Hello, good day. Thanks for joining us at the Rosenblatt Age of AI Technology Summit. I am Michael Genovese, the cloud and communications equipment analyst. Super pleased to be speaking with Calix today, and with the CFO, Cory Sindelar. Hi, Cory. Good to see you.
Hey, thanks, Mike. Glad to be here. Thank you.
I just should let the audience know that this is going to be a Q&A between Cory and I. If you guys want to put your questions in the mix, I do believe that in the upper right hand of your Zoom interface, there should be a little text box that you could type questions in, and they'll come to me, and I'll ask them to Cory. Let's get started, Cory. The stock performance has not been wonderful recently, so we're sort of looking for when we're going to get back on track. From a strategic level, my first question is that the move to this Google Cloud, the third-generation platform with Agentic, was important enough for you guys to sort of interrupt the steady march of software and services gross margins going up every quarter at least temporarily.
My first question, before we talk about margins getting back on track is, why is the third-generation platform, why are agents, why are the other capabilities of this AI-enabled platform important enough that it was important enough to do that, to impact the march on software and services? What new markets does it get into? How much of the TAM does it open up for you?
Yes, Mike. Thank you. Great question. We've been going along on our cloud platform journey now for seven to nine years. It was time to rearchitect the platform. At the same time, the whole world of AI just exploded onto the scene. We were very fortuitous in having the opportunity not only to rearchitect our cloud for additional scale, but to take advantage of AI and in getting that embedded into our platform. The third-generation platform is a culmination of that work. Not only was it architecting us for the next leg of growth, which would give us better scalability and better cost points for maintaining our clouds, but has the introducing AI and Agentic workflows.
The effect of Agentic workflows is to allow our service providers the ability to increase their productivity. They can go faster and to do more things. That's what we're looking to do. We think we're uniquely positioned with the fact that we have 1,200 customers already on our platform. We have deep understanding of their businesses, and creating a framework that takes advantage of A2A and MCP creates a framework that allows us to do it in a safe way to enable these Agentic workflows. We're in a whole new era in terms of that. The third-generation platform also allows us to move into larger and larger customers. It allows us to effectively provide a private instance of our cloud environment.
Large telcos can utilize their own Google contract, and set up our infrastructure inside of their data so that it does not go into a multi-tenancy cloud. It can stay in their own environment, but taking advantage of all the learning and the agents that we have built for them. By extension, going to large customers gives us international as well. You get some extension into the international markets. That's the benefit of the third-generation cloud. Obviously, to get our customers from the second-generation cloud to the third-generation cloud, we did a bit of an insurance policy. Obviously, our clouds run our customers' businesses, and to have an impact on our cloud and to have an impact on their businesses would be unforgivable.
For a period of time, not only did we run their data on our second-generation platform, we ran it into the third-generation platform for about a period of four months where we had those dual cloud costs. The great news about Q1 is you put it in the rearview mirror. We were able to complete the migrations in the first quarter, clearing the way for us to now start selling Agentic capabilities. From a margin software and services margin perspective, we put that one in the rearview mirror. Q1 is a low point. We said that Q2 would get us back to kind of a Q4 level. After we spend some time optimizing the current environment in Google, we can get ourselves back to record levels in terms of gross margins, likely in the third quarter.
As we progress through the back half of the year, all you're now doing is mixing in larger numbers of new subscribers that come in at very, very high gross margins and puts us on the path back to 70+ gross margins on the software side.
I think probably to me, the two most important numbers that investors ask about are the software and services gross margins and then also the RPOs. Was there something going on there in the quarter? You might have mentioned on the call, I don't remember, but this transition to the new cloud, did that impact the kind of new orders, and then do you expect to sort of re-accelerate those RPOs through the rest of the year?
Yeah.
Because of that?
It did. It was all hands on deck to get customers migrated over from gen 2 - gen 3. Whether they were interested in Agentic workflows or not, we needed them to do a few things. In order to kind of keep on the timeline, we basically were forcing them to do a few things. I would say you kind of created a hostile selling environment. It's not the best in place to be trying to sell them new contracts when you're trying to have them do things they don't want to do. We got that out of the way. We've got them over to the new cloud. Yes, I expect the RPO will increase Q1 - Q2, and continue then to accelerate in the back half as we continue to demonstrate the value of Agentic workflows.
Couldn't really start selling those until we got to Google Cloud, started in earnest in April. Obviously we're focused on taking those customers that were leaning forward with Calix, that were all in on Calix, and using our customer success organization to go in there and measure the productivity gains that they're seeing by the benefit of Agentic workflows. Those proof points will start to roll out as we go through these upcoming months and crescendos into ConneXions, which is our user group meeting in Las Vegas in October. I just think ultimately by the time we get to the fourth quarter, the value will be demonstrated, and we'll see kind of a large uptick in terms of our RPOs in the fourth quarter like we normally do.
It makes sense that the third-generation platform kind of the customer acceptance and rollout of it kind of happens in-- It doesn't all happen overnight at once, right? It kind of happens in stages. It happens incrementally. You target the first, most obvious customers and kind of go from there. Does the recovery in software and services margins and then the next move higher after that, does it kind of go incrementally with the Should the snapback maybe be faster than that? Does it kind of track the take-up of agents, or does it move along with something else because you've shut off the AWS? I'm just wondering if margins, if we should expect it to snap back or kind of come back incrementally.
If you're talking margins, you're kind of getting a snapback in Q2 as the dual costs are being eliminated, right? We're going from 55 something to 61, 62 something, right? That decline that you saw from Q4 - Q1 snaps back. If you're talking about RPO growth, I think that just gets back on track and then growing. It's a function of still subscribers coming onto the platform, right? Remember, our operating software, cloud, and managed services are all priced on a per subscriber basis. As these customers continue to add subs like they do every single day, the installed base continues to grow and you come around on these renewals. I suspect we'll get some acceleration in terms of RPO renewals with people wanting to add Agentic workflows. There's no snapback, right?
What you get is this nice gradual arc with a positive slope on the line that has a slight arc as more and more people are adding more and more subs as the installed base continues to grow and as they add more services, right? It's just a steady progression up into the right for software additions.
Adding more services, adding more experiences is important, and we'll, I think, dig into that more later. Probably, the biggest driver overall is going to be adding subs, right?
That's right.
Agentic, help me understand how I guess I'd say it helps the customers move faster. Is that the whole story? Is there more to it on how Agentic should actually accelerate the sub acquisition?
Yeah. You may be asking the question of how are you going to monetize your AI investments? For us, the main way to monetize it is an acceleration of our overall business model.
We're freeing up our customers to do more or to do things faster. If they're looking to add and increase their penetration rate, well, we should help them with that with our marketing cloud, right? The agents should find a better commercial fit for those subscribers at a faster rate than they were doing it themselves. You would see an acceleration of sub growth, freeing up capacity to do more, the people ought to have the ability to go add an additional service. If they've been focused solely on rolling out their growth in homes, well, maybe they'll take on small businesses now or MDU, or maybe they'll roll out SmartTown, right? Using workflow to free up capacity in terms of resources allow them to go do more, which would be an acceleration of our cloud.
There is some minor increment monetization, because what we learned is, yesterday when we were selling our clouds, they were oriented around people, right? We had a cloud that was focused on the customer call center based on the person sitting there. We had a marketing cloud focused on people helping with engagement with customers, and we had an operations cloud focused on network operations. Again, all geared to people. AI doesn't work that way. AI is a horizontal across the enterprise. For you to take advantage of it, we said to ourselves, "Well, you need to have all our clouds."So those customers that had three clouds, really there's no real incremental cost to move forward with Agentic workflows.
If I have two clouds, there's a small wee bit more you're going to have to pay to get to that third, you're going to want to understand the DV before you go forward. Obviously, if I have one cloud, then I've got to make a little bit more of investment in terms of upfront those costs in order to go to Agentic workflow. The primary way we're going to monetize it is a pull forward of the business model as these customers take advantage of the productivity gains to either go faster or do more.
Okay. How does this change the per-subscriber economics? Does it lift ARPU? The SaaS line, is it incremental? How should we think about that?
It'll raise the per sub a bit if you had one cloud or two. If you had one cloud, it'll raise it a little bit more. If you had two, just a little bit. If you had three clouds, it's not going to change the amount per sub. If you had three clouds, effectively, it's not changing your per sub rate. It's going to just accelerate the number of subs that you're adding.
I assume not a huge number of your customers had If your total experience as penetration is, as you said, 45%, that doesn't suggest that there was a ton of customers. In a way, as people sign on to this, you're adding clouds across the customers as they sign on. Is that the correct way to think about it?
You are. At the same time, we still have penetration to go in terms of our existing customer base, right? Calix at large has around 1,700 customers, of which 1,200 or so are on the cloud. There's still room to sell clouds or our platform in general. The goal of our Agentic platform is not to sell three clouds. It's really to accelerate their ability to do more with less.
Got it. Okay. Let me go to a client question, which is actually kind of a general question. It's kind of a high-level question. I'm interested to see how you answer it. It basically says: "Can you speak to the revenue model, average revenue per customer, and customer acquisition costs?" I think for the whole business, you're the CFO, how do you think about the revenue model, the average revenue per customer, and the customer acquisition cost?
If you're looking at that metric, we're looking at it ultimately from the number of subs. Yes, we are always wanting to land new customers, right? Because again, ultimately we are a land and expand model. Winning a new customer is just the beginning of the platform journey. It's really how do we accelerate their ability to deploy our platform across their enterprise at a faster and faster rate. Ultimately, our monetization is one of how many subs do they have and how fast can we add antennas. We're looking at it. We don't have true metrics of acquisition costs. We're not a enterprise. We're not trying to address enterprise, there aren't acquisition costs per se. We have a fairly narrow set of customers.
There's about 4,000 service providers globally, we have a direct selling motion that goes into working with those customers. It's not about acquiring customers, it's more about getting customers and ultimately getting to their subs.
Got it. Just from the same client, what is the average tenure of customers, and can you speak to the retention and churn?
Sure. We have virtually no churn, right? Once you're moving down the platform journey, you're on it. The value that's being delivered in terms of all kinds of operational metrics makes it incredibly hard to move off the platform. Whether you're looking at a call timed resolution or reduction in truck rolls, the operational efficiencies that you're gaining. On the appliance side, the amount of SKU reduction that we have makes it way more capital efficient. The overall model is super efficient. We rarely have any kind of churned metrics on our clouds.
Got it. Okay. If we go back to what you're doing with the third generation-
This acceleration of helping them win subscribers, do you see any other competitors that are doing this? What's the gap between you and anybody else in the market on this type of technology, this type of initiative?
We are by ourselves. It goes back to the years and years of investment that we've brought into the platform. We just didn't get started down this way. It starts with the operating system that we've built that truly allowed us to abstract all the functionality off the hardware and put it into the software layer. It's our clouds pulling all that first-person data off of the systems and using that information to help provide a better experience back. It's the managed services that get layered on top. So building that entire technology stack, we have evolved the company to where we're innovating on a 91-day base cadence. We've broken that old box ship model long ago. So in today's environment, where things are moving so fast because of AI, having an operating model where we can innovate on a 91-day cadence provides a strategic advantage to us.
We have 1,200 customers already on our platform, and that grows every day. Consequently, the rate of innovation is second to none. There's nobody that has the ability to innovate and bring new features and functions to the marketplace like Calix does. We are structurally just aligned to do that.
Let's talk about the current tone of the market. You raised your guidance for this year, I think to 15%-20%. Dycom recently said that they saw a pickup in deployments by a lot of customers. Are you seeing a broad-based pickup, or are you seeing it more concentrated with certain customers to the extent there's a pickup? How would you describe the overall demand environment?
The overall demand environment is strong across all tiers of customer. The raise in guidance for the year was broad-based, but it also has to do with the fact that we have increasing memory prices. Calix has elected to pass those memory costs on with no incremental margin stacking on top. Effectively, there's a de facto price increase, and that's giving rise to some of the revenue growth. It's not all associated with just the increase in memory pricing. You're seeing broad-based growth, and a lot of that has to do not only with growth within the existing subscriber customer base, but footprint extensions. We continue to expand our footprint, continue to take share.
This is a robust environment for us to continue to differentiate our offerings, because you're finally seeing that change in the marketplace where customers are realizing that speed doesn't sell. They need to focus on subscriber experience, and more so now than ever before.
On that topic, what would you say about Starlink? Which has obviously been getting a lot of attention. We were talking about it before as how much of the BEAD funding could they get, but now it's more of a, is there going to be an ubiquitous, easy-to-access commodity broadband out there for everyone? Why do they need these experience providers, I guess, would be the other question. Help us contextualize Starlink. On the other hand, is it playing into increased activity and part of the reason the market's strong? What's your view on Starlink?
We think it's a good thing.
Right.
It's actually a gift. Typically, service providers need a little bit of motivation to go do something, and typically it was competition from a fiber provider coming into town. Obviously they're down there looking left and looking right down the road and going, "Nope, I don't see anybody, no fiber providers today." They keep on doing the same old, same old. Now with Starlink and the attention that it's drawing today, everybody's talking about it. It's raised the level of activity as service providers are saying to themselves, "Okay, I do have this ubiquitous threat that's above in the sky. How do I make sure that I'm not impacted?" It's a focus on experience. Because at the end of the day, if I have a great experience, you're not calling Starlink. You're only calling Starlink if your experience is poor and/or you need a backup.
There are all kinds of other challenges with Starlink in terms of weather, capacity, and so forth. If you have the ability to move to a fiber line, you're going to have a fiber line service. You can still have fiber and be delivered a bad service, don't get me wrong. I have fiber here in San Jose, and I have to tell you, my experience is really poor. You still have to work on that subscriber experience and the Wi-Fi and everything else to make it a great experience, and that's what Calix is selling today. It's raised the bar on which service providers are looking at from that competitive threat. It's a great action for us. It really is helping us engage with our customers.
Getting a lot of good audience participation, a lot of good questions from the audience. I'm just going to go to this question. What's the split of the BEAD TAM between the customer prem equipment and the general PON? Let's just start with that quickly, and then we can move on to the meat of the question.
Yeah.
Yeah, go ahead.
Most of the BEAD is going to be on the access side, on the PON side.
Okay.
Unless they go into a one-box deployment model where the ONT is embedded with the Wi-Fi, it's going to be all access, all OLT and ONT. It does not cover Wi-Fi. That would be completely separate.
Okay. How much of the base have you passed through many pricing increases to so far, and what's been the reaction, and are they ordering more to get ahead of potential future increases, or are they slowing procurement in hopes of future price decreases?
It was across the board. In terms of who's been impacted, it's literally everybody. There's a few exceptions where we had some competitive takeouts, and obviously, we won them recently. It would be disingenuous of us to go back and change price on them. With limited exceptions, it's been around the whole base. You've got the emotional response to surcharges, as you would expect across the board. Some people are attuned to it and are sensitive and understand. Some are indignant and shocked. You've got a lot of emotion around the initial rollout. It's obviously something that impacts us all equally. Calix is no more affected or less affected than anybody else in the industry, and so we are rolling that through. We're trying to be, first and foremost, make sure that we have adequate supply, right?
For us, the most important thing is to make sure our customers have the ability to turn on subs when they get them. We do not want to be impacting our customers' businesses. First and foremost, our focus has been around acquiring the supply. Secondarily is how do you deal with the increased cost. We've been trying to do that in a partner-friendly way with our customers by not trying to put a margin on top of it. We've been just trying to pass that along to our customers. Albeit, it's zero margin revenue, has an impact on margins, but has really no impact on GP as you move forward. From buying behaviors, no, we haven't seen them front-run this, and nor would that really matter.
Like with COVID, with everybody was trying to hoard toilet paper, we have not seen that pull forward. If everybody tried to pull forward all of their purchases into the Q2, obviously, there's not the supply for us to do that anyway, and so it's still subject to our 12-week lead times. I would say it's still progressing in a very orderly fashion. If anything, they are buying smaller increments more quickly. Right? They're reducing their order size and just buying more frequently. That's kind of what we're seeing.
Got it. Okay, going back to what you said earlier on software services gross margins. You said, right, they were 55 last quarter, a little higher, and you said they could be in the low 60s this quarter, or the third quarter, I think. That's the third quarter, right? That number is, right?
This quarter.
Oh, this quarter. Are there further things to clean up in three Q and four Q? Because I think the prior peak was 65-ish, and then, of course, we want to see them go even higher than that. In terms of getting back to that two Q25 level, would we think about being there in the end of this year?
I think we can be there in the third quarter.
Okay.
Right. Build on top of that as we move forward and get back to the progression where the software margins continue to go up from there.
If that's happening and RPOs are moving in the right direction, that sounds pretty good. I guess from an overall gross margin perspective, right, as you've gotten more information, the hardware margins have come, the expectations have come down because of what the memory prices are doing. Has that stabilized where still the 100 basis points is right? It's not getting worse, or can you comment anything there?
Yeah. What I said was that we would have a 200 basis point headwind to the year due to memory. Obviously, given the guidance I had for Q2, it would imply that it's more back-end loaded. Right? The impact would be felt more in the third and the fourth quarters. I would say that memory prices have continued to increase. Our forecast is for them to continue to increase further. We have not seen any relief at the moment, so I would say that that trend's kind of continuing along those lines. I have no idea what it'll look like next year, other than I've got some high-level assumptions that the demand and supply dynamics start to normalize, and we'll start to see some relief in 2027.
If you take a look at the demand side of the equation, companies will go optimize their software stacks to minimize the memory footprints. We've done it. Right? We've taken some units that had two gigs, and we shrunk it down to one gig. We'll continue to look for ways to optimize the software stack and to reduce the memory footprint. Everyone else will do the same things. That's going to free up capacity. At the same time, on the supply side, the memory providers themselves talk about the capacity that's coming online. Some of it early in 2027, some other later in 2027, but you're going to bring more capacity to bear. I think you'll get to a new normal. When that new normal happens, we'll go ahead and adjust prices. That adjusting in prices might be just with our Wi-Fi 8 products.
You got Wi-Fi 8 right around the corner, right in the beginning of 2028. Consequently, it may be just a nice smooth transition where you just put the appropriate memory price into your Wi-Fi 8 products, and you let the surcharges for Wi-Fi 7 just sunset. If it happens faster, well then maybe we'll just do a price increase on 2027 and be done with it. You got to get to the new normal. Now we're not there yet, we're going to continue to treat that as a surcharge as we move forward.
Great. For now, the 200 basis point headwind though, that's still the guidance. For the year, gross margin's down about 100 basis points, total gross margins, right? That's what you said. That's.
That's the guide. Then, it sounds like it's sort of too early to call what would happen next year with that, with total gross margins, we can't really say yet.
Yeah. You tell me, and I can then tell you what it means.
All right.
Yeah.
The reason I kind of stacked the, is memory prices dropping, causing any pull-ins? I put that as the third question in my sort of, it seems like there's a lot of customer activity. The market is strong, there's increased activity, and this is all before we get to BEAD. Maybe just touch again on how busy you guys are responding to customers right now. The next question beyond that is, does BEAD still look like a go for 2027? What's going on behind the scenes? Investors haven't gotten a lot of solid BEAD news recently. It seems quiet on this side.
I don't think I'm going to be the guy that gives you new BEAD news.
I'll be just consistent with what I've said. I think our BEAD revenue in 2026 is safe. We said low tens for the year. I think that's still true. I don't think that's getting better or getting worse. I still set the stage for next year at BEAD being in the high tens of millions of dollars. I think that's still a reasonable assumption at this point. A lot of activity is happening. As we get into the fall, our customers will start planning their builds for 2027. That number then will firm up. As we start progressing into the back half of the year, we'll get a much bigger idea of what BEAD is. The best way to characterize BEAD, though, is that Calix will do well, whenever it happens. We're not counting on it for our growth. We're not dependent on it.
I look at that as it'll be a tail breeze to a tailwind at some point. It feeds access. It'll bring the opportunity to add more subscribers, which is ultimately what we're after. Customers aren't sitting around not building, right? They're building something. The construction teams are out there working. The same thing with BEAD, it's not 100% additive. To say it would be 100% additive, you would have to have construction crews sitting around waiting, and then I'm going to get going. They're not waiting. They're building this part of the network. They're doing this. When the money's all lined up and everything's ready to go, they'll pivot and start doing more BEAD. Great. Beautiful. The only way you grow the business is obviously bringing on capacity in terms of construction and jobs and get that going.
Look to Dycom and others like that to kind of lead you on how fast this whole industry can grow, as a result of Broadband Equity, Access, and Deployment. Teams are not just sitting around waiting. They're doing work, and they'll continue to do work.
Okay. You touched on the Tier 1 opportunity earlier. I wonder if there's anything else to say about Tier 1, but then also if you could dive a bit into MDU and talk about what that means for your business, basically.
Yeah. We talked about MDU at our investor day. It's an area of focus for us because we're pretty excited about that market opportunity. The TAM is about $10 billion. We like the composition of that marketplace. It looks very similar to our regional business. It's highly fragmented. There's thousands of MSPs across the country. The way these MSPs have been going about it is taking enterprise class solutions and trying to fit them into MDU and cobbling that technology together. We think we have a much more elegant package of taking everything that we've built so far, our SmartTown, SmartHome, SmartBiz, wrapping that into what effectively is SmartMDU in a managed way, coming from a consumer orientation. We've eliminated all that complexity. We think the market is ripe for disruption, and we can come at it that way.
Love that marketplace. Looking to do more with it. Very similar go-to-market strategy in terms of sales engagement. It has the added benefit of being served by large telcos. Half that market is served by large telcos, and what we're being told by a lot of these North American telcos is that they would like help. They would like to see what we can do with them, help them with MDU as a solution. We're pleasantly surprised by that and look forward to the opportunity of serving them, because this is still a land and expand strategy. Once you've adopted us for some solution. Whether it be MDU, small business, the technology stack is now embedded. Anything you want to do after that point is easy.
You can do things in days and weeks as opposed to kick off another $20 million IT project to go do something else. We look forward to it. We look for any way for us to penetrate a large customer, and if that happens to be MDU, fantastic, love it. MDU, we're super excited about. We think we have the ability to double the size of the business through MDU, time to get to work.
Great. Can you clarify for me or help me understand how the third-generation product opens up the international market in a way that the second-generation product you just kind of couldn't? Because I feel like that kind of speaks to everything can move faster with AI and with Agentic and the fact that you can actually now cover the international market and support them. Just help me understand that a little bit better.
Part of the re-architecture with the third-generation platform was to create that private instance in a cost-effective manner. Prior to it, the way we had architected our clouds didn't enable that. By creating that private instance, we can serve large customers. Because you're creating a single tenancy, you're not mixing their data with the rest of our cloud data. That was the whole piece of that. Large customers domestically and internationally, because a lot of those countries are served by a single large customer. It's not the same kind of marketplace here in the U.S. where you have lots of little service providers. It's usually dominated by a single large communication provider. Going into it, you need to be able to serve that large customer, and we now have the ability to do with our new architected clouds.
I'm sure you must have gotten this question, but the people who kind of have that cloud thesis, which is, couldn't all these customers just do it themselves? Have spend a few tokens and tell a different AI agent to write the code for whatever they need. Basically, the moat around the value of your agents versus customers that, I feel like this is kind of a very hand-wavy argument. Oh, yeah, people just can do this themselves. I don't really agree with that, but I'd love to get your view.
Part of having the agents and having them to be able to behave autonomously, you have to provide them the context, and you have to put around security layers so that you don't have hallucinations, you don't have them act in a way that would be destructive to your network. Otherwise, you're never going to go deploy them inside of a network. That's what we've spent the last year and a half building, is taking the network, taking advantage of A2A and MCP, so that you can have the data coming together that allows our agents, which have been trained on data that we've collected over the last seven, 10 years on our 1,200 customers, so that now they have trained agents that are now being deployed to customer-specific data, off they go.
It's a level of infrastructure and security that you have by us having put it together. That's what you're getting out of it. Could you go do it on your own? Sure. That's what they're going to go do. Large customers, that's what they do. They build it themselves. We think we have the ability to affect that buy-build decision and give them an opportunity to buy something as opposed to build it themselves. That's the opportunity we have in front of us with taking our Agentic workflows into large customers.
Great. We're just about out of time. Just any wrap-up statement in terms of. To me, if software and services margins are coming back and RPOs are accelerating, this is a hugely attractive stock at this price right here. You have to go run out and buy it. What do you want to say to investors in the last 30 seconds?
I couldn't be more excited about the opportunity ahead. Our core markets are growing and are strong. We are in a dominant position. The competitive landscape is kind of melting away around us, we now have the opportunity to take what we do very well and go into MDU and into large customers. Couldn't be excited about the different growth vectors we have for the company, look forward to just doing the hard work to get us there. Thank you for your support and your attention, we'll talk to you soon.
Thank you, Cory. Thanks for doing this. I enjoyed the conversation. Thanks for everybody out there and for all your questions.