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

Good afternoon, everyone, and thanks for joining us here with Synopsys. My name is Blair Abernethy. I'm a software analyst here at Rosenblatt. And joining us today are Shankar Krishnamoorthy, who is the General Manager of the Design Automation division for Synopsys as well as Trey Campbell. Trey is SVP of Investor Relations, has a long background in the industry as well, having spent a couple of decades with Intel. So thanks very much for joining us, gentlemen. Before I lead into my first question for Shankar, I just need to read off a quick safe harbor statement. Today's discussion may contain forward-looking statements related to Synopsys' current outlook, expectations and beliefs, which are subject to certain risks and uncertainties that could cause actual results to differ. Please refer to Synopsys' most recent SEC filings for a discussion of risk factors that they -- may materially affect these statements. So with that, welcome, gentlemen.

Thank you.

Thank you, Blair. Thanks for having us.

Shankar, I just wanted to set the context here for a little bit. And just for some of those in the audience that might not be as familiar with Synopsys in its current state. Maybe if you could just give us a brief overview of the business, sort of the core markets that you address and a little bit about your background and your role at Synopsys.

Sure. So just to start -- to kind of frame this thing, right? So Synopsys is right in the center of the semiconductor industry. And as we all know, we are in the midst of this Smart Everything transition, where cities are getting smart, automobiles are getting smart, phones are always smart and getting smarter. And what's fueling all this stuff is really the confluence of big data and AI. And the foundation of big data and AI is really semiconductors. And Synopsys is right at the heart of that semiconductor ecosystem, essentially delivering tools, IP, hardware to semiconductor companies, systems companies to enable them to build really complex designs, which are then used to fuel the progress in AI and big data and other types of applications. So Synopsys is organized along 3 lines of business. Our largest line of business is Design Automation, which is about 65% of the revenue. I'm the General Manager of the Design Automation team. The next line of business is our IP organization, which is about 25% of the revenue, which supplies all the building blocks for modern designs in terms of foundation IPs and the interface IPs like PCIe Gen 5 or USB and other types of IPs that most designs need. And then 10% of our business is the Software Integrity area where we are the leaders in application security testing and software composition analysis. My role as General Manager for the Design Automation side of Synopsys is to really set the strategy and drive the execution for this area. We essentially are very closely plugged in with all the leading companies in the semiconductor space and systems companies, and we work with them to essentially pave the way for them to realize their semiconductor ambitions. So basically getting ahead of the latest nodes, whether it is 2 nanometer or the transition to gate-all-around, getting ahead of their requirements in terms of emerging areas like multi-die designs and 3DIC designs and really constantly driving the technology to deliver the best possible PPA, power performance area, to enable these magical chips to get [ created ].

Great. And as we look at AI, which -- I mean, this is the third year we've held this conference, and some of us -- I'm sure yourselves, have been around AI for a lot longer. Certainly, things have really been taking off since last year. I think the emergence of large language models and the capabilities of these models. So maybe I'll just start off with, I want to ask, how is AI, which is a general-purpose technology and lots of different flavors. But how is AI itself impacting your customers and their businesses? And just maybe kind of walk us through what you're seeing as impacting your end markets.

Sure. So we see AI as a very disruptive force. Pretty much all our customers and all our target markets are looking to deploy and unleash the power of AI to either significantly improve user experiences, significantly improve efficiency. And when you look at how that translates to what this means for Synopsys, we are really looking at it from three angles. One is all the AI chips that are being built, whether it is for training or at the edge, basically need a lot of Synopsys tools, a lot of Synopsys IPs, hardware, in order to design these chips, to verify these chips, to manufacture these chips and so -- to test these chips. So essentially, a significant percentage of the spend amongst those companies really is directed towards Synopsys as the leading provider of those capabilities. The second kind of vector is really around what we call as Synopsys.ai, where we made a significant bet many years ago that AI would be a disruptive force in EDA and we have made several investments that put us in a position today where we have deployed AI end-to-end across our entire EDA stack from system architecture, to design, to verification, testing, analog circuit design. Pretty much every element of the EDA stack is being augmented by AI to drive a significant increase in performance as well as productivity for our user community. And then, of course, AI is also a powerful force for us to look internally in terms of how to improve our processes, whether it's engineering processes. I mean, we write a lot of code here at Synopsys, how to dramatically accelerate that with AI, our processes in finance and legal and many other areas. So we see this as -- it's as disruptive for us at Synopsys internally as it is for our customers and our customers' customers.

That's -- maybe we can drill into Synopsys.ai a little bit. The first kind of major component of that, that you put out in the market was DSO.ai. Maybe just help us understand how that has -- what kind of value that's been delivering to customers and what sort of the -- next steps from here for DSO.ai?

Sure. So we pioneered the application of AI to EDA. And back in 2017, 2018 time frame, we made our first investments in this area. And so that resulted in 2020, us launching DSO.ai, which was the first AI system for design. And essentially, what DSO did is it tried to encapsulate what an expert designer would do when they are trying to design a chip. Essentially, they would explore many different choices, whether it is different inputs to the tool, different parameters of the tool. But essentially, AI would do that exploration in a much more efficient way, thereby saving a lot of time for the designer to achieve expert quality results. So that was introduced in 2020. And it took us some time to essentially get users to buy into this new way of thinking and this new approach, but we really hit our stride, I would say, in 2021, 2022, where essentially the deployment started to happen. And early this year, in January, we announced the 100th tape-out with DSO.ai. A tape-out is a production design being completed using AI. And just to show you the trajectory and the pace of adoption, we are here in August, and we've already crossed 270 tape-outs. So this technology is being rapidly adopted by our users, both to push the limits of performance where AI is really searching and finding solutions that just a -- even an expert user was not able to find as well as productivity, where -- as the semiconductor industry is dealing with a huge resource shortage, how do you augment the next wave of entrants with AI-enabled flows so that they get that expert quality results much, much faster instead of spending many, many years to build up that expertise. So both these performance and productivity vectors are where AI is getting deployed. And then in March this year, we extended that whole vision to incorporate all aspects of the EDA flow. So we extended AI into verification, where we have shown some exciting results on how to achieve the ability to get a high coverage for your design much faster. We have extended it to areas like test, where we can cut down the number of patterns you need to test your chip completely, which results in big savings in test cost and test time on testers, as well as moving into domains like analog circuit design, which traditionally has been one of the hardest domains where automation has not really been fully adopted and here AI can be used to rapidly migrate circuits from one process node to another, given there is these wafer supply constraints that all semiconductor companies are facing, right? So we really see this as a broad disruptive technology across the full EDA stack and making both performance and productivity impact and thereby enabling the semiconductor industry to do a lot more in the next decade.

And these -- just to be clear, these tools are really -- the DSO.ai and Verification Space Optimization are really built on top to be used with your -- with Synopsys' regular design tools. Is that right?

Yes, Blair. So essentially, the way these AI engines work is, we've got our base foundation EDA tools, which are getting better and better. Our verification engines and our implementation engines, they are getting better and they are being constantly updated to deal with the Moore's law of progression that is happening down to 2 nanometer and below, but AI essentially wraps around these tools. And essentially, think of it as drives those tools to deliver much better results by intelligently searching the solution space using AI and arriving at optimal points more efficiently. So not only does it -- the underlying tools are an integral element of the AI solution. And yes, the coupling between the AI solution and the tools is really, really tight, which is how we can explore that space efficiently.

Okay. And on the Verification Space Optimization, we saw some really impressive numbers out of the DSO.ai over the last 2 years. What -- can you give us some examples of how the DSO is performing out there? And I guess when I look at it, I go, if I've been a customer using DSO and I've been -- it's worked really well, I'm probably going to not -- I'm probably going to look to VSO pretty quickly. I'm not going to sit back and wait for 2 years. Is that what you're seeing?

Yes. I mean I think we had to kind of spend that first period of time to convince the user community that AI had a place in their flows. But I think we are now moving into rapid adoption of the other AI engine. So just to give you a sense, we announced our VSO or Verification Space Optimization back in March this year. And already, we have like multiple deployments underway at leading the semiconductor companies, mostly because we are solving a very important problem, which is, with all this growth and complexity, the ability to verify designs to get that first-time right silicon is becoming very, very critical. And if AI is able to augment a verification engineer's workflow by finding hard to find bugs by essentially doing this autonomous exploration of the solution space. It becomes a huge arrow in the quiver to help deliver designs first-time right from a silicon perspective. So very rapid adoption. Similarly, on the test side, I mean, test costs, as you know, are very critical, especially for high-volume parts. And there, again, very aggressive adoption of our test AI technology to cut down the whole test times and test costs because it's so critical for any high-volume [ market ].

In an earlier discussion and on your call, you talked about AI being a $500 million business. Can you just describe that to us sort of what -- I think, that's a trailing number -- trailing revenue number or...

That's right. Yes. So essentially, I mean if you look at it, there are about 165 AI start-ups. And of course, there are several large AI chip companies out there, and we have a pretty disproportionate share with those companies in terms of both the tools and the IP. Because all those companies that are designing those chips, they need them to be extremely power-efficient because, as you know, the cost of training is skyrocketing, and their customers are pushing them hard on the total power envelope for the training tasks. So the power efficiency is important. So they are all using our AI technologies in EDA to drive that power lower. The IP is a very critical element since you need the interfaces for all that computation, both the memory as well as between the die. So again, the IP plays a big role. And so yes, our internal tracking is over $500 million of trailing revenue -- 12-month revenue assigned to AI-related designs.

And that's -- I guess, that's primarily from DSO, from the first component of this, is that right?

No. So when we say $500 million, we are really including the DSO piece, the underlying tools that are being used by those AI companies and the IP and hardware, too, because they need hardware to verify these very complex systems along with their software workflows. So all that put together is...

That's fantastic. That's quite an accomplishment in a fairly short period of time. I think, in terms of just uplift in pricing, we talked a little bit with Trey earlier about this, but maybe you can talk about your total addressable market, maybe in terms of in EDA and how AI is opening that up for you.

Yes. So from our perspective, if you look at all the software tools that we have in the design side of EDA, like every one of them can be augmented with AI to further improve performance and productivity. And so when we look at DSO, we went through a series of experiments to better understand how best to monetize this. And so we kind of converged on really two or three different choices. The first one is, of course, doing -- kind of really doing statement of work, proof-of-concept type of exercises and that's still available for customers that want to do a specific project with AI or want to see the benefits of AI on a specific design. But then as the customers pick this up and ran with it and started to deploy it broadly, we essentially evolved to a subscription model for AI, where essentially, if you have 100 blocks in a design, a customer may choose to apply AI to maybe 30%, 40%, 50% of those blocks and essentially, they acquired those subscription licenses, and then they need adequate amount of underlying EDA tool licenses for the AI to drive so that you can do that exploration efficiently. And then AI and compute sort of go hand-in-hand because essentially, there is a significant -- more amount of computation needed for AI. And so with our cloud offering, we have also bundled the whole compute element with the subscription license element to provide essentially -- think of it as AI untapped, for customers that are looking to get the benefits of AI without necessarily building up large on-prem compute infrastructures. So there are multiple ways in which we are going about the monetization. There are some several proof points to show that our strategy is working. So a couple of -- if I look at some of the renewals that have come in, in late '22, '23, where essentially, you see the AI getting deployed. We are seeing an uplift of 20% in the implementation business, which is where DSO is tightly coupled. And as the verification and the test renewals come in under that technology, we expect a similar type of uplift. Both from the AI subscription licensing as well as the additional licensing needed to run the EDA tools underneath.

How is AI impacting your IP business?

I would say significantly, right? Because firstly, as you know, we deliver IPs to pretty much every AI chip out there. So that's a big, big part of pretty much every AI company's designs. But then we are also using our own AI technologies we are building on the design side within our IP teams to improve their results. So whether it is DSO.ai for the digital IPs or using the analog AI technology to accelerate the design of the [indiscernible] and the custom mixed-signal IPs. So it's both an internal productivity benefit as well as, of course, delivering more IP to the whole AI ecosystem in terms of new designs.

Interesting. So you're using DSO internally with your own design engineers?

Exactly, exactly.

Is there opportunities to use VSO or TSO as well or?

Especially VSO, Blair, I think, again, very -- when our own internal teams have seen the benefits they've gotten in terms of verification. They have aggressively moved to adopting VSO as well internally. TSO, since we're not really in the business of manufacturing chips and doing product engineering around chips, TSO is really mostly with customers at this point.

Interesting. Okay. And then you touched on a little bit sort of the analog and manufacturing side of things. Where are you at on that front?

Yes. With analog, all the major foundries, right, who are doing advanced nodes are all talking about this analog design migration because what they are hearing from their customers is the need to rapidly move designs from node to node, and a pretty significant chunk of their designs are really analog mixed-signal circuits. So the ability to use AI to drive significant automation in that is being welcomed. We are having, again, a lot of interest in that technology. And we have several proof points that -- like, we recently had a presentation at the Samsung Ecosystem Forum on this technology, which was very well received. And similarly, with the other large foundries, again, at their forums, we are having similar presentations talking about using AI to drive that. And it's a big deal because for years, right, the whole analog mixed-signal area has not moved at the same pace as digital because of lack of automation and more manual work needed. But with this engineering shortage that's out there as well as the availability of AI, this could be the confluence of factors that drives more automation in that domain. On the manufacturing side, again, Synopsys is a leading supplier of TCAD and manufacturing software to foundry R&D teams and manufacturing teams. And here again, AI is very, very important because there's a lot of exploration that happens, whether you're trying to form up your OPC recipe for your computational lithography or whether you're trying to come up with the right process model for your -- as part of your TCAD process integration step. Again, AI can be applied to all these problems in order to both improve the quality, whether it's yield or whether it's the transistor performance or the productivity of the engineers to those tools.

That's great. Just to circle back on a question of pricing. When you look at a customer that maybe isn't using any of these products and they've traditionally been buying seats and so forth, they're coming up to look at DSO or VSO. How does the pricing model change? I think you said, is it more of a kind of consumption model to the AI tools?

Yes. So we have a variety of business models. So for -- we offer everything from a consumption model through a flexible spending account type of a drawdown methodology to subscription licenses where they basically buy subscription licenses for 1 year, 2 year, 3 year or SOW model, where they just do an outcome-based engagement with us, and we deliver the outcome and basically get paid for it. So we have kept it very flexible because, again, different people are looking at AI differently, and we want to be flexible in order to accommodate their needs. What I would say is that once a customer moves to broad deployment as many of them are, everybody essentially settles down into either a consumption-based or a subscription-based licensing model because they're looking to deploy this across a significant percentage of their engineering community. And so that's the business model that they gravitate towards. And then with the cloud we get an opportunity to also look at the consumption model, not just from a software perspective, but also from the underlying infrastructure perspective, because for a small company or a small group in a large company, you may not have the compute resources needed to do that exploration. And that's where the Synopsys cloud solution with the AI technologies in it enables that -- access to that as well.

Great. How do you size up the competitive environment these days for Synopsys? And I mean, there's definitely a tremendous amount of innovation that's happening in your space, but how do you think that's going to change the competitive dynamics over the next couple of years?

Yes. So firstly, we very much respect all our competition. And I think as you rightly said, it's a very vibrant space right now in our business. But we are the clear leader in all the 3 categories, whether it's EDA or IP or hardware, Synopsys is the clear leader. And also, we are the innovator in all these areas, right? I mean, if you look at the EDA space, we pioneered AI, we pioneered multi-die and 3DIC and the new approach to that. We pioneered the whole SaaS model for EDA with respect to essentially enabling EDA untapped. So we are always sort of -- we kind of are placing those bets, making those explorations and trying to drive the customers and the industry forward in terms of all these bets that we are placing. I would say that, I mean, we have, of course, these strong leadership positions, but we don't take any of it for granted. I mean, essentially, what I joke to everybody is the ground shakes under our feet every 12 months. And so if we are not agile, if we're not on top of our game in terms of technology and innovation, whatever lead way we have, it can fritter away. So we are just very, very paranoid in terms of staying on top and really driving a very strong innovation mindset and velocity within the teams.

That's great. That's great. Yes. Just shifting gears again, the -- maybe we could talk a little bit about some of your end markets. China, in particular, has always been -- it's been in the news a lot in the last year or so. Just give your sense of that market opportunity from your side. And in an earlier conversation today, we were talking a little bit about piracy of software. And maybe you could just touch on some of the things that you're doing on that side to protect yourselves?

Yes. So I think China has been very interesting. As you know, at the beginning of this year. We had -- we were dealing with lockdowns, then we had that big COVID surge. So really I mean the first 2 quarters of the year were quite rocky in terms of just how things were going there. And then, of course, all the export control and the entity list actions as well, I mean Synopsys is in full compliance of all those requirements. But I think we have definitely seen things picking up in Q3. I think that the markets have opened up and across all the lines of business, EDA, IP, hardware. We have seen some good momentum in Q3. And regarding the types of -- I think there's a lot of progress, I would say, in China, in the automotive area. I mean, they are definitely -- I think I saw some statistic, I think 1 in every 3 cars in China is an EV, and there are many, many players across the OEM and Tier 1 and Tier 2s, who are essentially trying to disrupt. And also their design cycles are 3 years as opposed to 7 years for many of the other automotive companies. So I think that's an area of high velocity, high agility and Synopsys is obviously playing a big part there with our IP portfolio as well as our EDA and hardware offerings for that market. As it -- specific to piracy and so on, so we do have license compliance programs that -- like most software companies and essentially, we sort of use that to keep it in check and have several technologies split throughout that.

Is it a problem that has been fairly stable for you for the last few years? Or is it getting worse? I'm just in a different space. Autodesk has really been -- they've put together a pretty systematic approach to dealing with and identifying and softly encouraging customers to get compliant. Just wondering what you guys have in place?

Yes, I think our strategy would be very similar to what you just mentioned. Is really -- have that soft yet firm approach around compliance. And I think this year, we've rolled out some new programs, which are showing some promising results. But we're still early in that game.

Yes. Okay. Is the -- the leadership change announced last week with Aart moving up. Any thoughts there in terms of the impact or change to the organization or just -- how are people taking it?

No, I think we are all super excited with Sassine moving to CEO and Aart moving to Executive Chairman. I think Aart and Sassine have been really a team for several years now. And I think the 3 vectors that they drove the company on, which is having this growth ambition, scaling the company to the next level and then maintaining that strong technology leadership and innovation pipeline have translated to 17% CAGR over the past 3 years, 700 basis points of ops margin expansion. So I think there's a lot of good work that has been done, of course, Sassine will bring his own perspective, and he will no doubt take the company to the next level. But there's been just an exceptional transition that has been planned out, and we're really excited about what the future holds for us.

That's great. That all takes place in January, I believe, January 1...

Yes, that's correct.

That's great. One of the things you touched on earlier was just ways -- looking at ways in which Synopsys internally can leverage AI technology. So you've got a lot of developers working for you. What have you guys been experiencing? What do you -- are you getting productivity boosts from using some of the LLM-driven technologies?

Yes. I mean, we are super excited about all these new and exciting ways to boost engineering productivity. I mean, as you know, that's really the core of Synopsys. I mean, we have several thousand software developers that build tens of millions of lines of code day in and day out. So it's one of the big areas where we've made a big investment. Now of course, the thing -- the key thing with this LLM-based solutions is really the accuracy and the ability to manage hallucination effectively, right? So they are -- so that the whole scaffolding that needs to be applied around an LLM to ensure that the outputs are correct most of the time and the hallucination can be controlled is really something that is very key to the success of these systems. So we are definitely very, very interested in deploying these types of systems to improve our productivity. And also, as I mentioned earlier, we also have a large team that develops mixed-signal IP and digital IP and again, bringing our own AI technologies to that community and getting them to aggressively adopt it is another way in which we are looking to boost the engineering productivity. So all in all, I mean, pretty much the same-store messages and the drive we have with our customers, we have the same drive internally because the AI offers us all these opportunities to get a lot more efficient in terms of how we build software and we build IPs.

If you look at your IP portfolio, Shankar, maybe can you describe sort of where you see the biggest -- what are the biggest growth opportunities over a little bit longer term you're seeing, around 3-to-5-year kind of time horizons because you really have to invest now and build now in order to deliver a couple of years down the road. So just maybe walk us through where you see the most exciting opportunities in the IP business.

Yes. So I think that -- I mean, the IP business is an area where we have to keep a very, very close watch on how the industry is evolving, right? In terms of standards, whether it's your PCIe, CXL, whether it's your die-to-die IPs, whether it is all the different types of interface IPs and also on the foundation IP side. So we have a very, very strong and industry-leading team that is on many of these standards bodies that is driving the next evolution of these IPs. And of course, Synopsys is always the leading silicon on many of these. So we put -- make huge investments to ensure that working silicon, that latest node for the latest protocol and have that be functional because our customers' road maps essentially depend on Synopsys delivering silicon-proven IPs to intersect their programs, right? So a massive investment there, and we have really kept that strong leadership position. I think one of the exciting things that has happened this year is we are essentially deepening our ecosystem partnerships with companies like Intel that you may have seen the announcement a few weeks ago. A very, very broad partnership, providing Synopsys IP at 18A as well as the 3 nanometer. And then a couple of months earlier, we had an announcement with Samsung Foundry talking about enabling the Samsung Foundry offerings with the Synopsys IP. And then, of course, our enduring partnership and collaboration with TSMC on all their latest nodes and providing our full IP portfolio. So I think essentially for -- with all these fabs and all these foundries getting built with the chips acts around the world, IP is one of the essential capabilities in addition to, of course, tooling and all the other underlying manufacturing capability that's needed to fill the fabs with designs. And so Synopsys is really playing a very important catalyst role in terms of enabling that across the different foundry ecosystems.

Your Intel announcement is broader than -- it really broadens your relationship too, right? You're doing more than just building IP for their customers. They're actually going to be utilizing some of your IP, is that right?

That is correct. That is correct.

Is that a -- I'm assuming that's a multiyear situation or contract?

Yes.

And then I guess if you -- again, just back to the beginning of my question, if you were to, say, pick 2 or 3 areas in IP that you think are really going to drive things, what would those be? I mean, I think interconnect -- multi-die interconnect, do you think that's going to be big enough? Or is it going to be niche...

I think multi-die is going to be very big. And it's not just a one-size-fits-all solution, right? I mean, of course, UCIe is very important and much of the industry is talking about UCIe, but there are many other ways in which companies are looking to connect their dies. HBM3 is an example of a protocol where basically compute and memory is being connected up. And then there will be other requirements. So the whole die-to-die connectivity, whether it is compute die to compute die or compute to memory and all the different requirements and the cost versus performance versus power trade-off points, I think, are going to essentially open up a lot of opportunities here. The other area where we see a lot of potential is really what we call as chip telemetry, right? So essentially, as you kind of build out the electronic automobile with many, many chips on it or a data center with the cloud silicon, you're going to have to really look at the uptime of these applications. And so having telemetry and telemetry IP built into all these designs is going to become critical so that you can read the signals from these designs and then use that to do predictive maintenance and those types of activities. So there are many trends emerging. And again, we are really humbled to be in the middle of many of these discussions with our customers.

The telemetry opportunity is that -- is this really what if -- we would call it digital twins, is that really what your -- what you going after there?

We call that initiative as silicon life cycle management. So essentially, the idea is to insert monitors and sensors into the device and then extract that data from those monitors and sensors across the whole life cycle from product ramp to high-volume manufacturing to in-field and then use that data to essentially optimize the performance of that device. Like, whether if it's -- the temperature is climbing up and throttling down the frequency or if the device is aging over time, then basically throttling down some of the parameters to ensure that it has a graceful sort of exit. So all these things are all becoming part and parcel of designing these high uptime systems. And again, this telemetry IP plays a huge role in that.

That's great. That's great. What about -- we haven't talked about the TCAD business at all. I mean, maybe just give us your sense of what are the major -- what are the major trends in that business these days? And I'm sure there's some angles in there for AI to help create more value for customers.

Yes, absolutely, Blair. So I think, as you all know, with the rapid progress that has happened, increasingly, it's getting harder and harder in a node transition to deliver the same 2x in terms of power, 2x in terms of die size, 2x in terms of performance that we had even a decade ago, right? Now the gains are more like 10%, 20%, 30% type of gains. But even those gains, they are coming about not just by a new transistor or a new type of material, they're coming about by a step that we call DTCO, design technology co-optimization. And this is where your -- the process R&D engineers sitting in a fab are essentially doing a lot of simulations, a lot of exploration to determine how to put that whole process together from a transistor perspective, from an interconnect perspective, from a substrate perspective and then basically build something that's differentiated. So the TCAD business is really, first and foremost, providing technology to that DTCO process because -- there are several papers that were published in the last 12 to 18 months, which talk about the disproportionately high role that DTCO is playing in terms of delivering that node-over-node improvement. And so obviously, we are delivering a lot of technology to that domain. But in addition, there are interesting new areas where DTCO can be applied, like power devices. I mean, it is a $15 billion industry. And you see a lot of the movements in semiconductors around gallium nitride, silicon carbide and other types of transistors being used for power devices. So there, again, we believe there's an opportunity, and there's a high level of interest in doing a lot of innovation to determine how the power device industry evolves from where it is. And then the third element of this is really that today, TCAD is largely contained in the process R&D teams. But as you look at applications like high-volume manufacturing and product ramp, there are a lot of opportunities for synergies between those departments in a fab and what's happening in the process R&D. So again, we are driving -- a part of bringing the whole EDA together as a stack, is we want to drive this concept of hyper convergence. Where we essentially connect up all these different EDA flows and to deliver the next level of value. So that's sort of some of the exciting things happening in TCAD.

When you talk about power devices, Shankar, how far up the power scale are you guys thinking in this whole world? I mean, automotive is a customer -- the automotive industry is a customer of yours -- becoming increasingly customers of yours. They're obviously doing a lot of work in power and battery management and so forth. Is that an area that would be of interest to Synopsys?

Yes, absolutely, Blair. So when we talk about automotive chips, we always talk about the most -- the most exciting high-compute type of chips like the ADAS subsystems. But the power devices opportunity in automotive is significant, both on the battery side as well as power devices on the whole drivetrain and digital chassis. And so I mean, all those customers who work with us for the latest ADAS or the in-vehicle infotainment subsystems are also talking with us on how to help them on the power devices aspect of that as well.

And is that -- how far along is that opportunity? Is work being done now? Or are they just sort of looking down the road at what their needs might be?

No, it's active work going on right now. There are -- in fact, there's a lot of really good partnerships that we have engaged in with all the leading suppliers in this domain. And again, it's an area where you can almost call it like maybe a sleepy area, maybe 10 years ago, but a tremendous amount of focus on innovation right now, because -- that's why our TCAD strategy is now being extended to incorporate that because a lot of opportunity for exploration, innovation. And again, AI is a great sort of overlay on all this stuff because all of TCAD is and DTCO is about exploration. So by having AI layering over there, again, we see a great opportunity.

That's fantastic. Just before we wrap it up here, we haven't talked at all about your system integrity, SIG side of things, the software side, I think, test side. I know it's not fully under your [indiscernible], but maybe, Shankar how do you kind of look at that business as it relates to your EDA solutions?

Yes. So it's -- as you know, it's a $0.5 billion business for us, right? So it's a pretty sizable business. And I know that the macro market conditions are challenging right now for enterprise software as a whole. But in the area of software composition analysis and both static and dynamic application security testing, I mean, we have a strong leadership position as evidenced by the Gartner Magic Quadrant and other types of organizations. One of the big opportunities here, Blair, is with the advent of gen AI-based code assistance. There's a great opportunity here to essentially take those code fragments or code snippets that gen AI generates and then compare them against a trusted open source repositories and basically flag anything that might be a problem. So that's something that our team is actively working on and has kind of announced. And so I think these types of things are showing how even in a world of gen AI and gen AI generated code, this Software Integrity has a big role to play and -- so even though the macro conditions may be a little challenging, I mean, there's still a lot of potential here.

That's great. Well, listen, we're at our time here. I really appreciate Shankar and Trey spending the time with us this afternoon to give us your perspective on AI. It certainly seems like just a tremendous amount of innovation happening both on your side of the fence and on your customers' side of the fence. So lots of opportunity to add value. I think...

Thank you, Blair. Thank you for the opportunity, and I really enjoyed speaking with you today.

Thanks, Blair. Appreciate you -- see you. Bye-bye.