管理層發言
Greetings, everyone, and welcome to the Calix Third Quarter 2025 Earnings Conference Call. As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Nancy Fazioli, Vice President of Investor Relations. Thank you, Nancy. Please go ahead.
Thank you, Latanya, and good morning, everyone. Thank you for joining our third 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 was 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 our actual results and trends to differ materially are set forth in the third 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 on 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 third quarter 2025 letter to stockholders. Unless otherwise stated, all financial information referenced in this call will be non-GAAP. With that, Michael, please go ahead.
Thank you, Nancy, and good morning. I just returned from the incredible experience of Calix Connections, where we celebrated customer-driven innovation on our unique platform and the success that the platform drives for our customers by enabling their teams to win new subscribers, grow revenue per subscriber and reduce churn. To stand in front of that crowd and celebrate NPS scores as high as 94 is a testament to the partnership and trust we have built with our customers and they with their subscribers and the communities they serve. At Connections, we officially launched the Calix Agent Workforce, our end-to-end integration of Agentic AI into everything that we do via our third-generation platform that will launch in partnership with Google this quarter. This marks the next stage in our company's ongoing evolution to help our customers simplify operations and go to market and innovate with our platform, enabling them to grow for their members, investors and the communities they serve. The ability of our customers to dominate their markets when armed with our unique and highly differentiating platform and managed services model was on full display in our results. In the third quarter, the Calix team achieved record revenue in our fifth quarter of sequential growth while guiding higher in fourth quarter. We set another gross margin record, our seventh consecutive quarter of margin improvement. We also had 20 new customers choose the Calix platform to dominate the markets they serve and RPOs grew sequentially. At the same time, the team maintained a rigorous focus on operational performance and the balance sheet with OpEx investments returning to the target financial model while yielding our 10th consecutive quarter of 8-figure free cash flow, ending the quarter with record cash. It was another great quarter of performance by our customers and our teams who support them as we launched the culmination of our vision where those with the Calix platform, managed services and access to data are best placed to succeed through the power of Agentic AI. Corey, over to you to walk through the specifics of Q3.
Thank you, Michael. During the third quarter of 2025, we delivered record revenue of $265 million, reflecting sequential quarterly growth of 10%. Our overperformance relative to our guidance reflected continued robust broad-based deployments from our BXP customers as they added new subscribers and footprint expansion as they continue to choose Calix for network upgrades, new builds and competitive displacements. RPOs grew 2% sequentially to a record $355 million and increased 20% year-over-year. Our current RPOs were $141 million, up 5% sequentially and up 28% year-over-year. This metric is a strong indicator of the strength we are seeing from our platform cloud and managed services model. The combination of our BXP customers winning new subscribers and strength of Access Edge deployments led to another record of non-GAAP gross margin of 57.7%, representing a 90 basis point sequential quarterly increase. Our balance sheet metrics were strong. DSO was 30, inventory turns were 3.8 and free cash flow was $27 million. We have produced quarterly free cash flow for over 5 years, including 10 straight quarters with amounts in the 8 digits. We ended the third quarter with record cash and investments of $340 million, an increase of $41 million sequentially. Moving to guidance. Given the broad-based demand picture and the rates at which our customers are deploying products, we believe we can continue to grow revenue sequentially even with the significant overperformance achieved in the third quarter of 2025. Our revenue outlook for the fourth quarter is between $267 million and $273 million, which at the midpoint would represent a 2% sequential increase in revenue and reflect revenue growth of 20% for the fiscal year as compared to 2024. Our non-GAAP gross margin guidance for the fourth quarter at the midpoint would represent a slight increase from the third quarter and reflects our expectations regarding customer and product mix. This guidance for the fourth quarter means our gross margin improvement for fiscal '25 will exceed the higher end of our target financial model of 100 basis points to 200 basis points. And regarding non-GAAP operating expenses, we expect OpEx will increase sequentially primarily related to investments in Connections and to accelerate the development of AI agents and functionality to our platform into the first half of 2026. That said, we expect to be back within our target financial model by the end of 2026. Michael, back to you.
Thanks, Corey. This is an exciting time for Calix and our customers. The pace of change that AI is injecting into the market is like nothing we have ever seen in human history. As I shared at Connections on stage and with the 350 general managers and CEOs who attended our leadership track where Netflix took 10 years to get to 100 million subscribers, OpenAI took 2 months. The pace of change is mind-boggling and for many, it's overwhelming. For Calix, that pace of change is our advantage and an advantage for the customers who leverage our platform and managed services. We've invested 15 years and $2 billion to put in place the foundational building blocks in the form of Access Edge, Experience Edge and Calix Cloud to make the most of the Agentic AI opportunity that is ahead. Our migration to our third-generation platform in partnership with Google allows us to support the success of our existing customers while expanding internationally into new geographies with local sovereign data centers and support large Tier 1 customers with a dedicated instance of our entire platform. We are already well into the fourth quarter. And while Connections is complete, thousands of attendees will be educated on the AI-enabled opportunity ahead as Connections on Demand, our virtual program launched yesterday. This is in addition to the on-demand replays on calix.com. At the same time, our customer success team is revamping how they support customers with the Calix agent workforce to speed transformation and expand our impact on our customers' ability to add new subscribers, grow revenue and reduce churn. In addition, our internal enterprise teams are focused on improving how Calix operates and improves with artificial intelligence. With the insights that our product teams have gained through 2 years of learning as they build AI capabilities into Calix Cloud, our enterprise teams are uniquely advantaged to aggressively embrace AI in everything that we do internally to improve our operational performance and help us scale at the fastest pace possible. As I shared in the Fast Company article, the 4 AI questions that every CEO needs to ask to succeed. Over the last 6 months, we have seen employees create 725 exploratory agents with 40 of those agents being selected by our AI steering committee to be scaled across our internal enterprise to drive operational gains across Calix. The next step in Calix's journey is here, and I'm excited to lead the team as we speed our ability to transform the broadband industry and enable the success of our customers and partners. I'd like to close by thanking our team, customers, partners and shareholders whose passion, grit, trust and partnership have brought us to this exciting next stage in the Calix journey. Nancy, let's open the call for questions.
Thank you, operator. We're ready to take questions.
分析師問答
Our first question comes from Samik Chatterjee with JPMorgan.
This is Joe Cardoso on for Samik. Maybe perhaps for my first one, very strong revenue performance this quarter. I believe this marks 3 in a row now with results tracking $10 million to $20 million ahead of the high end of your range. Just wondering if you can help contextualize what has been driving this outperformance relative to your expectations at the start of the quarter. And as we look ahead to next quarter, how are you thinking about the sustainability of those drivers and potentially implications to your outlook? And then I have a follow-up.
Thanks, Joe. Appreciate it. The overperformance is driven by a couple of things. One is just the broad-based demand that we're seeing across our customer base. But it's also a function of some of the competitive expansion of our footprint. We're seeing some of these customers, I guess, surprisingly, not only do cap and grow, but some rip and replace as well. And so when you combine that with the broad base from our existing customer base, that's driven the overperformance.
Yes. To provide some context, our customers are achieving more success. When they gain subscribers, we also benefit because it generates additional revenue for us. For instance, as they expand their networks and acquire new subscribers, it creates more opportunities for us to grow as well. We have nearly 1,200 customers using our platform, and their collective progress contributes to our revenue increase. This is why we have revised our fourth-quarter guidance upwards, as we are confident in the strong demand we are observing, as highlighted in our letter.
Got it. I appreciate the color there. And then maybe for my next one, we obviously saw the announcement coming out of Connections last week, lots of innovations being unlocked there, particularly in the backdrop of AI. However, if we kind of take a step back here, just curious how you're thinking about the implications of this innovation cycle for Calix and maybe more specifically on the investment side of things, particularly, it just sounds anecdotally like there's greater appetite to scale more aggressively. I mean, even if I think about you guys mentioning like international regions, like it sounds a lot more aggressive than what you guys have talked about historically. So just curious how you're thinking about the implications to the business model as it relates to more on the investment side, just given kind of this innovation cycle where we're still in the very early innings of. Appreciate the question.
Yes. From an investment perspective, we have a target financial model and we are continuing to invest. After 15 years, we've put $2 billion into this platform, and these investments, as mentioned in our letter, are the foundation for future growth. In existing markets, we see a significant growth opportunity due to our unique positioning. For instance, one of our customers is using the Access Edge, a consolidated network with great intelligence, along with our Experience Edge, which encompasses smart solutions for various customer needs. These elements come together in the cloud, enhanced by an AI layer. Customers with these foundational components are well-positioned to leverage AI, which processes their data and automates key business functions, allowing them to scale much faster than before. Previously, our customer success teams provided suggestions based on customer needs, but now they can present actionable AI workflows that simplify implementation without requiring additional staff. This shift from providing advice to enabling customers to automate their processes represents a significant growth opportunity in our current markets. As for our international expansions and larger customer engagements, these have always been part of our plan and represent major potential for revenue and margin growth. We're discussing this now because our cautious leadership has historically been careful about what we share. With the third generation of our platform launching soon, we can now see the path for this expansion. Notably, at our recent Connections event, I spoke with the Head of Google Telecom about the potential of our partnership, which could thrive regardless of geography or customer size, offering us a substantial scaling opportunity with minimal additional investment needed, aside from sales and marketing, since it’s a cloud-based solution that incurs costs only as it scales. This positions us well for both revenue and margin growth.
And Joe, I'll add on to the model question to that when you look at the AI investments that we're doing from an OpEx perspective, in the quarter, you saw us hit record revenue and at that time, get ourselves back into the model. But we also see an opportunity here to accelerate some of the development work that we wanted to do from the latter part of '26 into the first half. So you saw in the OpEx guidance that we took up the level of OpEx. It's for those investments in R&D related to AI functionality, but we will be back in model by the time we exit 2026.
Yes. Regarding what Corey mentioned, our cloud platform, Calix Cloud, which includes Operations Cloud, Service Cloud, and Engagement Cloud, is essentially designed to streamline the workflows of running a broadband business. We are uniquely positioned in the realm of artificial intelligence because we have established the necessary infrastructure to support AI development, including our data layer, knowledge layer, and orchestration capabilities that enable trusted actions. This was made possible by a $100 million investment since November 2023. With this foundation laid, the subsequent steps become easier and more rewarding. We will accelerate the development of our AI agents. Our Chief Product Officer has noted that creating these agents is straightforward; the challenging part was the extensive groundwork we’ve laid. Now, we can focus on the simpler tasks and significantly monetize this effort, as we can transform operational workflows in Operations Cloud into agent-based solutions for our customers, such as anomaly management. While others in the industry will be working on complex, custom solutions, we will leverage the insights we have gained over the past six years in Calix Cloud to implement these agentic workflows efficiently. By the latter half of 2026, our AI capabilities within our platform will be unparalleled, as we fully understand how to operate broadband companies, backed by our comprehensive workflows, data, insights, and the trust of our customers. Failing to move forward at a rapid pace would be unwise. Finally, many are discussing AI investments. I have observed a lot of focus on various tangential areas. Major players are developing large language models, which are becoming commodities. However, the true value will be derived from those who grasp the business intricacies. This is precisely what Calix Cloud offers—a thorough framework for managing a broadband company. We are committed to integrating AI across the board in the upcoming months.
The next question comes from Scott Searle with Roth Capital Partners.
Great job on the quarter and outlook. Mike, maybe just to dive in on the small customer front, it was broad-based. It was strong. I'm wondering if you could provide some other commentary in terms of if there were any pull-ins, the sustainability of that going into 2026 and kind of how you're thinking about that visibility building at this point into 2026 and the sustainability of double-digit growth.
Thank you, Scott. That's an excellent question. I want to clarify that we have broad-based demand. No single customer had a major impact on our results this quarter or will in the next. Our strength comes from up to 1,200 customers driving this demand. Regarding visibility, I'll let Cory discuss specifics, but our close partnerships with customers allow us to see the reasons for our positive guidance into Q4. Our customers are succeeding, and this success is not tied to any one customer; rather, it's about all of our customers performing well. As they succeed, we succeed and so do our investors. It's really that straightforward; the demand is broad across our customer base.
Yes, Scott. So I think the visibility we have into the demand profile, as Michael outlined, gives us that confidence that the sequential growth is durable. Now we had a large step up here in the third quarter. So I think the sequential increases will may be more muted as we move into 2026. So in terms of our target financial model of 10% to 15%, I think we'll be at the lower end of that range ex BEAD, right? And I'm sure I'll get a BEAD question. So I'll preempt that. I'll let that question be erased.
Okay. I'll pass on the BEAD question, Cory. But just to clarify, the sequential growth, so we'll see sequential growth from the fourth quarter into the March quarter. And then as my follow-up, Gen 3, right, we talked a lot about AI and Agentic opportunities there. If we look at the current quarter, international was down due to one customer, I think, in the European theater. But the Gen 3 platform is supposed to deliver private cloud capabilities, sovereign data center capabilities. When do we start to see that accelerate? And Mike, to follow up on your commentary on Agentic AI in general, as you start to think about that driving the flywheel within your customer base, does this poise you guys for actual acceleration of RPO growth as we get into late 2026 and the ability to really drive that recurring revenue?
Yes, Scott, great question. The way we think about the monetization of the AI agent is not so much a separate charge for it, but an acceleration of our business model, right? So our customers will acquire subs faster. They'll roll out more services more quickly. So that's how we'll monetize it. So yes, it will increase RPOs because as they go faster and adopt more of the platforms and the solutions, that will ultimately translate into a contract value that will get reflected into RPO. So ultimately, yes, that's how that will happen.
The key term you used is flywheel, which we frequently discuss in the context of moving from good to great. That's exactly what Calix represents. To address the previous question, our flywheel consists of 15 years of investment that has continuously strengthened our platform. This allows us to empower our customers to improve, which in turn enables them to operate at an increasing pace. This supports three key growth drivers: adding subscribers, increasing revenue per subscriber, and reducing churn. When churn is minimized, gross additions naturally rise, showcasing the power of our platform. Those foundational elements are integral, especially with our access Edge. If you've consolidated your network using our principles—merging B&G access aggregation into a unified system—you've created an extremely robust network with the necessary data and insights for autonomous operations. The same principle applies to the Experience Edge, and when integrated into the cloud with AI capabilities, we can assist our customers in optimizing their businesses. As they become more efficient and gain subscribers, both they and we benefit financially. Regarding international markets, we continue to receive similar questions every quarter. The data is inconsistent from quarter to quarter, but we maintain broad demand from all customer sizes. This variability doesn’t reflect a customer's strength or weakness; it’s merely a matter of timing. We experience consistent demand across small, medium, and large clients. For international markets, we anticipate starting our expansion in 2026. We're well-positioned with our current customer base, which has the necessary foundations to be exceptionally successful at a rapid pace. Your observation about the flywheel is completely accurate.
The next question comes from George Notter with Wolfe Research.
Can you hear me? I was wondering about your monetization strategy for the Agentic AI workforce. Did you consider charging customers separately for this capability and being more aggressive with monetization instead of giving it away and hoping these customers will find success in the marketplace? I'm curious why you didn't take a more direct approach to monetizing it.
We are taking a direct approach to monetizing it. It's a matter of perspective. There will be some elements that we charge for and others that we'll monetize by helping them drive gross additions. We have a clear monetization strategy and as a leadership team, we are confident that it will generate significant revenue and substantial upside.
Got it. Looking ahead at the company’s growth, I believe you mentioned a range of 10% to 15%, possibly leaning toward the lower end of that range.
Without BEAD.
Right. Okay. So of that growth, is most of it coming from same-store sales, meaning existing customers growing faster or spending more with you? Or is it new customer acquisitions that account for a significant portion of that 10% to 15%? How do you distinguish where the growth is coming from?
There was broad-based demand across all segments. Existing customers have significant unmonetized potential. For instance, some customers are at 65% to 70% market share, which is nearly the maximum in a competitive environment, while others have only 20% market share, indicating considerable room for growth. Additionally, we've launched our MDU products, and several customers have made substantial progress in this area. One MDU company manages around 900 apartment buildings and just completed their first two transitions to Calix, creating a clear opportunity for us. Approximately one-third of all tenants in the U.S. are in MDUs, representing a vast untapped market for our sales teams to explore. We are very optimistic about the MDU market. In the past quarter, we added 20 new customers, which are essentially new accounts. Year-to-date, we've likely added around 60 customers. We are successfully acquiring new clients and growing in this area. These growth figures feed into our overall strategy. Furthermore, our new market expansion is not yet included in the anticipated 10% to 15% growth rate because it is still in its early stages. We are planning to enter international markets and engage Tier 1 customers through dedicated platform implementations, which are also not reflected in those numbers. Additionally, while we haven't incorporated BEAD initiatives yet, we received our first BEAD orders this quarter, which are also not part of the current figures.
Got it. Okay. Regarding BEAD, can you share how much it might contribute in 2026 and 2027? Do you have any insights on the scale, potential, or timeline for that? That would be helpful.
Yes. So Yes, I'll just give you my thoughts on the BEAD. We are more constructive on BEAD than we were 91 days ago. The turnaround from the states on their preliminary awards was faster than we had expected. As such, a certain percentage of those customers have the ability to plan for a certain amount of the jobs next year. And so while we did receive our first order during the quarter, it's still too early to determine the demand dynamics for next year other than there will be some amount versus last quarter, I thought there would be like none. But to provide a little bit more color, here's what we know. Of the 49 of the 50 states reporting, California still hasn't submitted their awards yet. The total amount of dollars have shrunk by about 50%, right? So the original program being $42 billion, it's now going to be something like $20 billion. And when you include the matching funds, now we're talking about a $30 billion program. Fiber is still the dominant technology at 65% of locations and 85% of the dollars. Fixed wireless was 12% of locations and 8% of the dollars. Meanwhile, LEO was 21% of locations and 5% of the dollars. So as we have said, historically, we've done really well with government programs and that we expect to do the same with BEAD.
The next question comes from Christian Schwab with Craig Hallum.
It was quite clear that the double-digit growth is sustainable. Looking ahead to next year, should we expect to be at the low end of gross margin expansion year-over-year of 100% to 200%, considering this year's strong performance, Cory?
You've got that right. Thank you. We haven't talked about that yet. But yes, given the overperformance on gross margin this year, I think the increase next year will be more muted. It will be at the lower end of that 100 basis points to 200 basis points. So we'll continue to expand it? Yes, our software content grows every single day.
We're going to continue to guide the 100 basis point to 200 basis point, and we're going to grind at it, right?
That's right. But it will likely be at the lower end of that 100 basis point to 200 basis points next year.
Great. And then now that we found this new found enthusiasm for BEAD, if you think about aggregate dollars that could come to you over a multiyear time frame, when do you think that peak spending would be?
We still believe the deployment curve will resemble a lens shape. It will start to ramp up in 2026, then likely level off for a few years before declining. As we've mentioned, government programs often take longer to initiate than expected, and the funding involved is usually much larger than the initial projections, leading to extended durations for these programs. We still see it as a lens shape. However, it is too early in the award process to comprehend the buying dynamics, including the volume of purchases, timing, product mix, and the extent of building next year since it's late in the cycle. Currently, we anticipate some BEAD revenue next year, but we cannot accurately estimate the size at this point.
Well, I had many discussions about this. To Corey's point, we believe it will be much larger than expected, echoing what others said three years ago about BEAD funding being imminent. We have been very clear about this. One key takeaway is that our proactive discussions indicate that funding is imminent. In 2026, the funding will indeed start. It's important for everyone to understand that while building a BEAD network, they often overlook additional customers who have not been funded yet, which can be significant. For instance, if they are funded to serve 10,000 customers, there may be an additional 5,000 to 15,000 potential customers that they pass by and can later include. This potential isn't fully reflected in the BEAD total addressable market (TAM). Additionally, once the networks are established, the focus shifts to connecting with subscribers. Responding to George's questions regarding our monetization strategies, our priority is to increase subscribers. If we were merely a network company, we wouldn't concentrate on acquiring subscribers. However, in our model, the monetization potential of $1 to $10 per month in software margins is substantial, driving long-term growth and margin expansion. As they construct that network, BEAD funding will assist them, and our role is to help them maximize their market share from the BEAD funding while simultaneously reaching out to customers who aren't funded but are still served. We're discussing this now because by 2026, it will be tangible. We focus on topics that we can actually see and engage with. That’s why we are talking about BEAD now; it is within reach.
The next question comes from Tim Savageaux with Northland Capital Markets.
Congrats on the results. I have a couple of quick questions. Regarding the BEAD initiative, it seems that in the past we estimated the access infrastructure opportunity using about 10% of the award value or network cost. Is that still a valid estimate?
We're using more like 5% to 10% for that figure. Additionally, it's important to consider the incremental opportunity when you gain subscribers, which is separate from that, relating to your point about it being the access network.
You need to consider that these locations are more difficult to access, and additional investment will be necessary. Therefore, a figure of 10 is too high; the amount will need to be lower than that.
Yes.
Fair enough. Speaking of the $1 to $10 per month, any update on the timing for a more detailed breakout of the appliance business versus the software and platform business? I know you were end of next year was what you were discussing.
Yes. I mean that's still the stated goal by the end of '26.
Got it. So no change there. And if we look at the guidance for Q4 for sequential growth, the various quarters this year, you've had maybe small customers driving things in Q2, large and medium. Anything to call out in terms of customer segment movement into Q4 in terms of driving that sequential growth?
Yes. Broad-based demand across all segments.
Yes, nothing to call out, Tim.
Yes, I want to emphasize that we have broad-based demand across all segments. The value of our business model is highlighted by our nearly 1,200 customers. We do not rely on any single customer for 10% of our revenue. This diversification means that even though segments may fluctuate from quarter to quarter, it helps stabilize our overall performance. With broad-based demand, we can effectively plan our business.
We will take our last question from Ryan Koontz with Needham & Company.
Maybe a couple of topics, if I could. Maybe philosophically, relative to your 10% to 15% growth model and thinking about your SAM and your current customer relationships and how do you think about growth limiters in the industry that can allow you to outperform that, whether it's relative to the BEAD process, supply of fiber, supply of labor, supply of components, which doesn't seem to be that big a deal anymore. Like how do you philosophically think about risks and upside relative to things that are out of your control?
So from a risk point of view, at this point in time, we've actually been talking a lot about that as to where do we see the risks. we really don't see them with regards to limiters. So there's BEAD, the BEAD stuff, the reason why we're actually saying, hey, now we see it coming out is because of the fact that, obviously, because it's a governmental process, and as we said right from the get-go is that these things are a bit complex and takes longer to get here. But when it does get here, it starts flowing and it flows for longer. So there's some timing components affiliated with that, right? So that's the first one. And then from a systems and components point of view, we don't see any issues other than there's some memory prices going up, but we balance that out with some other areas, right? So but these things all balance out because of the fact that we have a broad-based business. There's nothing specifically from a leadership point of view that we're looking at as a risk. I will talk to upside, but are there any other risks that you see, Cory?
No, I think as it relates to the practical limiters on the business as you go and build new networks, it's permitting and labor. So that puts a natural governor on just how fast they can go.
In Washington, there's significant discussion about how to speed up permitting and access, which people are very focused on. With the midterms approaching next November, there's a push to get things accomplished ahead of the election cycle, likely speeding up the process to some degree. There are no major risks that Cory and I or the leadership team are discussing. In fact, we are quite confident about the strong demand, the operation of the business, the talent we have, and our relationships with customers. Everything looks positive. Regarding the potential of artificial intelligence, we are in a unique position because we understand how to effectively run a broadband company. We have integrated deep insights into our cloud and possess significant intelligence both within the network and at customer premises. This allows us to support our customers more effectively. When it comes to monetizing AI, we intend to do so meaningfully by helping our customers enhance their business at a much quicker pace. Understanding sales cycles is crucial, and from my experience of 30 years in enterprise sales, I'd prefer to help a customer significantly increase subscriber numbers instead of just securing a small price increase on a single cloud service. Maximizing revenue per subscriber is the goal. We aim for all of our 1,000 cloud customers to utilize all three of our cloud offerings, which AI can facilitate, driving more monetization. It's better to have all customers on three clouds rather than a smaller number spread across several offerings. There’s a tremendous opportunity here. Over my 9.5 years with the company, this has been our focus. We began testing neural networks in 2017 and recognized the potential but struggled to meet customers' privacy and security needs for AI usage. In the last two years, we have built the necessary infrastructure, and now we can engage with customers more effectively. AI will allow us to simplify complex tasks for them, enabling faster growth, which is why we are so optimistic about AI's future.
That's great stuff. One more quick follow-up, if I could. Nice to see the rebound in the small customer cohort best in like 10 quarters. But you're looking upmarket, the medium and large cohorts have grown over 50% year-to-date, clearly, a great indicator there. How much of that gain upmarket is competitive displacement versus your customers investing more and being more successful with your help?
Do you mean in small or in large?
Medium and large, medium and large.
A significant portion of our growth comes from becoming the essential go-to-market engine for our customers. Previously, they relied on other providers for basic WiFi services, but now they are recognizing the potential of our offerings across various clouds and the Experience Edge. They want to enhance their go-to-market strategies, and we are the best choice for that transformation. We are definitely replacing previous partners in the federal sector and assisting them in driving their businesses forward. Additionally, when our service provider collaborates with us and competes against others not using our products, their success represents a different angle of competitive displacement. It reflects how we are enabling our customers to outperform their competition, which aligns perfectly with our goal. This is how I view our competitive positioning on a daily basis.
We have reached the end of the question-and-answer session. And now I'd like to turn the call over to Nancy Fazioli for closing remarks.
Thank you, Latanya. Calix will participate in several investor events during the fourth 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. You may disconnect your lines at this time, and have a great day.