Awesome. Morning, everyone. Welcome to day two of the Citizens Technology Conference. My name is Nick Lee. I'm an associate on the software team. I'm really excited to have Bandwidth CEO, David Morken, and CFO, Daryl Raiford, with me on stage today to chat a little bit about the business. David and Daryl, thank you for taking the time and traveling over from your beautiful campus in Raleigh. Welcome to the conference.
Thank you. Truth be told, I traveled a little further. I'm 34 hours from New Delhi. Was in Jaipur when global incidents started happening. I'm delighted to be here with henna on my fingers, after attending a wedding of our chief operating officer's eldest daughter in Jaipur.
Oh, wow.
Awesome to be here.
Glad you made it here safely.
Thank you, Nick.
We're gonna take the next 20 minutes, 25 minutes to chat about Bandwidth with potentially a few minutes at the end for Q&A. To start, I wanted to ask what is known as Pat's trademarked question. You know it, you love it.
How's business?
You said it yourself. How's business? What would you say?
You asked it well, Nick. Business is epic. Voice is back. AI agents are emerging at scale, and we have a vital role to play, so business is epic.
The big thing right now is voice, as you just said, right? Can you help us understand at a basic level, you know, even if an AI agent isn't necessarily Bandwidth's, why is it good for you?
Yeah. Love the question. Every AI voice agent needs to communicate. They have a mission to accomplish, and to do so, they need to communicate. They live in the cloud. Their inference, their reasoning, their round trip of their tech stack all occurs at the data center level, and as a result, they need reach. For us, we're relevant to them because we have footprint across 65 plus countries, and they are able to originate and receive calls using our global platform and global network that's very low latency, in fact, ultra-low latency, high fidelity with lots of reach, and that's exciting to us. The next billion users of the PSTN, the next new billion users of the PSTN are largely AI voice agents, and they have needs that are unique in particular that we serve.
Do these agents need phone numbers?
Often, the use case requires them to be reached, to be called back, to have a bidirectional conversation that is persistent in terms of identity. So long as, the global carrier networks exist, that primary identity is a number. Often the use case will be bidirectional, requiring an address space, a namespace that is often a phone number.
Okay. Understood. Thank you. You guys reported 4Q earnings on February 19th. What were the key messages you wanted the Street to take away from the quarter?
Let's hand it over to Daryl.
Yeah, Darryl.
From the quarter, we reported record profitability and cash flow. It was a strong quarter exceeding our guidance. For the full year, we were pleased with our 10% revenue growth and cloud communications growth as well. We achieved record profitability for the full year as well. We were able to launch into some capital structure capital allocation initiatives, which we've already announced again yesterday, and we've been pretty pleased about.
Great. We'll get more into that later. I wanna start at the top line here. You called out two pretty big deals on the call, and with that, you called out partnering or, you know, using the use of Genesys, but also Google Conversational AI, which indicates more of like a freedom of choice strategy. How is your freedom of choice strategy, Bandwidth's freedom of choice strategy, helping you win large deals over other providers?
The orchestration layer that we provide is agnostic, which is important for the top 10 bank that we talked about and the household name insurance company that we discussed. They have a multi-vendor environment, whether it's Cisco in one case where we needed to integrate Google AI or it's Genesys in the other. That agnosticism that we provide through our Maestro Platform is vital. You don't yet have identified best of breed in a way that allows for lock-in or bundling. You have enterprises wanting to bring point solutions and to take a call flow and route the voice stream into sentiment or transcription or into fraud or into security simultaneously during the call. The Maestro pre-integrations, the partnerships we have allow an enterprise to be very flexible in how it changes its contact center call flows or its knowledge worker call flows.
That's been really important in winning, to your good question, Nick. The enterprise deal flow that we announced that was record-setting for us, Maestro's been critical, and we've got a 100% attach rate to Maestro for our enterprise deals, and that will be the case going forward, I think.
Great. Daryl, kind of on Maestro, it was called out that, well, including Maestro, your software services, which includes Maestro, Call Assure, and Trust Services are, quote, "producing recurring high margin revenue streams." Can you walk us through how the software expansion economics are altering your gross margin profile?
Absolutely. We exited the year at an exit run rate of $15 million annual recurring revenue just in terms of these new software services that we've launched. I think you can see the dynamics in the business model and how that's how that's improving and really contributing by looking at the total company for the year. For every incremental dollar of cloud communications revenue that we produced in 2025, we yielded $0.82 in gross profit. A gross profit yield of $0.82 on a record company margin that we achieved in 2025, non-GAAP margin of 59%. You can see that the gearing effect, the leverage, say, on the gross profit yield for every incremental dollar. We're really excited about that as we look forward into this next year as well, again, guiding profitability nearly 30% higher.
Yeah. Fantastic. You know, you guided to 14% next year adjusted EBITDA margin. What's driving these expenses, and what steps are you taking to achieve this? Is this all coming from that software services line item?
Well, as David said, in terms of the voice AI expansion, we're calling for voice, our voice solutions to grow at their highest rate. We doubled in 2025 our voice growth rate, and we're expecting to increase it again in 2026. We also expect contributions from messaging. We expect contributions from software services, as we said. We're guiding to 60% gross margin or greater. In terms of the non-GAAP EBITDA margin that you mentioned, we're looking for, you know, nearly 30% expansion to $120 million at the midpoint for 2026, with a very nice flow-through to free cash flow in 2026.
Very nice. David, I wanted to bring it back to you. We're seeing a shift from AI experimentation to more full-scale production deployments, especially among your customers. How is the Maestro AI Bridge accelerating this transition, and how does the deployment of Conversational AI act as a, say, multiplier for network consumption?
Yeah, it is a multiplier for network consumption because what used to be a single signaling and media call flow is now becoming forked into multiple call flows. We're a usage-based business. We're paid for each of those channels, which is exciting for us because, as I said earlier, sentiment, fraud, transcription, there are multiple use cases now on a single call. That's exciting for us. Maestro accelerates this adoption. There's really demand pull forward from AI to get companies into the cloud and simultaneously to get them away from the incumbents. Many high-volume enterprises globally have worked for 50 years or more with incumbent providers. Those incumbent providers have done nothing to embrace the current moment of voice AI agents.
Maestro as a platform allows a contact center, or any number of enterprise voice use cases to very rapidly be deployed, integrated to existing partnerships and vendor solutions and taken to market at scale into what we've seen over the last two years being experimented with, being prototyped. Those are starting to scale, and we see that both in terms of the use cases and the ways that calls are being routed and also in footprint. We have large new customer opportunities that are taking us into large new jurisdictions where we have a regulatory moat that is extensive, and so it's exciting to see how Maestro is also driving footprint expansion of a high-fidelity, ultra-low latency global network and software platform.
Fantastic. You mentioned that moat, and I wanna read a quote to you that you said on the earnings call that really stuck with me. You said last quarter, quote, "We have a moat that is a mile wide, filled with oil, and lit on fire." What did you mean by that? What are the moats that Bandwidth has to prevent this disruption from AI? 'Cause that is, you know, something that everyone is worried about today, especially in software.
When we started building this network in 2007, we had to stand tall in front of 50 state Public Utilities Commissions and qualify via interconnection to the incumbents under the Telecom Act in order to receive from the PUCs what is now close to $100 million domestic phone numbers about, I don't know, a fifth of the total in the North American Numbering P lan. It's even harder globally in 65 countries to move at the speed of government. The moat mile wide is both time and money. You would have to move at the speed of government across 65 countries where we have full PSTN replacement and over 80 countries where we have partial.
You would also have to deploy significant capital into the infrastructure required for the many thousands of IP interconnections, the hardware that our Universal Platform supports globally for the ultra-low latency global network. When I said it's a mile wide, that's time and money. Filled with oil, set on fire just illustrates the fact that we've seen exactly no new entrants in this space in 15 years, and I don't see any coming soon. There hasn't been private equity or venture deployed in our infrastructure space that's so vital to AI voice agents in a generation. I meant it, I don't really like the moat analogy because it's defensive in nature.
I would rather, like, have a trebuchet of a Maestro trebuchet that is taking out incumbent entrenched monopolies where they exist with new technology and reach for voice agents that need, vibrancy in their voice applications. I stand by that quote. Thanks for bringing it up again.
Absolutely.
Believe that we've got a very defensible future, but an offensive opportunity.
Just for clarification, these incumbents are your traditional, like your T-Mobile-
Verizon, AT&T, T-Mobile, Lumen, Colt, Tata. Then in every country, you have often, you have a primary incumbent in that country, yes.
Okay. Got it. We also saw that those incumbents are the big three actually increased prices. Is that affecting you guys in any way? Are you seeing any change in customer behavior because of those price increases?
The customer behavior not so much. We've increased our gross margin since coming public from 47% to now 60%, and that's a function of owning and operating our own infrastructure. The economies of scale have really made an impact. Even as carriers increase costs, if you're just renting their network, you won't have the gross margin opportunity that we do. In terms of customer behavior, we closed more large enterprise deals than we ever have in our history, and our pipeline is stronger than it's ever been before. I think that's actually a better answer to your good question, which is, yeah, enterprises are lining up more than they have in the past, and that may partially be driven by increases in price.
Understood. Daryl, I wanted to move to the balance sheet. You have $250 million in 2028 convertible notes. What's your plan to address this debt?
We do-- we did have $250 million.
You did have, yeah.
I n convertible notes that were due April 1st of 2028, so still more than two years away. On Monday, yesterday, we announced that we had repurchased $100 million of that, of the $250 at a discount. You know, it is pretty predictable in terms of our behavior since November of 2022, we've now retired over $550 million of convertible debt, leaving only $150 remaining in over two years. I think that we've been really successful, and why we've been really successful at doing that is that our business has been performing very, very well. In 2023, we set out a goal to achieve $125 million of organic free cash flow over the next four years through the end of 2026.
We hit that goal just over two and a half years in and have well exceeded it now as we move forward. It's given us a lot of confidence to take your question further, Nick. It's given us a lot of confidence in terms of a balanced allocation strategy. Not only are we growing our profitability almost 29%, 30% in 2026 over 2025, that comprehends the largest investment in technology innovation that the company has ever made in our history. We're not starving our business, we're actually growing into that profit profile and that profit expansion. Along with that investment, we announced an $80 million share repurchase program, which is underway. We just yesterday completed a $100 million opportunistic convertible repurchase.
We're able to do all those things and achieve our goals, all of our goals at the same time, and I think that it's a credit to the business and where we sit. I'm gonna put a fine point on what David said in terms of the mode. I wanna circle back. You asked about price. Right now, where we sit in our space is at a much higher value proposition level than what we would call the incumbents. Let's just call them legacy telecom carriers worldwide. We also are positioned well far away from being only a thin application layer that could be substitutable or perhaps without the infrastructure and the platform that we have, subject to some AI risk.
We're right in a really interesting sweet spot where we provide much more cloud features than the incumbents as a value proposition, but we charge less. We're already the lowest cost provider. We're providing all the software services on top of the platform, which is integrated through Maestro and the others, which our customers really enjoy and provide massive stickiness, and I'll come back to that, while not being exposed to simply a substitutable thin application layer like others. We don't resell our network. We have a global owned and operated infrastructure platform. I don't think folks recognize what that is in 65 countries. To David's substitution of, you know, 15 years and how to get there. Over the last two years in our fastest-growing segment, enterprise voice, we've experienced zero churn of our customers, our customer cohort.
We've had 100% name customer logo retention rate. Why that is because once a enterprise is on our network, they recognize the value proposition. We're already lower cost, we already provide more features, and the loyalty and durability of those customers is proven in empirical data. We feel really good about that, and that ties it all back to your balance and your repurchase and convert question and the like. Everything is coming together in a nice virtuous circle, and we feel really good about it.
Great. I wanna kinda go back to that zero churn comment and enterprise being your biggest grower and look at go-to-market strategy, actually. What is how are you attaining these new customers? What is your strategy for getting these new customers? You obviously have a product that sticks, right? I guess, what is limiting you guys in terms of speed of getting more of these new enterprises onboarded?
We've always been disciplined about return on advertising spend and ROI in our go-to-market. We've had a profitability commitment that we've spoken to for years and years. That is an important discipline. Our go-to-market motion has historically primarily been direct, and it's been through an enterprise sales force and through outreach of our own. Our channel, though, over the last two years, has really begun to pick up steam in two different ways and contributed to the record-setting enterprise wins that we spoke to. That channel, Nick, does two things: it increases your pipeline, your funnel, your reach, but it also, and this isn't something we originally understood, it accelerates the deal cycle. You're often with a channel at a stage with an enterprise decision maker much further down their decision tree than you are when you're direct and reaching out.
That's been exciting to see over the last couple of years. We expect channel to continue to grow. It was only something that we started focusing on a couple years ago, but it's contributing meaningfully right now to the guide that we put in place for 2026.
Great. This kinda leads me to my next question here, which is, we've seen the cost of developing software to drop a lot, compared to a few years ago, at least. Are the engineers at Bandwidth utilizing any code assist tools? You know, how does that affect the headcount of your company? Are you deciding, "Oh, hey, we're actually gonna hire less engineers and hire more of these, direct sales people to try to sell more to enterprise?" Has your philosophy on that changed at all?
Yes, everyone is using tools from Claude to Gemini to others, ChatGPT. I was in New Delhi, as I said, 48 hours ago, and met with 100 of our developers there, and one of the topics was the different AI tools that they're embracing and engaged with. The short answer is, to the second part of your question, Nick, we're gonna continue to invest, as Daryl said, a record amount in R&D and development right now at this moment and are pleased with the go-to-market and the sales trajectory. It's a time to make sure that we invest properly in the technology shift that is upon us that's so important. While we have developers that are using tools and enjoying breakthrough productivity and creativity, we're hiring, and I don't think that will change.
Got it. Daryl, I guess I wanted to ask, what are the one to two most important things that you get done this year?
The most one or two important things that we get done this year is to. That's a good question, Nick.
Yeah, that's a great question. Yeah. There's a number of strategic initiatives that we're focused on. We are, customer adoption of our Universal Platform. Universal Platform is very, very strong. I think for me, though, it's continuing across the market, across the customers, across our software development, across our engineering and our customer delivery and operations. I think the most important thing is to continue the gearing effect of our growing profitably. When I say growing profitably, I think that's become a really, you know, kind of phrase that people just dismiss. When we grow every dollar of cloud communications revenue with 82% gross profit yield on a 59% corporate margin, we are growing profitably. We are, you know, we're going to continue to invest in a whole bunch of different things
For the gearing effect to continue to grow our free cash flow, as a percent of revenue is really, really important, and we're doing that right now.
Fantastic. Before I open it up to questions for both of you, I would ask, is there anything I didn't cover that you would, that you wanna get out there, that you wanna address?
The only thing is if you ask me the top two this year, as Daryl answered, if I took a shot at that one, it would be achieve the P&L, exceed the guidance, and improve the culture.
What about the culture needs to be improved or can be improved?
You always accomplish your mission and love your people, and there are always ways to improve both.
Fantastic. With that, we'll open it up to questions if anyone has any. Yeah. In the back.
Daryl, I don't think the market quite appreciates the positive signal that the convert buyback represents. Did you use bank debt to... What portion of that was bank debt versus cash on the balance sheet? Is, can we interpret that as, that M&A is not something that Bandwidth even needs if we've, if we're, you know, if we've paid, utilized a lot of our balance sheet cash?
That, I think that's great. We do have a $150 million credit facility, and in the form of an undrawn revolver. We will from time to we have in the past flexed, we will from time- to- time flex that. I don't look for that to be any sort of long-term or, you know, modest term, six months sort of thing in terms of managing working capital. We did want to take advantage. You know, we've, in it, we finished the year with $111 million of cash in securities. Clearly, if we're going to buy $100 million, we're gonna flex a little bit of our on our undrawn revolver.
I think that's just a natural thing, and we've done that before.
On the second part of your question, yeah, our path forward will be opportunistic, but we have done two acquisitions in 26 years. We're pretty focused on organic.
David, you may have spoken to this, and I apologize. With the expansion of the R&D efforts, is John sort of the... Is that labor-based expansion? Is John guiding that lot, or are you... How much oversight? It's kind of what I always sort of ask you around how, where you're spending your time you know, giving people the freedom to, you know, run with those efforts?
John Bell, our Chief Product Officer, is essential to focusing on our near-term time horizon R&D. The further out is our CTO, Scott Mullen, and I'm working a lot with both, but both are really important and have important parts of our R&D budget.
Great. Well, David, Daryl, thank you so much for joining me today. Really appreciate the time.
Thank you, Nick.
Thank you, Nick. Grateful for you.
Thank you.