All right. Good afternoon. This is the 28th Annual Needham Growth Conference. Joining me here is Silvaco CEO Wally Rhines. Welcome, Wally. I've known you for a long time. Could never imagine me and you on the same stage having a fireside chat. But I'm thrilled to have this discussion with you.
Thank you, Charles. Very good to see you again.
Yes. All right. So to kick off this discussion, I think for folks who may not be familiar with what Silvaco does, can you tell us a little bit more about the major product lines of Silvaco, TCAD, EDA, IP? What are those groups? What do they do? What kind of customers do you have for each of the major product categories?
OK. Silvaco has been in business since 1984. It's the oldest of the current listed EDA companies. It is the oldest, and it has not grown as fast, but it has been successful in specific specialties in EDA, so Charles mentioned that TCAD has been a core business. In fact, for many years, Silvaco was the premier supplier of TCAD. Now they are one of two major suppliers of TCAD, along with Synopsys, and that's a core business. That's traditionally, even in recent years, that's been about 50% of the revenue. It's a very sticky business. People don't change often. The software is very complex. It involves a lot of advanced physics, solving partial differential equations, and you basically do that so you can simulate the results of process steps in manufacturing, so that's a core business. Another business of Silvaco is really more recent.
It's been the IP business, where you buy blocks of intellectual property or design information that you incorporate in an integrated circuit you design. That's been a small part of Silvaco's business, only about 10% over the last few years, maybe a little more, and it's one that is going to be a big growth engine in the near term. The third one is traditional EDA, the same things you see from Synopsys and Cadence and Siemens, except that Silvaco has tried to get niches in the market where they can be critical to customers, and so through a combination of product development and acquisition, a portfolio of EDA products has grown up, and that forms another portion of the business, so between the two, the IP and the EDA constitute roughly the other half of the business beyond TCAD.
Thanks, Wally. I think we have a lot more to discuss regarding how to grow the Silvaco business or the three groups. But before that, I do want to touch upon AI. This is one major topic of debate for investors looking at EDA because the semiconductor industry over the last couple of years has been kind of like a tale of two stories, the AI haves and AI have-nots. And EDA was kind of people thought they are the AI haves. But the problem is the growth hasn't been that pronounced since actually since the growth rate peaked out in 2022. It was a few years of a bit of a challenge, a few years since then. But it is a major, major driver for the industry. My question for you is, is Silvaco exposed to AI?
Could it become, if it hasn't, a major uplift in business for Silvaco?
Yes. So a major portion of our long-term strategic growth will come from AI. And the thing you highlight for the EDA industry is not that they haven't applied AI to EDA. AI is showing up at all sorts of places in the EDA industry and their products. But the fact is it's just replacing the old way of doing things for roughly the same price. What you really want and what Silvaco is looking at is where does AI come in that it creates a new enormous market rather than just replacing what we're doing and doing it better. And the answer is in manufacturing. Because today, the approach to manufacturing is unsustainable, not because it doesn't produce accurate results, but because processes are becoming so complex.
The time to run prototype wafers is becoming so long that you can no longer depend upon prototyping as a way to evolve a new process or improve its yield. You're going to have to do it virtually. You're going to have to do it on a computer. And there is no question about that. And it will be a total new implementation that is built on the foundation of TCAD, but for which you now move to instead of one simulation at a time, you do tens of thousands of simulations to build a model, a digital twin, and then you query that model. Then you say, what happens if I raise the temperature a tenth of a degree? What will the resulting thickness of the oxide layer be? Or what will happen to the implant?
All of that requires that you run enormous amounts of data to build reliable models that will transition the industry, go to a totally new model where people are willing to pay for a whole different kind of implementation. Instead of buying a $50,000 or $100,000 TCAD license, you now are using the models you build up and generating more data over time. And for a manufacturing operation, this is a trivial expense. You're spending $20 billion or more building a wafer fab. You're buying equipment that can cost $100 million a copy. You cannot afford to run it inefficiently. You cannot afford to have a three-month delay in the ramp-up of the new process. You will have to go virtual. It will have to be AI. It will have to be digital twins done from AI-driven machine learning.
And that's what Silvaco got into with Micron five years ago and began the transition with Micron, developed the entire methodology, the tools, and built up the database. And now Micron has demonstrated it. It hasn't permeated the whole company, but a large segment. And that's what we'll be doing with other companies with the endorsement of Micron, who is even helping us to do it successfully elsewhere, not with direct competitors, but with other companies because they would like to see us continue to be successful.
Yeah. Speaking of that, I believe that's the technology called the FTCO, Fab Technology Co-Optimization. That's how I call it, that joint development with Micron. And it was a pretty meaningful revenue contributor, I think, going back one or two years. But can you give us some sense? You said you want to go to more customers with FTCO. But it sounds like that's a yes. There will be more customers. But when will that happen? What kind of projection can you provide to us?
It'll happen as fast as we can make it happen. And Micron's not the limit. And by the way, as you mentioned, we didn't do that for free. Micron incrementally was funding additional licenses as we got new capabilities going. Now, the question is the other customers. And I think we've had the view that because we've done it once, we can do it very quickly with others. But I have to caution that it's not a traditional kind of sale. You don't just come in and say, how'd you like to buy our AI process development module? It's an engagement. It's a courtship. It's a, hey, we both agree this is where we've got to be in five years. Let's start working. Can you dedicate some engineering? Can you pay some money?
And so it's more like a service engagement where you're going through and plotting a roadmap and getting commitments. And it takes time. And these people are not sitting around doing nothing saying, "I need to develop my next generation method of process development." They are quite busy using the inefficient methods of today to do that. So getting the resource to do what you need to do is always hard when what you have to do is so pressing. But we are engaged with a number of customers. And there are a number of these in their infancy. We'll talk more in our earnings call about where we stand on that. But I believe it will have major long-term consequences for the company, that is, benefits for the company.
But it will be a gradual process as we engage with more and more customers and apply our limited resources to the ones that can make the most difference.
Yeah. We'll talk more about that in terms of resource allocation, your growth strategy. But I think at this moment, I do want to zoom out a little bit because, I mean, Cadence, Synopsys, that's household names. Mentor Graphics, I would say that's where you were the CEO for more than 20 years. That was household names as well. Now it's part of Siemens. So people are familiar with those names. And EDA has been as consolidated as it could be. But Silvaco is an independent company. You said you were founded in 1984. The company has been around for 40 years. And I think in one of the books you wrote, I think you wrote about in EDA, number one makes the most money. Number two, in terms of market share, makes some money. Number three makes no money.
So, tell us, you have so many, the top three there, occupying the market. How does Silvaco? I mean, obviously, Silvaco has survived 40 years. But how do you thrive? And how do you grow?
This is good. You read my book. I have given the prescription. And the fact is, EDA is a really different kind of industry. Small companies coexist with big companies because big companies dominate big market segments and little companies dominate little market segments. Everybody dominates something. The way the system works, as you said, if you divide it up, the 10 largest segments of EDA, things like place and route and logic synthesis and physical verification, the number one supplier in each of those spaces has, on average, 72% market share. They have a lot of market power. They make all their money where they are dominant. Everything else is either they're number two and they make a little money or number three and they don't make money. I joined Mentor. We had no number one positions. A big challenge.
1993.
Yeah. Today, most of their profit comes from two or three products. And they are very profitable. So you don't have to sell everything. You just have to do something really well. And then you say, well, you used to be king of TCAD. But now you've got a big company competing with you. How can you survive? And the answer is, first of all, we have a portion of the market that is already well engaged with us and is not going to switch. So we have recurring revenue from them. The second thing is we go and we concentrate on areas that would be viewed as niches by our larger competitors where it really doesn't move the needle for them, but it can have a big impact on us.
Specialties like simulation of power devices, gallium nitride and silicon carbide, we have particularly good models for those. New areas as they emerge, like the AI-driven process development, things where you're not face to face or head to head with someone who is able to overpower you with scale. Then you say, well, wait a minute. You're not number one. OK, you're a strong number two in TCAD. Where else do you have the strength to be number one? The answer is, wait a minute. You don't have to be number one in a broad segment. You have to be number one in a market segment that matters. Silvaco, number one in TCAD as it applies to display manufacturing, more than 80% of the market. They have to use the Silvaco product because everyone else does.
You can define other niches like that where you're able to have market power because you have a unique capability that is better than anyone else. And traditionally, that means larger market share. So if you're not number one in one of these big markets like place and route, that's a $1.5 billion market, you can be number one in specialty products. And there are still hundreds of little EDA companies that make their living exactly that way.
Got it. Got it. So that's a very important industry context, I'm sure. Wally, you've been in EDA 30 plus years, I guess. But specifically on Silvaco, how do you plan to grow this company? And more importantly, how you want to grow it in a profitable way? I think that's top of the mind for a lot of people looking at a Silvaco stock.
Yes. So I came in with a charge to get us back on the right track. We did an IPO and then proceeded to miss quarters of guidance. And clearly, something needed to change. So the first thing to do is get expenses in line. The company had dropped into a loss position, was no longer generating cash. There had been expenditures to do acquisitions, which are good in the longer term. But initially, they carry some cost with them. There had been other expenses that had come up. And simply, the spending had gotten out of line with the revenue. And so the first thing is get profitable, stop the drain. And so as we announced in my first and only earnings call, we're going to take an expense out of the company for at least $15 million annualized running rate.
And that plus revenue growth stops most of the bleeding. It doesn't quite get to profitability. But I'm expecting there'll be some revenue growth on top of that. So we think that net of all the other things, that we'll get modest revenue growth in 2026, as we said in our earnings call. And that as we go through the year, we'll continue to reduce expenses. And I think we'll continue to see the strength of the emerging business, particularly in IP. And we'll see some resurgence in TCAD because we had a weak renewal year. So I think the net result is near term, get the cost out. Next to near term or in parallel, get the revenue growth associated with the growth, particularly of the IP business, but several other special things. And long term, build the AI franchise.
That gives you the long-term big target to grow to and where we have all the basics, and we've done it once with a customer, so cost reduction, short-term growth driven primarily by IP, long-term AI.
Yeah. So I want to maybe double-click on this year's modest growth a little bit. Hopefully, you can provide a little bit of additional color. Yes. You started probably the first view on 2026 pretty conservatively, your first view, right? Modest growth. Anything we need to pay attention to why the growth is modest this year? What does it take for the growth rate to pick up again? Could that happen before 2027, maybe middle part of 2026? We could see something already.
So when we say modest growth in 2026, you have to look deeper in the numbers to realize it's much healthier than you would think because we had a big bluebird, a big renewal in EDA in the second quarter. I'm sorry, in the third quarter of 2025. That was from an acquisition. We acquired a company and had a major product renewal that will not occur again for three years. And that substantially drove up the EDA segment of our reported business in the third quarter of 2025. That won't repeat. We have to grow. We not only have to overcome the loss of that big injection of revenue, but also to grow on top of that. So I think we will do that. That's what we said in the fourth quarter we would plan to do to get modest growth. And you're asking, could it be greater?
I think I'd say let's talk when we give our earnings call for fourth quarter, and I think we'll talk more about where other upside might exist.
OK, great. So that's the growth strategy. Near term, long term. But let's say you came over and took over the leadership at Silvaco last year. But in 1993, I believe, you became the CEO of Mentor Graphics, right?
Yes.
Trying to draw some analogy there, maybe compare and contrast, what do you see at Silvaco today? Any similarities to Mentor Graphics, let's say that's 30 years ago? Because Mentor was a great turnaround story, right? And for both Mentor and I think for yourself, I mean, that's a career-making kind of journey, right? Anything you see today, whether some of the lessons learned from the Mentor turnaround could help? And what are some of the new challenges that you see at Silvaco that you didn't see really much at Mentor and how you plan to address those?
Yeah. So Mentor had almost been written off as a basket case, which I love. That's terrific because the bar is very low. Everything you do shows up as a success. Silvaco isn't in that kind of desperate situation. But there are similarities. Mentor was very troubled, losing a lot of money. Silvaco is losing money. Mentor had no number one market positions whatsoever in anything. In Silvaco, no number one positions in any major segments. A good number two in TCAD, but a little better than Mentor was. I think the commonality was Mentor did have a very loyal, mature, experienced workforce. And so does Silvaco. And Mentor had a great willingness to change because they knew they couldn't continue going in the direction they were in. They had fallen from being the leader to fallen behind Cadence and Synopsys. And so they had to focus.
They had to pick the winners and stop supporting the products that were not generating growth, were simply collecting maintenance. Well, that's pretty much the same as what we need to do here at Silvaco. We need to be profitable, first and foremost, because we need to generate cash. One of the things that at Mentor they needed to do was because they had missed an entire generation of product, they needed to go out and jumpstart with acquisitions. Well, that's what Silvaco needs to do. We've gotten a little start. But we've got to replenish the cash engine to continue that because it's a great model. It's the way EDA has traditionally grown. Cadence grew 100% by acquisition, simply having a big sales force and acquiring companies with great products and no sales and then helping those companies grow.
I think the major difference between Mentor Graphics and the situation I faced in 1993 versus Silvaco today in 2005 is I had 20 years to fix Mentor. I only have a short period of time to fix Silvaco. Single-digit years, not decades, and so I've got to do it faster, and we will do it faster.
Great. Thanks, Wally. On the financial side, I know your CFO isn't here. And you said you'd rather keep him at work. But asking you to speak for him, at the time of IPO, Silvaco did have some financial metrics and financial targets longer term, maybe not revenue, but margin targets were all there. Do you have any thoughts on how you want to set the long-term revenue targets, margin targets that you can share with us today? If you cannot, it's fine. But kind of give us a sense of when you can share something with us.
Yes, so I asked for a 90-120-day reprieve when I joined just to figure out what was going on, and so we'll provide more visibility into that kind of thing. But as generalities, small companies ought to be able to grow at a faster percentage than the big companies, and the big companies are growing in double digits today, so that's the kind of standard I think people look to for a company like Silvaco. It should be able to grow faster than an industry that is growing at 10%-15% a year. The way you do that is to have a base business with organic growth that is at least in the double digits, as we described at the IPO, and then augment that with acquisitions.
Now, as I mentioned before, the perfect acquisition in EDA is a company with a great product and no marketing or no sales joins with a company with a big sales force and a lack of products. If you go back to the early days, the Cadence acquisition of ECAD was joked about because they said, "ECAD, here is a company with a great product and no marketing. And Cadence with Joe Costello is a company with great marketing and no product." And the two went together. And the ECAD sales soared. That was a product called Dracula. And it was the basis for Cadence to then grow on a bunch more. Well, you take a look at Mixel. I mentioned IP company, one salesperson. Now, Silvaco, 80 salespeople, including FAEs. So all of a sudden, you bring a sales resource.
Then the question is, OK, how fast can you increase capacity? Well, with Mixel, you can increase capacity. They're located in Cairo, Egypt. That's a great source. I built the group from scratch at Mentor. They have over 1,000 employees today. Universities turn out real engineers with real degrees and more than they can consume in the economy. And so you have the opportunity to hire really outstanding people and grow as fast as the demand for your product grows. So I think you've got all the dynamics here to move toward a set of measurable metrics. And I think the ones people care most about is, what will your growth percent be and what profit goes along with that? And for profit, while we haven't done any revised projections, the fact is once you get to profitability at any level, then incremental revenue falls through at 70% or 80%.
So any growth improves your profitability. So it's a very scalable model. Then whether it's an acquisition or organic, you can increase the profitability to the industry levels. Today, Cadence and Synopsys are in the 30%-40% operating profit level. It will take a while to get to there because there is a certain amount of fixed cost. But I think it's the kind of business that lends itself to very high operating margins because the cost of goods sold is so low. You're basically selling ones and zeros with a little bit of support. And because of that, the flow-through is very high on incremental revenue.
Maybe let's dig in a little bit more about M&A. You mentioned about Mixel. I know since the IPO, Silvaco leveraging the cash balance have done quite a few, right, and I think I've heard you talking about some of the acquisitions that you particularly liked, like Mixel being probably top of the list, it sounds like. So can you kind of provide a little bit of assessment on how the M&A has done so far, but it sounds like you're going to want to replenish your cash first, so M&A may be on the back burner for now, but once that replenishment is done, what kind of M&A are you looking at? And maybe along the way, give us a little more color on why you like Mixel so much. You touched upon a little bit.
Yeah. So Mixel, I think, will be a home run, partly because it's this perfect model, a company with a great product. So great, in fact, they are a standard for quality. They are respected throughout the industry, through the history of their company. Over 25 years, they have never had a customer bug in a product. And so they are a premier supplier. And they're doing very complex IP, mixed signal, PHYs, for example, physical interfaces and things. So tremendous skills, great reputation. They are sought after. I have a premier relationship with TSMC and the other foundries. Actually, their IP is supported in 12 different foundries. Very great reputation, very broad, terrific kind of acquisition. We're very fortunate to get them. Now, another one was OPC. This was Cadence wanted to get out of the optical proximity correction business. There are three large competitors.
They were sort of subcritical, and they had only been very successful with memory companies and just did not see a lot of success later, well, you look at Silvaco. Hey, memory companies, that sounds familiar. We're really successful with memory companies. Look at Micron. Look at others, SK Hynix and others, big customers for Silvaco. That's a pretty good match, and we don't need to go and sell the product to the broad market. We don't need to invest a great deal, and so that acquisition remarkably had more than 100% payback in the first year. That's largely because of a large renewal, but still, it's paid for itself, so not an enormous contributor to growth, but a very good contributor to cash and profitability. The third one, Tech-X, great potential. Right now, it's an expense. There is very little revenue, a few million dollars of revenue.
It has the potential to be very big. It's premier simulation capability to do what we have done. One example, for example, is Plasma Etch and Plasma Deposition. They take simulation to the next level, another order of magnitude of performance. It allows the manufacturers to simulate things they couldn't simulate before. But once again, you go through a long qualification process. So far, so good. But it's not an instantaneous one. So you did three. One's going to be a near-term home run. One's a great cash generator, maybe not a big growth generator. The third one, still up in the air, but very promising looking, and you ask, what would we do in the future? We've got a whole list, lots and lots of companies that have good technology, good products, good people, basically break even or even profitable.
They're too small to move the needle at one of the majors. A company that has $5 million, $7 million, $10 million of revenue is just not worth the trouble for the big EDA companies to spend going through an acquisition process. It just doesn't change things that much. For us, it makes a big difference. $7 million of revenue is a 10% growth instantaneously. So there are a lot of those we could go after. It's a business where startups can start very easily. It doesn't take any money. You buy a computer. You're a smart person. You go into the garage and you start writing code. And you get a really great product. And you sell it to one customer, maybe two. And then you say, oh my gosh, I've got to hire a salesperson. And then I've got to hire another salesperson.
Then I've got to provide support. You say, you know, this is going to cost me all my equity. Why don't I go cash out? That's what they do.
Yes, please.
Do you really have the capital, though, to go after other deals that are out there?
Not now. Not right now. We're going to have to raise more, and there are lots of ways we could do that, but not with a $4 stock. I mean, the stock is not at a, it's at two times revenue, not at a value that's representative of a company with this kind of potential, so if you go to the filings, you'll notice that I personally, on my own, have purchased 50,000 shares in the last two months, and the CFO, Chris Cereghetti, has purchased at least 20,000 shares, so yeah, we're buying it.
And just on the renewals, maybe what did the company do wrong in terms of managing expectations around renewals? So it seems like this is a business that has generally high renewal rates. But there would be unpredictability in terms of when you'd actually recognize renewals and whether they're this quarter or that quarter. And usually, it seems like there'd be a rush to get renewals done at the end of the quarter.
OK, you've got a whole bunch of issues there. The first one I'll say, renewals are very predictable. Renewals almost always occur. If you lose a renewal, well, you're probably not going to totally miss on a renewal. You might have a renewal that doesn't grow very much, but a very predictable part of the revenue. Now, you're into the issue of revenue recognition. Silvaco has the same revenue recognition as Siemens, Mentor Graphics, and others. Essentially, this is the 606 accounting. Synopsys and Cadence are able to do ratable revenue. They can build a two-year backlog. They can always know exactly what their revenue will be. We can't do that. We have to recognize revenue when the software is shipped to the customer. We reserve about a third of that for the maintenance that they will pay for ratably over the life of the contract.
And so we have more volatility. So that's what creates the uncertainty and volatility associated with those renewals. Is that what you were talking about or something else?
Is there any better way that you can manage expectations around that?
Yeah, what you can do is set your contracts up so they intentionally don't trigger all the revenue up front. You can have add-ons that are only available over time and things like that, and get more and more like a ratable revenue. And of course, that's what we try to do. In the Cadence and Synopsys case, they just went dark for three years while they transitioned from one model to another. I'm not a fan of doing that. But anyway, you've got a question, sorry.
Yeah, so first of all, I just want to say I really appreciate the fact that you've taken on this. I think it's a great thing for you. I think it's a great thing for me as a stockholder. And I was an investor in Oasis Design Systems. And I was also on the board of Numascale, an IT provider out of Norway. And the challenge that both of those guys faced, and I think the challenge that you faced, is the all-you-can-eat model from Synopsys and Cadence and those guys. And you say you're a strong number two behind Synopsys. But how are you? What is your value proposition that you're giving to guys like Intel? It's 25 years with Intel, so.
Yeah, so here we're talking.
That was an example of TSMC, those guys.
Here we're talking sales dynamics and EDA. And if you knew what really goes on, it would be an eye-opener for most people. So the way it works is these companies have, as I described, the areas where they are number one, 70% share. That's what drives the contract. Then you have a dynamic with the salesperson and the purchasing manager. The purchasing manager wants to get a lower cost per copy. And the salesperson wants to get a larger contract than last time around. So they start throwing in stuff for free, free licenses. And in the extreme case, if it's a really big customer, even, say, what about unlimited usage for the product? Now, underneath in the dynamics, it makes very little difference because the engineers don't change. I mean, engineers believe there are bugs in every EDA product. I know the bugs in the one I'm using.
I know how to work around them. If you force me to use a different one, it's a six-month delay, and so I don't want to switch, so the engineers don't switch. But the salespeople do this dance around how many licenses do you get for how much money, and so if you've got an area where your product is really weak, you offer it for free, and if your product is really strong, you hold the line on the total amount, but you can reduce the cost per copy by throwing in all this other crap, and so that's the way it works. A company like Oasis is far, far smaller than Silvaco, so they're not in this kind of game. Silvaco is in the good stage that they are big enough to have a worldwide sales force. They are big enough to have negotiated big contracts.
And they are big enough to play the same game the big guys play. But they have to use the cards where they have power. And they know how to do that.
How are you using the cards? How are you using the cards? Are you able to?
Absolutely. I mean, that's a necessary part of running the business. We out of time?
We're out of time. But appreciate Wally dropping EDA insights here, there, throughout this entire Fireside Chat. And thank you so much, Wally. And we look forward to hearing from you. And thanks, everyone.
Thank you.