. But challenges are opportunities. And so we are in the midst of that because Synopsys has invested in the lowest level of verification and simulation at atomic level for chips. We work at the design level. We work at the system level. We work at the intersection between architecture and software and we work in the software level, and bringing all of these together in a fashion that will support the ambitions of the world going forward is precisely where we see great opportunities. That makes the question, of course, well, is it not true that Synopsys also is doing AI itself. Yes, indeed, we have been fortunate to invest at an early time, so about 6, 7 years ago in a fundamental push on AI as used in the design of chips. And the results have truly been fantastic, meaning that on entire design flows, we've been able to reduce the time of design from literally months to week -- weeks, while at the same time improving the quality of results and quality of results on a chip means how fast is it, how low power it is, how small can you make it. These have all improved. And at the very moment where there's a sharp shortage of designers in the world being able to drive productivity forward substantially is of very high value. And as a matter of fact, for this decade, our intent is to deliver about 1,000x an improvement. And by the way, for the previous 35 years, we delivered 10 million[x]. So at least we know how hard it is, and so we will try again to deliver on that. There's one difference between the AI that you see now frequently in the press, sort of the generative AI and ours -- and that one difference is that with generative AI, you get fantastic context and ideas and so on. But maybe everything is not completely correct. We have no such freedom. There can be only zero mistakes. A single transistor doesn't work, a whole system can fail. And so the way we have approached this is to make sure that all of our tools are working well together, and the AI is sitting on top of that literally as a governor to run the tools and get better results than where possible before that. This is still at the beginning, and yet we see a very rapid adoption. Our AI capability, which in late January, had about 100 chips already going to manufacturing. Just last month already moved to 200. So you can see a very rapid move with that. And going to manufacturing is a very big decision for companies because if they fail, it costs them many, many millions of dollars. So in the midst of this grand story with many interesting technology challenges all of which we say we're going to link, but it's going to be some work that will proceed over the next couple of decades. So a great outlook. And so you may ask, of course, well, how is the company doing? And so I'll pass it on to Shelagh to give you a sense of how we close the quarter.

Sure. So just a quick update. We just finished Q2 earnings. We were able to beat the high end of both our revenue and EPS inside the quarter and then that confidence allowed us to raise the full year 14% to 15% revenue. And then importantly, improvement of operating margin about 150 basis points. So that was an improvement, and then that flow through 21% to 22%. And as Aart said, we're excellent -- we're executing very well with superior financial results in a very big opportunity space.

Perfect. I appreciate the opening commentary. Let's start off at a high level. There's been a lot of questions around the adoption of AI. I will talk about Synopsys' usage of AI, but as it relates to the semiconductor industry in the tech industry, is it high [indiscernible] is it a reality? What I can say is covering 22 semiconductor companies and looking at their chip design activity, it's real, right? It's going to be manifesting itself at some point in revenues, big revenues at some point, but the design activity on AI processors, but AI processors pulls in the need for more high-performance networking, more high-performance storage, Smart Nick, DPUs. And so we're seeing a big acceleration not only in chip design starts, but for some of your hyperscale and cloud titans that are doing their own chip designs, they're pushing their ASIC semiconductor partners to move up their programs, right? And so I'm wondering -- and we try to do our best to sort of track chip design activity, but I'm wondering if this is manifesting itself in terms of higher levels of accelerated business engagement and so on with the Synopsys team? And is that part of what's driving the strong sort of full year growth outlook for the team?

Well, the outlook, obviously, is driven on one hand by the customer side. On the other hand, what are the capabilities we offer them. And on the customer side, there's no doubt that despite all of the ups and downs and outlook on economy, that is to say the least, a little nebulous. The semiconductor industry is impacted by that, but it's in no doubt that investing in the new chips is absolutely essential to be there as the next positive wave of the economy happens. And in that context, there's an additional driver while for many decades, semiconductors have always been driven technology up. For the first time now, in many years, you have a massive pool the vertical markets down because everybody who wants to do smart everything in their domain has different data, has different needs, different speed requirements, different safety requirements. And so a lot of individual architectures are being developed or optimized specifically for these markets. And this is all visibly in front of us starting to happen in so many different dimensions. They all have something in common, and we like to sort of put our mission under the banner of smart, secure and safe. And of course, security is not only in the software that it traditionally has been the place where things get hacked, but also increasingly in the hardware, in the building blocks. But the word safe is actually an increment on top of that, which specifically for industries such as automotive that are massively going in this direction as well is absolutely essential. And so we have made sure that we have invested in all three words from a value proposition. At the same time, the technologies that underneath have to play very closely together.

You brought up a good point about the move to multichip chiplet-based architectures, right? All the leading-edge chip companies are moving to multi-chip heterogeneous packaging architectures, right? It's a way to drive Moore's Law like performance scaling without having to rely solely on semiconductor process technology scaling Here, the team benefits with your 3D IC Compiler platform for the actual 2.5 million 3D layout and packaging design, but that also pulls a lot of critical IP, right, like your chip-to-chip connectivity portfolio. Can you just give us an update on the revenue contribution, momentum of your solutions sort of targeting this new and up-and-coming trend?

Yes. This is a super promising area because it's, in many ways, the natural place to go when it becomes harder and harder to cram more transistors in a single chain. And by the way, I'm by no means saying that, that is over. There are people now going to your single angstrom numbers size transistors, which is really, really small. But at the same time, the notion of taking certain chips and splitting them in multiple ones or actually bringing, let's say, processing and memory together with maybe some sensors and some communication chips. And if you can bring them really close together, you have essentially a super chip that is made out of different chiplets, which opens up, by the way, also market opportunities for companies that may be focusing on just providing certain chiplets that others can then integrate. And the word integrate always sounds simple, but making things work really well together in a very high degree of proximity is one of the areas where we have seen just tremendous advances in just the last 5 years, I would say, by making more and more connections smaller and smaller and make them also faster and use less energy. And it's all about these things that have to work together. So in that context, we have massively invested in providing a complete design system for this. And this will evolve for years to come because there's so many capabilities. We're working with the absolute top advanced people working in this with around 140 or 150 already in action multi-die systems. And I think that the opportunity for us is great because it's actually complicated to do. And if you had asked this question 2.5, 3 years ago, most exact would say, that's not going to go there. It's too expensive, too difficult and so on. And now here, the other wave is coming.

And it does -- and you do have quite a number of IP building blocks, right, to support that as well?

I'm glad you asked because I mentioned this notion of how do you connect these things? Well, just like you connect your computer to your printer, you have a cable and you have this thing called USB on both sides. Well, there's a USB equivalent-ish thing in multi-die. It's called UCIE. I won't go into the alphabet soup here, but these are essentially the connectors that make it possible to have very high exchange rate between chips. And Synopsys is absolutely the leader in providing the building blocks for that. Think of 25% of our business as -- we're super LEGO blocks for electronics. All the main interfaces we provide. And for example, in the process of the world, [ ARM ] would be the leader with providing a large collection of those and a good partner to us. And so these things have become way, way more complex in the last few years, and having them work well together is absolutely essential for this type of problem.

Your overall revenue is expected to grow at a double-digit CAGR. I mean clearly, it's going to grow strong double digits this year, grow at a strong double-digit CAGR over the next several years. And you're benefiting from the emergence, right, of the custom chip or ASIC chip design adoption by the large cloud titans, large compute, networking, industrial, auto OEMs, right, Google, Microsoft, Meta, Cisco, Apple, Nokia, Ericsson, right? All these guys have amassed significant design teams engaged on a lot of their own chip design activity. This is catalyzing obviously some acceleration in design starts. And so it's also expanded the SAM opportunity, right, for the team because all of these guys need design engineering teams -- so of the double-digit percentage annualized growth target that the team has said, how much of the incremental growth is coming from these systems, cloud titans, OEMs that are engaged in their own ASIC opportunities? And roughly what percentage of your EDA mix to these cloud and systems OEMs represent.

I'll let Shelagh answer the numbers part a little bit. But just to give some color, all the names that you mentioned are, of course, close customers. And it's been quite remarkable how system companies that in the past were very much software only, are starting to push down, and they all realize the same thing. The game is won at the intersection between software and hardware. Because at this point in time, we have long passed beyond the early days where hardware was built and then once you have that, you do software on top. Now it's almost flipped over. And to give you a sense of how much of that has flipped over. Today, a car is referred to as a software-driven device. And so -- or software-driven vehicle in this case. And so the software is actually now essentially the description of what you want to optimize for and don't do anything else, just make it as fast and low power as possible. And so that intersection is really where this all comes to bear. And I always like to cite a small example of how crucial this is because the CTO of Mercedes gave me some simple numbers. He said, for every 100 watts of computation in the car, you just reduced the reach on the battery by 10 to 15 miles. Can you imagine, right now, people are talking about 1 or 2 kilowatts of computation. You can just barely make it to the neighbors. You'll drive very safely to there, but there's an issue there that of balance, that optimization all sits between the software and the hardware. And the thing at the intersection is called the architecture. How do you design this? And the architecture just got this new thing multi-die. So there's a whole set of new opportunities, and that is why people in automotive or the hyperscalers are so much you need.

Yes. And I would add in about 45% of our revenue comes from system companies. So some of these new entrants. But we see growth actually in both traditional companies and in systems companies. And the one thing that's not changing, which is the other opportunity for us is hardware engineers in semiconductor engineers. So I think Aart touched a little bit on AI. We'll go a little bit deeper, but the capabilities that we're giving allow our customers to get more productivity out of those key talent people that they have. And I think that's going to be an ongoing trend that's going to help continue to drive our growth.

Your software solutions and hardware based platforms are touching, I believe almost all of the highest performance SoCs that are being designed in the market today, right? And therefore, your customers -- these are 40 million, 50 million, 60 million, 70 billion transistors on a single piece of silicon is very, very complex, right? And your customers' chip design cycle times are extending, what used to be 18 months, 24 months, maybe extending to 36 months. So they're looking for as many efficiencies on the chip design cycle chip design flow as much as possible, right? And just sort of leads into your prepared commentary, which is the Synopsys team has been developing and integrating AI and machine learning-based methods into your tools for many years now, your AI-based design space optimization platform called DSO.ai was the first solution you rolled out back in '21 talk about the use cases, adoption of DSO.ai solutions. I mean then at your recent Users Group conference, you rolled out a few more solutions that leverage that AI framework. If you could just maybe talk about some of those new products as well.

Yes. So to put it in perspective, our history has been a history of automation and Synopsys has sort of changed digital design by luck or hard work, whichever way you want to look at it 35 years ago. And in an interesting way, I think we're seeing a very similar phenomenon happening. This time, it's not an individual tool doing a step better. It is a whole set of tools that have been harnessed together. We call it fusion, meaning they work really well together. They share information. And on top of that, sits testing DSO, which stands for design space optimization, which may be a mouthful, but it essentially means look at all the opportunities, how you could design something and rein in to the ones that are best and then optimize them until you have a really good design. And a really good design is measured typically in at least two metrics, which is the speed of the chip and the power and very often also the area because if you can make it smaller, you'll have higher yield, you get better economics. And so our capability of DSO that sits on top of now an entire design flow past the Bassmaster already a few years ago. And I think this was an AI pioneering step of being able to get quality of results that was better than what the traditional design flow with the best tools in hand could achieve by quite a number of percentage points in both speed and power. The other step I did, though, it did this in a matter of weeks versus months. That is a very big productivity increase. And we're quite familiar with those because in our history, we did about 10 million of x of that. And so we have a big ambition to keep on that track by necessity, a; because the complexity of the systems will continue to increase exponentially with the new opportunity space, but at the same time, and you heard it earlier, the design community has not grown all that much and therefore, there's actually a shortage worldwide, not just in the U.S. of engineers. And so not only can we move these steps from months to weeks, we also do that with typically on the customer side, only one or two engineers versus a dozen or so, including many specialists. So these are steps that are revolutionizing the adoption rate has been very, very rapid. And the only way to measure that really is to go all the way to, well, did they go to production, right? Because it's easy to say, "Oh, well, you do -- you have successful experiments, you do some test chips and so on, no, no, what chips have gone with DSO. AI all the way to manufacturing. And there, I think we passed the number of 100 in late January. We passed 200 already in April, and we see a continuation of adoption which is faster than normal because customers are very prudent. Remember, there's no room for errors. So they want normally to test out things for many years, before they use them in production. In this case, the value proposition is so great. And so much of the testing is built into what we do, that the adoption is very fast.

Automotive sector, there's obviously continue to be a lot of focus on accelerated compute and AI. But automotive sector is a market where there is a significant step-up in silicon content, right? As you mentioned, I mean, ADAS, autonomous functionality is driving more leading edge compute and networking chips into a car and a significant amount of software that sits on top of that. And then on the analog side, power side, you've got vehicle electrification, strong step-up in power transistor power control. So silicon and software complexity is always good for the team. But where specifically does Synopsys have strong differentiation in automotive.

Well, first, your description is correct. There are two major changes in automotive. One is sort of obvious by now and everybody is going as fast as they can, which is electrification. And by the way, one has to realize that such a change has massive impact on the entire industry because automotive supply chain has the OEM, the automotive company and on an ease a whole set of Tier 1s that do subsystem assembly and then you have the semiconductor vendors. And as of a few years ago, suddenly, there was more and more direct contact between the OEMs and the semiconductor companies because they realized at that point in time, hey, we need to know more about chips and software because the initial fear after hearing about autonomous driving was hey, we're never going to sell cars again to individuals. People are just going to call for their car to come -- for a car to come, it's going to be rentals and selling to fleets. Now obviously, things have not developed quite like that. But the fact is still there that the push for autonomous driving has been quite substantial and is computed in Level 1, 2, 3, 4, 5, where 5 is really truly autonomous, truly safe and go to sleep at the wheel, so to speak, and you should be fine. We're not there. We're not there. And so aside of electrification, this focus on autonomous driving has proven to be an extremely complex problem requiring a very large amount of computation because the car also absorbs a large amount of data on an ongoing basis. All, of course, having to function in real time and all, of course, having to function with minimal energy consumption as it trades off versus the battery life. And so in that context, Synopsys started to invest in automotive, I want to say, well over 10 years from a very different angle, which is in our IP collection, all the building blocks. And the building blocks have to be safe. And automotive has always had a very strict set of rules and standards that you have to pass, and we have now a large collection of IP that is -- that satisfy that. Secondly, we started to invest in security in the software, which initially was not really on the radar scope for automotive. But in the context of SAFE, you can imagine, is actually very essential. And lastly, we have now become more and more a central catalytic point in between the semiconductor people, the OEMs, some of the Tier 1s that will be playing in this as conglomerates are being formed to actually develop the next architectures that actually will make it to probably Level 4 by, I want to say, 2028, '29. That's sort of the sense I get from the automotive folks. And the race is absolutely on because it's going to be challenging. But we're in the midst of that. And it's been interesting because Synopsys had the benefit of literally touching way down the foundries. We know everything about the silicon being provided and way up in the software, some of these companies. And they themselves are all in, let me call it, active learning mode. Because for an automotive company doing their own chips, it was completely inconceivable a number of years ago. And now the question is how much should they do and how much can they actually delegate.

Any questions from the audience? If you have a question, just raise your hand and we'll get -- right over there.