Hello, everybody. My name is Matt Ramsay. I'm one of the semiconductor analysts at TD Cowen, and just wanna thank you guys for coming to this session with Amir Panush from CEVA, the CEO of the company. I think Amir's gonna do some presentations, slides for us because they're making some changes to the way they're communicating the story, and I think that probably takes precedence over my nonsense in asking questions. We'll do that first, and then leave a little bit of time for Q&A at the end.
One thing that I did wanna point out, and sort of my disclosure language there, disclosures around, CEVA's filings and whatnot, and also, TD Cowen's research practices, et cetera, are all in your conference materials, that used to be in a book, but are now online. So, if you need to find any of that stuff, you know where to find it. But, really appreciate your, your time and attention, and, Amir, really look forward to hearing, things about the story.
Thank you very much, Matt. Welcome everyone. Thanks for joining us today. Amir Panush, I'm the CEO of CEVA since the beginning of 2023. I'm very happy to share with you the recent updates of what we are doing as a company, and of course, a little bit time for Q&A. So first, just the big picture for CEVA, for you guys that maybe don't know the details of our company, we've been, for the last two decades, in the marketplace, with more than 70 billion devices powered by our technologies. Last year, in 2023, more than 1.6 billion devices powered by our technology. We are about $100 million in revenue annually, have about $160 million in cash, no debt.
A very good customer base of more than 50 licensing deals annually, more than 70 companies, royalty-paying customers, and more than overall, 100 active customers. Lots of investment going into R&D. 75% of our workforce globally, both between US and Europe, are driving the R&D with more than 200 registered patents. Our mission is really to enable our partners in the semiconductor ecosystem with a transformative IP, driving innovation, with a very strong focus on the smart edge.
With that, what we mean by the smart edge is basically everything that is connected around us, that is then connected to the cloud, to the data centers. We see more and more devices going everywhere, with more than 80% semi-revenue growth. These markets, from an IP perspective, is more than $10 billion, and we have lots of TAM and SAM potential for us to grow, and of course, very large, more than 30 billion devices by end of the decade, and overall, a large market.
In addition, this market is really all the devices are becoming more and more connected. They need more and more AI capabilities and infer, and this is our focus as a company, driving and supporting the smart edge. We have two business models, for the guys here that don't know us very well. We have the hardware IP licensing model, where we license the IP first, and then we get royalty per unit as that goes to the semiconductor companies when they go to production.
But also, in the last few years, we added software IP licensing business model, where we license our IP directly to the OEM, to the end product manufacturers, and that's where we are- we get higher royalty per units, more value, and more stickiness with the OEMs in the marketplace. All that, we're focusing on six key market segments within the smart edge: consumer IoT, automotive, infrastructure, industrial, and mobile and PC. If we look historically at the company and how we have done, historically, we have known as a player in mostly in the handsets, in the mobile marketplace, and we have shifted dramatically in the last few years to a much more diversified markets and customer base, what we call the smart edge.
You can see that basically, most of that revenue got replaced with that much higher diversification of our technologies across the different market segments that I just shared. We expect it now to get back to growth as we go through 2024 and beyond, with overall the cycle of the semiconductor and the portfolio, which is much larger of our technologies moving forward. With that, in talking about portfolio and customer base, again, originally, we've shipped large volume, more than 10 billion devices, in the handsets markets, but in the last few years, we have diversified very, very nicely into the different markets.
More than 6 billion devices in consumer IoT. Already more than 100 million devices in automotive, and this will be a growing market for us moving forward. The same for the industrial and the infrastructure, wireless, market segment. In terms of customers, we have basically three major customers. One are the smaller fabless companies, where they drive innovation, and that's our way to get new technologies out to the market and innovate with them. The large semi, where we drive much better relationship, long-term relationship, and also more value to augment their internal capabilities.
And then, direct OEMs. You can see in some names here, like LG, Samsung, and others, that where we drive our technology directly to the OEM. Overall, a very good mix of products. You can see great brands, a lot also that, not necessarily are shared here publicly. With that, you can see also the different product categories, that this is just from the last quarter or so. All the way from the wireless, heavy infrastructure, to the rings, and the wearables, and the earables, and the cameras, mobile devices, light bulbs, pretty much almost every consumer device out there, we see our technology plugged into that and enabling this technology in the marketplace. We truly appreciate the partnership with our customers.
As we look at the smart edge, with our technology portfolio that has really changed significantly in the last 2, 3 years, we support 3 very clear use cases, and all the 3 of them really enabling what is coming to the new frontier of the smart edge. It's the connect, all the wireless connectivity technologies. It's the sense, different type of audio and sensing capabilities to create more immersive use cases and experiences to the consumer. And the inference, running basically neural networks at the edge, and this is the new frontier for us to grow moving forward, combined with all the other technology that I just talked about.
Talking of which, specifically on connectivity, we have the largest portfolio in the marketplace. We are the leader de facto with these technologies, spanning all the way from Wi-Fi to Bluetooth to UWB, Zigbee. We have all the different protocols like Matter. We are the number one IP licenser in that, in that market segments, with very high volume shipments already, many years of investments, and very, very large diversified customer base. In addition to the so-called short range wireless connectivity, we have the long range wireless connectivity, all the way from 5G to 5G Advanced, and then later on 6G, where we deploy our technology in the wireless infrastructure, in the mobile devices.
Putting them all together, all the different wireless technologies, this is really helping us to drive leadership, more value that we can bring to our customers. With that increase, the value per deal and the larger ASP or price per unit that we can charge for our technologies. On the sense side, we basically have two product categories, one on the hardware IP, as a soft IP that we deliver, and the other one is the application software. On the hardware IP, we have a very good mix of audio and voice DSPs, sensor fusion DSPs, as well as computer vision and radar capabilities.
Then on the application software, this is a recent acquisition that we did last year. Now, we have spatial audio for head tracking, with the motion tracking all combined together, and AI voice assistance. Where we're dealing directly with the OEMs, we see much higher royalty per unit. We bring more value and more stickiness to our technology, and that's really help us to drive more and more royalty into the mix of our business model, and overall, as we move forward to grow the royalty base.
Now, talking about AI, this is a major investment that we are doing in our companies. We have started a few years back with consumer and industrial vision, AI capabilities. We'll see some of these technologies already go to production towards the end of this year, and then ramping in royalty. That's the middle here in the picture, that goes into these different type of devices. But now with the AI, where we see generative AI really goes to the edge in a hybrid model. This is where we launch, about a few months back, a new NPU called NeuPro-M, and this is a technology that scale up all the way to several hundreds of TOPS, teraops.
And with that, supporting all the different high-performance gen AI models out there, from Llama 2 to Stable Diffusion, and so on. And this is right now in heavy qualification with multiple customers. We expect to start licensing this technology in the second half of the year, and much more in 2025 and 2026. In addition, you should expect that as we go also towards the lower end of the spectrum, to the so-called MCU ecosystem for IoT, this is where we'll come with additional technologies to support and complement it with our wireless connectivity that already is very entrenched in this market space.
And that combination together will bring much more value to our customers. We truly see the AI, so-called to run neural network and do the inference at the edge, becoming more and more valued at and a need in the marketplace, and we are in the front runner of these technologies to enable this in the marketplace. An example that we would like to share with you today is a customer called boAt. This is an Indian OEM. They are actually number 1 in volume shipments of earable wearables in India, and number 2 worldwide after Apple.
We are very fortunate and very happy to work with them directly, and basically provide them all the three use cases and technology that we have in our portfolio: the connect, the sense, and the infer. On the connectivity, we provide them the wireless connectivity for this headset. On the sense, we are providing them 3D spatial audio, this new technology that we integrated with our motion tracking, and this is embedded into that headset. You can see how they're basically emphasizing head-tracked spatial audio as part of the capabilities of the device.
As well as we provide them the DSP and the core technology to run so-called the AI neural network for audio systems, embedded in that, the same device. So this is a great example. We can see basically a one device out there, consumer device, where we can generate royalty and access for our technology across both connect, sense, and infer, including AI capabilities going really to the far edge. If we look at the overall portfolio, we have done lots of developments in 2023 and 2024 around the connect, sense, and infer.
Connect, we announced many, many so-called more advanced Bluetooth solutions out there, now going into 6.0, and later on into 7.0. Wi-Fi, that goes from 6.0 to 7.0. A UWB for consumer and automotive. In addition, in sense, as I just showed, additional capabilities in 3D spatial audio and motion tracking, a new solution for radar. And in the inference, we announced the new NeuPro-M for generative AI, and you should expect more to come from us this year on that portfolio as well. Now, a little bit into our strategy and outlook.
So if we look at really what is our strategy as a company moving forward, there are four key elements to the strategy. First, to maintain and operate a pure IP business model, where last year we divested the design services Intrinsix that we had in our portfolio. We are back to about 90% gross margin. We believe that we know very well how to scale, sell, and deliver value in the IP domain, and this is our focus moving forward, both organically and non-organically. Also, within the wireless connectivity portfolio that we have, we really have a very, very strong incumbency and leadership.
We are de facto the supplier for all the semiconductor companies out there that want to add new connectivity they don't have internally. We are combining more of those assets, we are enhancing those assets, as I mentioned, to the next generation, always, always being ahead of the curve. And this year, we also launch a new combination or combo of Wi-Fi, Bluetooth, and 15.4 Zigbee, all in a one solution IP, that we already licensed to one of our customers. This is where we see our ability to drive more value, and with that, extract basically, more dollars per deal, as well as higher royalty moving forward.
So that leadership will really propel a really great growth for us moving forward. But on top of that, the more so-called linear organic growth that we have, the investment that we are putting in AI is a major, major focus area for us. Not just delivering that as a standalone, but as a combination with our connectivity and the rest of the product portfolio that we have, that will provide for us the non-linear growth as that market evolve. From, as I showed in the previous slide, from the MCU, low, so-called low tier of that market segments of the IoT, all the way to the highest performance of ADAS and generative AI systems.
Last but not least, we have about $160 million in the balance sheet, with no debt. In the IP domain, there are multiple opportunities for integration and increasing value to our customers, and this will be also a focus area for us moving forward. Just few data points related to our strategy and how we made progress, and several announcement and things that we have done in just the last quarter or two. We were first to market, and we launched the Wi-Fi 7 and Bluetooth 6, and we already have early customers.
Last quarter, Q1, we announced two new deals related to that with early customers. We announced UWB, and we have already licensed that to the consumer devices, not just to automotive, where we started. As I mentioned, we introduced the combo chip of multi-standard techno wireless communication protocols. And we also announced, and we are delivering a strong partnership with Arm, based on their CPU and our 5G wireless communication capabilities.
On the sense, a new radar platform that we are in discussion with customers, I mentioned boAt and the spatial audio that we are making very good inroads in the marketplace, and overall, great milestones of our audio DSP in high volume with multiple customers. On the inference or AI, as I mentioned, NeuPro-M that we launched in the market and is right now in heavy qualification with multiple customers. Two of the three largest automotive tier ones are basically going to launch their ADAS with our AI-enabled technology, and that will start basically bearing royalty and go to production towards the end of the year.
And we keep expanding our ecosystem of AI with multiple partners. As we shared in the analyst day that we had, late last year, this is our long-term financial model, where we are looking about to 8, between 8% and 12% revenue growth year-over-year, maintaining the 90% gross margin as a pure IP play. With the scale and the investment that we already put into the business, we expect the operating expenses to grow in a lower range than the revenue, at about between 4% and 8%.
And we also mentioned that in 2024, we'll keep the OpEx roughly flat to 2023, and we are on track on that in Q1. And with that, we're expecting leverage in the model and better earning towards a $1 EPS, in the coming few years. To summarize that, and then I will be happy, to take any questions. Our strategy and what we are doing right now as we execute 2024 and beyond, back to pure IP business model. We are really addressing high growth, smart edge market segments and focusing very much on those markets. Very diversified customer base and product and technology.
This is really helpful as an IP company to be able to propagate our technologies and costs. Very innovative mindset, with lots of investment in AI and how to connect it with the rest of the portfolio. Financial disciplines. This year, again, OpEx is about flat to last year, and we will keep basically good control of our expenses. Investigating different M&A, as I mentioned, in the so-called pure IP play. We have great global, scalable IP platform, that we can go and add other technologies to the same platform and drive the growth moving forward. With that, thanks very much and open for questions.
Awesome. Thank you, Amir, for the great presentation. And I have the mic here, but people have heard me talk enough today, so if you have questions, please let me know. I have really two. One is sort of a more specific one, and one's a bigger picture one. The more specific one being, you gave the example of the work you guys are doing with boAt, and how it kind of ticks all the boxes as to what the strategy of the company is, and it's kind of a marquee account.
Maybe if you're able to, let us know a little bit more about the economics of that relationship, so from a royalty perspective, from a license perspective. Like, how does it come together, and how should we... If you're able to land other customers of, of similar, content, like, just any, any kind of royalty per unit or anything like that would be helpful. The second question's a much bigger picture, which is, as you've come into the organization as, as a new CEO and looked around, and you guys made the decision to divest Intrinsix, what do you think the, the organization learned through that whole process, and, and how you might apply that going forward?
Yeah, definitely. So on, on the first one, without getting into the very specifics, I would say first, of course, from royalty per device, we get royalty for each of the technology. So on an aggregated basis, it's much larger. But even more importantly, since we license the software and the value add directly to the OEM, and they look at the so-called price of the end products and the specific value that we bring to them, they really advertise the additional capabilities as part of selling the product, and they do it to their direct consumers.
We're able, basically, to bring up the value of our technology and get higher royalty base per unit. It's multiples of what typically we get for just so-called, the silicon IP. So it's really multiple X higher than the typical silicon IP, to answer that. On the Intrinsix, and going back to pure IP, first, with me coming on board, it's very, very strong focus on what we believe we know to do very well. So with that IP and how to sell IP, we also want to look how we can create better synergy between the different IP products offering that we have to go to the, to the customers directly, and also to the OEMs.
And also, operationally, to make sure that whether it's organic or non-organic, those teams can execute well together, and, and related to communication, ability to go to markets, the customer base. So moving forward, whenever we look at acquisition, we look basically, do we have a good synergy across the customer base, and does it apply very naturally into our business model, so we can really scale it up? I believe there are multiple opportunities for us to go and basically execute it that way.
Can you just talk about the maybe time to both licensing and royalty revenue for some of your edge inference products, and what you see as the, I guess, cadence throughout the, you know, next couple years, and how quickly your customers are both... I would assume that there's a lot of licensing interest up front, but how soon do you expect that to translate into, you know, products on their side, and of course, then royalties on your side?
Yeah. So the licensing process typically takes between a few months to, you know, a few quarters. Depends on the type of market segments. Automotive typically is longer. In terms of how quickly it goes to royalty, so all the consumer activity that we are doing today, within 1.5 years to 2 years, typically, you start seeing the royalty ramping up nicely, and then, of course, the 3rd, 4th year, it's even larger in terms of the ramp. Automotive, you can typically add another year or so, a little bit longer. Depends on where it goes into the automotive system.
Thanks a lot for the presentation. Just curious, when you're talking to generalist investors, what do you think the biggest thing is that, that they're missing?
I would say, one thing is really understanding the value of our leadership in wireless communication. I think... And I get lots of questions about it. Can we basically so-called charge more? Can we bring up the business? And the answer is actually definitely yes, and we're already seeing that happening. With the leadership that we created, the portfolio that we are putting, putting together, and I think more and more the investors community will understand truly the leadership that we have in that domain, and the potential of that bringing up both the licensing deals up, as well as the royalty.
With that, basically, a transition, so-called, from a high volume, mostly dependent on Bluetooth to the Wi-Fi, to the other wireless communication, that will drive much higher average ASP or royalty base per unit, and I think this is not yet well understood by the community. And the second, really, how AI, beyond the high-end, how the AI more towards where the volume shipment is very large, so-called in the MCU, larger IoT ecosystem, how we can so-called utilize that and monetize that over time.
As we get more of the proof points out there, I strongly believe that the community will understand that better. Last but not least, is really how we can scale up with more M&As, but that will come as well.
I guess, Amir, just maybe on top of that question, the largest company that does your exact business model, Arm, is also raising prices a lot. I wonder, and I mean, I've been of the view that that's gonna take some time to happen, but it's sort of justified in happening. Maybe they undercharged for a long period of time, and now they're trying to correct to that. Are there parallels there with CEVA? You know, I mean, it seems like there are.
Going back to what I said previously, what is parallel is that where we bring strong value to our customers, and I always believe partnering with the customer and bringing true value to them, as we do that, and we're truly doing that today with our very broad and very capable wireless communication portfolio, we are already doing and able to basically bring up the value, and with that, how much we charge. More and more over time, which already started, you will see per deal, as well as the average ASP, to keep going up.
So that's definitely a good, basically, parallelism or equivalent to what Arm is doing, but it's done differently and with different type of technologies. Maybe I will add more on that. Just a few comments is that, again, if we look, I didn't talk about Q1 and all that, but overall, from volume shipments, what we've seen is a very good ramp, quarter-over-year-over-year in Q1. We expect that, overall, generally speaking, as a comparison between 2023 and 2024, to continue. The market is doing better.
Our customers are gaining market share. So overall, with the exception of wireless infrastructure, we expect volume to keep growing in 2024 versus 2023. And in terms of the licensing, we expect... And we talked about total revenue for Q2 to be between 6% and 16% higher than Q1, so definitely we are looking for growth quarter-over-quarter. And with that, basically, having 2024 a stronger year from... compared to 2023, and then, with that AI and the more value for our wireless connectivity to add on to build the growth for us moving forward.
Excellent. Thank you so much, Amir, for
Definitely
... for the presentation, and taking our questions, and for the partnership over the years. Thanks very much.