Right. Hi, everyone. I'm Anand Agarwal with JPMorgan. I'm joined here today by Amir Panush, CEO of CEVA. Thank you for joining us today. We have 30 minutes together. We'll leave the last five minutes for some Q&A. Maybe to get things started, Amir, if you want to give the audience here a brief overview on CEVA and for folks that are hearing about the company for the first time and who don't kind of know the evolution and the story, it'd be great for you to provide them with an overview.
Yeah, definitely, Anand. And thanks for having me today. I'm Amir Panush, CEO of CEVA. At CEVA, we are an IP licensing company. The best way to think about it is like ARM. We are licensing our technology, and then we get royalty for every unit sold. But contrary to ARM, we are really enabling three use cases for the smart edge: the connect, sense, and infer. In terms of connect, we have broad-based wireless connectivity IP. IP is like Wi-Fi, Bluetooth, cellular, IoT, 5G technologies. For sensing, we have lots of DSP capabilities and audio and other sensor capabilities. And then for infer, it's all the things related to NPU or AI accelerators that go into the edge devices. We have been very fortunate to enable more than 20 billion devices so far, about 2 billion devices every year, with strong leadership in all the different wireless connectivity.
Overall, we see great momentum and growth engine for us. If you want to think about CEVA, CEVA has been around for more than 20 years. The first phase of the company was all about mobile communication. The second was about the IoT and wireless connectivity all around us. Now the third is really the smart edge with inference, AI, smart devices, enabling the future growth of the company.
That's great. You mentioned Smart Edge and Edge AI. Could you spend a few minutes maybe talking a little bit about the company's specific offerings around that space?
Yeah, definitely, Anand. First, for us, smart edge is really the combination of those three use cases that we are enabling: the connect, sense, and infer. Every device needs to be wirelessly connected to the cloud to provide the communication. Every device needs to be able to be aware of the surrounding with the sensing capabilities. Of course, the ability to infer that and to have the smartness inside the device. Specifically for our technology on the inference, the NPU, all the technology that people are talking about, the AI, AI is well known to be a technology deployed in the data center, super hyperscalers for training and now for inference.
We are truly enabling that into the edge devices with the combination of technology that's extremely low power from our heritage in mobile low-power devices and very scalable, going from basically the use cases, the simple use cases for audio, predictive maintenance, medical devices, and those types of things, all the way to the high-end NPU or AI inference that goes into smartphones, tablets, ADAS system, autonomous driving, and so on.
When it comes to all of these solutions that kind of cater to AI, how do you differentiate from your competitors, and why do you win?
Yeah, very good questions. There are really three factors related to that. One, we always have a very, very strong combination of the performance and the low power. Again, that comes from the DNA of CEVA for more than 20 years, enabling extremely low-power devices on the edge. With AI, performance and ability to run the network efficiently with low latency is super critical. The second is the scalability of this technology. If we work with MCU partners, if we work with SoC partners, what they are looking for in AI is the ability to serve all their product line, from the lower-skilled devices to the very high-end devices. We basically have one architect that has the scalability to go multi-engines, multi-core, different types of configuration, but the baseline is the same. We can serve them in all the different configurations that they need.
Scale is super critical for AI at the edge. The last piece is really the integration of the hardware and the software. In every device and every technology that we go and enable with our partners, we provide both the silicon IP as well as the software IP from the graph compiler to all the software development tools and so on, and the integration of these two things together. If I can give an example, we just won a major design win. We have Nextchip, a Korean automotive tier one that supplies to Hyundai. When we asked them at the end of the day why they picked our technology, what they told us is that they were looking basically to have one solution that has the accelerator of AI, so running the neural networks, but also doing all the pre and post-processing.
You need to get the sensor information, be able to decode it. This is a DSP functionality. We provide all the DSP functionality plus the accelerator, so called the NPU, the neural processing unit, integrated with one software that runs all things together and scheduling that, all in a very power-efficient manner.
Got it. Got it. Given the breadth of end markets, and I know you kind of segment your end markets as smart edge and mobile potentially, could you maybe provide a little bit of an overview around some of the end markets that you serve and then essentially your outlook on some of those end markets, maybe for this year and then essentially next year as well?
Yeah, definitely. We serve and target six end markets related to the smart edge. It's PC, mobile, automotive, wireless infrastructure, consumer IoT, and industrial IoT. Specifically, let's start with mobile. Actually, mobile, we see a tremendous opportunity for us this year and going into 2026 and 2027 of accelerating and improving our market share with a new large U.S. mobile OEM that decided to deploy with their 5G modem our technology. That's a great tailwind for us as we go towards the rest of the year and then 2026, 2027. The other piece overall, if we look at the consumer IoT and the industrial IoT with the different wireless technologies that we have delivering those IPs like Wi-Fi and Bluetooth, we see significant market share that we are going to gain.
Already, we are shipping today more than 1 billion devices of Bluetooth annually, more than 200 million devices of Wi-Fi, around 200 million devices of cellular IoT. All of them are expanding within this year and even more growth in 2026.
All right.
Just one more piece. In wireless infrastructure, we are already basically supporting two of the four large OEMs out there, and we have more than 30% market share.
From a mobile perspective, do I recall correctly that you have more than 60% market share? Is that right?
Yeah. So today we are around, let's say, around the 15%+ market share, and we expect to go more than 20% and above as we go to next year.
Got it. Could you maybe talk about what does your competitive environment kind of comprise of? Because there aren't a lot of companies that are essentially catering to your end markets from an embedded IP perspective, right? Essentially, what is the basis for competition in this space? Is it price? Is it time to market? Would love to get your thoughts on that.
Yeah. So overall for our technology, I would say the first things to keep in mind that our competition many times is so-called the make versus buy is our customers need to decide whether they want to do these technologies in-house by themselves or to basically license from outside supplier like CEVA. In terms of wireless communication, whether it's Wi-Fi, Bluetooth, cellular IoT, and so on, we really don't have that much competition anymore as an IP supplier. We are pretty much so-called the leader in that space. The competition is, to the most part, the make versus buy with the NIH potentially of our customers. Having said that, what we see more and more, and the best way to put that is if we look 20 years back, right, in the semiconductor industry, companies, they have their own fabs.
Over time, they mostly or almost completely moved to a fabless model. They go to TSMC and other fabs, right? Building a CPU, companies did by themselves. Today, ARM pretty much dominating that space, right? We are looking to basically be the leader of wireless communication around those CPUs to be really de facto the standard to help companies drive their success in the marketplace, and we are as an IP provider. In order to do that well, back to your questions, what are really so-called the metrics or the KPIs, the things that our customers look at? Definitely time to market. They need to make sure that we can really help to accelerate.
We believe that as we make that progress in the marketplace from market share and capabilities, we really see how we accelerate our customers versus their own internal developments in terms of getting product out to the market. The second is really the cost of ownership or the price of the cost of developing that. We have the economy of scale. Basically, that's the beauty with IP. Once we develop that, to the most part, we can replicate it, and then we can make it much more economical to our customers instead of each of them developing on their own. With that, I really view that wireless communication will become basically an IP supply to all those different companies, and they will innovate in how they integrate all the different components together.
If we look a little bit at the rest of the product portfolio that we have, whether it's NPUs or DSP, so-called more the processing capabilities, definitely there is the make versus buy, but there there's also some competition that we see sometimes here and there, Synopsys or Cadence, depends where.
That was a great analogy, by the way, the whole transition from fabs in-house to from an IDM model to essentially like a fabless model. I think that's definitely true and relevant today. Could you talk a little bit more about maybe what is, besides that, besides the transition of people outsourcing more and more of the IP, what do you see as some of the short-term and longer-term growth drivers for the business?
Yeah. First, short- term, if we think about 2025, while of course tariffs and all that can create some jittering in the marketplace, fundamentally, we see a significant growth opportunity for us in mobile. We see a significant growth opportunity for us in terms of Wi-Fi. From a licensing perspective, we see a large number of customers, tens of customers that are already licensed for our Bluetooth or other technologies coming and licensing the next generation, so-called from Bluetooth 6.0 to 7.0 and from Wi-Fi 6 to Wi-Fi 7, plus the integration of this technology. We have lots of customers that have licensed Bluetooth, and now we license both Bluetooth and Wi-Fi and other cellular communication technologies. From a royalty perspective, we just talked about the two other tailwinds and the engine for the business in 2025.
Then beyond that, if we look in 2027, 2026, 2027 and beyond, it's really this revolution of devices are becoming smart with AI inference capabilities. We literally see now the transition from cloud-based to the most part to hybrid with edge devices capable to run from small models all the way to LLMs, so-called the large language models. That will drive significant growth opportunity for us.
It's almost like a land and expand model.
Yeah.
From an end markets perspective, where do you see the growth coming from in 2025 and 2026? There's a lot that's been said around how some of the end markets, specifically consumer, et cetera, are going to see fairly muted growth in 2025. In your view, where is a lot of this growth going to come from?
We are overall in a very good position that even if at the end of the day the market itself does not grow significantly in 2025, whether it is mobile, whether it is the industrial IoT, whether it is a different wireless connectivity that we have for the different consumer IoT devices, we are still in a gaining market share position, very strongly in mobile, very strongly with wireless connectivity, all of them actually. We are still gaining market share. What we have built in the last few years is a large number of customers that license our technology and ramping. We will keep seeing that ramp coming in the next two, three years.
The last piece I will mention is the automotive, where with our vision AI, DSP technologies, those technologies we licensed now almost four years ago, and now it's really starting to ramp into cars that are going on the road.
Because of the cycles essentially.
Because the cycle in automotive, and especially the more complicated vision AI, the DSP, SoCs. Now it's in production, in car production, and we will see in 2025 and 2026 nice growth in the automotive space as well for us.
Yeah. So maybe, and correct me if I'm wrong here, maybe a good way to think about this is like you've done the investment that they've embedded your IP within the vehicles. Irrespective of how fast or if the auto market continues to decline through the end of the year, you're still going to win because your IP will start shipping this year essentially. That's kind of the way to think about it.
Yeah, exactly. That our technology will start shipping as well as those cars are becoming, they have more and more of those so-called ADAS, different ADAS system enabling the.
More content.
Exactly.
Yeah. Yeah, that makes sense. Over the last several years, you've been investing in your software portfolio, right? Could you maybe give the folks here an overview of what that portfolio, one, comprises of, to essentially what are your areas of focus for the future within software? Have you seen those investments that you've made in the past kind of pan out?
Yeah. First, even before I go to the specific investment, I would say that our approach, and it goes back to what I mentioned earlier, is that for any silicon IP that we offer, any product, we are really going all the way from the physical layer up to the upper level of the software stack. We provide it as a complete solution. What we have found is that our customers, in order to really be able to get the first time to market and have a good solution that works well in the field with interoperability and all the other capabilities, they require basically the complete stack, including the software. We have invested quite a bit, and whether it's Bluetooth, Wi-Fi, Cellular IoT, 5G, and so on, we provide complete functionality.
The same for AI as we do today, as I mentioned, Graph Compiler, all the software tools. That is why, again, Nextchip and other customers really like our technology and start basically adopting that. On top of that, what we have invested very specifically is adding so-called embedded software application capabilities for sensor fusion, for contextual awareness, for 3D spatial audio to provide more immersive type of user experiences that differentiate overall our silicon IP offering. With this type of technology, we go directly many times to the OEM, not necessarily to the semiconductor provider, because they really drive so-called the user experience. They value that. They can integrate it into their own devices. With that, we get higher royalty per unit than typically you would see with silicon IP.
Got it. Is there a % attached rate of software that I know you do not publicly disclose this, but is there a directional sense around what is it a vast majority of customers that are kind of adopting these software solutions along with the silicon IP? Or is there still something that is kind of gaining traction in the market today?
Yeah. So overall, as I mentioned, first, most of our silicon IP when delivered with software, but more the software application goes on top of that. We are not really talking about % of market share, but definitely we have enabled more than 300 million devices already with these type of technologies. And we see that still going very nicely for us.
Got it. A lot of conversations that we've had over the course of this week and the last several weeks, actually, companies are increasingly selecting geographies to sell into, whether it's if they're specifically the U.S.-based semis companies, they either want to sell into the Western Hemisphere or they want to sell into Asia, right? And by Asia, we're kind of you can almost segment China into its own little bucket, right? What is your one view on that? And two, do you feel like you need to kind of bias your strategy for the future to align with that sort of sales motion Asia versus the U.S. or the Western Hemisphere?
Yeah. First, I would say as a tech veteran, I would say that overall, free access to technology and ability to collaborate, it's something that generally speaking we prefer, right? Having said that, with so-called more the localization right now, fabrication of needs of technology, actually it's a tailwind for us as an IP company. We do not develop the final products. We do not ship the final products. We do not have that so-called limitation or impacts related to tariffs and so on. Actually, as an IP provider, because of the localization, we see in China and in other regions, basically each region wants to make sure that they have access to the best technology out there and to maintain that access. With that, as an IP supplier, this is actually a great tailwind that can help us to keep driving the business.
We serve both regions, or the Western and the East sides of the world, very, very successfully. I would say that what we do see is acceleration, basically, of the need to go and get access to the IP. Also, I would say to some degree, it even drives more innovation, surprisingly, right? Like the DeepSeek and other models that came along and all that, they came from the needs that the harder, to some degree, harder access to the innovation that comes worldwide, and then everyone tried to innovate on its own. Again, that is where propel more IP needs.
Do you feel that despite all the noise with tariffs, do you feel like just given that response, it feels like you are not going to be impacted by tariffs as much as maybe some of the physical silicon companies?
Yeah. In terms of direct impacts, we really have today literally no direct impacts. Again, we are basically not shipping any goods. From indirect impacts, the way that we look at that is really whether it will impact the end demand by the consumer. At the end of the day, if tariffs will create negative sentiments of consumer in terms of consumer demand, that will eventually translate for consumer people buying less goods and less products and technology. That propel and create indirect impacts at the end of the day to our shipments or royalty revenue. Overall, to some degree, potentially the level of innovation, at least for a short period of time. In the long run, it will also correct itself. The innovation, the need to get new IP and access to the new technology, it just will get accelerated. It's not going to slow down.
Yeah. That makes sense. As a business, you've done extremely well at maintaining your gross margins above 85% fairly consistently. How do you see this trending forward given the discussion around tariffs, given the discussion around end market demand, all of those factors?
Yeah. So we are really an IP company. And within being an IP company and how we manage the company, definitely our goal, and we see that continuing, that we will be so-called above the 85% gross margin and an average between 85% and 90% gross margin. What we do see is here and there, big projects where we basically provide the off-the-shelf IP that we have that really drive the 90% gross margin. But on top of that, because we are going after really the top players in the different markets and they need some help with some developments of the technology, we are also enabling some customization that we work with them. And with that, the gross margin can go single point up and down. But overall, we are going to sustain and maintain about the 85% gross margin.
Two years ago, we had a business of design services that was a lower gross margin. We divested that. I'm a strong believer in the IP business model and that will drive. It drives great gross margin. From there, it can drive great profitability as the scale goes up.
Got it. You talked about scale going up and driving profitability as you kind of increase scale. If you think about the growth required to attain a certain scale, let's say $1 billion, right? What are some of the compromises you're willing to make in terms of margin for that? How do you think about that general growth versus margin paradigm?
Yeah. So first, we are definitely a growth company, right? And we would like to be a growth company. And that's our goal. With that, in the IP business model for us to grow, we are doing all the things organically. We talked about the several tailwinds, the mobile, the Wi-Fi, the AI that is coming more and more. On top of that is also inorganic growth and definitely M&A opportunity that we are looking at. That's all within the domain of IP. Within that growth, we will still maintain so-called the 85%+ gross margin.
With the one thing that I mentioned, that when we see the opportunity to really drive new innovation to the marketplace with the leaders out there to create a technology that we can own and even scale up, we will do that additional customization on top of the off-the-shelf type of an IP, which can impact a little bit the gross margin and single point. We are still basically the focus like with that is really growth-minded company.
Got it. Got it. I think you kind of answered the next question I had, which was around historically, you've done a phenomenal job of kind of managing organic and inorganic growth at organizations that you've been at in the past. The focus here is going to be on acquisitions or inorganic opportunities related to IP. They're going to be basically you're not going to venture into the device space. Would that be fair to assume?
That's definitely our focus first. We have done very well. If you think about CEVA, what is the strength of CEVA? Definitely strong inheritage of more than 20 years being as an IP company, serving large number of customers in all the different edge markets. This is a great platform as an IP company to go and expand more and add to that portfolio. We already have done multiple acquisitions and very successful one, as you mentioned. Our Bluetooth and then Wi-Fi technology came from acquisition that we grew up extremely, extremely well over the years. We have done acquisition on software. That's how we added more of the software application capabilities that drives for us really great royalty and revenue overall. The last piece recently, we acquired also a company to develop RF IP.
Now we can combine it all together with really, as I mentioned, to be even more content and better economics of the deal to our customers. We have done multiple steps through the journey. Definitely, I see in the coming years that we will add more IP capabilities, more technology into the portfolio to the most part around smart edge, becoming even more relevant and more content to our customers. What we see is that, as I mentioned, the whole history of semiconductor, right? We can see those systems, the so-called SoCs, are becoming more and more complex with more and more ingredients. It's a very challenging task for all our customers to do it at scale up and down through the different tiers of the product line that they need to support.
The more we, as an IP provider, as another, can provide them the large portfolio of technology they can embed in, and then they can basically go and put it all together, the more value we create and more synergies.
One-stop shop.
Yeah, exactly. That's a nice way to put it.
All right. We have a couple of minutes left and wanted to see if anybody in the audience had any questions. If not, Amir, I will thank you for your time. Appreciate you going through the fireside chat with us. Thanks so much.
Thank you very much, Anand. It was my pleasure.
Great. Thanks so much.