Hey, good morning. Thanks for joining us again, for Session number 3 of Fireside Chats on Monday, at the Roth 30th Annual Conference. I'm Scott Searle. I'm the Communications Wireless AIoT Analyst at Roth. What we try to do in our Fireside Chats are highlight companies that are going through undiscovered growth phases, something that's unknown about them or transformational, and trying to really shed some light on that for institutional investors. With that in mind, and really fitting key into that theme, it's my pleasure that we have Adeia here today. Presenting on behalf of the company is CEO Paul Davis. Paul, thank you so much.
Well, thank you, Scott. This is an amazing conference, and we're very thankful for being invited and being able to get to talk to you and all the investors here today. It's very exciting. Beautiful setting and very exciting, and we can kick right into it.
Well, good. Yeah, you were out surfing this morning, I take it, right?
That is not what I do. No. Maybe getting lots of coffee and getting ready for this conversation.
You've got a story that is very stealth in a bunch of different ways in terms of what's going on in the world. Let's take a step back. Why don't you give us a quick overview of the company and kind of what's happened over the past couple years to get to this point of where we are today?
Sure. It's been a pretty fun journey. You know, we separated from Xperi in October 2022. We've been a standalone public company for a little over three years. Xperi and Adeia were two sides of similar types of technology, but we monetized in a different way. Xperi was on the product side, and Adeia is really a technology R&D company that then monetizes through our patent portfolio primarily in two specific areas. The media part of our portfolio, which our legacy comes from TiVo and Rovi. On the semiconductor side of our portfolio really comes from Tessera Technologies, Ziptronix, and some other technologies. We invent. We continue to invent. Invention is at our core.
It's one of our key values at the company. We've continued to add to our portfolio. At the time of separation just over three years ago, we had about 9,500 patent assets, which is really the lifeblood of our company. This is how we monetize. Today we've got 13,750 patent assets, plus. I'm looking at my CFO. Beyond that. We are, you know, we go beyond, you know, just you know monetizing through patent portfolios. We partner with the industry as well. That's a big part of, you know, our value proposition, that we go down.
I've tended to try and focus investors on some of the excitement around semiconductors and not intentionally neglecting the media side of the business.
Yeah.
Why don't we start there? Because media is 90%+ of the revenue stream today.
Right.
An incredible patent portfolio there. Take us through where we are now. It's been on this secular decline in terms of pay TV, but the investment that you've made in other sectors, such as streaming, e-commerce, social media, et cetera, are now at a point where we're starting to overcome that decline in pay -TV.
Yeah, absolutely. You're not alone. There's a lot of focus on semiconductors, and I know we'll get into that, because there's tremendous growth there. Media is a really important part of our business and continues to be. Today it represents, you know, over 90% of our revenue. Pay -TV is a big part of that. It's gone from when we separated to being really the bulk of our revenue, almost the entirety of our media, you know, revenue, to this year we're projecting it to be about 35%-40% of our revenue. This was anticipated. What do we do? With the time of separation, we said we gotta focus on OTT, we gotta focus on adjacent media markets.
In addition, we have a very strong consumer electronics part of our media business and social media as well. We have had a concerted effort to increase our revenue streams in these other areas. OTT's been the biggest success when you look at that. We got a deal done with Amazon in 2024, at the end of 2024. Great deal. Kind of an initial big licensee in OTT. We had some other licensees in OTT before, like Starz and DAZN, amongst others. But Amazon was the first big proof point, right? In terms of, okay, we can get these deals done with the big guys. Then, at the same time, we ended up having to file litigation against Disney, which is not our preferred approach.
We have a tremendous track record, especially on the media side, of getting deals done without litigation, and that is our strength. Occasionally, you do have to go down that path. At the time we got Amazon done, about a month before that, we filed litigation against Disney. Last year, in December, we were able to get a deal done with Disney. Just 13 months after filing litigation there, which really demonstrates the strength of our portfolio in the OTT space. We have those two proof points, one with litigation, one without litigation. As we go and talk to others that are unlicensed, that are significant, those are great way to juxtapose, like, the different paths we can take.
Our preferred path, which is doing it without litigation, or if we need to, going down that path.
Couple quick questions before getting into semis. I always get asked, what are some examples of media IP that you guys are licensing?
Yeah. We have a very broad portfolio in our media portfolio. The media portfolio alone is over 10,000 patent assets. You know, people think of us kind of in that pay TV space with the legacy of Rovi and interactive program guides, but it's much broader than that. Where we've been investing is in areas where it goes beyond that. Search and recommendation has always been at our core as well. That still continues to be a strength. When you think about, okay, how do you find and watch entertainment? Well, if you've watched The Diplomat on Netflix, for example, it's probably gonna recommend to you The Night Agent, 'cause they're similar genres and you're gonna see that. I'm basically telling you what I'm watching on TV right now. I try to keep it relevant.
It used to be I'd give Baby Reindeer as an example. I realized that was pretty dated. This is something that, you know, that's at our core. When you think about TiVo and your DVR that you had, that was kind of like one of its strength early on. We've continued to invent in that area. The technology's gotten better. The algorithms have gotten better, the technology behind that. In addition, we have really strong portfolio in CDN. The content delivery at the back end, how video is being delivered, that has gotten more and more deeper into the tech stack. Then a lot of the user interface, like how the people find shows. Again, going back to that at the strength.
That is something that how you navigate through a smart TV or an OTT platform is something at our core. Imaging. I could go on and on 'cause our portfolio is quite long. Those-
No, that's perfect though, 'cause I always think of Janet Jackson and wardrobe malfunction with TiVo. Before I get myself into trouble. Now with pay -TV, right? Pay TV is now gonna be down to 35%.
Yeah
of the mix. We take that step down in the first quarter. Now from this base going forward, now with what you've done from an OTT standpoint, from e-commerce, should we be thinking about this business on a growth trajectory now, where those growth vectors overcome the secular decline in pay -TV?
Yeah. I think that's true. I think couple things that I would also note. I think now that we've had this step down in pay -TV, we had a very large pay -TV provider that was paying in more of a fixed structure. That deal expired at the end of last year, which we knew obviously was going to happen, and that helped moderate some of the secular declines we were seeing in the pay -TV industry. Now we've taken that step down in this year, and we've anticipated that all along. Now we anticipate our pay -TV business will really mirror most of the pay -TV industry writ large, and so you'll have some moderate declines.
I also think, just as I look at the industry, that there is going to be some stabilization in the pay -TV industry. You're seeing that already. Charter, for example, had actually an increase in their video subscribers in Q4 for the first time since 2017. You know, so you're seeing that. They've gotten really smarter about how they think about skinny bundles and making their offerings tied to some of the hybrid OTT subscriptions as well. You have companies like YouTube TV, Google as a licensee of ours, who have actually seen consistent subscriber increases, right? So the way people want to watch entertainment, there's actually a going back to almost what pay -TV was before.
We think there's always gonna be some subset of the consumers that are gonna want this bundled offering that pay -TV offers. We think those declines will continue to moderate.
I've been harassing you for a couple of years now. There's been a demonstrable change in improvement, I think, in the backlog in this area.
Mm-hmm.
Not the backlog necessarily. The customer engagement.
Yeah.
It seems like it's much larger.
Yeah
... it's much more diverse. Could you just comment on that quickly?
Sure, yeah. I mean, it's a great point. You know, our pipeline of customer opportunities has increased substantially over the last few years, in part 'cause we've diversified, you know, who we're trying to attract, right? On the pay -TV side, it was these kind of large, you know, pay -TV operators, right? Even semiconductors, there's only a handful of really large, meaningful semiconductor opportunities, and we were primarily focused on memory, right? Well, if you think about memory, there's really kind of five guys in memory that really matter, right? All of them who actually are a licensee of ours, but only two we're getting revenue from. We wanted to make a concerted effort of how do we diversify our customer base. OTT was part of that. We've talked about that already.
E-commerce was another one that was really has just increased the number of customers in such a significant way. Search and recommendation being key, 'cause how you try to find an a jacket that you might wanna wear, right, and buy. Well, if you go and search for that, it's often gonna recommend, well, that jacket might pair with these pants or this shirt, right? That technology behind that recommendation algorithm is actually very similar in the entertainment space as it is in e-commerce. Not to mention our imaging portfolio becomes more relevant. Video try-on is becoming more relevant. Media is coming to e-commerce in a significant way, and so our portfolio is becoming more relevant as those sites are getting more and more sophisticated.
Yeah, we've gotten four e-commerce deals done so far. They're relatively small in dollar value, compared to some of the others, but we see kinda continued progression there. Our pipeline of opportunity in e-commerce alone is, you know, over 100, you know, customers. And that continues to grow.
Now let's get to some of the fun stuff. You've taken half your time on media.
Sorry. I apologize. Hybrid bonding. What is it? Why is it relevant?
You know, we like to say all roads, you know, lead to hybrid bonding in the semiconductor industry today. It is really a platform technology that we think is relevant in logic, that we think is relevant in memory. It's currently in high volume manufacturing in NAND and image sensors. AMD, who was our first, you know, big logic customer, really pioneered hybrid bonding for chiplet architectures as well. When we look at, you know, inventions, right? When we try to think about where do we want to invent? How do we wanna differentiate ourselves at Adeia?
On the semiconductor side, we think, "Well, what are platform technologies that are gonna, you know, be widely used across the industry?" If you step back in time, way back, till 1990 when Tessera was first formed, they created chip-scale packaging. Well, chip-scale packaging was a platform technology that was widely adopted across the semiconductor industry and actually is still used today, right? In almost every semiconductor device out there. When we looked at, well, where could we invent that could be the next chip-scale packaging, we invested in a company called Ziptronix in 2015. Ziptronix had invented hybrid bonding. They're a small company out of North Carolina, and we saw this as a real opportunity to make it the next platform technology. Why is that?
Because challenges with Moore's Law, people wanna have more interconnects, right, in their chips. Wafer-to-wafer bonding was a thing then, especially with image sensors. Die-to-wafer bonding, this is something that Ziptronix really hadn't worked on, but our team of engineers said, "Hey, we could work on die-to-wafer bonding, and that could be relevant in logic. That could be relevant for memory. And there's a lot of invention still to do in hybrid bonding." We started in that space, and now we're seeing that in a big way come to fruition. AMD obviously being a huge proof point. We got that deal done, and I know you're gonna ask me, I'm sure, about that and all the details, which I will not be able to share.
Nothing about AMD.
Yeah. They were really a leader, right, in that adoption, and now we're seeing Broadcom, Intel, Marvell have it on their product roadmaps. It is coming for NVIDIA. I know everyone asks me about NVIDIA. NVIDIA is gonna go down that path at some point. We are sure of it. It's a matter of when, not if, because this is the way that people are gonna be able to scale from a platform standpoint.
See, you're stealing my thunder. 3D NAND, right? You've had traction. You've been generating revenue. Logic was that next big step.
Yeah
With chiplet architectures. AMD is deploying it currently into the marketplace, but it was a little bit of a soap opera, right, in terms of getting this across the finish line. How important is AMD, not just in terms of the revenue contribution from itself as we start to look out a couple years from now, but in terms of being the leader for the rest of the industry and driving other logic relationships?
I mean, it's huge. I mean, it's a seminal deal for us. It is really the proof point that we were looking for in the logic market. This is the first real logic deal that we have gotten done. Our technology on the semiconductor side had really not been, you know, utilized from logic players in mass, right, in the past. This was the proof point we were looking for in terms of getting to that target we had talked about of $100 million in recurring revenue from our semiconductor business. Today, well, last year, our semiconductor revenue was $26 million, right? The year before that, it was $18 million, right? We have been growing that steadily.
We think we can now really make a huge jump to get a lot closer to that $100 million. This deal with AMD gives us that path really just with the logic market, and not getting 100% of the logic market done, but you know, another couple you know, licenses, and we're gonna be there. Then we can get to that next level. People always ask me, "Well, $100 million, like, really?" I'm like, "No, it can be a lot more than that." I know Scott thinks it can be a lot more. I read his reports, as I'm sure all of you do. It can be significantly beyond, you know, $100 million for us and it can be getting more of the logic market done, right?
Getting the memory guys done, right? That will be really that proof point. AMD sets a great precedent for us. It shows that our technology is relevant. It is valued. We did have to go through litigation, right? We got the deal done in four months from filing litigation. That's kind of unheard of, right? That's the thing that excites me as well, is that, okay, they made us put our cards on the table. That's sometimes what you have to do. Our hand was strong, and that's what it showed, and that's what excites me about the future.
I know you can't provide details on this, but just generally speaking.
Yeah
...the AMD economics attractive? Is there some way we should be thinking about this looking out three or four years, how big this could be?
Yeah, I mean, I think there are a couple things. I'll have to, you know, dance around this a little bit 'cause, yeah, all of our agreements with our customers are confidential. It's one of the things that we pride ourselves on and being, you know, very, mindful of because, you know, that's our reputation is key when we go and talk to other customers. First of all, we said it was a multi-year license agreement, let's start there. What does that mean? When we talk about multi-year versus long-term, there's a distinction that we make. Disney, the Disney deal we said was long-term. When we say long-term, we mean greater than five years, right? Not five years, but greater than five years.
When we say multi-year, we mean five years or less, and our average length of our deals are five years, right? That gives you a sense, right? Why was it multi-year versus long-term with AMD? Well, think about how much the semiconductor industry has changed over the last few years and how much it is going to change over the next, let's say, five years, right? It is gonna change dramatically. Trying to price that in and getting a deal done with that change in the marketplace, having sometimes a shorter deal helps in those negotiations to get to the right, you know, total value that you're looking for.
The other thing I would say is that with Disney, because of the accounting that we need to take into with our pay -TV business, there was a significant portion of that was attributable to the back, right? Or catch up as I think you like to refer to it, Scott. We have more flexibility outside of our pay -TV. There is some portion of it that is attributable to the back, but there's more flexibility in how we structure those deals. That was the case with AMD. It's more forward-looking, I would say, if that makes sense.
Do you get more logic deals this year?
Well-
What have you done for me lately?
Yeah. Listen, I think right now we're looking at the market and looking at adoption, right? AMD was a leader in this space by quite a significant margin, right? They've been in the market with a chiplet architecture for, you know, almost four years now. They launched it in 2022. Broadcom just started shipping in volume with a partnership with Fujitsu, with a chiplet architecture that utilizes, we believe, hybrid bonding. We're almost certain it does. They've said that they are going to take that product and use it for other customers as well, and have those other customers be shipping in the second half of 2026, and then they're gonna get to 1 million units by next year. These are not cheap, you know, chips.
These are very expensive. You think about, you know, what that means for us. This is what gets us excited, right? In terms of what that opportunity could be. They're just one other logic player. We need to continue to see that market kind of adoption and see that through. This year might be a bit optimistic, right? Certainly we're watching that, you know, market evolve and seeing that opportunity.
Let's shift the conversation to DRAM.
Yeah.
Any annual report you read for the last 6-plus months talks about the pull in the migration to HBM4, HBM4E, and HBM5.
Yep.
I think it's generally believed that hybrid bonding comes in 4E and 5. More recently, there was some discussion out there from JEDEC, one of the industry bodies.
Yep
about reducing or relaxing height requirements.
Yeah.
One of the benefits of Hybrid bonding is small form factor. It seems to really be missing the whole point, I think in my opinion, of Hybrid bonding is better electrical performance, better thermal characteristics. How do you see that playing out in the marketplace? Where do you see Hybrid bonding really coming into play in the DRAM market?
Scott, it's like you've talked to Laura Mirkarimi, our head engineer. By the way, if you haven't watched her interview with Laura, you should. It's fantastic. She goes deep into hybrid bonding and a bit into RapidCool, and it's far more than I will do here. I will say that when we saw the JEDEC news, we certainly took note of it. You're right, it is missing the point. The pitch, you know, benefits of hybrid bonding are one aspect of it. The real benefits, especially when you think about HBM 4E and 5, is going to be in the thermal management and performance advantages that you get.
If you just look at what Kioxia and SanDisk put out about the performance advantages they're seeing in NAND, and the thermal management advantages, I mean, they're talking about, you know, 80% advantages on performance and 30%-50% advantages on thermal management. I might have had those reversed. Anyway, they're huge advantages on both. That is like what we think HBM can get as well in terms of those types of advantages. Yes, the JEDEC specs, it's not surprising, first of all. One, they did this with HBM4. The industry starts getting like, "Hey, can we really continue to hit, you know, what these requirements are?" and they start loosening it.
What we see is likely gonna happen is with HBM4 or HBM5 is you could have both. You could have an HBM4E that maybe has a micro bump you know packaging, and then you could also have some that is hybrid bonded. That's what that JEDEC spec really allows for. It allows for both to happen. We think the first memory guy to really adopt hybrid bonding in HBM is gonna have huge performance advantages. That's what we're looking for. We're still bullish on it being adopted in kind of that same HBM4E5 timeframe that we mentioned before.
2027, which will also coincide with the expected timelines of renewals?
Yeah. I think 2027, 2028, yeah, is probably what I would expect. Yeah, our renewal timeframe for our key memory guys is right around that 2028 timeframe.
I wanna hit on real quickly RapidCool as well.
Yeah.
It's something that you guys have started more recently to talk about. What is it, the timeline, and, you know, how basically important is it in terms of potential data center architectures going forward?
Well, I mean, I'll just start back to, like, the invention story as well, because this is what we do really well. We can think about big industry problems in a way that a lot of other companies can't. Like, we have 150 employees at Adeia, right? We were number 66 in the number of issued patents in the U.S., right? We are ahead of small companies like AMD, right? Or, you know, Verizon, or, you know, Comcast. Like, I mean, these are massive companies, right? That we are inventing, you know, far ahead of. Why are we doing that? 'Cause that's what we focus on. We focus on solving big industry problems, and RapidCool is a great example of that.
Our engineers sat around a table back in 2021, 2022, and they started thinking about what's the number one problem in the industry today, or what's a huge problem in the industry today? And thermal management was something that we're seeing in data centers, we're seeing with AI workloads, and they're like, "Well, how could we uniquely solve that, right, given our capabilities?" What are we strong at? We're strong at advanced packaging, we're strong in hybrid bonding and the idea was, well, what if we could bond a cold plate directly to the chip itself, and what would that result in, right?
We started experimenting with that, and we bonded a cold plate using a silicon-to-silicon bonding, and what happens is you get rid of all the thermal interface material, or what's sometimes referred to as TIM, and taking away that TIM just had huge efficiency advantage. We're talking like 70% out of the gate. That was the first thing we noticed. We started realizing we could design the cold plate, because we're bonding it directly to the chip, and address certain hotspots on the chip. This is what our customers get actually most excited about because these chips have certain hotspots on them and that are significant, and a standard cold plate's gonna reduce the heat kind of evenly across the chip.
With our solution, you can design it, so it's gonna pinpoint where the liquid cooling is going to go through channels and direct it towards the hotspots on the chip. That's what gets people really excited when we talk to it. From a revenue opportunity, it's still kind of mid to long term for us because these things are we're in kind of like the late-stage development cycle. We're talking to customers. We're trying to get, you know, adoption as well. It's plug and play in a data center. It can go into any kind of liquid cooling data center today. You can replace a standard cold plate, so that's also a huge benefit.
We've gone over, but just real quickly.
Yeah.
Capital structure, you generate a tremendous amount of free cash flow, debt, buyback, otherwise.
Yes. All of it.
All of it.
We have a balanced capital allocation approach. Our debt is gone from $759 million at the time of separation to right around $400 million today, so we've paid over $300 million in debt in three years, which is tremendous. We at some point feel comfortable with around the debt that we're at right now, maybe a little less, so $300 million-$400 million is really our target, and we'll refinance that into kind of more of a fixed debt structure. At the same time, we're going to, I mean, not pay down. We're gonna buy back stock, we're gonna pay our dividend, and we're also gonna reinvest in the business in looking at tuck-in acquisitions.
Perfect. Paul, thank you so much. Absolute pleasure having you here. Thanks again.
Thanks, Scott. Appreciate it.