Good afternoon, and welcome to the Needham Technology and Media Conference. This afternoon we're excited to have Bandwidth presenting with CEO David Morken and CFO Daryl Raiford. David, you just had a pretty big quarter here, and we'll get to all of that. Just start off with an overview of Bandwidth and how you compare and contrast to a traditional carrier.
Thank you, Josh, and appreciate being here, and thank you to all the Needham folks. Bandwidth is based in Raleigh, North Carolina. We are a global provider of communications and have an owned and operated voice network in over 60 countries, and a software platform on top of that network that allows enterprise customers and large conferencing customers and large unified messaging customers, and now recently, large CRM customers, do voice calling around the world and also text messaging. We are very different from an incumbent carrier such as Verizon and AT&T and Lumen, from whom we win most of our new enterprise business, because of two important factors.
The first, our footprint is around the world. Theirs is domestic only. Second, we have a software platform on top of that network that allows our enterprise customers to orchestrate call flows, to provision in real time, to configure services in ways that are both, through a GUI, through an API, and then recently, through a command line interface for the emerging agentic voice agents that we serve.
Got it. Al right. One of the common questions I get is why is it important as a CPaaS vendor for you to own your own network versus kind of this fully aggregate others' networks, and how does that kind of create a competitive advantage for you?
The utility and value of owning and operating your own voice network around the world, a network that's regulatorily compliant, really has two primary dimensions. The first is quality/reliability. The second is cost. On quality and reliability, by being able to control the hardware and the interconnected circuits around the world of this network, you really do have visibility and can manage the service delivery and plan and promise through an SLA the level of that service to your large enterprise customers. When Microsoft, for example, has a QBR with us and comes to the office rolling deep with 30 people, you're talking about your network's performance according to the parameters that you promised them, and you're also showing them where they may have emerging issues that only you can see throughout the guts of your network.
That's the quality component. The reliability component is you get to manage your uptime by deploying capital against needs and priorities in a way that you don't get to see if you're relying on a Lumen, for example, and you're just renting or leasing the network from them. Not, I don't have anything against them, I'm just using them as an example.
The second part is the cost advantage. When we went public in 2017, we had gross margins of mid-40s. We are now just below 60% gross margins for our voice service, or for our overall company, and the voice services are very similar. The reason we've grown our voice gross margins over that period of time is because we get to put every additional 1 billion minutes of month of voice traffic on our fixed infrastructure costs. The economies of scale emerge and allow you to grow your gross margin from the mid-40s toward 60%. Quality, reliability, cost, those all allow you to go to an enterprise customer as a single source around the world at a very good value.
Got it. Another question I often get on you guys is why haven't the legacy carriers done more to keep up on the voice AI support kind of opportunity there? What do you think is their strategy around voice AI, and how has that static development helped your competitive dynamic?
Yeah, static is a compliment. That the incumbent providers domestically have focused on the mobile opportunity for consumers. They have focused on broadband and rolling that out, they have not at all paid any attention to voice and what it means to support either cloud-based voice during the last migration from the premise to the cloud, nor the voice agent AI moment that we're in now. I've been told that they are the place where innovative software developers go to die, and because that's true, they have very little innovation or vision for what the future may or may not be. I don't see that changing. We're about to be 27 years old as a company in August, and that certainly hasn't changed for those 27 years.
Got it. Al right. Before we get to some product and financial questions, maybe you could just rewind the last five years for investors since the pandemic. You know, what caused the voice revenue growth to slow post-pandemic? Now, how do you think about the re-acceleration here? How much is driven by AI versus some other market share gains or other factors? That's a lot there, I know.
No, it's okay. When the pandemic hit and everyone was required to work from home, we were forced overnight to scale our network and service massively in support of customers such as Zoom and Microsoft Teams and Google. We were able to do that with our model effectively, and yet what that really did at that time was pull forward a huge amount of demand for moving from the prem to the cloud. That demand pull forward happened and we scaled massively. When everyone returned to the office or began returning to the office, that growth slowed, and you had more of a tempered pace of migration to the cloud in the intervening period.
Going back now about a year, what has emerged is this moment of voice AI agents, their effectiveness, and their vitality, and that's turned voice into an enormous new opportunity as a critical infrastructure layer for the primary interface that we are all beginning to use for AI.
Got it. Al right, you just had a significant may understate it, win with Salesforce Agentforce in their contact center offering. Why do you think you won that deal? How are you thinking about the ramp there? Then any other relevant information that you can share regarding the deal at this point?
We are fired up to be Salesforce's partner with their announcement of Agentforce Voice Agents, and won that opportunity to partner with them precisely because of the global reach and the orchestration capability of our network and platform. The team there is very innovative. Their vision that every call should be a conference call with an intelligent voice agent as part of it really resonates deeply with our team. We think that that intelligence is meaningful in every sales conversation, and so the ability to support that is a great opportunity. They are a visionary team. They have enormous scale. They have a vision for Agentforce and Voice Agents that's compelling, and we have an asset base that was so unique globally and has high such high value that we were able to win out and be their partner and are excited about that now being GA and scaling.
Al right. In terms of product innovation, you've done a lot around Maestro, trust and verification to drive total cost of ownership, but what specifically has been some of the priorities over the last few years in terms of product investments?
We have invested in R&D to bring that Maestro orchestration layer to life, to have pre-integrations with all the major CCaaS providers, many of the UCaaS providers, with AI leaders, effort into that integration has been key. Footprint expansion is something that isn't valued appropriately, but getting to be able to provide regulatorily compliant PSTN service in Turkey, for example, with emergency services, is highly valued by our enterprise customers as an edge case jurisdiction, we have others that we've been building toward that we look forward to announcing in the future. Supporting the AI use cases that are emerging now has been a key area.
When you have your OpenClaw or Hermes agent and you want to be able to have it, make a phone call or receive a phone call, we have a command line interface that you can go to or actually that you could point your agent at and it could sign up on your behalf. Efforts like that have been really exciting over the last several years, and seeing those come to fruition to provide a low latency global voice network for voice agents has been very rewarding.
Right. One area is underappreciated is your global capabilities, which you highlighted a little bit here, and now you've continued to add a number of new countries under support. How do you think about adding new countries and ensuring direct carrier connections in these regions to support these low latency use cases? 'Cause that's really critical with the AI, the voice AI stuff.
It starts with having a universal platform of the shared same components in each jurisdiction. If you are trying to stitch together various networks that you've acquired over years, that's a very challenging thing. We are excited about having come through and having a universal platform everywhere so that when we expand to a new jurisdiction, it has the uniform stack of gear and the protocols for security and jurisdictionalizing the traffic. There are countries we are not yet in which we're excited about, India being among them, where we're aggressively working toward offering the same level of service that we offer elsewhere.
When we have customers, whether it's a large global enterprise or Salesforce, it's incumbent upon us to make sure we have the appropriate reach at the same level of quality and fidelity and to navigate that on their behalf. There's an enormous high value. AT&T can do that for you in the United States. Colt can do that for you in EMEA. Tata can do that for you in some jurisdictions. Nobody can touch the scale and scope of our country by country geography we serve.
Got it. Maybe just touch on, you know, with voice AI, isn't this becoming an even bigger competitive advantage for you because of the low latency dynamic with all the direct carrier connections you have?
How do you think about that aspect of it?
It's vital. That the tech stack for inference and reasoning by itself is already 400 milliseconds to round-trip a prompt and a response, and yet the PSTN latency is understood to be about 400 milliseconds by itself. You've already spent your entire time budget for a conversation you don't get ticked off by just in getting an answer from the model. We have an advantage in our network when we can yield back to the reasoning model even 50 milliseconds or 20 milliseconds advantage. We have a thesis that there are component parts of the tech stack on the reasoning side that may be more effectively deployed closer to the core of the network where we have an advantage.
Owning and operating the infrastructure and being able to drop components of the tech stack closer to where the communication is happening may yield back even more latency savings, and therefore accuracy. Those are things, even just adjusting jitter buffers to get 10 millisecond gains are vital for the improvement in accuracy and responsiveness. That's an advantage we have that others don't.
Got it. That's a great point there. Al right, as you think about the progression of voice AI use cases across the ecosystem, right now they're heavily focused in customer support, and you were talking about this a little bit earlier in your meetings, but how do you see this evolution to sales and marketing use cases and other areas for enterprises developing in the next few years?
It's very early in the growth of voice agents, and they originally became rationalized to save OpEx in a cost center like customer support. It was very easy for Daryl and his peers to say, "We know we can use a voice agent to render service more effectively, in a customer care context, resolving tickets, doing ticket deflection." What's happened just in the last 90 days, voice agents have gotten better and better, not just at handling customer care, but now in doing demand generation, lead creation and qualification, prospecting.
That's new, and that's only in the last quarter. Those are revenue-generating upside cases. Small and medium business are deploying voice agents now during times of day when they don't have employees on station, doing roles that they otherwise wouldn't staff, whether it's appointments or calendaring or reminding. These are use cases beyond customer care that small and medium business now have access to that they're using that they would never otherwise be able to do.
Got it. Al right, one concept that's gaining traction with investors right now is around the whole theme of conversational AI or guiding a call with a voice chat bot to messaging as needed, depending on the use case. How do you see that opportunity playing out over the next few years? We'll start with that.
Yeah, a voice agent will be very good at tool calling. If we're having a conversation and you say, "You know what? Will you text me that?" It'll actually volunteer, "Hey, I'm gonna send you that picture," or, "I'm gonna send you that boarding pass," or, "I'm gonna send you that reminder." That'll be during the conversation, and it'll just be a tool call. You can have a messaging point solution as part of a conversational AI solution, just like it can send you an email, and it'll be a great and effective tool call during a conversation, and it's gonna be important, and that's something that we certainly have natively by providing messaging ourselves for the voice agent services we support.
Al right. You now have about $25 million in ARR from high margin software add-on revenue. Can you from a high level discuss, you know, what's included in this figure and how you see that metric maybe evolving over the next few years, without giving too much guidance?
You want me to take that?
Go.
We have, with the Maestro platform, a platform fee. You have a fixed fee for using that orchestration layer. You have more usage-based components for trust services that help large brands verify that they are who they say they are when they call. We have other services within the orchestration layer that are integrations done to every single CCaaS platform out there, pre-integrations on UCaaS, integrations done on AI. All those component parts are factors into that $25 million number and are all very gross margin accretive, and we believe that they are growing fast. Let me pause and ask Daryl to remind me how fast they're growing, the $25 million.
Absolutely. Thank you, Josh. Really appreciate you invited us to the conference. It's been a wonderful event. We did exit the Q1 at a ARR exit rate of $25 million for software services. That grew from $15 million at the end of December in the fourth quarter. As David said, that's an exciting basket of features and capabilities that are primarily recognized on a monthly recurring charge basis. Those are it. We're real excited about that. We think the uptake is going to be increasing and improving, and as David said, the gross margin profile is extremely accretive to our consolidated gross margin.
Got it. Al right, final product and strategic question. As we move more towards orchestration of AI across different channels being the key to success, versus simply aggregating messaging and voice volumes, how does that impact your pricing power and competitive position relative to, I don't want to say the old way of doing business, but the old world of a year ago?
Some of our call flows used to just be a single call between a hotel receptionist and a guest. That same call flow with AI involved really now routes what was one call to a sentiment analysis engine, to a transcription service, to a translation service potentially, as well as to potentially recording. That becomes four legs of the call and monetization of each for us, that's a much more lucrative call flow. Our pricing power is unique in our owned and operated model, but the value of these kinds of call flows to the outcomes that the enterprise customer is looking to achieve are so high that it isn't yet, and I don't anticipate it being for a long time, a price conversation at all.
That's exciting to us. There's an enormous amount of value for us to add to these new use cases. That's evident in the pricing history we've disclosed where it's favorable, and we're actually increasing price.
Right, that's an important distinction, right? Because the traditional carriers, they've increased prices, but they've kind of been under price pressure themselves for a number of years.
You know, you actually have pricing power going forward now relative to what maybe investors from the outside would maybe not understand that dynamic, right?
It is interesting. If what David was saying, you look back at least over the last eight years, potentially more, the company every year has, on an aggregate basis, both grown its volume and its price in a price volume analysis. I'm not addressing one of 10,000 SKUs, but on aggregate, we've always been able to do that. How have we been able to do that? One is through the attachment of higher value software services and the like, as well as growing our mix towards higher value products, in terms of pricing with Enterprise, in terms of pricing with messaging. We've been very successful with that. I Obviously you've seen that with our growing margin. It's really driven our growing gross margin as well.
I agree with David. When you are already How we win from the incumbents is a really nice surprise to any particular customer prospect. When they realize that we are bidding as a low-cost provider against a higher cost product that they have with the incumbent, and we're providing substantially more features, you will not believe the surprise and delight on their face when that occurs. When we are already the low-cost provider providing more features, price and commoditization is not an issue.
Right. Al right. I'll just throw out there that we can leave some time for questions at the end, so if anybody wants to prepare questions in the audience, we'll leave some time at the end for that. Moving on to some financial questions, you previously gave some 3-year targets in 2023 with some ranges on revenue growth and margins. As we think about the guidance for 2026, how are you thinking about these targets now with kind of greater clarity on how the year is shaping up?
W e think that we're really pleased with where we're coming out. We're guiding to revenue growth of approximately, total revenue growth of approximately 18%. Our Cloud Communications growth growing at 10%. Embedded in that growth, embedded in that guide is a 60% target gross margin, which we set out on a mission to achieve in January of 2023. Embedded in that guide is a 20% EBITDA margin, again, that we set out in early 2023 to achieve, with a 15% free cash flow margin. We're really pleased to be growing for, into those medium term targets this year with an 18% total revenue growth, 31% EBITDA growth, and a substantial growth in free cash flow.
You've impressively added some, large $1 million plus customers in the last few years. I guess a couple items. Are you winning these customers from legacy carriers? As we think about their ramp in 2026, what process are you using to include some of these, volumes and guidance?
All these voice wins are from legacy incumbent carriers. We had a large messaging win. It was from a CPaaS competitor, where that customer just needed more deliverability insurance and reliability. We're very successful in demonstrating our capabilities with those large voice customers, whether it's in and many of them are enterprise related. We feel like we have a really pretty good pipeline moving forward on that as well.
Got it. Just for some perspective, if we look at, you added 5 or 6 million-dollar plus customers last year. How would that have compared to the kind of the three or four year period before that?
That was the six. The number was six.
Was six.
That was the, that's a record. That was a record for us in $1 million plus deals in terms of contract value. Before that we might've been 4. You know, in the law of low numbers, that would've been a 50% growth rate. How about that? Importantly, it's like, what are you doing with those customers and what is your time to revenue? We're real excited about the time to revenue and what we're seeing in terms of those contract values. By the time the Q1 occurred, one of the six had fully deployed, and fully deployed at 120% of our initial estimated total contract value. Very happy about that. The other ones are on track to deploy this quarter and into the third quarter. We're seeing a very nice, calling for a very nice, healthy, growth, acceleration rate in our Enterprise Voice category in the second half.
Awesome. You mentioned this either on the call or the call back, but you have a nice pipeline of million-dollar plus customers that you can add even this year, right? You added two so far that you publicly announced during the day.
Two in the Q1. One substantially over $1 million. Two in the Q1. We, you know, while we don't really guide that bookings detail, that we're very much on track to achieve what we did last year, if not overachieve.
Right. That would be a positive indicator for next year's revenue, right?
Absolutely.
Maybe discuss, it takes 6 to 12 months. A wide range, right?
The deployment rate for the largest enterprise, the largest banks in the world, the largest healthcare concerns, it does take, it will take a ramp of between three and nine months.
Right. That would be a positive indicator for next year's revenue growth as a starting point. Al right. As we think about the ramp for Agentforce in your business, you know, I know you've discussed this in some of the meetings. You're taking a pretty conservative approach to the outlook there, how do you kind of anticipate that the project will be rolled out over the next couple quarters in terms of customer usage?
The partnership is in place and our partner, Salesforce, has launched their offering. They have customer dialogues in place right now which lead to customer dialogues with us, and we feel like there's good traction and momentum growing with the just recently launched, it was in middle March timeframe. We feel pretty positive about that. Importantly, we have not played, we have not gotten ahead of our customer and tried to forecast adoption rates. We have not played a significant amount of financial benefit or favorability into the guide for this year. To the extent that customers with Salesforce are able to launch on the Agentforce platform, take advantage of the capabilities with the voice AI agents to the extent that that embeds itself into our second half financials, that's really all upside to what we're calling for the year.
Got it. Could Salesforce have done what they need to do with this integration and Agentforce with a traditional carrier, or would it only have worked? That's maybe more of a your question.
Yeah. No, they could not. It's that straightforward. The ambition is global. It is not just domestic, and the orchestration for call flows is essential to integrate with their tech stack and to deliver quality and reliability, so no.
Got it. Al right. In the last quarter, Global Voice Plans revenue growth rate accelerated. What is driving the higher growth rate with those customers? Maybe you could just explain how this, that segment differs from Enterprise Voice.
You're right. In the Q1 of 2026, our Global Voice Plans growth rate over the Q1 of 2025 tripled. The tripling occurred from really two dynamics: price and volume. If I take the latter one first, in terms of volume, we are seeing in Global Voice Plans, our largest customers, the power platforms, the hyperscalers, as well as the CCaaS and UCaaS providers, we are seeing favorable volume profiles from those large customers where we believe that volume is being driven by their own initiatives in terms of AI solutions being released. Those AI solutions embedded with Genesys and embedded with others in each of those customers will improve ultimately our volume. We believe we're seeing that. The 2nd thing is price.
In terms of price, it's not that a SKU gets improved or increased, it's that part of that software services that we spoke about with an annual recurring revenue exit rate of $25 million in the Q1 of 2026, part of that is in Enterprise Voice and part of that is embedded in our Global Voice Plans customers as they adopt Maestro, and they're using the integrations and the like for their customers as well. We're able to embed the software services both into Global Voice Plans as well as into Enterprise Voice. That's the dynamic in GVP. It did triple its growth rate. We think that those are the two principal reasons.
Turning to Enterprise Voice, Enterprise Voice is characterized as a market category where we sell directly to large enterprises. I encourage you to look at our investor deck and the like. We highlight many of those. We highlight many use cases that you can access to describe our product offering for those enterprise customers and who those are. That sector is characterized by 15% higher, percentage point higher gross margin than the company's aggregate growth, gross margin rate. Part of that is due to the global nature of our offering and the ability to enable AI as they're thinking about their use cases.
As you think about, selling the voice Or not the voice, but just the software add-on products back into the Global Voice Plan customer base, because you have had more penetration thus far with the Enterprise Voice. Maybe you could explain to investors the dynamic there of the uptake of the you know, the software add-ons into that piece and how you can expand it now more into Global Voice as well.
Yeah, we talked about Maestro attach rates being 100% Q4, Q1, but we have an install base of enterprise customers we've sold to prior to that, and even prior to the rollout of Maestro generally, that we need to upsell to with the use cases that are now native and are, and are more expansive in AI than they may know about or be used to. That's a deliberate motion by the sales force, and we're excited about what that represents over time.
What about selling, software add-ons into the Global Voice customers? Is that an opportunity as well?
Yeah, same thing. Parts of Maestro are available to the GVP customers as well, just like CRM with Salesforce contains components of orchestration. That's also something that we can upsell the GVP partners to as well.
Got it. Okay, cool. Al right, as we think about capital deployment, I've pointed out to investors my expectation for a pretty healthy ramp in free cash flow in 2026 and with a nicely de-levered balance sheet. What becomes the capital allocation priorities here going forward?
We have, we have expressed a very what we believe to be just a balanced and prudent capital allocation strategy. First and foremost, we, our mantra has always been, over these last five years, to grow profitably, and we've done that both in terms of gross margin, EBITDA margin, and now you see it in terms of free cash flow. To grow profitably, especially in the face of the growing Momentum and wave of voice AI-enabled usage patterns. We have been investing in R&D. This year we're making our largest investment in R&D as we roll out more features and capabilities, primarily in software services, again, at a very high margin and what is appearing to be a very nice attach rate. That's balanced against our primary capital strategy, objective of de-leveraging.
We have, over the last four years, reduced hundreds and hundreds of millions of dollars of long-term debt. We finished the 1st quarter at $150 million long-term debt even on convertible notes that are due April 1st, 2028. That's against our guided EBITDA midpoint for 2026 of $122 million. We've essentially hit a 1.2 times leverage ratio on long-term debt. We're pretty close to declaring victory on that particular aspect. Recently our board has announced, as a third prong, a $80 million share repurchase program for equity, which we deployed towards that aim $11 million in March, after it was announced in the 1st quarter. We're tackling all of those things. We feel real confident. You're right, Josh, our cash flow is growing quite a bit, and we're really, really pleased with that. It gives the company a great deal of flexibility.
Yes. one of the questions I get, when I'm doing a teach-in or people who are new to the story, they ask, "A s the voice and revenues grow on the network, why does that not lead to, like, higher capital expenditures at a similar pace? Why are you able to scale the growth in the network and not necessarily have a lot more CapEx?" That might be helpful for people to understand.
In terms, it is worth, you know, having a bit of perspective. It, we have all, last five years essentially, we have maintained around a 3%-4% of revenue, CapEx rate, which means for folks listening, around $15 million-$20 million a year in terms of CapEx. Our CapEx is actually a fairly light model. This last year in 2025, we did announce in 2024 towards the end, and in 2025 we invested in about $10 million-$12 million more CapEx for network expansion, primarily, some geographic expansion in Asia, as well as United States network efficiencies and optimization. That put us in a really good place for the next few years in terms of what we think we're looking at.
It's important to know that even when we invest $10 million or $12 million, we're not investing $100 million. We're investing $10 million to $12 million on, you know, free cash flow of something that looks like $100 million. It's not really of any substantial hardship to the company to ensure that our network stays bleeding edge.
Right. Exactly. Al right, any questions from the audience?
Going once.
Going twice. Al right. With that, we are up on time here. I wanna thank the Bandwidth team.
Joshua, thank you.
Thank you, sir.