All right. Good morning, and welcome to Harmonic's 2026 Investor Day. I'm Walter Jankovic, Chief Financial Officer. For you all that are joining us here in New York and over the webcast, thanks for being with us. Today's a special and exciting day for us. It's been 12 years since we've done an in-person investor day. More exciting, this is the first time as the new Harmonic, a pure-play broadband company with a larger market opportunity, a sharper strategy, and a growth framework that we're really excited to talk to you about today. Today, you're going to hear from our leadership team. Nimrod will walk us through vision and strategy. Asaf and Yaniv will talk about the markets we serve today and the markets we'll serve in the future.
I'll provide a brief financial overview. Then Jeff will host a customer panel. Then we'll wrap it up with a Q&A. Some housekeeping. We're going to make forward-looking statements today. Actuals may materially differ. Therefore, please review our SEC filings for risk factors. Now, with that, I would like to introduce our President and Chief Executive Officer, Nimrod Ben-Natan.
Thanks, Walter, and thanks, everybody, for joining us here today and online. October, next month, 30 years ago, I joined Harmonic. Quite a journey. When I look back, what I really see is a series of transformations that we've gone through. We started as an analog optics company. We transitioned into digital video. We were at the forefront of that. We moved the industry to high definition, digital ad insertion, and more recently, the transformation to virtualized broadband access. Today, we're excited to talk to you about the new Harmonic and the road ahead. Let's jump right into it. Many of you really know us about what we do in cable. You may remember we started with cOS, and the leadership that we've got in the DOCSIS cable market. Today, we're really going to talk about the broader vision of how we see the opportunity ahead.
In cable, the market is expanding and we're as well with fiber-to-the-home. Lots of cable operators are doing fiber for edge out, for new development, and surgically in competitive markets and MDUs. I'm excited to share with you that this year, in 2026, we're going to end the year with $70 million of revenue just in this category. With the momentum that we have, with the maturity of the product line, we feel ready to go after the broader opportunity of the telco fiber. Same technology, different customer base, but we are at the point where we are really ready to tackle that market.
We announced last week our cOS SensAI, which is our intelligence platform to serve the broadband operators, cable and telco, DOCSIS and fiber, to really address the opportunity, which is the toughest challenge that they have right now to maintain customer satisfaction, reduce churn, while maintaining their operating cost. Another very interesting development in the market related to the fiber technology is a new market, we call it PON beyond fiber- to- the- home, which is using the PON technology in data center markets and in the enterprise market to really connect devices. In data center, you have tens of thousands of devices to manage what is called out-of-band management, and in the enterprise market, this is an alternative to a point-to-point Ethernet. Very interesting market that the market started to follow, and we will talk more about that.
That is rounding up the connectivity, cable, fiber markets, as well as the intelligence. The wheel is effectively complete, but there is one other thing that is very interesting that we started to focus on, which is really the edge compute opportunity in the AI market. This is really a very interesting opportunity where both cable and telco are going through the modernization of their architecture, where they free up a lot of space across their network. Facilities that used to be central offices or cable hubs with legacy equipment now become empty. They have power, they have space, they have fiber connectivity, and they have very strategic location in the network, milliseconds away from businesses and customers. We are not the first to talk about that. I think last week at Goldman Sachs, Verizon talked about that. Sievert talked about that the other day.
What we really want to do in this market, which we have not sized yet, is really enable that through the longer tail of customers and the opportunity, and we will talk more about that, is really how to leverage cOS as an orchestration platform and help customers to bring that to the edge of the network. When we look at the opportunity, we see clearly the connectivity is the biggest market opportunity, about $5.4 billion, including cable, fiber and cable, telco, as well as the PON in the data center and enterprise. We will break that down as we go through the day today. The intelligence opportunity as we define it, a market between $1 billion and $2 billion, and yet to be sized opportunity across the edge AI compute. By 2030, we see an opportunity of about $7 billion ahead of us.
When we look into the future, one of the things that we believe is that everything that we see AI today that is generated in data centers will get experienced, consume and experience in the home, obviously on the go. But when you go home, you offload into your Wi-Fi, and lots of connected devices will evolve with the way we see AI, agentic AI, everything that is every day we read about, will clearly change the way the network is performing. Will really change the shape of the network. We see growing demand for intelligence, and we see growing demand for fiber, both in cable and clearly, as we go into the telco market. We talked about the compute opportunity, which again, is an opportunity in development.
Really what we want to achieve is to be the broadband platform of this transformation across the connectivity, the intelligence, and the compute based on the cOS platform. You know what, this is not the first time that we do identify an opportunity and work towards bringing that to market. Quite often, we do it in collaboration with key customers. We've called it before, the way we did it about 10 years ago with the virtualization of the broadband cable, we really pioneered that. We identified the opportunity, and we executed on that. We did it even before with those of you who remember the EdgeQAM market that enabled video on demand over cable networks. I think this will be a good opportunity to hear from one of our customers, how is it to collaborate and partner with Harmonic.
The partnership that we have with Harmonic is strategic and critical to everything that we want to do going forward. My name is Elad Nafshi. I am the Executive Vice President and Chief Network Officer for Comcast. In 2018, we really started to partner together on creating the blueprint for how do we, as an industry, scale broadband services into the future, and that led to the creation of the DAA, Distributed Access Architecture, and the vCMTS, the virtualized CMTS project, which we've been working on with Harmonic ever since. We have been operating as one company, and we have been innovating as one company for the past eight, nine years. We would have not been able to deliver this much innovation in such a short period of time without that complete partnership.
About four years ago, we were able to be recognized with a Technology and Engineering Emmy Award, that we were very happy to jointly receive, both Comcast and Harmonic, for the creation of the DAA architecture and the vCMTS. That just signifies just how critical and how transformational this relationship has been to the industry. With DOCSIS 4.0 FDX rolling out across our footprint, delivering multi-gig symmetrical services in great scale, and we're super excited with our accomplishments. We're just getting started. How do we continue to scale seamless fiber delivery and continue to grow the network at the scale that we are? We're adding 1.2 million homes to our network today. All of that is fed with our joint development, DAA platform, and we're very, very excited to do even more. We connect 65 million homes. We serve 39 different states.
Our network is huge. It's over 1 million miles of fiber and cable, and how we bring all of that together and leverage the technology that we're inventing for the industry in order to deliver even faster, even more reliable, and even more scalable services, is what this partnership is about. It's what we've delivered over the course of the last eight years, and is what I cannot wait to continue to expand across the footprint.
Thanks, Comcast. Thanks, Elad. When I took over about two years ago, the CEO position, we set four strategic imperatives for the business that we kept updating you on the progress. Diversification, leading the DOCSIS market on cable, growing the fiber, and getting into the intelligence. I am happy to update. While this is not a victory, this is a great milestone to indicate we have grown the diversification percentage of what we call rest-of-market customers from 20% to 37%. We continued the leadership on the DOCSIS market. We have the 90%+ market share, and we were the first to introduce the unified architecture. We got JR from Mediacom on the customer panel here, and he will talk more about that.
On the fiber, I am very excited with the update we are sharing with you today on the revenue. We will achieve $70 million in 2026 on that product line. Number of customers have grown. On the intelligence, we gave you updates along the way. Elements of the solution, like the Beacon and the Amply. We got more than 30 customers now using the technology. Last week we announced the cOS SensAI, which is our intelligence platform, a very comprehensive intelligence platform, and we are going to expand on that in more details. That execution got reflected on the financial performance. We achieved record revenue, record backlog. We executed on the share buyback. With the sale of the video, we are in a good position to keep funding our growth initiatives as we execute on that.
When we look at the market itself, hypercompetitive. It used to be cable fiber. There is a fixed wireless, there is satellite, and every day there is reports on how challenging this is for the customers. I think the key point here to remember is that broadband operators are doing a really good job in converging services.
Just the other day, one operator introduced a new security capability into their Wi-Fi at home. They do the mobile wireless, and they will expand the portfolio of capabilities. All of that is running on the same broadband infrastructure. That broadband infrastructure is changing. There is way more upstream. Latency is very important. As we look into the future, the number of what we used to call connected devices that we just got connected over the Wi-Fi now will become AI speakers. They will communicate back and forth. Every camera will talk back and forth over the network to get AI guidance what to do. When you look at all of that, we believe that the broadband industry will continue preparing for the AI era.
We see that across cable. You got to go high-split and beyond. You got to take care of your low latency capabilities across the network end-to-end. Fiber will do 10G and even 50G, which is becoming available. At the same time, heavily invest in automating the network, making sure you address the customer experience, quality of experience, while managing your operating costs, which is really where cOS SensAI is targeting. When we go back to the core platform, what we introduced 10 years ago as a virtual CMTS was effectively a platform with an application that is doing virtual CMTS. There was a forward-looking about what the platform will be able to do over time.
And in fact, other than being a mission-critical platform that can work in a high availability mode, what we really designed back then is an edge compute platform that can evolve over time. Indeed, when we introduced fiber, it was yet another application on the same platform. As we go after the telco market, we bring couple of differentiating capabilities, such as the open ONT architecture, which is not easy to achieve, and I think it's becoming very critical today, especially with supply chain and memory costs. You really want to have a choice of supply. Unlike the DOCSIS market, which is well-defined, you've got CableLabs, the market of fiber is more open, so maintaining that interoperability is not easy. We bring our advanced, hardened, remote OLTs into that market.
We do have a differentiated 50G roadmap, and we have new provisioning capabilities that we bring and complete our FTTH capability. Asaf will talk more about the opportunity beyond FTTH and the data center management and the enterprise market. Intelligence is a big topic, and really the strategy that we have is to bring something that can work across any network, not exclusive to cable. It's applicable equally to FTTH, and it can work even if we don't provide the access network. It can work in a multi-vendor environment, and we will explain that in more details, what exactly we do that, and how do we achieve that. Really, the goal is to achieve reduction of churn while lowering the cost.
We'll talk more, although this is in development stages, but we clearly see an opportunity for this edge AI inference, leveraging the space and power that got freed up in conjunction with our orchestration capabilities. I did have two asterisks on this previous slide, which is really a tuck-in, very focused acquisition that we did of a company called Int6, based in Brazil, very focused on the Brazilian and Latin American market. They serve about 7 million subscribers, 40 customers. What's unique about what they do, they work in a multi-vendor environment, and they really help both our fiber complete our provisioning and management capabilities, and on the intelligence cOS SensAI, they really give us actionable workflows and working across multi-vendor environment to really allow us to bring intelligence not only to where cOS is deployed.
In summary, three engines. Connectivity is clearly the core market. DOCSIS will continue to be a major engine for us, expanding on the remote OLT, getting in a serious way into the telco market. Then the intelligence, the $1 billion-$2 billion market, we'll talk more about that, and the upside with the edge AI. A total opportunity that is way bigger than what we're addressing today, and as we go through the day today, Asaf and Yaniv will expand on these engines. With that, I will introduce Asaf Matatyaou, our SVP of Product, to talk about the connectivity. Asaf?
Thanks, Nimrod. Good morning, everyone. A little bit about me first. I started my career in data about 30 years ago as well, not at Harmonic, but at a different company. Back then, there was dial-up. What we were trying to do is provide an always-on connection, and the speeds weren't fast, but it was an exciting time going from dial-up and having always on. Now, obviously, we look at that as something that's a utility, something that's always there, something that everybody expects. I'm excited about what's next, even more so today, because back then, it was purpose-built, it was focused. It didn't have the upside, it didn't have the opportunity, the growth potential of what our vision for broadband is.
What we built our connectivity platform on and what Nimrod just shared with you as a foundation of connectivity into intelligence, and then ultimately into compute. As you heard from Elad, we're only at the start, we're only at our beginning of where we go, and we've planted those seeds, and we're going to watch them grow. Two connectivity types before we go further. DOCSIS, for the folks who don't know what that is, the first three letters say it all, data over cable. It's basically broadband over the coax cable that exists in people's homes. PON is the fiber equivalent. It's fiber that leaves a facility and splits passively to many homes. Let's start with DOCSIS. It's a market we lead, and it's the footprint that pays for everything after that. All right. Let's see. That's a stronger click. Okay.
If we look at the market, and we'll go from the market and dive deep into a few different segments here within DOCSIS and how we grow that. We look at where we started a decade ago. If I recall, a decade ago, the market was filled with legacy CMTSs. The transition we've seen over the past decade is, as legacy CMTS has diminished, what's replaced it is the virtualized CMTS, the DAA approach that we've seen. If we look at the transition, not only is the market growing between 2026 and 2030, but we also see that the larger portion of it is going to be vCMTS and DAA. In fact, $9 out of every $10 being spent in 2030 will be in this architecture that we pioneered and we led. Our credibility is, well, we started this about a decade ago.
We own over 90% of the vCMTS market share, 60% of the DAA footprint, and over 160 operators globally, and expanding. We feel very confident with the ability to understand what's necessary to continue to execute in this marketplace. After all, we created the category, we hold it, and we continue to raise the bar. We'll talk a little bit more about that in a couple of slides. Let's look at the vCMTS portion and the decade of value it's delivered. A decade ago, I'll just share with you a story. A decade ago, when folks wanted more speeds, what they had to do was either replace their chassis or replace their line cards. Not only was this operationally disruptive, it was also a purchasing event.
Every time somebody wanted to replace their line card with a different one, it was a chance to upgrade their CMTS to something else. We've replaced that. No longer is the core. Now that it's software, it's all software upgrades. The proof is that 100% of the operators who've migrated to cOS over the last 10 years have stayed on cOS, all through software upgrade. It's a very exciting opportunity for them to be able to deliver actually more speeds, more capabilities, all on the same software and platform that transitions with them as their subscribers need more services, faster services, lower lag services, and the connectivity they expect. We'll talk a little bit more about what that buys them going forward. But the fact is that once that software and the solution is there, the opportunity is to connect fiber to use it for edge compute.
This decade of leadership has led us to a compounding effect of opportunity. The 90% of vCMTS deployment in that footprint, the 350,000 connected devices that grow every day, and the operators that we continue to expand globally, and every sort of footprint you can imagine. What's left to transition is very exciting as well. From the operators that already work with us to the prospects that we're working with, there's 75% cable modems out there that will be transitioning to DAA and virtualization over the next few years. Each DOCSIS generation that we've talked about renews itself on this core as a software upgrade rather than replacing it. That was transformational for the industry and something that we've seen with over eight generations of Intel servers in the industry. One of the key reasons the cOS core is a durable investment is the pace of our differentiated software.
In fact, it's one of the reasons I joined Harmonic 15 years ago. When I joined, I was inspired by the culture. I joined, I met with Nimrod and other folks, and that culture continues to define us. We are not standing still. It's always what's next. What can we do to advance? As you've heard, we were number one in EdgeQAMs, and we're number one in the cable broadband industry today. That's not enough. We need to continue growing and pushing and delivering better and faster connectivity as subscribers expect that, differentiated services, better quality of experience, putting the intelligence layer on top that Yaniv will talk about, and then augmenting that with the compute opportunity that's in front of us in the world of AI.
Most recently, just to give you an example of some of the quality of experience advancements that we provide all through software, faster speeds, software upgrade, simpler operations with PTP-less, software upgrade, lower lag, software upgrade. Improved quality of experience is incredibly important to retain subscribers on a network. Going from ideation to production is simply a matter of waiting for the next release a quarter away. No longer are you waiting for a hardware cycle. We move from ideation to production. We have sprints that are two weeks long, and in some cases, one week long. We go incredibly fast, and we work with our customers to make sure we understand what they need to be successful and deliver that. What also is important in software is to continue in the scope of sustainability.
As you deliver more, you want to be able to free also, free space and free power. No longer was it double the speed, double the facilities. We've actually reduced that dramatically, up to 70% less power consumption relative to the older legacy technology. But when you compare ourselves versus ourselves a decade apart, we're actually 85% less watts per gigabit, less power against ourselves a decade later because of our software advancements. Some of our customers report up to 95% space savings. That's important. Unique capabilities we've added in software that I just want to talk about briefly, the effect is, the result is happier subscribers, better NPS, and higher QoE. Beacon ISM adapts dynamically to plant conditions. Less tickets, less trouble tickets, less truck rolls, less OpEx for service providers. Pathfinder. This delivers software that adapts to an individual's home.
PTP-less. Being able to deploy DAA simply and quickly, more cost effectively, giving higher network reliability, incredibly important. L4S, really cutting working latency where it matters the most. Video calls, gaming, and cloud, soon enough with AI. So QoE is one in software, and that's where we continue to differentiate. We continue to raise the bar every single quarter. Let's look at a case study. We'll be focusing on Mediacom's transition to DOCSIS 4.0. We started on this path with Mediacom about a year ago, and in fact, if you remember at D.C. last year at the Tech Expo, there were congress people coming to the show floor. They were excited to see symmetric multi-gigabit speeds practically and easily deployed on existing infrastructure. Symmetric with no new hub space is what I should say as well. The outdoor platform you see here today is actually right over there as well.
I encourage you to go look at it as well. It's actually configured in the same way. It's called a platform because, in fact, you can configure it in so many different ways with different modules to adapt to different footprints out there. Over here, we're talking about DOCSIS 4.0, and it was 1.8 GHz. It was delivering, I think, 8 G down by about 3 G up to some subscribers, and it was able to do that on that PTP-less network we talked about, which simplified the deployment and accelerated it. The ability to put in a fiber module in there, in this case, a remote switch, a Jetty, provides the ability to tactically deploy fiber when it's needed. This is incredibly important. After folks came to our booth a year ago, that built the momentum. This is simple. This is easy. This is possible. It's available.
We see that momentum building in our ability to deploy DOCSIS 4.0 with other operators around the world. Thank you, JR, for your partnership and collaboration. We truly appreciate it. This truly matters when you think about the footprint of cOS. We already have deployed with nine out of the 10 largest U.S. operators in cOS. This opportunity, it's right in front of them. I'm going to transition now out of our fiber and PON section. A few years ago, we recognized we needed a more complete portfolio that's end to end. We've made the investments and we've executed, and we're counting the results, and Nimrod has shared over the past few quarters. Let me share with you a little bit about our expanded portfolio as well as the different market segments that we'll be focusing on.
The three segments within fiber that we will be looking at is fiber and cable, where we are already the vCMTS in the facility, the telco FTTH segment, and PON beyond FTTH. It is important to look at this as an opportunity that leverages these opportunities and segments off of the same software and the same base of investment that we have made. There is a lot of growth in fiber and cable, as well as PON beyond FTTH over the next few years. As you can see, the telco fiber-to-the-home segment is already quite large. All the same software that we are looking at will connect to a breadth and variety of OLT devices that these investments that we have made will come to fruition. Let us look at that portfolio.
Two years ago, when we started looking at this segment, we really focused about fiber in a cable space, and that really took off. That was successful, and we used our Fin pluggable. If you start in the far right corner, you will think about the Fin plugging into maybe a DOCSIS node. One of our customers actually called it a unicorn at the time because they did not think it was possible to have DOCSIS and fiber being fed by the same platform. Once we did that and that was successful, we built dedicated modules, whether it was a remote switch, a remote OLT, different enclosure types, small, compact form factors, ones that are specialized for MDUs, ones that are a little bit larger, like the Ripple, that houses multiple modules.
We look at remote switches like the Jetty-3 expanding to hub consolidate and have more fiber out there, saving fiber and giving longer reach, and our 50G Pier-2, which is over there as well, showing you the expanded portfolio we have. Basically, any form factor for every PON generation, as well as the opportunity to expand this for other sorts of footprints. Let us talk a little bit about that. We have actually used PON as a backhaul to connect to a DOCSIS customer. That is an opportunity with the pluggable that goes in.
Let us think about MDUs with our SeaStar, which allows us to address low-density MDU markets in the most cost-effective way. Let us look at the advancements we have made, because even if you are on a fiber network, there are power outages in different regions of the world. How can you support making fiber even more resilient? Well, with our Oyster+ right over there, we are able to do so. Every one of these elements in our portfolio has a lead customer, is differentiated, and will be successful. Whether it is urban, rural, fiber pockets, high or low density, we have a portfolio, and the software is versatile for every architecture.
Now let us look at that first segment we talked about, the cable segment for fiber. This is our tailwind. It is not a threat. When people ask fiber or DOCSIS, we say both. When you have a cable segment that has a transition and ability to move over customers and subscribers to fiber, you want to make sure you do it in the least disruptive way, making sure you retain those subscribers, minimizing churn. This is exactly what we do with our software.
Ten years ago, a little bit more than that, we had the idea that our platform, it wasn't just the vCMTS, it was a platform with an application on top. We actually looked at having other applications served by the same software, and that's what we did. Over time, we've been able to connect many ONUs and OLTs and fiber subscribers out there, especially on the cable footprint. One platform, two access networks fed by the exact same infrastructure. DAA allows that to happen. The transition is accelerating and whatever pace service providers take, we're there to assist with that. You got your cable modems out there, and as they transition to fiber, it's just a simple configuration and a piece of software. Our advantage right now is that we're already in those facilities. Our software's already running there.
It's as simple as plugging in a Fin and configuring that fiber. It's really hard to replicate that, and the reason that is you need to have that footprint. Our footprint is in over 90%. That's where we're already there. That transition to fiber in the most cost-effective time to market way is to stay on cOS. DOCSIS or fiber, the same platform is the one that powers cable's network evolution. Let's look at an example of exactly that, what izzi in Mexico has done to unlock a national HFC-to-PON migration.
If you look at the diagram, they already had an existing trunk, and they really leveraged that and their fiber assets by just expanding that fiber-to-the-home. Allowing DOCSIS subscribers to migrate with the minimum disruption out there using the exact platform that you see over there, our Oyster with our Pearl inside. This allowed them to expand their network and allow, whether it is GPON or XGS or Combo PON, be able to serve those fiber opportunities in front of them. They are the largest operator in Mexico. It is an example of other FTTH migrations that we anticipate others will make as well. We are going to hear exactly that from Carlos Eduardo. Just here we go.
We moved from traditional supplier or partner to a more strategic kind of partnership. My name is Carlos Eduardo. I am a General Director for izzi. We are the largest cable company in Mexico, transitioning from cable to fiber these days. In order to update our network, we decided to find a technology that could support that transition and evolution for our network. In relying on this partnership with Harmonic, we decided to pick one of the greatest technology you guys have to offer. What we are doing, not building or rebuilding our network from the scratch, but we are reusing part of the other work done and infrastructure done for the cable network, but complementing with some equipment and transitioning from the cable footprint to a more FTTH or fiber-to-the-home deployment.
I think the main differentiator for Harmonic and the other guys from the fact is this capacity to understand our needs and come with a solution that fits exactly to what we are looking for. We are not only talking about high tech, and we are not talking about the single solution, but we are talking about the architecture itself. The izzi view for the future is to become more than just a cable network operator, to become a telco operator. This network reconstruction and rebuild the network to be 100% fiber based. Harmonic is key for our success in such a thing because we believe you guys have the right technology, you have the right people, and you have the right solution. Yes, I am 100% confident that with the right company to help us achieve our goals.
Thank you to izzi and Carlos for the wonderful video. Let's look at our big telco market segment now, and we believe that the investments we made in our portfolio will allow us to disrupt and compete and differentiate. The demand is real. We know subscribers are moving to fiber over time, and we believe we can differentiate with open ONT, our hardened DAA, things that we built our enclosures since the 1980s, and our differentiated 50G. Out of a $2.8 billion market, we see that growth in the 50G spend, and we'll be there ready as that market takes off. We also have the Pier-2 over there to show you exactly what that 50G platform looks like. One of the things we talked about that enables us to succeed in this telco marketplace is our ability to do open ONT.
I wanted to focus a little bit about some of the differentiators. Our software connects to a wide breadth of devices that we talked about, shelves, switches, and modules, all delivering fiber. This open ONT culture that we have is the ability to connect any ONT, any vendor, and we have over 20 vendors and over 75 qualified interops. This is incredibly important as we try to unlock the vendor-bundled and bookended solutions that are out there. Keep in mind that rising memory prices out there put the most pressure on CPE devices like ONTs. Operators need to have the ability to choose the right CPE for their business, and the ability for us to have open ONT gives them that opportunity of diversity and supply chain.
When you look at this software and our connectivity allows telcos to have every PON generation, every form factor, every service connected to our software and our broad set of devices. The last segment I'm going to focus on is the PON beyond FTTH segment. There's no better place to be than to differentiate in a new segment that is growing. Where it is today is just as that start. We didn't create the category, but we know we have the ingredients, the devices, the software, the solutions in place to be successful here. Let me describe a little bit. In the data center today, as Nimrod mentioned, it's the connectivity and the management to those devices that's out there.
If you're doing point-to-point Ethernet, there's a lot of wiring, there's a lot of active switches, there's a lot of OpEx associated with that if you think about data centers that are thousands and thousands of racks. The opportunity and the value of having PON in the data center is fewer devices, less power, and longer reach for out-of-band management. The same value holds true for enterprise. If you think about a WeWork, in fact, Nimrod shared with me a video when he was walking through a WeWork and all the cables above him, and we both looked at that and said, "Wow, this is a great opportunity for new enterprises. If they had PON in the enterprise, how much less wiring would they have?"
How much less gear would they have? This is the exact same software, the exact same differentiated solutions that we have that we ship today that would serve those two opportunities. This is a new fiber segment with dramatic growth, and we are confident we can differentiate as we move forward. Let me leave you with this. The connectivity that we have built is a foundation for the future. We will talk a little bit more about that after Yaniv shares with you about our intelligence, and I am excited for what the future holds. I will now hand it over to Yaniv, who will share with you our intelligence layer.
Thanks, Asaf. Good morning. You have been hearing this morning from both Nimrod and Asaf about our accomplishments so far. What I am here to talk about, it is what is coming next. I am not only excited about what is coming, I am actually really excited about what we are already doing. We see a $1 billion opportunity that is already coming into focus. In the next few minutes, I am going to talk to you about the market, how we are addressing it, and in fact, share with you two incredible stories about one operator in Alaska that implemented our platform and is seeing the benefits in the field already, and another one in the Southern East part of the United States that shows how we are going to scale.
I am really excited about what I am about to share with you, so let us talk about it by talking about the market, the opportunities. First, let us start about talking about the challenges. An operator, it is not a secret. Broadband operators need to do more with the same revenue. They are facing a competitive landscape like nowhere before. They first need to secure their base, and what is winning is reliability and not speed. At the same time, due to this competitive landscape, they are locking their subscriber to multi-year contracts, which put even more pressure on their ARPU. They need to grow. They are investing in DOCSIS, in fiber, they are making acquisitions, and every one of them is adding more and more tools that they need to buy, pay, and manage in order to operate those networks. Of course, customer expectation is only increasing.
With more and more devices to manage, and as Nimrod pointed, those devices need broadband connectivity, not just to show you the time, but in order to operate and reason. With more network to operate, higher expectation to meet, all of that with the same revenue. What is the solution? It is a vendor-agnostic, intelligent operation platform that will help operators increase reliability, all of that without adding cost. In fact, even reducing cost. Let us talk how big it is. We see a $1 billion-$2 billion market that is priced per subscriber against cost we are saving. We looked at it from three different direction. The first one is an independent market research that was done by Analysys Mason that put the AIOps software at about $2.2 billion annually, excluding tier ones.
We also looked at the savings that broadband operators can get from an AIOps platform, and we see a minimum of a $1 billion market for us. That's down from $72 billion annual OpEx in the saving from an intelligent platform. Of course, there is the lifetime value of a subscriber that is staying with you and not churning. All of that puts the intelligent market into a $1 billion- $2 billion opportunity that we are unlocking, that we price per subscriber annually against cost that we're saving. Let's talk about what we built. cOS SensAI is a broadband intelligent fabric. It's a vendor-agnostic, network-agnostic that start with data collection. We collect data from the entire footprint, starting from the back office, the different network elements of the broadband, all the way to the equipment inside the customer home.
We build the first intelligent layer, what we call the network graph. That's the layer that provide the context, provide the connectivity between different element and different department that's define what is a customer, what is a service, what's the SLA and the constraint we need to meet. Of course, there are the different tools. Tools that people are using today to operate and manage their broadband network, either DOCSIS or fiber. The agentic systems. Those are the seven agents that are working 24/ 7 without sleeping, without eating. As you can see, they are mimicking the different departments at the operator. We have the subscriber care agent that has a 360 view of the subscriber that helps the customer support representative. The network operator agent.
The field ops agent that helps the field engineer that goes to the right place and fix the right issue. Daily intelligence that look at trends, long-term trends, and moving operators from being reactive to proactive. Planning agent, provisioning, and maintenance. All of that feeding with a OpenAPI the existing customer application that every operator is using. We don't pretend that operators would want to see another UI, although we have a beautiful UI that you can see. But we assume that operators want to keep using their own application. We're just making them smarter. We have a video to show you how it works.
[Presentation]
So that is cOS SensAI. You may think, "Okay, Yaniv, everybody is talking about AI these days, right? Everybody is presenting his own platform. So why customer choose you?" You are right. Customers are taking meetings today with every company that talks about AI. They are taking meetings with what we call the generalist AI, who build beautiful platform, but they do not speak broadband. They take meetings with the equipment manufacturer vendors that build AI platform, but in their own silo. Of course, in the AI era, what we call the honeymoon era of AI, everybody is building his own application. Some of them tried, most of them has not succeeded to do that. When we come and present cOS SensAI, what they really love about that it is coming from us, a company that understand broadband, that has the experience in broadband, that understand fiber, that understand DOCSIS.
They love the fact that it is multi-vendor, that we are agnostic to what system we operate and manage. We build a platform that is hardened. We have experience in building cloud for more than a decade. What they really like is the fact that every deployment we have in cOS SensAI, the rest of the footprint is learning from it and is becoming more and more intelligent. As we identified signature and issues in one operator, it trickles down to everybody else. We can do that thanks to a very special team that we establish, what we call forward-deployed engineering. Those are the engineers that rolls in the truck with the field operation. They are sitting and listening to customer call to understand the challenges, and coding and improving the product on the fly. Just to show you the type of velocity we are talking about here.
Let us tell you the first story about GCI in Alaska. GCI is the largest broadband operator in Alaska, and they started to implement our platform a few months ago. We help them with their NOC operation that sees, identify issue, and escalate internally or to another agent that helps the field team to identify what are the tickets that they have, and as you can see, hundreds of tickets. cOS SensAI identified and open hundreds of field issue tickets. However, it is intelligent enough to also prioritize them based on severity, and you can see the number of trucks that you actually need to dispatch are a few dozen. To tell you more about that, let us hear from Troy Goldie, the CTO of GCI.
Having those tool sets that allow us to be more proactive, to allow us to have an infrastructure that can self-heal with auto tech dispatch, gets us to the point where our customer experience is actually realized. Time to task has to be first and foremost. The ability for a system to direct a technician where to go and what to do, it helps us with the time to task reduction, but then also the real-time validation that task has been completed successfully enables us to have a repeat ticket percentage that is drastically less than what it exists of today. Because we get the systematic validation that the network is where we want it to be, versus it has been fixed at that moment, but then it is brought back out at some point in the future.
Our repeat rate will absolutely decrease utilizing the system that you have delivered for us. Harmonic is a true partner at the working level, at the strategic level, at the business level. We are able to have open and honest conversations about performance levels, both on the GCI side and on the Harmonic side, so that we can then fail fast, resolve it, and implement those fixes that are required. We are already starting to talk about where that broadband component can even extend into the home, where we can ensure that the customer experience is everywhere from their PC, their TV, through the Wi-Fi, into the modem, and back into the network, such that Harmonic is part of the entire customer experience that we can then deliver a differentiated experience in this market.
Great. Next, a different story. Hotwire Communications, that I am proud to say that today Hotwire can find and identify service degradation issue before customer are actually calling them. What I really like about Hotwire, that first of all, it is a pure PON customer, 100% fiber, and second, none of the equipment they deploy is with a Harmonic platform. 100% non-Harmonic. I know Jeff is working very hard to change that as well, by the way. What they are doing, detecting the problem, and it has been a journey. Learning, and we are now in the process to automate things, making it the path to autonomous network.
One example that I love to share with you today is the fact that, cOS SensAI identified a service degradation, realized that it had nothing to do with the infrastructure, not the PON network, not the ONT itself, not the Wi-Fi, and identified as a video degradation issue, that required some CDN configuration, which fixed the video problem. Okay? Without understanding the network graph, without understanding the topology, this will take weeks, if not months, to find that. Let me hand it over to Pragash, the CTO of Hotwire, to tell you more about that.
The work that Harmonic and Hotwire is doing together, I believe this is going to break some barriers and introduce something that we never thought of. Pragash Pillai, CTO, Hotwire Communications. I have been in telecom for about 27 years now. We have been working with Harmonic for a few good months now. We are ready to deploy the solution in production. Part of the cOS SensAI product, what are we looking at doing with the solution is gathering all the data and telemetry about our customer and providing insight either to customer or to us. In a quick example use case, we found a customer with one access point with 70 over devices connected to access point. There is two problem we can solve there.
First, we can solve the technical problem. Second, this is an upsell opportunity for us to improve those customer experience by having the whole home connected. In a traditional way, typically, we try to run databases to crunch the data. It takes days, it takes weeks to get to the information that we want. With AI and the ability to structure the data with advanced data lakehouse platform, we can get to the answer much, much faster. That is what I think going to drive our industry further.
If you look at Hotwire history, 20 years ago, when our CEO started fiber, nobody was doing fiber infrastructure. We are the first one. Hotwire was one of the first fiber company out there in the U.S. And now we are looking at network intelligence, AIOps. I hope when we get a few years forward, we look back, we say Harmonic and Hotwire pioneered this AIOps approach that everybody else is leveraging. We value the partnership with Harmonic and look forward for the future.
How are we going to go after this market? How we are going after this market is first land the base of customers, win new operators, and also sell more services. What do I mean by land the base? Today we have more than 160 customer that are already using our first generation cloud, and the idea is we are outgoing and upgrading them to the cOS SensAI experience. As we win more and more operators on cOS as well as non-Harmonic customers, we are going and offering them our intelligent platform. That opens also an opportunity for new business services for us.
Now that we have cOS SensAI, we can offer a NOC-as-a-service or operating the service for the customers. What is that doing for us? It is actually enhancing our recurring software revenue stream, but furthermore, it is diversifying our customer base. We have talked about the second engine.
That is the intelligent platform. Asaf talked to you about connectivity, our first engine with our beautiful software and hardware products and solution we developed for cable and telco in order to accelerate their deployment. With cOS SensAI, we are bringing the intelligence to operate those network in a much more cost-efficient way and in a smart way. With that, I will hand it back to Asaf to tell you one more thing. Thanks, Asaf.
I am back. Hope you did not miss me too much. Thanks, Yaniv. Let us talk about the third layer, the third engine, the compute side of this. We spent time on the connectivity layer, sharing with you the footprint that we freed up at the different facilities. Let me walk you through a story of what that means. If you think about the simple diagram of the headend and the hub and the homes that are connected today, and if you think back to the integrated equipment that was there, the purpose-built chassis. With cOS, we talked about the freed space and the power that becomes available now. In some cases, some operators do hub consolidations. In other cases, those hubs exist. The cOS and the RF gear that pushed out into the network DAA, there is an opportunity there. We are all familiar with physical AI.
Some analysts point out to over 100 million robots by 2030, all kinds of different humanoids, cameras, and sensors. Physical AI. Imagine the challenge they are going to have when they need the latency, the predictability to run mission-critical operations and services for all those different opportunities that they have to service their purpose and their mission for their companies.
We can connect our existing infrastructure with that freed space on the GPUs. The opportunity that we've built with our orchestration, the freed space that we've enabled now can be used, and we've already done it. We're actually in development now where the GPUs and the different types of compute can leverage the platform. The platform is there, the ability to connect it to GPUs and to run local tokens through that or to do it in a very intelligent way. It's a very exciting opportunity. When you look at other opportunities as well to connect cell towers backhaul, all on the same fiber. This will give the performance they expected that the operators already have for their existing subscribers to connect all these things.
We look at all the different opportunities for those different facilities, thousands in the network, and we'll show you a little picture of what that actually means. The modernization we sell frees the space and the power that e dge AI inference for all these physical AI needs. We call this cOS Edge AI. This is a real map of North America, obviously, with the different sites that we have today and in the future, thousands of sites. Globally, it's even larger. If you think about that's where cOS runs. If you think about the diagram of the transition for other facilities that we showed a few slides ago, the opportunity is there.
This is an example, just to put words to pictures, of a site with over 20 racks of equipment which have been reduced to one rack. It just gives real perspective on how much space and power is available. The freed space is already there. Our customers own that, and our software is already in it. We look forward to expanding on this in the near future and providing you more information. It's now time for me to hand it over to Walter, who will provide you our financial overview.
All right. Thank you, Asaf. A lot of excitement in terms of the addressable market. Today, what I am going to focus on is talking a little bit about how we are thinking about growth, considering all of these new opportunities and expanding markets available to us. Before I do that, let me just start in terms of what we have delivered over the last several years. When you think about our midpoint of guidance for FY 2026, we have delivered a 19% CAGR growth rate on revenue over that period of time. I will address 2025. It was a down year for the industry. We had identified that prior to the start of 2025, in that with DOCSIS 4.0, operators were pausing decisions in terms of investment. We had said at that time, we expected a rebound in 2026, and that is what we are seeing.
We expect growth for this year-over-year on revenue, to be 43%. If you look at the purple line, that is our rest of market non-top two customers, and you can see the increase in revenue. Year-over-year for 2026, we are expecting rest of market to grow by over 50%. If you look back two years ago, it is more than doubled. That has been a priority. You heard that from Nimrod earlier. The team has continued to improve, and there is still more work to be done, and I will talk about that. From an operating margin standpoint, based on our guidance for this year, we will achieve 20%+ operating margin based on the business model that we have. The key message is we have held our share. You heard that from Asaf earlier.
The only deltas are market timing in terms of the investment, in terms of the modernization of these networks. Since I have talked about 2026, I want to take the opportunity to reaffirm our guidance that we provided back on August 12th for Q3 of this year as well as full year 2026. This is the slide that Nimrod showed earlier, and it is so important, I wanted to bring it up again because it really signals how we look at the market and the expanding addressable markets in terms of additional adjacent markets that we are focused on. As Nimrod mentioned earlier, we are traditionally known for the cable access market, which is $1.3 billion, and obviously today from the discussions, that is a much larger available market that we are now targeting.
The key point that I want to make before I jump to the next slide is really how we are thinking about the growth, what we are including in our growth rates, and what we are excluding at this point in our growth rates. I will share with you our three-year view of our growth rate in terms of CAGR, and then walk you through what is included and what is upside based on all the presentations today. When we look at the midpoint of our guidance for 2026, we see 10%-13% CAGR over the next three years. Let me be clear, that is a floor in terms of our business, and I am going to walk you through and give you additional color in terms of what is upside to that number based on the discussions that we have had so far today. I will go category by category.
In DOCSIS, we are assuming the market is going to be flat from where we are today after peaking in 2027, 2028. You saw in Asaf's chart earlier that the market view is that it will still continue to grow. However, we are going to be conservative in our assumption around DOCSIS. In cable access fiber, we are assuming a 40% CAGR over that three-year horizon. You heard of the progress we are making already. We have disclosed that in prior calls in terms of the wins and the momentum building in that area.
In telco fiber-to-the-home, you heard Asaf talk about our diversified portfolio in terms of our differentiated portfolio. What we are building into the growth rate is a 40% CAGR from where we are today. Let me be clear, we are not including 50G PON or any tier one wins in our base CAGR that I am showing here.
The reason being is you saw in Asaf's chart that 50G , we are expecting that market to start forming in 2028 and 2029, but we are going to be conservative because we have seen market transitions that take longer in time. So that is not included in our base growth model. PON, beyond fiber-to-the-home data center enterprise, that is all upside to the model. In terms of intelligence platform that Yaniv walked through, our expectations is to grow that at 100%+ CAGR a year over this three-year period. The more important metric that we are leaving you with today is that we expect that portfolio to be 10% of our revenue over this three-year horizon. Then finally, on edge AI compute, that is still a to-be-determined, as we talked about earlier, and therefore, that would be upside to the actual model.
In summary, we are setting a floor of 10%-13% growth with significant upside as I have walked you through and given you the color so you can understand how we are looking at the market opportunity for the company. I will make it very explicit just to make sure that I covered it off clearly, is that there is four key upsides. We are working on those today. You heard from Yaniv in terms of our 50G PON. We are making those investments in a differentiated solution that will be available in 2027, but we are not including the revenue in terms of the CAGR rate that we are showing you over the three-year period. But the key message is we are funding these investments organically.
Now, it is more than just building a bigger business. It is building a more durable business. In speaking with many of you, both our analysts as well as our investors over the last couple of years, top of mind is always customer concentration, lumpy revenue. From our plans today and what we have shared with you, we see the opportunity to build a more durable business. First of all, from a rest of market standpoint, we want to increase to 55%-60%, and we see that in the three-year horizon window from the 40% we were at today. Recurring revenue, bumping that up to 22%-25% of total revenue, and that will include the intelligence platform that we highlighted at 10% of total revenue. Then finally, with regards to operating margin, 20%+ is our view. Let me say two things. One is we are already doing that in 2026.
But the key point I want to leave you with here today is that those organic investments that we're talking about from a product, from a go-to-market standpoint, is factored into our thinking in terms of ability to execute on this plan and continue to deliver 20%+ operating margins. The key point here is we're building a more diversified, a more recurring business with more growth vectors for upside. Let me just finish here, talk a little bit about our capacity and our ability to execute on our capital allocation priorities. With the sale of the video business, it's put us in a great position in terms of financial flexibility. We have over $300 million in liquidity as of the end of last quarter. We expect to continue driving strong free cash flow over the next three years.
And with the record backlog that we have, we have confidence to commit our capital. You've heard us talk about these priorities previously in the organic area. The one that is new is the funding of these investments for the expanded market opportunities, in addition to investing in the ramp-up of our business, including the purchase of inventory to secure supply and to be ready for customer demand. Return of capital, we've executed $122 million of our $200 million program. We've set a floor that we'll buy at least enough to offset equity compensation, and we'll look to do more.
Then finally, on M&A, we've been very consistent with our comments on M&A. If we can see an opportunity that will accelerate our strategy or de-risk it, we'll prudently look at those options. And right now, we're holding cash as we assess options. But I want to be clear, we're already moving in terms of our organic plan to drive the growth into these markets that we've talked about today. The net, just to summarize, we've set a floor in terms of growth rate. We've talked about where the significant upside is to those growth rates, even within this three-year period. And we've talked about the capital that we have and the team to go and execute on all of this.
And so from my perspective, we see profitable growth, we see a lot of upside potential in the business, and that's all great. But what really gives me confidence about the future is about what customers are saying to us.
Recently, we were at a fiber trade show where we were presenting some of our new products, including our hardened fiber products, and we were discussing use cases in suburbs, rural settings, and the customer stopped us and said, "I can use this in high-density areas in cities. I'm paying a ton of money for putting my fiber equipment in leased facilities, and those leased facility costs are increasing dramatically. I can take that out of there and leverage your solution to remove those costs." I think that's a great example of the versatility we have in our portfolio in terms of our differentiated portfolio and the ability for customers to unlock value with that portfolio that we have out there today. So with that said, I would like to transition it over to Jeff to introduce the customer panel and go from there. Thank you.
Thank you so much, Walter. Thank you so much, Walter. Good morning, everyone. My name is Jeff Glahn, and I am the Senior Vice President of Sales here at Harmonic. Today, we've got a great treat for everyone. This morning, you've heard from Nimrod, you've heard from Asaf, you heard from Yaniv on our vision, our mission, our technology, and innovation. Now for the most exciting part, we're going to talk to some operators, our customers of ours, that are using all these innovations, not only to help their networks today and optimize their networks and drive smarter and more intelligent networks, but also how they plan to work with Harmonic well into the future. With that, I'd like to invite the panel to the stage, and a round of applause, please. Please have a seat. Let's start off with some introductions.
Next to me, we have the Chief Operating Officer of Bluepeak, Cash Hagen. Next to Cash, we've got Eric Svenson, VP of Engineering and Operations and Technology at Armstrong. On the end, we've got JR Walden, Chief Technology Officer of Mediacom. Welcome to the panel.
Thank you, Jeff.
First question we have on is fiber migration. We've talked a lot about fiber today, and I wanted to talk with the panel here about the journey of the panel. The journey to fiber, excuse me. Many of you are on different journeys, on the different paths, and I wanted to start with Cash to talk about fiber migration. Our teams are doing a lot of great work together, Cash. We'd love to get your insights on fiber migration today and what your plans are for the future.
Sure. We are on a shot clock here, and we are all appropriately verbose as we decided. We will keep it short. Just by way of backdrop, Bluepeak, as a service provider, operates in eight states, High Plains, Midwest, into the South. About our 70/30 split of pure play fiber-to-the-home, fiber to the premise, fiber to the business, and traditional HFC. We have been on the journey of greenfield fiber launches now for the last, call it five years. We operate in 80 distinct cities. Of those, we call it 30 of those are our greenfield fiber-to-the-home cities that have been built within the last four and a half or so years. So lots of activity just in a pure play fiber way.
On the HFC side of our business, we, like many, DOCSIS is evolutionary, not revolutionary, and we have followed the same path as many of our peer groups have with DOCSIS and started off on the path to DOCSIS 4.0 a handful of years ago, I would say, and did quite well in partnering with Harmonic on that. Then we took a bit of a pivot within the last, I do not know, year or so, and those were some fun conversations. Obviously, these are capital-intensive networks as you all in the room certainly know. The pivot was not in a negative way about HFC or DOCSIS. It really had nothing to do with that as a technology medium. It was bang for the buck value creation for us as a service provider, and made a decision in call it 80%- 90% of our traditional HFC footprint.
We will not go down the DOCSIS 4.0 upgrade path, and we will instead overbuild ourselves effectively. These are all competitive markets. There are other providers there. There are some that have been more aggressive on their own fiber plans, and it was a pretty big shift for us. Where we were at the time with Harmonic, Wael and Jeff and team said, "Well, how about this?" It was effectively kind of repurposing a lot of the technology that we had learned a lot about with the products that are all sitting over here, and to be able to do it. One of the real benefits in that strategy was optionality.
It is really a key point of we have a DOCSIS network, we have large HFC networks that have been out there for 100 years, and the ability to offer fiber to the prem services within the same footprint is incredibly valuable. That was a unique kind of delivery option that Harmonic provided, versus your traditional kind of more telco-oriented fiber-to-the-home vendor partners. That was a pivot. It was a big one. 100% privately funded. We are a provider that we have not taken any BEAD funding, any other state or fed funding. So it was a big decision to do that from a cost per passing standpoint to do that. We have a roughly four-year plan. We started, really first kind of shovel in the ground was in, call it March of this year.
And we will have overbuilt roughly, call it 30,000 passings by the end of this year. Over the course of the next three years, about 90% of that traditional HFC footprint will be fiber overbuilt. So exciting, lots of work. Not rocket science hard, but incredibly hard to do what we are doing.
Excellent, Cash. Our teams are doing exceptional work together and looking forward to a bright future. Next up, Eric Svenson. We are, at Armstrong and Harmonic, at a little bit of different point in the journey, more emerging. Could you please tell us a little bit about the work that our teams are working on together?
Yeah, absolutely. We first started looking at Harmonic when we were looking to upgrade our HFC network, our traditional broadband network. A lot of the presentations earlier talked about the space, power of all the traditional CMTSs that were in our facilities, and looking to modernize that as well as bring new and faster speeds to our customers to be able to compete. The start of the journey was looking at how can we upgrade our network and work together, and work on that long-term vision. Because at Armstrong, we are focused on fiber and delivering fiber, but we also need to upgrade and leverage what we have in our existing HFC footprint.
Working together, yeah, we were able to deploy the Harmonic solution and be able to upgrade and get similar to the fiber speeds with gig symmetrical services that we could provide our customer, and looking to do actually multi-gig solutions with that. Someone mentioned, Asaf mentioned before, the unicorn. It really is because on our path, we do see doing HFC and to what Cash said, but eventually get to fiber. Having the same housing, the same equipment, the same management plane, to be able to transition to that fiber space long-term really simplifies that journey for Armstrong. That is really been important to us, and really appreciate the partnership and kind of understanding what our needs were and helping us on that long-term plan and evolution. We appreciate that.
Yeah, we love the partnership and looking forward to the future, Eric. Switching gears, we're going to talk about DOCSIS 4.0. JR, you were a bit of a celebrity at last year's SCTE in Washington, D.C. I know a friend, a senator, made friends with you quickly, the state of Illinois. Would love to hear, JR, about your journey on DOCSIS 4.0, the work we're doing together, and what do you see for the future?
Yeah. Mediacom's also, like some of the others up here, a rural market cable company, lower density, spread across 22 states. We have 3.1 million homes that are fed with HFC today. We wanted to find a path to bring higher capacity to those customers, but maybe more importantly, higher reliability and more network intelligence into those homes. We felt like we needed to do it really quickly, right? That was one of the things that really kept us attracted to the DOCSIS roadmap. So, of our 3.1 million homes, we've already upgraded north of 650,000 to multi-gig symmetrical services. If we had adopted a fiber-to-the-home first strategy, overbuild strategy, we don't think we would've gotten to nearly that many homes. We're north of 700 nodes now deployed with, I'll call full unified DOCSIS 4.0.
So those homes all have a capacity north of 9G down, which is frankly faster than our 10G XGS-PON goes in other markets. Yeah. Being a rural market cable company, Senator Tammy came by and was very excited that we were deploying, of course, in her state of Illinois. Also, I think there is a lot of concern among politicians and others, and maybe appropriately so, that they don't want a new digital divide or a continued digital divide. So, we see a lot of representatives from states that have significant rural markets concerned about whether the next generations of technologies and the capabilities that AI and other things will bring to the country leave out a certain portion of their constituents. Being a rural market company, these technologies deploy exceptionally well in the densities that we are at.
Like I said, they're very quick to market. The other two highlights I just give, we absolutely run fiber-to-the-home markets as well, and running XGS-PON. I can tell you, an upgraded DOCSIS network is every bit as reliable as a PON network, at least the way we see it and monitor. So, we're very excited about that. The upgrade process is very customer-friendly in that we're the incumbent internet provider, broadband provider. We have the majority of the penetration. All of our existing customers, they can keep their CPE, they can keep their current connections. We don't have to swap anything. It's only if they want to take one of the new capabilities and new speeds that are being enabled with these upgrades, then we can swap it.
But for a lot of the customers on day one, they like to hear about investment in the community, but as long as that investment doesn't disadvantage them today with what they're doing, that's a great part of that technology.
That's great, JR, and we're proud of the partnership, and we're proud to partner with someone as visionary in this business as you and the Mediacom team. Cash, you talked a lot about fiber, but I understand your strategy is a hybrid approach with a bit of fiber and DOCSIS 4.0. Would you mind telling us a little bit about that?
Yeah, 100%. It's interesting. It's a bit of a religious debate, right, or a philosophical debate. Fiber over here or HFC over here. One's good, one's bad. At the end of the day, we've all been around a very long time and that is certainly not the case. There are situational reasons why it may be a pure play fiber or it may be upgraded HFC. Had lots of conversations with our board and our investment group on DOCSIS, and there are times where it comes with, there's a negative association, not with DOCSIS, but maybe with the coax cable. But a lot of education on really where DOCSIS is at today, and it's the point that JR just made. Customer experience, reliability of service, products offered. Today's DOCSIS is far in a way different than anything where we have been. Again, it's an evolutionary path.
I don't see a scenario where we'll ever have one or the other. It's going to be a combination of. I think we talked about it briefly in one of the presentations. There are certain use cases, MDUs are a good example of a use case, where HFC and the utilization of the internal wiring inside of an MDU where HFC is very critical because you may not physically be able to get fiber to the unit. Costs, operational complexity, or other, where in most buildings, certainly in the U.S., there is coax that had been pre-wired when that facility was constructed. So to be able to utilize that, and again, the technology sitting on that table over there, so your fiber in, but still DOCSIS kind of within is a very critical part of our go-forward plan.
A balanced approach. I think many operators are looking at that, and that is interesting words. Next topic is AI. A bit of a hot topic these days. Enormous attention across the board. Network operators are looking at, as Yaniv described, everything from homegrown to what do we do? I love your insights of the panel here. Where do you see AI creating the most measurable value in broadband operations? We are going to start with Eric.
Yeah, absolutely. So I think the most important for us is the customer experience. What is that customer's experience using our services? It is interesting. There are some customers, they may have some issues that are totally fixable, but they are not even calling us up. So being able to leverage a platform like cOS SensAI to really know what is happening in the network, and it is not as simple as just one location. We have to look at what is happening in our hub sites, what is happening in our networks.
You heard the example of pointing out the CDN problem. Collecting all that data, we have access to all that data at different points in the network to understand what is happening. But to be able to put it together, we were not able to do that before these AI capabilities were there. So that is really important.
Another thing is, as we are, whether it is someone calling into the call center or someone in the field looking into something, to be able to present to that agent or that rep in the field to better understand what is happening, we are able to identify issues faster as well as be able to get out there and fix things faster. We are not all graced with lovely days outside. Folks are out in the rain, up in the bucket truck in inclement weather. So to be able to identify faster and rather than drive to different locations and find out what is happening, to be able to really put the pinpoint on where the issue is and target that to resolve the issue and best serve the customer. That is really the thing that drives and excites me towards these opportunities that are out there now.
Great insights. AI is a mission, a game changer in AI operations. That is interesting.
Yeah.
Excellent. JR, how about you?
Yeah. I think Eric stole probably the number one or number two answer. But there are more. There is a lot of things to be delivered. Let me hit on one I think is also adds some value. Designing networks is a lot about balancing the capability and performance of the network against the reliability of the network. And often it is a sliding scale. You become more reliable by sort of backing off from the edge capabilities of the technology.
When you have large, diverse, multi-state, multi-technology networks, you pick a sweet spot someplace, and you visit it every once in a while, but you are not visiting it in real time. It is just not practical. AI really changes that for us. We see so much from the monitoring telemetry of the network. We know the impact that that has. But when you have to equate that over tens of thousands of nodes and service groups, the AI can come in and say, "Okay, great. While we are getting things fixed," and that is a great thing, and improving the network, we can adjust the capability of the network and sort of the fundamental engineering of the network to slide it more towards reliability, for example. We do not need maximum performance probably 20 hours out of 24 hours of the day.
Even if we are at that point where we want maximum performance, offline is far worse than a little bit slower but fully functional. You want to be able to adjust that slider in real time and always, the way we look at it, reliability has to be number one, performance is number two, and we want both at all times. But being able to adjust in real time really gives us a lot of capabilities to manage that network and engineer that network in a way it is hard to do today.
Excellent, JR. Subscriber experience is everything. That's excellent. Cash, how about you?
Now that they've stole all of it. T here's probably not a topic that I've spent personally more time on in educating and learning and talking about than AI. There's all the buzz associated with it. Really, my view is it's not just AI. There's a whole umbrella of just technology enablement systems, platforms, processes, how we operate the business under that umbrella. AI is a component of that. What I talk a lot about, we just got to run the railroad better. We got to be smarter. We got to be better, faster, cheaper, more efficient. We can't just throw more bodies as our businesses scale. Those are hard things to do, but there is volume of work, and we got to be more efficient in how we do that work.
We've all been around a long time, and customer care reps are great, and NOC techs are great, and field techs are great, but they're not always great. To have technology platforms that help them, that educate them, that help our customers, and do as much as we possibly can hands-off. I know there's a lot of debate on digital assistants or AI call center reps and how people react and respond to that. It's a lot of fun if anybody wants to listen to some recorded phone calls on a Saturday night and how people engage with that and then fine-tuning that to make it. But it's all about just delivering a better and cheaper experience through technology means and smart kind of implementation of AI.
It's not the take over the world, because we very much believe human in the loop, AI is where we are and where we need to be, to help us and to enable us to do better, faster, cheaper, and more efficient.
This is great. A lot of optimism on the stage, and I think this panel is a great barometer for our industry, and their insights as well. Excellent. Question. Over the next five years, what are you prioritizing as your network investments, and what is driving those decisions? We are going to start down at the end with JR.
Yeah. So, we are really prioritizing our investments on experiential-type things. Right? I know we are doing a lot of upgrades, and so you might argue, well, that is really about adding capacity. But we really look at it as the experience. Right? It is a lot about the reliability, it is a lot about having the network capabilities allow us to roll out new products, new capabilities very quickly.
Look, we are happy to have more capacity as well, but that was not the principal driver for it. We would have picked different technologies if that was the most important thing. We are constantly reminded, and it is maybe the hardest thing for a CTO, customers do not buy technology. Politicians get excited about technology, and engineers get excited about technologies. Customers buy experience. If you are investing in experience, you will win. If you are investing in technology, you will not.
This broadband business is a great business. There is a lot of people who want a piece of the pie. We think that investing in the experience is the way to go.
Excellent, JR, thank you. I got some great notes for our final question at the end. Cash, insights.
Yeah. I think similar, we talked about just from an upgrade standpoint, but there is an aspect of ours, return on invested capital. We are spending a lot of money. This is a capital-intensive business, as everybody certainly knows. What do we get from that? Our prioritization isn't about upgrading networks for the sake of being cool or because that is where the industry is going. We need to keep pace.
All these are competitive markets. I think every market of any size, and maybe some smaller than that, will have competitive providers. Fact. I just think that that is the world that we are going to live in, and delivering solutions and investing in solutions that do improve the customer experience, because as JR said, they are not buying technology in a very hyper-competitive space. Again, back to we got to run the railroad as efficiently as we possibly can, and protect every dollar of investment, and measure the return.
Excellent. We have got time for one last question. JR, you made a comment, customers buy experience. So why Harmonic? Why do you all buy from Harmonic? Why do you choose to partner with Harmonic? Eric, why don't you start out?
Yeah. I think Harmonic has been an easy company to deal with, and I think the innovation and the collaboration are really what are key. I feel like when I am going to work on a project or have a problem to solve, that Harmonic is there to help us, and a great portfolio of products, and are really willing to listen to us and understand our needs. You see that in the development of your platforms, that innovation coming through in the partnership of working together on how do we solve the problems that service providers are facing today. So it has been a good relationship. Support is always key. Making sure that when there are things that happen, we have support to take care of it. So yeah, it has really been a successful relationship.
Thank you, Eric. JR?
Well, I often like to say we buy on technology, then relationship, and then price. I think, look, you guys had the right technology for us at the right time. So kudos to your development team. That wasn't so much a partnership as it really was you guys had, I think, good vision and invested in the right place and were appropriately rewarded for that. That's part of what brought us to the tent. Then now that we're here, we really appreciate the ability to work together on some of the technologies, right? We're a rural market, mid-tier cable company, and our needs are not radically different, but at least somewhat different from our bigger brothers. It's appropriate that you're going after that business, and we certainly need their dollars helping to feed the R&D.
But we need partners who can listen to what some of the smaller providers, and not just smaller, just the type of markets we operated in. There are some needs that are slightly different than the big ones, and we need some flexibility there, and you guys have been very flexible.
Well, thank you, JR. And Cash?
Yeah, short and sweet. I think it is you and a lot of the people in this room and a lot of people that are not here. Nimrod, 30 years later, there is a lot of tenure. You guys are operators. You are a technologist, you are operators. You understand the business, you understand the landscape, where we have been, where we are going to. You have an eye towards the future at times that we do not, because we are waking up every day running a business. I do not often have time to think about in 10 years from now, where are we going to be? Having a partner that is thinking about that, but how to help us get there. Again, it is not just about the best tech. That does not always win. It is how that works in our ecosystem.
I think you guys are certainly an exception in a really evolving industry that we have all seen lots of change. You stay true to who you are and what you are and what you are focused on, and that comes through at every level of your engagement with, I think, our respective companies. It is not just the senior team, there is a give a shit factor in your team in taking care of our business and ultimately taking care of our customers, and we appreciate that.
Well, Cash, Eric, JR, thank you for giving us the opportunity to earn your business every day, and we love the partnership. Round of applause for our panelists here.
Thank you.
Thank you all very, very much. Excellent. I would like to hand it over to our Chief Financial Officer, Walter Jankovic.
Thank you, Jeff. Thanks to our customers or the customer panel. That was terrific, very insightful. We are going to have the management team come up and do a Q&A session for our sell-side analysts, and it looks like we are right on time here to take your questions, so feel free to give us one or two questions each, and we will go around the room. [Scott] and [David] will have the microphones, and we will just get set up here. Give us about 30 seconds. One other point just for everybody in the room and over the webcast is that we will put the recording as well as the slides up on our investor relations website subsequent to the event. Okay?
Are we ready?
Oh, yes, we are ready. The team is here. Simon, please go ahead.
Thank you very much. Simon Leopold with Raymond James. Appreciate it. Hadn't realized that 12 years went by that fast. It's been a blur. I wanted to first ask about the expansion into the addressing the telco opportunities, and this, I think, this is for Walter, in terms of the expenses associated with sales and marketing, and maybe for Nimrod, the aspect related to strategy in terms of getting the right people and the kinds of relationships. That would be a change. Even if the same headcount, maybe different people. Then I've got a quick follow-up.
Okay. Certainly, I can kick it off. In terms of the expenses, we think about our operating model, and I mentioned it when we were setting up some of our longer term targets. I think really the key element is that we are planning in terms of what is it going to take to go to market with regards to Tier 1 telcos. As we model out different scenarios for the company and evaluate it's something that we want to get started on immediately in terms of developing that. Especially with the transition to 50G being a pivot point in terms of that overall transition.
We are thinking about that in terms of building it into our model, and that is why I gave the comments I did today in terms of what our operating model still looks like as we go through a phase of investment to not only that market, but others as well.
Yeah. In terms of the go-to-market expertise, I think we have a growing percentage of them that are not coming from pure cable, whether it is vendors that are known to be pure telco. I think over time you see the technology is changing in a way that these guys can move from one vendor to the other. I think this is a capability that we already have some of that and we are going to expand that with Jeff's leadership. I think that engaging with these customers is not something new for us. I think tier ones is a different ballgame, but it is not yet what we are targeting initially. As we make progress and we think we have got a differentiated 50G story, we will certainly go up the value chain of customers, and we will build up the capability.
Thanks. Then just as a follow-up, this longer term around the opportunity of edge compute, I want to make sure I understand it. It sounds to me similar to what some of the accelerator vendors have talked about in this concept of AI- RAN. The idea that you can put compute in the edge of the radio network in mobile, and I think what you are articulating is we could do the same thing with broadband solutions. First, am I understanding it correctly and drawing an appropriate parallel? If so, what do you envision as the timing for that market to really develop?
Conceptually, it is exactly that. I think the difference is that putting GPU compute in a cell tower is much more challenging to put that in a facility that has power and space and potentially even more power from the local power network. Conceptually, yes. In terms of a timeline, we did not invent the category. Some others are doing that. I think that it will take some time to not only develop the technology itself, AI is available everywhere. I think this is more of an edge inference. Sometimes it is not about the quantity and the capacity of hundreds of thousands of GPUs, it is more about what is really important to be at the edge. You clearly have the GPUs, XPUs. There is new technology that is being introduced as well. As well as what exactly is the role that we are going to be taking.
Clearly, we see the orchestration opportunity, but there is also the opportunity to cross-connect a lot of these smaller facilities into more of a marketplace, if you will, especially within a state or a wide area network. It is in development. I think others are looking at that. I think tier ones are looking at this. We see an opportunity with a longer tail of customers to really aggregate them. Customers that cannot spend their in-house R&D to do that, we are going to aggregate that for all of them together.
Thank you.
Okay.
Walter, thank you. You gave a number for recurring revenue in three years, 20%-25%. Level set with where we are today in recurring revenue.
Yeah. Certainly. Today's level is in the lower teens and factors in predominantly support contracts. Now, we do have a cloud offering. You heard from Yaniv, we are selling our services on the platform, so that is part of that recurring number as well. The movement from where we are today to where we are headed in that three-year horizon is really the bring up of a broader intelligence platform that is driving that addition in recurring revenue.
Okay. One more level set. Asaf, when you talked about 75% of the market for the cable modems you are addressing, is that within the 160 customers you have today, or are you saying you basically today own the market and the future is beyond that in these other opportunities we talked about today?
Yeah. Thanks for the question. So what I am saying in that graph, is that there are today approximately 190 million modems, of which 48 million of them are operating with cOS today, and the opportunity remaining is for the remaining cable modems that exist in the world to transition to cOS over time. Some of them with our existing customers who are rolling out cOS and some with new prospects. So it is a broadband subscriber count. I would add that some of those cable modems will transition to fiber under the cable fiber segment that we talked about. For us, that is just a simple configuration, as I mentioned, connected all to the same software. Does that make sense? You're welcome.
Hi. How you doing? Tim Long at Barclays. Two as well, if I could. First, could you guys give us any updates on BEAD program and what you're seeing from federal funding from your operator customers? Second, I wanted to dig a little bit more into the intelligence business, kind of similar to Simon's question. Can you talk a little bit about go-to-market for that? It's obviously going to be a little bit different. Maybe Walter, how are you looking at kind of incentive programs for some of these higher value, margin accretive and recurring revenue types of businesses? Thank you.
On the BEAD, number one, we do have a product ready for that, made in America. Number two, our hardened architecture and the different form factors that we have, in fact, is very cost competitive for the rural market. Sometimes you get a remote village that really all it needs is one or two ports, so you can really put a big chassis. Our architecture is a perfect fit for that. Comcast is one of the biggest beneficiaries or targets to go after the BEAD, and they use our solution. The timing, we gave up on the politics. The timing is unclear, so it's not really a near-term revenue that we're counting on. If and when that will go forward, I think we're in a great position.
Just on the go-to-market, maybe I'll hand it over to Yaniv to take a moment to explain the go-to-market plan, and then I'll come back to your final question, Tim.
Yeah, sure. So when it comes to go-to-market, we are talking about sales marketing that we already have in place. Thanks to Jeff's preemptive, we already have those people in place as well as collaboration between sales and R&D that I already explained about how we work together in order to not just sell it, but also to implement that in a very efficient way.
As we went through it, Yaniv presented today, it is a significant opportunity for us. And Jeff is already bringing in resources from an overlay standpoint that are experts at selling this type of solution out there in the marketplace and helping to up-level the team for this type of sale, and obviously to accelerate it as well. In terms of any specific compensation plans around that, we are developing that with the factors of we have got overlay sales teams involved in order to help us drive faster with the customers that we engage across the globe and ones that we want to engage with regards to the solution. Because you had heard earlier, this is applicable across cable, it is applicable across the telco market.
Question here, thanks. On the R&D side, how should we be thinking about cOS and your DAA nodes and where those are in development relative to your roadmap? We did not hear a whole lot about that today, so I assume that development is slowing down, allowing you to invest more in cOS SensAI and some of these other opportunities. Can you maybe expand on that idea? There is always maintenance and things you have got to do, but we did not hear anything about DOCSIS 4.x or anything like this.
Yeah. We don't break down the way we invest in different categories. There are clearly maintenance requirements and few other areas that we have to complete, but we're certainly over the hump.
Yeah. Okay, great. Maybe a follow-up, Walter. You've had a couple years experience now in fiber. Can you update us on where your thoughts are around profitability around selling fiber versus HFC? Any updated thoughts there?
Well, I think it comes back to the value that we're providing in the marketplace. You saw some of the examples today in terms of the differentiated portfolio and creating value for customers. The customers see the value, the economics of that value, and I will pivot to that as being one of the key drivers. As we talked about how large fiber is expected to be this year, Nimrod had it in one of the slides, and when we look at those solutions, and you heard from some of the customers talking about them specifically, it's about the value that we're driving into the market.
Great. Maybe one last one if I can squeeze it in on edge compute and cOS edge AI. Where do you see your insertion in the value chain there? Is it at the OS layer as an orchestrator, or any thoughts there around?
I think the orchestration for sure. We see a couple of other layers on top of that, but I think it's still in development stages. I think it's subject to what exact compute and the use cases of the compute.
Hey, thanks a lot. George Notter from Wolfe Research. I guess I was just curious about just the R&D investment here. You guys are in the high teens in terms of quarterly R&D investment, and I'm just thinking about the other competitor in the marketplace that's done a lot of this intelligence stuff would be a company, Calix. They did a lot of development around Calix Cloud and agentic, but they did this spread out over many, many years. I think for Calix, it was a 5-10 years development to come to market with that product, and they were also fairly R&D constrained. I'm just thinking about, again, do you have enough resources R&D-wise to get all this done and support the existing product portfolio? I know you guys are also committing to a 20% operating margin, so that's another constraint in all this.
I guess I'm just trying to understand that do we have enough resource here to get all this done? Thanks.
Yeah. Thank you for the question, George. We're doing it today in terms of building out the solution. You heard earlier in terms of this isn't what we plan to do, we're doing it today. So it's a great point to say, do we have enough. We've been putting resources on this for some time now in order to develop the overall solution. When we look forward in terms of our modeling of what we will continue to need, because obviously it's a continuation of R&D support as you span out, but also go-to-market investments that need to be made. We feel we can work within our envelope. Okay, we got a minute left here. A couple minutes.
I think George has a follow-up.
Oh, follow-up. Here we go.
Yeah, sorry, just as a follow-up. For the intelligence product and the initial customers that you guys have, I'm just curious how much. Any sense for what pricing looks like there? Anything you can share? I assume it's obviously on a per subscriber basis, I would guess, but is that $1 a month? Is it $0.50? Is it $5? Any sense for where that shakes out and what's reasonable there?
Yeah, I don't think we'll give a precise number around it, but it is based on subscriber, so as Yaniv presented earlier on one of the slides. And we've looked at various different models. We've worked with outside companies as well, looking at. Because other people have done other things like this in terms of these recurring models. And Yaniv and the team have taken some good feedback from external sources in terms of value created, how these models look to bring them into our commercial model in terms of what we will charge customers for that service and value. Okay, I think we hit the zero on the clock, so I think we're going to wrap it up.
Well, thanks again for joining us today here and remotely. We covered a lot. We truly believe that the broadband network will have to keep up with what we see out there on AI. There is a ton that is being invested on data centers that will have to get consumed, and that's going to challenge the network. We did go through the three engines. Connectivity is expanding. The opportunity is expanding. In cable, we see a lot of fiber, and we see the opportunity to go to telco, and we even see a bigger market for the same technology. We talked about the PON, beyond fiber- to- the- home. We're super excited about the intelligence opportunity we're entering into. We think we bring a differentiated vision and architecture. And by the way, George, I think we also do spend a lot of tokens to make up for time.
I think whoever started later could actually speed up because we don't only develop AI, we also use AI for the development. And finally, the compute is intriguing. It's in development. We're going to keep focusing on this because we think it can make a difference, and we're going to update you as we make progress. Thank you very much again.
Thank you.
Thank you for the amazing work that the team here did on the back office preparation here. Yaniv, Jeff, Asaf, Walter, thank you again.
[Marketing]. Thank you.