Greetings. Welcome to the Calix Second Quarter 2025 Earnings Conference Call. [Operator Instructions] Please note, this conference is being ordered. I will now turn the conference over to Nancy Fazioli, VP of Investor Relations. Thank you. You may begin. Nancy Fa

Thank you, Darryl, and good morning, everyone. Thank you for joining our second quarter 2025 earnings call. Today on the call, we have President and CEO, Michael Weening, and Chief Financial Officer, Cory Sindelar. As a reminder, yesterday, after the market closed, Calix issued a news release, which was furnished on a Form 8-K, along with our stockholder letter and also posted in the Investor Relations section of the Calix website. Today's conference call will be available for webcast replay in the Investor Relations section of our website. Before I turn the call over to Michael for his opening remarks, I want to remind everyone that on this call, we will refer to forward- looking statements, including all statements the company will make about its future financial and operating performance, growth strategy and market outlook, and that actual results may differ materially from those contemplated by these forward-looking statements. Factors that could cause actual results and trends to differ materially are set forth in the second quarter 2025 letter to stockholders and in the annual and quarterly reports filed with the SEC. Calix assumes no obligation to update any forward-looking statements, which speak only as of their respective dates. Also in this conference call, we will discuss both GAAP and non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures is included in the second quarter 2025 letter to stockholders. Unless otherwise stated, all financial information referenced on this call will be non-GAAP. With that, Michael, please go ahead.

Thank you, Nancy. As I stated in our last earnings call, the investments we made in 2024 to manage through the post-pandemic period are yielding results and the market is evolving as we predicted. Cory will cover our exceptional second quarter, which is a testament to our long-term strategy, our Calix team, the incredible customers we serve and the partners who support us. As I continue to state, the industry is at a crossroads. A broadband provider must decide if they will retain the legacy mindset of a speed-based network operator who tries to cost cut their way to growth while being commoditized. As our homes passed is not a measure of success as it does not guarantee a revenue-generating subscriber, while the successful broadband provider is the one that delivers incredible experiences across all segments, thereby establishing a brand that is loved by the communities they serve. The result is higher revenue per customer across all segments: residential, business and municipality, higher Net Promoter Scores, which yield customer loyalty and lower churn. That experience-based success is what Calix enabled as we always think subscriber in, not network out. Our mindset is grounded in the experiences beyond the undifferentiated speed of a pipe via our unique appliance-based platform, cloud and managed services model that helps our customers transform with the support of our industry-leading customer success team. Our press releases frequently highlight our customers' ability to use differentiated experiences to deliver value for members and investors as evidenced by a discussion I had with a medium-sized customer last month. That customer launched SmartBiz in late 2024 and saw a 250% increase in revenue per small business subscriber by delivering a better experience. Experiences are the future of broadband, and our platform is unique in that it can address residential, small business, MDU and municipal needs with the same appliances cloud and managed services, which brings me to the second even larger disruption that the world is paying attention to, artificial intelligence. Since late 2023, our team has believed that the long-term impact of AI is significant and represented the next critical component of the Calix platform. More importantly, we have ensured that our teams do not underestimate the pace of change. This is not a normal technology curve. The pace of change in AI is staggering. The difference between short term and long term is going to be radically shorter than any technology before it, as evidenced by the fact that Netflix took a decade to get to 100 million subscribers, while ChatGPT took 2 months, is now approaching 1 billion users. It is useful to note that both Netflix and ChatGPT stand on the shoulders of the platform known as the Internet. Since 2007, we have been investing in our platform, cloud and managed services to transform the entire broadband market. In 2016, we launched our second-generation platform, which introduced several key components, 2 operating systems for network and premises appliances that are extracted from the chips, enable local applications with rich telemetry and policy management capabilities. We also introduced our persona-based cloud and the managed services model, which now serves residential, small business, MDU and municipal use cases on the same appliances. We have invested more than 15 years and almost $2 billion into our platform. As of the second quarter, we have enabled 1,116 broadband providers to deliver a differentiated experience to ensure their brand is front and center as they delight the communities they serve, including a new large cloud customer who selected us in the quarter. In late 2023, we recognized the emergence of AI would be an incredible opportunity to address the largest constraint that our customers face, the capacity to transform across operations, marketing and service. To meet that need, we began investing in our third generation of the platform. And in second quarter, it went into preproduction for a second half launch in 2025. Beginning with the upcoming release of our third generation mobile application in August, CommandIQ, which is a valuable brand portal to end subscribers. The third-generation Calix platform has 3 goals: First, we expand our platform, cloud and managed services to allow us to meet the needs of local geographies through sovereign data centers. Second, we gain the ability to serve large customers with private clouds. Last and most important, we evolve our platform architecture to speed our capabilities with agentic AI across our solutions to allow the Calix team to move from success advice that a customer may or may not implement to success advice enabled through execution capacity with the legion of Calix AI agents. While we remain the only organization in this industry with a substantial investment and customer success teams to support transformation, it does not overcome the very real capacity challenge that our customers face. For example, many leaders prioritize new installs over adding a new experience campaign such as outdoor Wi-Fi, despite the very real truth that this campaign could add $10 to $50 of incremental revenue per month per subscriber and is wildly sticky, which reduces churn. They do not have the marketing capacity to build and execute campaigns nor installer capacity as they have not embraced the high satisfaction driving self-install model. Agentic AI changes that as it will enable our customers to move faster, a force multiplier for action. Calix AI agents learning across our unique end-to-end platform and aligned with our customer success teams will allow a capacity-constrained customer to LEAP over the case. Our fast-growing legion of agents will speed opportunities to simplify, which improves margin; innovate, which increases revenue and reduces churn; and grow to meet the financial goals of our customers, members and investors. In short, more than 15 years of investment, building a unique end-to-end platform does more than enable an incredible second quarter. We are poised to enable an industry-wide transformation that we have always envisioned for all broadband providers regardless of size or geography. Platform- based Agentic AI will enable network operators to become experience providers that dominate the markets they serve as the concept of customer success moves from advice that a BXP needs to prioritize and build capacity to implement to the push of a button by our BXP team member to approve the actions of a quickly expanding and evolving legion of Calix agents. With that, I'll hand it over to Cory who will cover our second quarter financial performance and third quarter outlook and ongoing investments to transform and lead the broadband industry. Cory, over to you.

Thank you, Michael. We saw a very strong and broad-based demand environment during the second quarter, which allowed us to deliver revenue of $242 million, which represented 10% sequential quarterly revenue growth. Our record RPOs grew 2% sequentially to $347 million and increased 30% year-over-year. Our current RPOs were $134 million, up 5% sequentially and up 30% year-over- year. This metric is a strong indicator of the strength we are seeing from our platform, cloud and managed services model. This strength led to another quarter of record gross -- non-GAAP gross margin of 56.8%, representing a 60 basis point sequential increase and is related to customer mix and our BXP customers winning new subscribers as they continue the adoption of our platform. In the second quarter, we added 18 new BXP customers that were largely competitive takeaways as we continue to focus on landing new footprint. Our balance sheet metrics remained outstanding. We marked our fifth year of quarterly free cash flow and generated record free cash flow of $36 million in the quarter, our ninth consecutive quarter generating 8-digit free cash flow. We ended the second quarter with record cash and investments of $299 million, even after utilizing $33 million for share repurchases during the second quarter. DSO was a record 24 days, down 6 days sequentially and down 14 days from a year ago. Inventory turns was 3.4%, down from 3.6% in the first quarter. As we noted last quarter, we have a diversified supply chain and manufacturing presence. The data and direct relationship we have with our customers, combined with our strong balance sheet, allows us to make intelligent investments in critical areas such as component inventory and incremental finished goods, thereby ensuring supply for our customers. So far this year, the impact by tariffs has been minimal. And if the situation in this dynamic environment changes, we will do our best to minimize the impact to our customers. Moving to guidance. Given the broad-based demand picture, we believe we can continue to grow sequentially even with the big step up from this quarter, specifically for the third quarter of 2025. Our revenue outlook is between $243 million and $249 million, which at the midpoint would represent a 2% sequential increase in revenue. Our non-GAAP gross margin guidance at the midpoint would represent a slight increase from the second quarter and reflects our expectations regarding customer and product mix. For 2025, we anticipate annual gross margin improvement will be at the higher end of our target financial model from 100 basis points to 200 basis points. And regarding non-GAAP operating expenses, we continue to restrain our OpEx investments until we are back into our target financial model. That said, we expect a slight increase in the third quarter as we made some incremental investments in sales and marketing. However, as a percentage of revenue, operating expenses will continue to decline as our revenue grows each quarter. Michael, back to you.

Thanks, Cory. Nine years ago, I joined Calix because Carl Russo painted a vision, where Calix would transform from a network system company into a platform company that would get ahead of the disruption he saw in 2007, the end of legacy network operators and the rise of broadband experience providers. With agentic AI on our unique end-to-end platform, an important piece falls in place, the ability to automate action to drive the success of our broadband experience customers as they expand across residential, business and municipal in the communities they serve. This next step will see Calix leverage our platform in more than 15 years of investment to evolve into an AI as a service platform company. I'm excited to lead the team forward at this truly amazing time. In closing, I'd like to thank our team, our customers, our partners and investors whose passion, grit over 15 years, trust and partnership have brought us to this exciting next stage in the Calix journey. Nancy, let's open the call.

Darryl, we are ready to take some questions.

[Operator Instructions] Our first questions come from the line of Scott Searle with ROTH Capital Partners.

Great job on the quarter and the outlook. Just a quick clarification. I'm not sure if I heard a normalized number as you reclassified one of the customers from a smaller to a larger customer. I wonder what the growth was sequentially for the smaller customers on an organic basis without that was being adjusted. And then Mike, maybe to dive in on the agentic front. I'm wondering if you could talk a little bit more detailed in terms of the impact from an OpEx standpoint. It sounds like you're going to keep that pretty constrained in the near term until you get to some target operating margin levels? But how do you expect cost to trend on that front? And the actual impact then from an ARPU standpoint, as you start to look at the customer base and how you monetize that? I know it's going to come from improvements in general in terms of your efficiency. But in terms of rollout of new services and adoption, are we starting to think about ARPU levels from value-added services at a higher level as you start to implement agenetic on a more broad-based basis?

Yes, Scott. We haven't really specified what that impact is, but it's mid-single digits.

Now with regards to -- okay, you asked a lot of questions in there. So I'm going to -- but I wrote them down, so I'll parse them out. So the first one is that you asked about what's the impact of cost. I'm assuming the question you're asking is internally? Or are you asking the cost with regards to our investment in artificial intelligence?

Sorry about that, Mike. Your investment, right, in terms of what's going to happen to your R&D and OpEx, does it start to require a little bit more of an inflection as you start to move down that path.

Good point, okay. Good question. So with regards to investments, so have a model that we've identified, and Cory mentioned now we've constrained OpEx with regards to it because until we get our revenue to a certain point and Cory, why don't you take few seconds to remind all investors about our model with regards to R&D, it is?

29% of gross profit.

Correct. So as we think about it, sure there's going to be opportunities where we have to go a little bit faster because the transformation, the rate of change is going on with AI is something we've never seen before. But you have to understand that what we're doing with artificial intelligence is not a new thing that we plunk on top of a legacy -- as a legacy company that we plunk it on top and hope that AI does something magic. You have to recognize the fact that this is the third generation of our platform. We are building on a significant strength that allows us to drop this capability in place and really accelerate what we're doing for our customers. And the best way to think about the times and how things are changing is that Carl started to vision this back in 2007. But it took us to 2019 to launch our -- and so we started to play with it. That was the first generation, then in 2019, we launched the second generation. That basically took us over the last 6 years. And now we're going into the third generation, which has had a development cycle of like 18, 24 months. So the pace of change here, you can see that the investments over time that we have made are making us much more efficient at each subsequent state. And so as we look at this, it really drops onto our platform in a very unique way that allows us to help our customers monetize the opportunity. And that opportunity is transforming from a network operator into a broadband experience provider without investing a ton of capital. So when you ask the question around how does this monetize out? For our customers, it monetizes out in the form of I am -- as I stated in my opening comments, I want to launch a new campaign, but I don't have the marketing capacity to even do it. And now all of a sudden, this capability pops up in our cloud that allows our organization to go to one that was coaching our customers on how to do micro-based segmentation, how to build campaigns, how to win that new subscriber to push the button and the campaign launch. It's not only does the agentic AI, one agent will build out the segmentation, the next will then do propensity to buy. The next will select what are the right social media opportunities for them to advertise on. The next agenetic agent will use A to A to reach out into HubSpot and actually run the campaign. The next agent will grab the response back and then present it to the end user inside the broadband company and say, here was the campaign that we launched. And so in that scenario, we go from people who are running for broadband, these broad-based campaigns to truly what we've always talked about, which is small micro segmented campaigns that frankly were too difficult to run because if you run a really good marketing campaign in broadband, it should cost you dollars, not hundreds of thousands. So for example, I should be able to micro segment down to 50 or 75 customers and run a $3 social media campaign and get an ROI on that through upsell, cross-sell or net new subscribers. But then I want to run thousands of those, and that just doesn't scale because I can't hire the employees to do it. So when we talk about the monetization, that helps our customers radically transform how they operate today, which then -- and as we've always stated, we only make money when our customers make money. And so it helps them do the things as we go faster. They win more subscribers. They sell it at a higher ARPU by doing higher attach than they cross-sell, upsell. So all of these become force multipliers, which is something that Carl and I have frequently talked about. In fact, he sent me an article about 2 months ago, and he talked about how private equity is shifting into looking at legacy businesses and investing and saying, if I drop AI into that legacy business, can I get this massive multiplying impact on that business and in essence, finding undervalued assets using artificial intelligence to turn them into radically higher value businesses. And frankly, that's been our thesis right from 15 years ago, that broadband is undervalued, that it is a business model that is inelastic that once they make the infrastructure investment, that there's a huge monetization opportunity on top of it. And what this allows us to do with the third generation of our platform is help our customers make a lot more money faster and crush their competition, and we get a portion of that. So that's how it all comes out.

Mike, then just to follow-up, and I'll get back in the queue. You're accelerating your customers' time to market in terms of their ability to deploy value-added services and campaigns. How quickly can that get deployed then in the customer base with the third-generation platform? And when do we start to see that, I guess, ramping up as part of your current RPOs?

Yes, so that's 2026. And the platform rolls out, as I said, it's in preproduction right now. We put it into preproduction in second quarter. We have the first component of that launching in August, which is our third generation mobile app, which actually there's multiple iterations of that. There's a consumer mobile app, there's a small business mobile app and then there's installer. And we just showed it to 30-plus CEOs 3 or 4 weeks ago and walked them through where we're going with the mobile app to support a brand portal, frankly, for customers. And they were blown away because all the things they've been looking for us to do are implemented in that mobile app. And on top of that, that becomes the portal upon which AI has a profound impact on the end subscriber. So we see that impacting through 2026. And then the other part of that is, as I said in my opening remarks, there's 3 facets: One is helping our existing customers succeed. The other two are facets of our platform, which we have been very thoughtful about what markets we expand to. This evolution of -- to this third generation also enables us to do private cloud for large customers and allows us to eliminate the geographic constraints that we have because privacy and data control and sovereignty of data has been a significant constraint with us with regards to our ability to go into new markets. And we will now be able to set up sovereign data centers and eliminate that issue as we go through 2026. Long answer for a short question.

Our next questions come from the line of Samik Chatterjee with JPMorgan.

Congrats on the quarter and the strong RPO. Maybe if I can start on the RPO growth that you had quite robust at 30 plus but I'm just trying to sort of match this up relative to the revenue growth acceleration that you've had this quarter. The RPO growth, in fact, decelerated modestly, while still at a healthy level. Was there something different in terms of attach of platform services, et cetera, this quarter that would explain why probably the RPO numbers are a bit lower compared to the last quarter while your revenue growth, in essence, is quite significantly better than the last quarter? And I have a follow-up.

So it's worth reviewing again how RPOs work. So with regards to RPOs, when we sign a contract with a customer, we actually go through and we'll sign them up for a minimum. And I'm just going to use really simple numbers. So we sign them up for a minimum of 1,000 subscribers a month. And then they would sign up for a 3-year contract. And let's say their growth is, they actually do 1,500 subscribers, 1,600 subscribers, 1,700 subscribers. Those incremental 600 or 700 subscribers per month actually dropped straight to revenue. They don't show up in RPOs until the customer comes up for a renewal, so they can either renew at the end of that 3-year contract or if their growth is significant, as we saw in some of our lumpy RPO numbers where you see big jumps where the customer would say, I want to renegotiate because I'm twice my volumes, therefore, I would like a better per month subscriber charge. And so for us as a company, we're not going to push for a renewal because the -- why would we not allow just to drop to revenue, we're not focused just on driving RPO growth. So that negotiation of the contract is what -- then they go to 2,000 subscribers as the minimum over 3 years, and that bumps up the RPO number. So that's why you see it's lumpy, but that also is why RPOs are a portion of the growth that we see in our cloud.

Okay. In that sense -- sorry, go ahead.

In the quarter, the acceleration that you saw really was on the appliance side. As you know, a large number of legacy vendors in our industry have been challenged. We anticipated that this would present us an opportunity to land new footprint, and it has. What we didn't anticipate was the speed at which these new customers would want to roll out Calix appliances.

Okay. Great. And maybe for my follow-up, if I can get a quick update on the current supply situation that you have? I think you had updated us that you're looking to move the supply capacity to Mexico. But what's been the progress on that front? Any update on that?

I mean at this point, we are, knock on wood, in a stable environment. We are okay where we're at from a manufacturing process perspective. And so we'll continue to build that. It will take time. So like we said, it feels like anywhere from 9 to 18 months to complete having the capacity in multiple locations. But at the moment, there is no need to make any real changes. We're okay as it is. So unless something changes, we're going to continue to do what we're doing with no real big modifications to our manufacturing footprint.

Our next question comes from the line of Michael Genovese with Rosenblatt Securities.

I just understand on this third-generation platform. This is just a software update, correct? I mean do the appliances change at all? Or do you just press the button and then everybody has the third-generation platform?

Yes, I know it's exactly that. So it's a cloud upgrade and like it's a transformation of the architecture of the cloud. And yes, everybody gets it, it actually goes back against all systems. That's the value when I talk about the 2 operating systems. We have a network operating system that's unique in the industry. It allows customers to virtualize their networks, so access, aggregation, subscriber management and other provider edge functions into a single platform. And then we have our operating system that sits on top of a WiFi appliance, but we're agnostic to the hardware. So for sure, it's just press the button.

Sounds good. And then the other two things I want to ask on quickly. Last quarter, there was some pull-ins with a Tier 1 customer, Tier 1 category was strong this quarter. So was that customer better than you thought it would be in 2Q? And I mean I'll just ask my other question about BEAD. What's the confidence that BEAD is going to be a -- I mean, obviously, at this point, I don't care at all, but I just want to get your opinion, a confidence level that it will be a significant program to close the digital provider in this country or not -- end up being that or not.

So on your first question, what's your first question?

It's the Tier 1 pull-in.

Yes, pull-in. It's not strength from that customer. It was largely reclassification from the small customer segment to the large customer segment. So when you factor that in, that was -- the actual large customer strength in the first quarter is actually down in the second quarter, so that's kind of predicted. But you had the strength across the entire broad base that just kind of pull that up, right? As customers continue to cross the chasm and add more subscribers, that's broad-based, and we're seeing it across our customers.

Second one is BEAD.

On the BEAD thing -- so in terms of BEAD, there's really no update. It's still in the state of flux, as you know. They're going through a rebidding process at the state level. I'm not going to speculate on when that gets itself sorted out. I think it will take time. What's important to know is that it's not in our numbers. And when it ever eventually happens, which is likely later than you think it's going to be, we'll do well.

It's the same thing as we always say. I always quote Carl on this, right, our Chairman, which is it's always going to take way plenty longer than you ever thought it would. When it does arrive, it will take longer to roll out and it will be bigger than you ever expected. We do believe that this program will go forward, but we don't have it in our numbers because it's a political program. And the good thing is that at its core, it's a bipartisan goal because the country is filled with red and blue states, right? So those people who want to get elected or reelected in the midterms definitely wants something to happen. But this process has been a bit more arduous than others. Although there's lots of people who would argue this is exactly as arduous and as it was before. So the good thing is not in our numbers. Yes, it will come at some point, and when it does, we'll be ready to -- well, our customers will take advantage of it, and we continue to engage with them and support their rebids.

Our next questions come from the line of George Notter with Wolfe Research.

I just wanted to go back to the agentic AI discussion. I'm just curious about what elements of this are here to benefit the end subscribers? From what you've described, it sounds like it's more about operational benefits for the service provider. I'm wondering if there's also new offerings, new capabilities, new ARPU-enhancing things for the end subscribers that drive this as well.

Yes, that was my initial statement is that the first iteration of this for the first component that drops into place is the mobile application, which is a brand for the end subscriber. It's a brand portal for them, for our customers to get their brand out to the end subscriber. And within that, that's where a lot of these new incremental capabilities will show up. The agentic AI capabilities for an end consumer subscriber are going to be around cyber capabilities, enhancements with regards to performance and analysis of everything that's going on in the home to provide a better capability. We evolved from machine learning to agentic AI and things like troubleshooting. And I'm sitting by the pool, my laptop is not working so, well, what's going on, and then it'll come back and identify it. So yes, there's lots of opportunities. And then it does, as I said, expands out into -- there's really 3 elements to a broadband business: Operations, which drives down -- which drives efficiencies and improve profitability; innovation, which drives new subscribers through partnering -- sorry, innovation drives new subscribers through new capabilities that attract new customers. And I gave you the example where we had a customer put out SmartBiz and the 2.5x revenue, it was incredible per subscriber. And so yes, there will be all kinds of customer-facing, subscriber facing and customer enhancement.

Got it. And then just as a quick follow-up. Was there any pull forward do you think from your customers around tariffs? I was just looking at how strong the hardware business looked in the quarter. Any benefits there?

No. We work closely with our customers with regards to like this concept of pull forward. I shouldn't have used that word because it's not true. What we do is we work really closely with our customers on how they manage their inventories and then we just move things around to help them meet the demands of their subscribers.

Our next questions come from the line of Christian Schwab with Craig-Hallum.

First, congrats on the better results and the material earnings leverage. My question has to do with given the better-than-expected results here this year as well as what appears to be improved visibility, as I look to 2026, last year, you guys talked about being able to drive double-digit top line growth. Should we assume that still is the expectation, given the better-than-expected results in the near term here in '25?

Yes, that is correct. We still think we can drive a double-digit growth next year and improving margin and cash.

And improving margin. I got on the gross margin point, but thanks for reiterating that. My last question then, as we move to the different layers of agenetic AI applications and helping your customers over -- now I understand it's kind of being rolled out here and kind of finalized throughout 2026. But on a multiyear kind of CAGR basis, would you expect this to have a positive impact on your long-term growth rate? And if it does, could you give an aspirational expectation?

No aspirational expectations. But yes, because, as I stated, there's 3 elements to what agentic AI and our third generation platform do for Calix. The first one is our existing base. It allows us to help them add new subscribers at a faster rate. And as per the previous question, that comes out from a customer benefit in 2 ways: one, that they can radically improve operating costs and therefore, their own margins, which is important. And then also win new subscribers and selling at higher ARPU. So that's the first part. So that's our existing base. So we continue to see -- we have strength in that base, and we'll continue to dominate those markets and grow with our customers. But the other key element that I identified is that the growth of incremental segments. So we announced that we launched MDU. We have a goal to move into the medium-sized business segments. And then agentic allows us to eliminate the geographic constraints that we have with regards to our clouds due to data sovereignty and privacy rules and move into new markets. And then on top of that, we've never dealt with the very large in our industry primarily because of the fact that while we won a large customer this quarter, who bought our existing business, most of the others actually have a different business model. And so part of the evolution of our architecture in this third generation was also to allow private instances big companies. So those are TAM expansions.

Fantastic. And then if I could sneak in one last question. Your commentary about security and software and cloud and being in a position to expand into international markets with a stronger presence. Can you just elaborate on what that means? Does that mean that you would expect over time to have a stronger presence in, say, Europe, for example, than you currently have? I guess that wasn't clear to me, I'm sorry.

So the constraint that we have is that if you look at the rules with regards to -- data sovereignty is something that most governments have a lot of -- care a lot about it, right? And if we look at some of the polarization of what's going on politically across the world, one could, I think, safely hypothesize that data privacy and sovereignty is going to become more and more important by a company -- or by country. And so in the past, we've had 2 data centers, 1 in the United States -- two data instances, I guess, but not truly just data centers, but 1 in the United States and 1 in Canada, which has served like U.K. and places like that. We now have the capability with this third- generation platform to leveraging our cloud partner to set up instances in a local level by country, if necessary. So not necessarily in the EU, although if there's fragmentation in the EU that might become possible. But we look at -- this whole conversation with regards to where we go with this back in late 2023 started with, we looked at the United States, which is a very state-based country, and we hypothesized out that, hey, what would happen if every state started to put data privacy rules in place and wanted sovereignty in a state, which is not frankly -- that could happen. And so we started down this path in 2023, investing almost $100 million to make this happen. And that's the point of this. So we can set it up in the EU. We can set it up in the Middle Eastern country and not have to worry about the data privacy. That has frankly held us back in a lot of markets. And now is the time.

Our next questions come from the line of Tim Savageaux with Northland Capital Markets.

My congrats on the results and outlook as well. A question about some specific customer segments, even correcting for the reclass. Your large carrier revenue is up pretty good over what you saw last year. And that's true on the medium side as well. I wanted to kind of get your view on a couple of dynamics driving that. We heard from Verizon yesterday that they're ramping up their fiber build per their plan. And I assume that's part of it and also have a really second half loaded CapEx plan for the year. Of course, CityFibre just raised a bunch of money. I assume that's part of what's driving your dynamics near term. I wonder how both of those situations might affect your outlook for the second half. It seems like there's a lot of tailwinds there in those categories. And I'd be interested in your thoughts.

Tim, we're not going to get down into the customer-specific details of what's kind of happening at each level. Again, I'll come back to the demand environment we're seeing is broad-based. So we're seeing strength across the board. The revenue from those large customers that you've seen price in CityFibre and so forth are inherently lumpy and they come back and forth. And so we don't comment on kind of the quarter-to-quarter fluctuations. We just know that the demand environment is broad-based. And even with this large step-up in the second quarter, we can continue to grow sequentially from here. So it's not relying on either of those two customers.

And with regards to demand, I'm going to go back to the smart broadband providers making bets on who are their partners that are going to walk them through the significant disruption that the entire -- every industry faces. The power of what's going to happen with artificial intelligence is frankly underhyped. This is going to -- there's a great Ted talk on YouTube with Schwartz, the Google CEO, the ex-Google CEO.

Schmidt.

Schmidt, and he basically talks about how everybody has been hyping it out, but it's massively underhyped. The impact of what's about to happen over the next couple of years is frankly larger than industrialization. There's not a faceted society that's not going to be impacted in some way shape or form. It's profound. And anyone who thinks that it's kind of status quo is completely wrong. And the winners and the losers are going to be completely different. And in fact, the ones who we think is winners and which AI model is buying today are not the ones that are going to be in the future. And that's kind of what we actually really have to think about is that who are the companies, and back to that hypothesis, that article that I talked about a little bit earlier, that Carl and I were talking about is that there's these private -- these investors out there, who are looking at legacy companies and saying, who are the companies who are poised to take a legacy industry or an infrastructure play that has a certain valuation, pop AI on top of that, execute effectively because in the end, there's so much BS out there. It's about execution because there's so much hype in nonsense, it's about those who can execute through this opportunity, and those are going to be the winners. And when we had the 30-plus CEOs together, we finally unveiled to them exactly what we're doing. And there was a collective sigh of relief, frankly, as they left because they were all reading carefully about what's going on with agentic AI and they're saying, do I need to go build this out myself, who are the right partners, and what they all left with was Calix has got this. And the platform that we have bet on the second-generation platform since 2019 has been the right bet. It's helped us grow our business, but this next one is so transformative that I need to be at the forefront or I'm worried that my company is not going to succeed. And they all left, like I said, with that collective sigh relief saying, wow, Calix has got this. And we got this. And we're uniquely positioned because we've been doing it for 15 years. So when you want to talk about demand, this isn't something that we woke up and said, "Hey, on our crappy boxes, we should go pop an AI engine on it and see if it can do some autonomous networking." We systematically rebuilt everything that we do from a platform top to bottom, starting in 2019 with the launch of our 2 operating systems. We've evolved on it for 6 years. And over the last 18 to 24 months, we've invested significantly to update that architecture for AI. So with regards to where growth is going, I have to be really explicit. This is what I joined 9 years ago for, to do this, and it's all been building up to this point. And as I said, with regards to where the growth opportunities are, we're no longer just in our existing businesses and the constraints that we had with regards to geography, sovereignty and size of customer, frankly, who will demand a dedicated instance inside their own private cloud are no longer going to be constraints. And so with regards to where demand is going, this is the next big step for us.

If I can just do a quick follow-up here. Michael, I think you mentioned also with a large cloud win with a 2 large customer or...

Brand new large customer.

I think you mentioned a new win, I'd love to get some more color on that. Is this a net new customer, whatever you can tell us.

Yes, brand new customer who's a very large, very large customer. And so they became a cloud-only customer, which is great. They didn't -- they haven't been buying other technology and they bought our cloud. So for us, I think that's our second-generation platform. But they know where we're going. And I think that's a good indicator of what's coming.

Our next questions come from the line of Ryan Koontz with Needham & Company.

Just to clarify what Tim was asking out there, this large customer win is not related to the reclassification of small to large you talked about or is it?

No comment.

Okay. Fair. The great gross margins here and especially with the strength of market in medium and large obviously, a strong appliance shipping quarter 2. So what are the mechanics behind that, Cory, in terms of your continued gross margin expansion? Is this purely software mix, anything going on in terms of hardware mix we should be aware of driving margins?

Yes, it is primarily just the continued adoption of the platforms and the growth that you're seeing there, and a little bit of favorability on customer mix.

Got it. And lastly, great DSOs there. It sounds like the quarter was done practically before you started. On the small customer growth rate there, maybe a little below expectations from investors. Do you expect that to improve as we go through the second half of the year, small customer cohort, at least on an organic basis?

For sure. I mean, again, my comments are that demand is broad-based, and we're seeing strength across the board. And so even with the reclassification of one small customer to large had the impact of still growing quarter-on-quarter, albeit at a more muted rate, but still, it's growing and it's broad based.

Thank you. This now concludes our question-and-answer session. I would like to turn the floor back over to Nancy Fazioli for closing comments.

Thank you, Darryl. Calix will participate in several investor events during the third quarter. Information about these events, including dates and times and publicly available webcast, will be posted on the Events and Presentations page of the Investor Relations section of calix.com. Once again, thank you to everyone on this call and webcast for your interest in Calix and for joining us. This concludes our conference call. Have a good day.

Thank you. Ladies and gentlemen, thank you for your participation. This does conclude today's teleconference. Please disconnect your lines at this time, and have a wonderful day.