Good morning, everyone. It's Blair Abernethy here again, Software Analyst with Rosenblatt, at our Fifth Annual AI Conference August Session. With us this morning is Babak Taheri, who is the CEO of Silvaco , being a small-cap EDA and TCAD software, so it plays in the technical tools space. Babak, I'm wondering if you could just give us a couple of minutes on your business, just a general overview, sort of the end markets that you address, and a little bit about your background as well for the audience to understand where you're coming from.
Absolutely. Thank you very much, Blair. Good morning, everyone. We have been in the industry since 1984, and the company hired new managers in order to make it less of a scientific, I would say, scientific lifestyle company to a real company. After five or six years, when I joined in 2019, we took the company IPO around 2024, back in May of 2024. Matter of fact, in 2024, our bookings were about $65.8 million. Our revenue was about $60 million, 10% growth for revenue, 13% booking. We were participating in $3.8 billion of SAM for our industry. As you know, we are considered to be an EDA company, the likes of Synopsys, Cadence , and Siemens and intergraphics. We have three business units.
We have a business unit that sells the traditional electronic design automation tools, such as any electronics software that you'd need in order to design a chip or an IP. We do have what's called a technology CAD software that enables you to design any new thing in terms of process or any new devices that require new material analysis and translation of that into models that you can actually simulate. We call those nowadays digital twin models and also a semiconductor IP company. Also this year alone, we've acquired three new companies based on that and our growth in organic together. Right now, we've added another $710 million of SAM to our market. The company is expected to grow double-digit this year and be neutral in terms of profitability. That's putting things at the high level, Blair.
Okay, that's great, Babak. Maybe talk, let's start off on your core TCAD part of your business, which is, you know, give or take half your business, I guess. How are you positioned there versus your major competitor, Synopsys? What's the value proposition here for customers? Why do they buy this software? You know, how do you grow within that marketplace?
That's a very good question, Blair. Our technology CAD software is a fundamental software that we started the company. The company started based on that technology that was licensed from Stanford University. The technology is such that any new, as you know, the chip complexity and photonics complexity is increasing. You need to pack more chips into it. You need to add new materials into it. You need to get to market faster. When you start adding things like new materials, like if you're going from bulk CMOS semiconductors, as they call it, down to technology nodes down to nanometers, if you travel that journey down to FinFETs, gate all around nanosheets, and also angstrom level simulation, you find out that new materials come into picture to make the sandwich layers of a wafer that form transistors and circuits.
Whenever you do have additional material, their properties are not well understood. Technology CAD is, if you think of it, a fundamental simulation model, which does physics simulation. We call that multi-physics, not only physics, but also in terms of thermal management, optical. Actually, believe it or not, there's a lot of optical characteristics of these materials that need to be looked at, thermal and material properties that need to be looked at. TCAD fundamentally enables going to the next node, going to add a quantum dot on your LEDs to make a screen that's called QLEDs, and getting into new power devices like GaN or Silicon Carbide . Any new materials that come into picture, TCAD is a must-have in order to do simulations, not run expensive wafers in order to do the design of experiment analysis. TCAD is fundamental to any new material or anything that comes up.
Our software enables our customers to do that. As a matter of fact, we have lead positions in the power market, GaN or silicon carbide, if you will, photonics in terms of displays and other photonics like image sensors. When you think of memory as well, we are pretty much leader in terms of supporting that market going forward. In terms of our TCAD as a whole market positioning, we are number two in the world. There are no other companies that provide full solutions besides us and one of our bigger competitors. It turns out that we have such a vertical market leadership that the accumulation of that right now positions us in number two in the world. However, we've introduced what we call FTCO, Fab Technological Optimization, that uses AI.
TCAD historically has been used for the design portion of chips, the designing of the process to design the chips. Traditionally, that has been the case. What we have done as of last year when we announced, we've actually moved TCAD to be very useful in terms of augmenting it with statistics and analytics of AI. We're able to help manufacturers go through all this data and be able to simulate enhanced yield, get to market faster, not run so many simulations that take months to years to actually do, but rather do them quickly, not run wafers that you waste a lot of sunk cost for the manufacturers, which is critical. We've taken it to this new market we call basically Fab Technology. We are the only ones in that white space using TCAD for production-level simulations and these software models.
As a matter of fact, when we talked last time, I mentioned that we've added another $500 million of SAM to our market based on that. That basically doubles the total market for TCAD. As we progress, we are pushing very hard to get our products accepted both in design and Fab so that we can, hopefully, with our efforts, we can become number one in the world.
Thank you. The FTCO has, you've really had that in the market for, what, less than a year almost. What's the first couple of customers, the first customers, what's been their experience with it? Are they driving value off of it? I guess also, you're typically selling TCAD into the early, early stage experimentation or design, the research side of product development. Is FTCO a little further along in the food chain?
Oh, much, much, much more. As you recall, we went to IPO in May, June. In May, we announced we had Micron as our main and first customer for it. We showed also that our relationship got extended because every time you have a new device that you want to mass produce in a new Fab or a new lane within the fab, you require additional R&D. That's what we follow up with. The first thing that, matter of fact, FTCO requires you to do is use our TCAD, what we call design technological optimization tools. You add analytics and statistics, and that combination makes an FTCO, which is an enabler for manufacturing. As a matter of fact, also in our last call, we mentioned that the first stage of FTCO customer acquisition is doing R&D with them, not for them, but with them.
We said that we do have advanced CMOS fabs that are in the R&D phase with us. We mentioned that we have also power electronics companies that are in the R&D phase of FTCO with us. Going from the R&D phase, which makes sure that all the bells and whistles of the product is there for manufacturing, takes it between one to two years typically. Beyond that, you go to what we call manufacturing phase, which is where those are the much larger deals we do because we've changed also our business model from selling seats to designers, one seat per design. When we get into Fab manufacturing, you sell seats per lane or per fab, which commands a much higher complexity of simulations, digital twin modeling. That's where we are at with FTCO. We consider it to be very successful.
A lot of it, some of the delays that we've seen over time is because of the fact that these tools to get adapted and to have a manufacturer, a Fab lane, which could cost in the orders of hundreds of millions of dollars to make a chip, it's very expensive to just experiment. The R&D phase of what we do with FTCO is to make sure that they run examples of what's in the Fab before they launch this in the fab. That's the process we're learning. That's the process our customers are learning. We're the first one in the market with it. That learning cycle has taken a bit of time, but we have quite a bit of tailwinds in this technology. It's very unique. It solves a lot of running wafer-based, masks-based.
As you go down the FinFET and below technology nodes, you end up paying tens of millions of dollars per mask layer. Each of these, depending on the chips, could have hundreds of masking layers by a thousand processes in terms of the memory. You find out that it's very expensive to make these wafers and experiment to see which parameters you need to yield them. However, if you have a simulation software that can do that for you, which is the premise and actually delivered results of FTCO, that's what they're looking for.
I know it's very early still with this FTCO product, but do you have any indications as to what kind of value the customers are getting out of this? How much do you think they can save? You know, with software you've always got to look at, if I'm going to spend $1 million on it, what's my payback on that investment?
No, that's an excellent question. Frankly, you find that manufacturers, people in manufacturing, think about cost very differently than people in design. People in manufacturing, if you are making a large memory company, you look at every penny that contributes to the cost of making those chips. Unless ROI is justifiable, they won't even consider this. I'll give you a feel for it. Think of a manufacturer that runs, let's say for now, a 20 nm, not even go down below that, 20 nm wafers or 10 nanometer wafers for that matter. As an example, you find out that the wafer cost is about $10,000 - $20,000 depending on the product. The mask cost exceeds tens of millions of dollars. Manufacturers, in order to fine-tune their processes and the tools, right, they have tens of tools that they have to process these, as I said, a thousand process steps.
They have to process these wafers through, and it's expensive. You find out that if you can save them 10 wafer runs and two or four mask sets, that could add up to hundreds of millions of dollars of payment. The ROI for the software we sell is very justifiable to them and to their management. As a matter of fact, if they can get the product out six months sooner, and if this can help them put the next generation out six months sooner, I would call it a piece of cake, for lack of a better term. It's very easy to justify to the company, to the management, to manufacturers, to their COO, to get this software.
That's really helpful, Babak. Just to help us understand what you've publicly disclosed so far, you know, Micron was the first customer to take this down. How many are sort of piloting or in your pipeline? I guess trying to understand how big this opportunity could be for Silvaco.
Yeah, we've said historically we've added half a billion dollars of SAM, and we would like to capture anything in the order of 10% - 15% of that in the first three to five years of the introduction. We are in the second year of introduction. We introduced it in May last year in manufacturing, not in R&D. I've said before we were in three years with R&D with Micron before we got it into manufacturing. Now that we've done our first one, and we're doing our second and third one in power and advanced CMOS, those learning cycles have improved drastically. The R&D phase is shorter, and that's our goal. Our goal is to get to 10% - 15% of that market within three to five years.
Okay, that's really helpful. Shifting gears to inorganic growth, this year you have made three acquisitions. Maybe you could just walk us through what you bought, why you bought it. I guess it's a little early to say how they are going because two of them have just closed in the last 60 days, but you've made some strategic moves and maybe give us some color behind that.
That's a great question. We've actually done three acquisitions in the timeframe of three months. It's unheard of, right? I mean, before when we went to the roadshow, they asked us whether we could do acquisition, how fast can we integrate it, right? The proof is in the pudding. We've done three acquisitions. Four of them are fully integrated. The third one we just acquired a couple of weeks ago, it's getting integrated. I expect the integration of these three should be completed by, I would say, end of September. When I say integration, you know, the staff are on board. Everyone's using the same IT baselines that we have. Everyone's using the same tool methodology. We found out that acquisitions, especially if they are non-public company, cash-based, challenges come when you have to go and convert those into their financials into ASC 606.
That takes longer time than the technical portion. We're in the middle of that and getting that done. As a matter of fact, I'll tell you about the strategy of acquisitions. It was very key for us to get those done because let me see if I can share one screen video, Blair. By the way, this is already on our page. I don't know if you can see it right now or not.
Yep, yep, that's good.
If we make this bigger, you'll see that I've always said this. To make a final chip, you have to go through a design step, a wafer Fab step, a packaging step, and a testing phase. The final chip, whatever else you need to do in order to, I mean, if you have requirements for automotive, you have to burn them in. You have to test them differently. Those things happen at the final chip level. You'll find out that we have and other EDA vendors have focused for the past 40 years on design. One thing that we added just most recently, I'm starting acquisitions from most recent to first, was Mixel. It's a mixed- signal IP. They're one of the world leaders in making MIPI IP for automotive. That acquisition was completed. We've added about $110 million of market share for us.
That's enabling us to grow our IP business unit at a much faster pace. This year we've said we would add anywhere from $2 million - $4 million of revenue just because of Mixel, because we just finished that acquisition. There's only five months in the year left. That actually helps us increase our presence in the design space. Before that, we acquired a company called Tech-X . Tech-X, we showed in our last slides, does two things for us. One, in our FTCO, it helps us with the wafer Fab portion, digital twin modeling, because Tech-X enables us to not only do digital twin models of smaller components, but also can actually simulate the performance of things like how do you etch this wafer.
Etch, for example, plasma etch, which is the most popular etch profile used for advanced nodes, how this profile looks like over the full wafer, not a die, not a transistor, but over the full wafer, which takes us to the next level of FTCO, which we are incorporating into our FTCO. Not only by itself can we solve those products, but also integrating them as part of our platforms enables us to go after solving bigger problems, which is the reason for our existence. We always love to solve the most difficult problems that come our way. Tech-X, besides the FTCO combination, Tech-X also fills out, I would say, the LEGO blocks for our photonics offerings. We had mentioned many times that we are in the photonics market. We had mentioned many times that photonics integrated circuits are big.
The component we were missing was the component that was provided by Tech-X , which is a finite difference technology, or I would say the algorithm method by which you can run simulations on AI-based microprocessors, graphics processors that can really compete well and surpass everyone's performance in terms of accuracy and speed. That's the component we were missing. Adding those helped us complete our DIGSOP puzzle for photonics. You'll see our presence more so in the photonics market. These are partly to get revenue this year from the existing customer, land more customers, and expanding them. We're basically, with these two acquisitions I've shown you, teed up really well. We're teed up well for the next year because that's where a lot of the new customers, new expansions come in.
In photonics, as you know, it takes the same steps as electronics, except you have to use different kinds of software, different kinds of behavior. Photonics and electronics behave differently, and at times you have to simulate them together. That's what we call multi-physics, going from photons to electrons and conversions back and forth, dealing with waveguides, dealing with circuits. That's what this is all about. This alone by itself has also added a total of $216 million to our market. Part of it, as I mentioned, is photonics integrated circuit, which is a $115 million market. The FTCO portion, we should call it Fab here, is another $145 million. These are the latest two acquisitions done in the last month and a half. Prior to that, we acquired what we call PPC, or it's basically proximity correction for masks.
That's where you transfer designs into wafer to a mask, which happens in that step. We actually acquired the OPC technology from Cadence . That gets us into the market where we can actually take that technology into the next stages of evolution of OPC or PPC, as you call it, to make masks. All the mask corrections happen in two dimensions, X and Y. Those are done because when light goes through the mask, hits a wafer. It comes back, reflects, refracts, and those basically call for aberrations that create problems. What you print out doesn't come out as what you want to be on the wafer. Those are real problems. Mask corrections like PPC and OPC take care of that in two dimensions. Anything that needs to be adjusted in order to correct for those reflections and refractions, that takes care of.
Where we are taking this now, since we have TCAD, we are enabling in the third dimension correction. With TCAD, since we can simulate actually what you print, what shows up on the wafer, we actually can pre-simulate this and see the impact of it before they make wafers. We can actually do the correction in the mask in the third dimension, which is where we are taking this technology. Very excited about it. Every time I see how progress is being made here, combining TCAD with OPC in order to go to the third dimension is very powerful for us. That is where we are taking it. We already have customers in that domain. We've had over $2.5 million, $2.6 million, $2.7 million of revenue from it.
We continue from the existing customers who have more revenue from it and expand it into the new markets with the new technology and new customers. That is the ballpark of what we did for the last three acquisitions. I want to emphasize, this was not only to buy talent and technology, but also buy revenue with the potential of high expansion. We are hoping that as soon as we finish our integration of these three companies by September, the next quarter when we do our call, we can also keep track of, which is very near and dear to my heart, keep track of ROIC, right? How much ROI you get from the cash or from the investment you've had with these acquisitions. That is a metric we have never reported.
As we finish our acquisition integration, I'm hoping to get that done also, if not in Q3, then in Q4.
Interesting. The Cadence acquisition you made, I want to ask you a couple of things here just to clarify that. Was it Cadence that came to you guys or you came to them? You both sort of saw that maybe it would work better with the TCAD owner? I would think also though that the proximity compensation correction is almost more of a manufacturing side of the equation than on the design team, the R&D team. The TCAD is, can you just help us understand that?
Yeah, yeah, as I mentioned, we had historically focused only on the design portion. If we did not have FTCO, and because FTCO takes TCAD into manufacturing, we can do certain things about process, you know, the actual manufacturing process, which corrects everything. The first step of manufacturing is what? It's making masks to expose wafers, you know, photoresist to different kinds of light sources to make, you know, manufacturing of wafers is like brick laying, right? You put a layer of cement, then you put the brick, you put a different layer of cement, you do your changes if you want to make a window or door. The same thing in manufacturing of wafers. You put down the photoresist, you know, we call it depannage. You deposit something, photoresist, and then you'll expose it through mask and you etch it.
Depannage, those are the main steps of manufacturing of wafers. The first step being mask, you find out that there's a lot of aberrations and errors that get introduced by masks when they get into advanced nodes. We said we have the design portion done, we have the manufacturing portion done, all the process and device steps. What we don't have is the first step, which is mask making. We've never been in there. Unless you have other technologies to complement what exists, you can't take it to the next level. You have to be an ASML, right? The biggest, I would say, provider of masks is ASML, right? Or mask corrections is ASML. For us to compete and be able to get to that, we have to have had complementary technologies to enable us to help correct the first step of manufacturing, which is mask making.
With TCAD combination, as well as existing PPC and OPC and the new algorithms we are justifying, is basically where we are taking this. If we did not see a way for us to basically go and expand a way to improve and enhance based on our technology, we would not have done this acquisition. As the first portion of your question, did they come to us? They did come to us. Matter of fact, the nice thing about this technology is three of the largest customers turned out to be our customers as well. When Cadence had told them they want to sell this product, their customers recommended that they come to us. That's how the process started. There were other suitors, but it turns out that we were the best match and the best way the team thought we could expand what they have.
That's how it worked out.
That's great. Thanks for the clarity. Maybe I would take the conversation up just a level here and talk a little bit about some of the market trends in EDA and in simulation, multi-physics simulation. What are kind of the major drivers out there now that you're seeing and how are you responding to them?
No, absolutely. You know, we have talked about this many times, but it's always good to show a slide that I've also shown in our presentations in the past. If you think of it, I mentioned already increasing complexity of chips and photonics, new materials introduction, go-to-market challenges. You think of wanting to manufacture very high complexity chips and photonics. That's challenging, right? That's the first problem we're solving, right? In order to do that, you find out that this complexity of requirements surpasses the memory, the high-performance compute, photonics, compute, automotive, every aspect of our life, I would say, right? I've named some. The acquisitions that we did actually improve on our presence in the markets that I've done dashed orange lines on. You find out that for complexity, all those markets get impacted. For new material, all these markets get impacted.
Any of these devices or circuits or photonics that need to get to market require you to have a new set of rules, new set of simulations, new set of digital twin models. As you know, in the old days, it started at one transistor simulation. Now it's gone to the full chip. Now it's going to wafer level. Now it's going to basically what I would call a system level simulation. You find out that all of these are needed, capabilities are needed to be able to do, I would say, digital twin and simulations of bigger and larger systems in order to get us where we are.
Okay, great. If you kind of, I want to shift and talk about one of your markets for a minute, which is China. China's had a lot of change in the market for EDA vendors in the last two years. What is Silvaco seeing in that market and where are the opportunities for you there?
That's a very exciting and dynamic market based on the geopolitics of it, based on the performance of it, based on the technology part of it. As you know, there were quite a bit of news a few months back in terms of involvement of the government in this space, prohibiting EDA software tools to be shipped. It was reversed, and then certain things were prohibited to be shipped. Now it's reversed. The good news is that this technical collaboration, I think, is very healthy. We've always had about, I would say, 15% - 20% of our revenue coming from China. We see that still going and perhaps growing based on the less limitations that we are exposed to. The nice thing about it is that market is growing very fast, and we are part of it. We will end up to be a bigger part of it.
Matter of fact, if you think of our TCAD and automotive, as well as photonics and power, those are the areas we are growing in China. We will continue to grow, and I think it's a healthy market to be participating in as long as we follow the processes that are acceptable to everyone.
Has that market been a better market for you this year than last year?
If you had to ask me this question six months ago, I would have said no. Now, yes. Yeah, it's very dynamic. That's what keeps us on our toes, right? You've seen our bigger competitors, they're dropping their estimates based on the fact that they didn't look good six months ago, but now it's different. We'll all see how that goes and monitor it.
Okay. You know, the other area, I don't want to touch on this briefly, but you do, obviously, your TCAD business is very advanced technology, very, as you said, working with new materials at a very early stage. You do a lot in photonics. Is there an opportunity in quantum computing for Silvaco? Is that something that you're watching or you can do something with your TCAD technology?
Oh, absolutely. Not only TCAD, but also EDA as well as, as a matter of fact, we have two customers that do quantum computing. One of them not only uses our TCAD, but also our modeling services for IC design. As it turns out, a lot of these companies require new material. You know, not many circuits could have been modeled at 4 degrees Kelvin, right? The models did not exist. We had to generate those for those customers. Not many circuits can go to the boundary of optics to electronics because you find out that at some point in time, the frequency band is required to go to that boundary of electronics to electronics. It becomes so high and so power hungry. We had to provide solutions in terms of TCAD and IC and coming up, I think IP also solutions.
It's been always, we've been there for the past five years with them. We will be there for the next 30 years. Hopefully, they have a quantum computer with 2,048 qubits that everyone can utilize and succeed in. To me, it's a long-term journey. It's not something that's going to come and go. It's been going on for a while and it will go on.
Have you seen increased investment there in the last couple of years?
Yeah, I mean, look at the companies that are private and public. You can see from their stock. You can see how much capital was raised. As a matter of fact, there was an announcement a few months back. There is a quantum computing company in Palo Alto that raised over $450 million just recently. You'll see quite a few of them are doing it. It's a very hot topic. God help whomever that falls behind that technology. It's something that if we don't keep tapping into it and don't grow at it and solve it, we would fall behind drastically. That would not be a good thing for us.
Yeah, interesting. On the Tech-X acquisition you made, a lot of it's in the simulation market, the technology. What would you be, what's sort of your strategy in simulation? If I look at Synopsys acquired Ansys a month or so back, Cadenc has been building a lot of simulation capabilities. What's Silvaco Group's positioning in this space?
Yeah, because they're kind of our bigger competitors. They tend to focus on system-level simulations. For example, Ansys was really good at simulating engines of a car, right? Their bigger competitor is Dassault. They simulate or generally model the whole airplane, right? They actually make airplanes. They're getting into bigger and bigger systems. Cadence also acquired BETA CAE. I think that's what you were referring to. BETA CAE enables them to do system-level simulation as well. The difference between them and us is we've always said, if you look at the physical world, if you look at the city, you know, the airplanes, you know, the cars, those are the sort of things that other companies are focusing on. What we really focus on is things at the wafer level and below, to the atomistic level. We call, you know, our tagline is from atoms to systems.
Our system right now stops at the chamber level in a manufacturing site or at an optical module. That's where we are at. Our premise or philosophy has been always to build that up to a level that basically gets us to a very deep understanding of, and I mean it by atomistic level. Physics simulation is, you know, doing green functions, doing mass transport functions, all the way up to what a wafer work looks like, mechanical, thermal. If we are not there yet, we are probably 80% there. We have 20% more to go to really get to where we can solve a lot more problems in manufacturing. When we're comfortable with that, we look at where else can we focus on that we can lead, right? We have to have, we have to have moats, right? FTCO is one of our moats.
Our next moat is where we are taking PICs to, right? The photonics integrated circuit. We are looking for the next moat that hopefully we can announce some next year that we can differentiate and have products there. That, you know, like I call our bigger competitors or bigger brothers, we all are working towards expanding the market, not necessarily have a head-on competition. That's my goal. My goal in the industry is to be able to expand markets, solve newer problems, more difficult problems. That's what we're here for.
Got it. No, that's very clear. That's helpful too, because, I mean, you're really, because of the TCAD core technology of your platform, there's lots of simulation required around that level of technology research. As opposed to, as you said, you're not talking about wind tunnels for aircraft wings or cars. Can I ask you this same sort of question around your IP business? Silvaco has always had some IP. It's been anywhere from 5% - 10% of your revenue. You had a relationship with NXP and their portfolio. Just talk about, Babak, a little bit about, you know, what is your, what's Silvaco's strategy around IP and how much, how important is it to, or how big do you think it will be?
That's a great question, Blair. You're right. Our IP business has been 5% -1 5% depending on our deals during the year as we share. Our strategy was initially to, as a matter of fact, our IP business unit is the youngest business unit we've had. TCAD is over four years. You know, EDA is over 30 years. IP is merely since, I would say, 2019. 2018 is when we've had it. Historically, it was to partner up with companies that have their IP and they can't go sell those, but the IPs are already tested and proven track record in manufacturing. Given that, what we did initially, we partnered with NXP, which we still have that partnership. As a matter of fact, we renewed it a year ago. We actually partnered with Samsung. We announced we were the first company ever to be able to commercialize Samsung foundry IPs.
When we looked at that, we found out that there's an upper limit on how much of these things the companies are willing to commercialize. There's a new need for the next level of things. What we've done is we've expanded some of those relationships to the IPs that we know need future expansion, and we can provide that expansion. We are doing that. We found out that the future expansions that our customers are asking for don't exist there. What we decided to do is add external of your acquisitions like Mixel, that again, we focus on very unique modes, right? MIPI IP, you find out maybe there are only two companies in the world that can do it to the level that Mixel does. I think technically, in my opinion, Mixel is probably the best technical MIPI 5 companies that exist.
Again, we focus on the modes that differentiate us. We focus on the areas that we are not necessarily competing head-on with our larger competitors, and we can coexist and expand our market. That has been the next phase that you've seen. The third phase is, of course, we are adding to the team. Now we have a large team after this acquisition. As I mentioned before, we've added over 100 people. We can do more different kinds of IP, different flavors of it. That's where we're at. Creating our own IP that's differentiated is where that comes up next. Don't be surprised if within a few years, our IP business unit will surpass our TCAD business unit because this has a much higher TAM than the TCAD market itself.
Great. Just coming almost up on our time here. Let me just ask one question back on the topic of AI. At a high level, where is the, you know, the opportunities for Silvaco in AI in terms of adding into your product line? If you look at Cadence and Synopsys, they've added additional products on top of their stack and so forth. Are you guys building AI into the product or are there net new products that you can charge for that will incorporate AI?
Yeah. Historically, I've said there are four areas that AI is being utilized and will be utilized in EDA. The first one is what we call design assist, changing the graphical user interface so that the designers can have it easier, and the AI engine behind it learns their behavior and their habits and helps them get things done faster. EDA companies are all there practically. It's pretty much the first step that was done. The next step is trying to create portions of the design, like IP blocks, and how much of it can be automated. That's where everyone's working at. The third stage, which is to me the ultimate state, is can you put in a paper design spec into this EDA tool, like you do ChatGPT with text prompting, but much more intelligent prompting?
Can you provide that prompt, that design spec prompt, for lack of better terms, to this system, and it generates a whole design for you? You don't have to have manual, and I think that is 20 years, 30 years away. That would be the holy relic of AI in the design portion of it. The next level, as we said, we've already taken AI to manufacturing, and we are the leader in that. That's where we are at. The manufacturing portion, again, I think it's in its infancy in terms of AI for the sort of thing that's needed to do. That's another long-term journey. I think the potential of simulations, digital twin modeling, integration of AI will surpass well beyond my lifetime. That's the sort of thing that I take pride in participating in and helping move forward.
Excellent, excellent. Listen, Babak, thanks very much for joining us here this morning. Thanks for getting up early. We appreciate your insights.
Thank you very much for your time. I appreciate it. Bye-bye.
Bye.