Hello, everyone. Thanks for joining. I'm Joe Gallo at Jefferies, and we're delighted to have Cadence Design Systems today. We've got John Wall, the CFO. I believe you just hit your 27th anniversary, so-
I did. Yeah.
It's fair to say you know a little bit about Cadence. We also have Richard Gu in investor relations as well. I know, Richard, you have a-
Yeah
- a disclosure you'd like to read.
I'll be really, really quick. So today's discussion will contain forward-looking statements, including Cadence's outlook on future business and operating results. Due to risks and uncertainties, actual results may differ materially from those projected or implied in today's discussion. Back to you, Joe.
Awesome. Thanks. I know there are a lot of generals in the room, John, so maybe just to kick it off, a quick overview of Cadence, and, you know, what you want investors to know about you.
Yeah, sure. Cadence is largely a computational software company. It's engineering software. It's distinct from enterprise software. Engineering software typically is purchased out of the R&D budgets of our customers, whereas enterprise software would probably be an IT spend. But the company does provide IP as well as hardware, but predominantly, it's a computational software company. And computational software, when you apply it to silicon, that's EDA. When you apply it to things, physical things like phones and cars and buildings, things like that, that's system analysis. And when you apply computational software to data, that's AI.
We kinda sit at that kinda sweet spot where all of the engineers that are designing all of those gadgets that we can't live without are using Cadence technology to build those chips and those products.
Awesome, thanks. It certainly feels like the golden age for semiconductors. Nobody wants to talk to me as a software analyst. I think most of the people here are here for semiconductor companies. Even the software meetings are asking about CapEx because they're trying to gauge the demand for semis. Yeah, you're one of the few software companies that should truly see a benefit here from the semi rise, and maybe just talk a little bit about the secular growth drivers and how they pertain to your business.
Sure. I firmly believe that I think history will remember our semi customers really, really well in time. I think they'll be described as the Renaissance artists of our time. And Cadence really is the paintbrush company. We're providing paintbrushes to these artists. And you know, thankfully, they're all painting masterpieces with them. But effectively, you know, in this era, now, in this era of semi design and chip design, that when we look at our customers' budgets, their R&D budget, they're probably spending about 88% of their R&D budget on people and about 12% of that budget on tools.
And while there's a secular growth trend in the design or the proliferation of silicon, I guess, you know, as chip design becomes more democratized, that's really, really good for Cadence because there's more people using our tools to design the to design chips. But the, as they're doing that, it's more and more of our—more and more and more of the heavy lifting is being done by tools. But, so, you know, as it's 88/12 now, I think with the advent of AI and with AI doing more heavy lifting, I mean, the growth in complexity of design is far outpacing the growth in human population. So the engineering headcount cannot keep up.
Like, if I go back 10-20 years, started in the 2000s, that you probably had a 500-person engineering team, taking 5 years to design a chip. Nowadays, it's probably a 40, 50-person team, and it's taking 6-12 months. I mean, that's a 100x improvement, productivity improvement. And, and people that buy our tools generally measure them for power, performance, and area, and improved productivity. They're trying to optimize the, the value they're getting from, from each chip. And I think as we provide that productivity, we're, we're getting more of the value. So as you see, semi companies and systems companies do more chip design, they should spend more on R&D. And as R&D, more of that R&D is done by tools versus people, we should outpace that growth in semi R&D spending.
Who are your biggest partners or customers? Like, you know, who's, I know you love all your children equally, but is there any kind of poster child for, you know, how you're helping somebody?
Well, we're in all the things you can't live without, right? So, the big names that you can think of that are designing chips, that they'll all be customers of ours. But, now, we do have a concentration, like 40, top 40 customers, we generally generate about 55%-60% of our revenue from just 40 names. It's probably close to 80% at 200, but, so there's a big concentration that, those are customers that we treat them all like Jensen, right?
Okay.
That they're used to concierge service, white glove treatment from Cadence, and they're partners of ours because, you know, their next generation of products are probably dependent on our tools. That, as we expand more into the system space, we find our customer base expanding, but that's a new challenge for us because we're dealing with kinda low-value, high-volume-
Yeah
... transactions, and we're building our channel partner program and our e-commerce offering for things like that.
Makes sense.
Yeah.
Can we unpack AI a little bit?
Yeah.
I mean, you talked a little bit about how it's making it tougher for your customers because they can't keep up with just headcount anymore. But, you know, how is that impacting your business? Are you seeing more design starts, more licenses, better ASPs? Like, where are we in the AI cycle for Cadence?
Well, we're certainly seeing more design starts, and I didn't mean to, for it to sound like AI is causing problems for customers. I think it's actually helping.
Yeah.
But, I think what's, I think what's challenging for our customers is that the pace and growth in, in all of our demand for technology, is, is just far exceeding population growth. That, the growth in complexity design is, is outpacing population growth. And, and, as, as I said, even without AI, you go back, like I said, 20 years, to go from 500- a 500-person engineering team taking 5 years to design a chip, to, like, 50 people doing it in half a year. That's huge productivity gains, and AI is like that. You know, it's, I mean, we're, we're always dealing with that. I mean, Cadence is a very disruptive company, very innovative company. We spend 35% of our revenue on R&D, and, and new product innovation. We disrupt ourselves.
That when we launched our Z3, our new emulation system, just back in March, it took us three years to go through that production cycle to upgrade Z2, that the previous cycles were six years in length. That so the two best systems, two best emulation systems on the market right now are probably Z3 and Z2. And I think that AI is helping. I mean, the AI is helping doing more heavy lifting that for customers, you can do more with our AI copilots. And we're always trying to help people do more with less, so to speak. And generally, the more you spend with Cadence, the more you should save. That...
I think we're in a sweet spot from the perspective that, like, when I talk to and when I meet a lot of other CFOs, when we're doing this kind of network, that they're quite envious that our customer spend is out of R&D. 'Cause they invest in Cadence technology as opposed to spending in Cadence technology. The more you buy from Cadence, the more you'll probably sell yourself. And when you're, like when people are focused on CapEx budgets and things like that-
Yeah.
It's an investment because they think there's gonna be a return on that investment, and we should benefit from some of that investment ourselves.
It's great to hear about the disruption. I mean, since you brought up Z3, can you just talk a little bit more about Z3 and X3 and what the implications are to your financial model?
Sure, yeah. Yeah, we probably gave ourselves a little bit of an air pocket in Q2, 'cause launching Z3 so quickly, like I say, we disrupt ourselves. Z3 is... On the verification suite of tools, that's our emulation system. An emulation system is, it's a hardware machine that allows our customers to bring up a chip onto a hardware platform and simulate the chip so that software, the software apps can be developed for the chip before the chip is even hardened. That's what causes the design cycle to shrink, 'cause, you know, back when it took five-- I mean, the reason it took five years, a long time ago, was because people used to have to bring, design the chip, and then the software folks would design the apps to go on the chip.
But now they try to do that in parallel. Before the chip is even finalized, software developers can start developing apps for it. Not only do they do that, but they can influence the chip designer with our emulation system, 'cause it's not hardened yet. But if a software developer thinks that the chip that they'd like to make changes to the chip, they'll discuss it with the chip developers, and they'll collaborate on a final design for the chip. And it means that by the time you have your chip in silicon, that you already have your apps developed for it. It's quite disruptive that...
And that, the launch of that product alone kind of disrupted the flow, 'cause like you say, in the past, chip designers did their thing, software designers did their thing separately rather than in parallel. Now they're communicating and collaborating more. Our prototyping system allows you to take that. Normally, with our emulation system, when they're done with that, when they've hardened the chip, they'll transfer it over to the prototyping system, and the software folks will continue to develop, and then they'll free up the emulation system. But we have a suite of tools in our verification suite on the software side. When we launched the systems, and it's a great acclaim, we launched it at CDNLive.
I think Jensen said that, that it was the most important system, in his life, that more important than his refrigerator. That, but he said that, they used our previous version, Z2, to design, Blackwell. And, and it, it stands to reason that you need to have a supercomputer, essentially, to design another supercomputer. And with Z3 now, it can handle five times the size of, of, Blackwell. But, so I, I think we're, we're ready for the next few years with, with that. But it'll, from our guidance perspective, because it's so popular, it takes us time to build those systems, so we took the guide for Q2 down. We didn't take the guide for the year up, for our hardware systems.
Our verification group have had multiple record years in a row, but we're expecting that to grow about 7 or 8% at the midpoint of our guide right now.
Is it the Millennium supercomputer that's the, the competitive differentiator, or, you know, how do you think about the, the EDA competitive landscape and what your points of differentiation are?
Well, that's another great disruptive system, 'cause no one's ever approached things like that before. Actually, I'll let... Will you talk to Millennium?
Sure.
Yeah.
Be happy to. So, Millennium certainly is a groundbreaking kind of technology and innovation. Historically, in the past, when companies are dealing with the physical simulation in the automotive or, like, aerospace simulation space, they only get to simulate 20% of those, okay? Versus on the chips design side, we simulate almost 99.9%, so it's a truly digital twin. The challenges and impediments historically in the past are twofold. One is accuracy, because the solver's algorithm is not good enough in order to simulate the entirety of the very complex eddy flows. So we acquired this company called Cascade a few years ago, founded by the scientists, PhD students, and professors in Stanford.
They, over years, they developed this really high-fidelity eddy simulation model, where, you know, which it solves the problem beautifully and elegantly. But then comes the next challenge, which is the speed-up is not enough, because once you have a very complex solver, obviously it slows the machine down. So that's where the GPUs and the accelerated computing come into play. Where we kind of, over time, over the past couple of years, our R&D team has been working really closely with different companies like NVIDIA to really co-optimize GPU together with the high-fidelity simulation solver to solve the problem. Now, the speed up is almost 32,000 versus x, versus a previous a typical CPU. And the third piece of puzzle is the AI model, okay?
So we use our own internal developed Optimality to really orchestrate the entire workflow. So now, all of a sudden, you have this tightly integrated, co-optimized, three-layered stack, really go there and solve the historically unsolvable problem. Now, we launched the product about a few months ago and got raving, kind of really, kind of testament and validation from companies like Honda, like Boeing, like GE. They are using, proliferating the tools across the board. So we have two different models. One is through SaaS and cloud, another is on-premise. So it's truly a plug-and-play and really solves one of the key knotty problems historically. Thank you.
Yeah, and Richard raises a great point there with the... On the system analysis side, that's a good area for growth for us, and we've had this Intelligent System Design strategy for the last 6-7 years with Anirudh. When you think of what we've done on the chip design side of things, I mean, people don't physically make chips anymore. It's like it's so automated, like 99%+ of the process is automated. On the system analysis side, only about 20% is automated right now, so I think there's much more room for growth and disruption there. And then on digital biology side, it's probably less than 1%. There's probably another leg of growth there at some point.
Maybe just zooming back to the competitive landscape, I mean, what does that look like? Has that changed, you know, since kind of the rise of AI or is it still the same?
Yeah, kind of. Well, still the same. I mean, like I say, AI is not new to us.
Mm-hmm.
I mean, we've been using AI for a number of years. But the pace of change and the pace of disruption-
Mm-hmm
... if you like, is not new to us either. But, I think when you look at the competitive landscape for us, that the area that we've played in has been quite narrow. It's been chip design, chip verification, packaging and board, and system analysis. That's generally the area that we service our customers in, a little bit of IP. But, the company that's probably closest to us is maybe Synopsys.
Mm-hmm.
But Synopsys were heading down. They were much, they're much bigger than IP, and they had a software integrity business. And they weren't in systems. They've pivoted a little bit now, but they—I think they've offloaded the software integrity business, and they're and they bought Ansys. So, so I think they're inclined to follow us now into that whole systems area.
You recently closed the acquisition of BETA. I mean, how do we think about the strategic rationale behind that, that deal?
Yeah, so BETA CAE, great company. It's like a gold standard in structural analysis. I mean, we think that our simulation capability is a superpower, but Anirudh says that you measure the capability based on the number of variables your matrix multiplier can handle, and our matrix multiplier can handle, like, 50 billion by 50 billion variables. But I think the next closest was, like, Ansys had a product that was doing 50 million by 50 million by comparison. I mean, it's just huge. Orders of magnitude difference. But this capability that we have means that for things like, say, crash testing a new car or something like that, that in a lot of countries, the laws require you to physically crash the car and put a dummy into the car and stuff.
You do it more accurately for all different body types with simulation software like ours. But, so I think it's a good marriage between ourselves and BETA CAE. Also, when I talked to Anirudh about it, what he told me was that he thought BETA CAE was, like, the last piece that we needed to be able to effectively, compete product per product, product to product against Ansys. Now, we have a broader view of system analysis than Ansys does, 'cause we take in digital biology as well. But, so he views Ansys, what Ansys does, as a subset of what Cadence does. But for probably less than $3 billion over the last few years, we've added some small tuck-in M&A into the Cadence portfolio, 'cause we like to make rather than buy.
We spend 35% of our revenue on R&D, so Cadence is like an innovative company. It's R&D-driven. Our leader is an engineer. But we'll make more things, and then we'll add some tuck-in M&A when it basically allows us to accelerate our strategy. But I think BETA CAE is the last piece of the puzzle, essentially-
Yeah
... to be able to compete effectively with them. In contrast, Synopsys is inclined to imitate and follow that they purchased Ansys, spent $35 billion on that. That's, I think that's good validation of that what we've been doing for the last six to seven years has been good value for our investors.
That's really cool, and really cool to hear on the crash test. Sounds like you might put crash dummy suppliers out of business if they no longer have to be used in actual tests.
Well, we could help a lot of people, but,
For sure.
Yeah.
And maybe just tying, you know, that acquisition back to the financial implications of the model, right? Like, how do we think about the, the margins and the growth profile-
Mm-hmm
... and how does that layer in over time?
Yeah. That's a great question. I mean, over the last 7 years, we've prided ourselves on, we scale really well. For every $1 of revenue growth, we're dropping more than $0.50 to operating income, and it's probably closer to $0.60 organically, when we do M&A, we're never able to find companies that are as profitable as ourselves, generally, and so it takes a while to get them up to our level of profitability. So generally, when we acquire a company, it'll be a headwind for incremental margins and for operating margin, generally. But so what you see for us, like last 7 years, we've achieved that 50% incremental margin. Actually, just a little bit more than that, probably. Probably closer to 54%-55%, maybe.
The organically, I'm confident enough that we'll achieve that again. Well, I looked at the end of Q1, we gave the guide that compared where we were at the end of Q1 in the previous 7 years, where we had achieved 50% incremental margin, and at the end of Q1 this year, we're ahead of 5 of those 7 years. So I think we're still on track to achieve that 50% incremental margin. BETA CAE is a big acquisition, though, so let me put it that I'm pretty confident we'll achieve the 50% incremental margin if I can exclude BETA CAE.
It might take a bit longer to integrate BETA CAE, and then if we can't achieve it by the end of 2024, including BETA CAE, what we'll do is we'll measure 2025 against 2023, so you can see how we're doing it, including BETA CAE.
It's good to hear on the incremental margin.
Yeah.
Sometimes we investors keep things overly simple, and just like we managed a rule of 40-
Yes
... and you guys are actually well above that.
Yeah.
So, yeah, how should we think about the growth versus profitability profile, and what can the long-term margin be for you guys?
Oh, look, I mean, if we're achieving more than 50% incremental margin, naturally margin will levitate towards that, towards that level. It depends on the amount of M&A that we do. Now, I do think that we don't need to do any more M&A for that, area that we're focused on. OpenEye has been a bit of a science project for us on the digital biology side, but, if we're successful, the more successful we are there, I guess, might lead to some other tuck-in M&A over time, but I don't see anything like that in the near horizon. But, but from a, from a margin perspective, we're very focused on profitable and sustainable revenue growth. And Cadence is really a compounder, that, you we're not gonna blow the lights out in any one year.
But, if you can pick a winner in AI, you're probably better off going after that one. But, but if you're not really sure who the winner's going to be, you know that we'll win with those winners. But, over time, we tend to like win with the winners, and that's... We'll be servicing a lot of companies, but there'll be some come through, and they'll end up spending more with us eventually.
Can you just double-click on capital allocation? Like, obviously, you just mentioned M&A, small tuck-ins going forward, not needed for growth, but you also have really strong cash flow. So what are you-
Yeah
... what are you gonna do with everything?
Yeah, I mean, generally what we do is 50% of free cash flow is used to repurchase shares. We want to try and reduce the share count over time. And then in the past, it's been 50% of cash flow has been used for investment in the business and in M&A. From a capital standpoint, we're probably undercapitalized a little bit because we've been carrying about $1 billion in liquidity and cash, but we were doing... That used to be, like, 5% of our,
Yeah
... of our market cap. Now, it's probably 1% of our market cap. But, so we could probably do with a little bit more cash on the balance sheet, so we'll build up some cash. The... But, but our, our approach is to use 50% of our free cash flow to reduce share count over time. Yeah.
I imagine you guys are probably mostly immune, given, you know, you're helping companies in their R&D, but how do you think through macro and the implications? Have you seen any tightness of budgets or... 'Cause, like, at some point, your customers' customers might have some impact, so.
Yeah, so, from a recession standpoint, say, I mean, a company like Cadence is not immune, but we would be somewhat resistant in that, that because our spend is... Our, our customer spend is our, their R&D budgets, it's, it's normally the last thing to get cut. We have felt some impact from slowdown in China last year in our numbers for this year, but we're already seeing that improve in terms of design activity. Last year, we had a customer, OPPO-
Mm-hmm
... that they closed the design group and, and let 3,000 people go. And what tends to happen with Cadence, of course, is that you, you had that ratable revenue coming for that group, and when the contract expires, when it renews, there, there's no need for them to renew. But what will happen is that those 3,000 engineers will end up at other customers over time, and then those customers will purchase tools from us. The impact of things like that are as kind of a slow, kind of some semi-weakness and some China weakness last year, has turned up this year in our recurring revenue profile. Recurring revenues dipped from- it's typically around 13%, low teen growth.
It dipped to about 9% in the first half of this year, but we're projecting 11% second half of the year, so it's already starting to recover. But yeah, we're generally pretty resistant to everything. It's like-
Good business model
... cockroach of a company. We'll be around forever.
That's, that's a great business model.
Yeah.
That's what investors want. Maybe just final question to close out, one for each of you. Richard, maybe, what, what do you think is most misunderstood by investors? And then, John, if, if you had, you know, a magic wand to remove one constraint from the business to turbocharge growth or, you know, alleviate, you know, an issue in your life, what, what would it be?
Well, that's a great question. Yeah, for me, what, well, one of the things people talk about is, you know, Cadence is always a great quality asset, but it's premium priced. I think I just encourage all of you to just look at the long-term growth drivers and tailwinds for the business for the next 10, 15 years, you know. And none of that is gonna change, you know, over time. You know, you see all these long-term growth drivers going to stay intact. So I think that's probably the most important thing just to keep in mind.
Mm-hmm. Yeah, I think for us, our customers basically are changing the planet, and we hope they use it for good. We think they're doing great things. Like you said, I really believe our customers will be remembered as the Renaissance artists of our time, and all we're trying to do is give them the best tools. So we're just helping everybody do their jobs better.
Awesome.
Yeah.
Thank you for the time. Thank you, everyone.
Thanks.
Thank you.