Thanks for being here. I'm Jason Ader with William Blair. I'm pleased to introduce Paddy Srinivasan, CEO of DigitalOcean, and Matt Steinfort, CFO. Before we begin, I'm required to inform you that a complete list of research disclosures or potential conflicts of interest is available on our website at williamblair.com. Paddy's gonna go through some slides, and then we'll have some time for Q&A, and then the breakout session will be upstairs in Adler.
Great. Thank you so much, Jason. Good morning, everyone. Thank you for spending this morning with us bright and early. Appreciate that. I'm here to introduce a terrific company called DigitalOcean. I'm sure many of you have heard of us, but I'm here to give you a deep dive on what we have been up to, starting with an introduction. H ere is an obligatory safe harbor slide. I won't go into the details, but this is me. I joined DigitalOcean about 100 days ago in February, and I'm super thrilled to be here at DO. My background includes a lot of time spent building infrastructure for developers.
It was a great fit and something that, I was super thrilled to embark on as part of, reimagining what DigitalOcean stands for, especially, in an era where cloud computing and AI and all these, new emerging trends are converging. F or those of you who are not familiar with DigitalOcean, our mission is to simplify cloud computing so that developers can do what they do best, which is to change the world with software. As we know, Marc Andreessen famously said, "Software is eating the world." That was about 15 years ago, and there have been many different iterations of that.
Our core founding mission has been very grounded on that, to take away the complexity associated with building software so that we can really enable developers to focus on things that really matter to their business. D o that in a way that democratizes access of cloud computing for the smallest of the development shops. Where do we play? We play in the public cloud market, and the numbers you see here are specifically for the small and medium segment of the public cloud market. Just as a reminder, the public cloud market, as per IDC, is around $360 billion today, of which $114 billion or so is spent by companies that have less than 500 employees.
That is our sweet spot, and then there are a couple of other slices, as you can see here. T he point here is that it is a large market, and it is growing robustly. I t is a terrific place for a company like DigitalOcean to operate in. Y ou might say that, "Oh, I'm familiar with this space. There is Amazon, Google, Microsoft, even, the resurgent Oracle Cloud and so forth, s o is there actually room for a different cloud provider?" T he answer is absolutely yes, because, as you see, in the landscape, all of these companies, their cloud divisions alone are running into tens of billions of dollars. AWS famously reached $100 billion run rate last quarter. Azure, I forgot the exact number, but it's $60 billion-$70 billion run rate for Microsoft Cloud Services.
Even Google Cloud Platform, I believe, is in the $30 billion+ run rate. A ll in all, they have to chase the large enterprises, and that is totally understood. W hat happens when you start building for the large enterprises, that the platform gets very bloated. I know it from my personal experience, having worked several years at Microsoft and then subsequently at Oracle and more recently at Amazon. It is really hard to do both and do both really well. B y the law of large numbers, the hyperscalers, quote-unquote, are required to focus on the largest of their customers, which means the platform is super complex. The pricing gets very complex too, when you have a long menu item of services.
One thing that we have repeatedly heard from our customers is that it's really impossible to decipher the cloud bill from one of the hyperscalers, for any of the hyperscalers, and it is really, really hard to make that predictable and consistent. The next thing is lack of support. Unless you are Procter & Gamble or Eli Lilly or one of those companies that are spending billions of dollars on a cloud platform, it is really hard to get support directly from these hyperscalers. You have to go through a variety of different partners and things like that, and if you're a small company just trying to get off the ground, it's really, really hard to navigate through this complex maze of technologies and get things to work and scale. T hat's where we come in.
W e take that as a big opportunity for us to serve the underserved, to arm the rebels, so to speak, with our simplicity, and it's easy to put that on a slide. It's really, really hard to do it in practice, and do it at scale. W e have over 640,000 paying customers. It's really hard to make things as complex as cloud simple, and that's what we try to do on a day-in, day-out basis, on all aspects of doing business with DigitalOcean. Starting with the core platform, we make it super, super simple for people to sign up and get started. We make it extraordinarily easy for them to be onboarded and then grow with us as their business needs evolve.
The second aspect of this is, that support and community go hand in hand. We have, I believe, and I was, in the early days of the Microsoft platform, even before it was called .NET, so, I know how hard it is to build a developer community. Some companies have done it really well. Microsoft has done it really well, and I believe that we have a very, very powerful, vibrant developer community, something that we take really seriously. A lot of people ask me, "Okay, how do you build a developer community?" It's really hard. It is all the little things you have to do, like speaking engagements in developer conferences, contributing to open source, going super technical and doing some forks of specific, Linux versions, and conference.
Giving back to the community and making heroes out of the community developers, and they really provide a lot of support to our customers. It is just the pride of being part of that DigitalOcean ecosystem that really differentiates us. F inally, the concept of multi-cloud was only for the few and the privileged until a few years ago. Now, when I talk to our customers, it just turns out that even the smallest of technology company wants to have optionality. They don't want to get boxed into a hyperscaler cloud, because once you are there, it's really hard to get out. T hey want to have optionality, they want to do business with multiple clouds, and that really talks to the success of our platform.
Here are a couple of examples of customers. By intent, we picked a couple of logos that you might not recognize because that's who make up our 640,000 customers. H ere is a little startup, and you can read it for yourself, and they trust us to scale along with them. Like, we don't we have a fairly narrow cloud computing platform, and we take pride in having 100% of what companies need to run digital infrastructure on our cloud, and very little of what they don't need, so that we can keep our platform simple and delightful to consume. H ere is another bigger company called Tango. This company is in the hundreds of millions of dollars of ARR per year.
They started with us when they were super small, and they're still with us, and expanding their footprint every year. I t's a great example of companies that start very small as a startup and then grow with us. One of the important things that I want to underline is that, when you look at our customer composition, some of our bigger customers, we call them scalers, a lot of them, if not most of them, are tech companies. These are independent software vendors. These companies run software on our infrastructure, and software is their business. We don't have a lot of companies that host internal business applications on us, like a local florist or an orthodontic practice running their calendaring system with us.
Yes, we do have some of them, but a vast majority of our big customers are tech companies that rely on our infrastructure to run their mission-critical, revenue-generating software. W here do we operate? We operate all around the world. We have 16 data centers across nine regions, and we have customers in pretty much every United Nations country. T he composition of our revenue, interestingly, is 2/3 outside of North America. W hat do we actually offer? We offer three sets of things. The first one, the white box, is our core DigitalOcean offering.
We offer the essential compute database, virtualized infrastructure using Kubernetes and Docker and things like that, various storage primitives, including just pure data storage, volume storage, but also an increasing number of database as a service and things like that. T his really appeals to the developers in small and growing companies because it's super, super easy to get started. Y ou might not have heard of DigitalOcean, but you might have heard of this concept called Droplets. Droplets are our single unit of infrastructure, and that's what made DigitalOcean the company it is today. We also acquired a web hosting company called Cloudways. We did that two years ago. That's also part of our platform. It's really aimed at e-commerce and digital agencies.
Then finally, on the AI side, we acquired a company called Paperspace last year, and that is getting aggregated and included as part of our core DigitalOcean offering. F inally, before I turn over to Matt for our financials, no technology presentation these days is complete without a click-through of AI, and we are super bullish on AI ourselves as a platform provider. We currently have both platform as a service and infrastructure as a service, and we have very, very big plans in terms of AI in the future. I'll pass it on to Matt, so that he can go through the financials.
Thank you, Paddy. I'll just spend a couple of minutes going through the financials so we can get to Q&A. The company is a very solid mix of growth and profitability. We had a very strong quarter in the first quarter with $185 million of revenue, which was up both ARR and revenue, up 12% year-over-year. We had a very large improvement in our profitability, driving EBITDA margins from 34% in 2023, in the first quarter of 2023, to 40% in the first quarter of this year. Y ou're also seeing that flow all the way through to earnings per share.
We had a 55% increase in earnings per share, and for the first time, we were GAAP net income positive in a material way in the first quarter. A lot of progress on driving operating leverage into the business. If you look at the growth story over a longer period of time, it's very clear, like many software providers, that our growth has slowed in the last several years, coming out of the pandemic as the software slowdown kind of hit the industry. We're down to a 12% year-over-year growth projected in 2024. T o offset that, we've been able to drive material improvements in our operating leverage and drove EBITDA margins all the way to 40% for the full year 2023.
We've stepped back just a couple of points to take advantage of the investment opportunity and the tremendous growth opportunity we have with AI. We're investing a couple of percentage points of both EBITDA and free cash flow to take advantage of that opportunity. If you think about it from a shareholder return standpoint, one of the most important metrics from our perspective is free cash flow per share. A s you see, we've delivered about a 50% CAGR in free cash flow per share over the last year and projected through the end of this year. That's on the combination of the strong operating leverage that we drove into the business in 2023, but also a regular and committed share repurchase program.
We repurchased about $500 million of stock in 2023. We continue to have a $140 million authorization over the next two years, and we'll continue to have share buybacks be a part of our capital allocation strategy as we look to drive free cash flow per share growth in a healthy way over time. If you think about the growth story, and Paddy talked about a lot of things that we're excited about, b ringing Paddy on is a big part of this growth story. T o reinvigorate the pace of innovation at the company and the ability to take advantage of not only the core platform to continue to grow customers in our core cloud service, but also to take advantage of the tremendous AI opportunity we have.
We're very focused on improving our net dollar retention, which dropped to its lowest point in July of last year, as we lapped a price increase that we had done in 2022 and saw the software slowdown impact our growth. We've been steadily improving that quarter-over-quarter. We expect to get above 100% at some point in the relatively near future, although we're not banking on getting above 100% net dollar retention in 2024. We are seeing steady progress, and the progress that we're seeing is on the back of the increased pace of product innovation that Paddy has brought to the company and driving up ARPU across our existing customer base. W e'll marry that with the tremendous growth opportunity that we have with AI. Our AI platform, Paperspace, that we acquired in July of last year.
W e've augmented that with other AI capabilities, and we saw a 32% increase quarter-over-quarter in our ARR around our AI-related services from Q4 of 2023 to Q1 of 2024. W e're excited about the potential there. J ust to close before we get to Q&A, we feel we're in a very strong, differentiated position in a market that's growing and is vast. We feel the business that we've built is incredibly durable. Almost all of our revenue comes from a self-serve notion, where customers hit our website to find our rich content, to learn about how to develop. They sign up, and they grow on our platform. W e have a very low cost of sale. We've demonstrated the ability to drive material operating leverage and profitability in the core business, and now we're sitting at the very early innings of a generational growth opportunity with AI.
We feel like we've got a very strong foundation with a very big upside option around the AI impact on not only our business, but the industry. The priorities for the company are investing to take advantage of those two opportunities, to accelerate growth through product innovation and the enhancement of our platform to to enable customers to scale and to democratize the AI capabilities to the developer world so that we can make it easy for them to leverage that technology in their business models as well. With that, we'll close and turn it over to Q&A.
All right. Thanks, Matt. I wanted to start out the questions just with the thing I think is on. The thing I hear the most from investors, which is: you guys are great at onboarding customers. You have 600+ thousand customers. I t seems like it's more of a challenge that once they get to a certain scale, you know, to keep them on board. H ow do you think about the things that you need to do, especially in your first year as CEO, to to change that reality? I'm probably exaggerating. It's probably, y ou probably have plenty of customers that stay on, but-
Yeah.
It does seem like there is some graduation from, in terms of certain accounts once they get to a certain size?
Yeah, that's a good question. I'm glad we are starting with this question, because that's, one of the areas that is overestimated in terms of, us having a graduation problem. If you look at our overall churn numbers, they've been fairly steady over the last several quarters. Do we have customers that actually outgrow us and, move to other clouds? Yes, and it's just a natural part of the evolution as companies get bought over by someone else that has a much bigger footprint on a hyperscaler, or companies that are, requiring, some geographic expansion into areas where we might not have a footprint and things like that. T here are obviously cases where that happens.
There are also cases where companies, as I said, pursuing a multi-cloud strategy, move from one of the bigger clouds, and that's a very underreported thing in my mind, that move to us as well. W e absolutely take it super seriously, right? It's hard enough to go acquire customers and get them to a point where they are scaling with us. We obviously want to do our best to not have them be required to graduate to a different cloud just because we don't have a few things. I f you look at t hat was my number one priority in my first 100 days, is to look at what those specific gaps are. T here are a few that we are addressing very aggressively.
If you look at the product velocity just in the last six weeks or so, you can look at our blog, and you can see how much we have started shipping on a week-over-week basis. T he common thread in all of that is to help companies that are scaling, utilize the breadth and the depth of our platform more effectively. F or example, we just announced, Premium Optimized, Memory-O ptimized, and Storage-O ptimized Droplets. E ssentially, it is like, if a company is using an in-memory database, they need more storage, on a single Droplet to take advantage of caching, for example, right? W e announced that. We also announced a lot of other enhancements to our core networking stack. We announced auto-scaling, in many different, parts of our platform.T hese are all advancements that are intended at keeping our biggest customers happy and on our platform.
If you look at the roadmap over the next 3 months-6 months, that is also geared towards closing the gap on some of the granular security requirements or management console visibility and observability of our platform. M any core networking between our data centers and things like that. Y ou will see a steady drumbeat of innovation aimed at keeping some of those large customers within our family. T hat's not to say that we are doing that at the expense of even more simplifying the onboarding. O ur learners and builders are the funnel that eventually get to our scaler platform. W e are very aware of it, and I think it is. It's not as acute as one might think from the outside, and the numbers bear itself out.
Matt, what is the breakdown in the business between the learners, builders, scalers?
It's interesting that 480,000 of our 650,000 customers are learners. Those are the people that spend less than $50 a month on the platform, but they only represent 13% of the revenue. 87% of our revenue comes from our builders and scalers, which are about 157,000 or so customers. I t's interesting to support Paddy's kind of high-level view. The churn rates go down as our customers get bigger. They don't go up. T he average lifetime of the customer goes up.
It's as you expect, as they get larger, they're buying more products from us, their attach rate is higher. They're more kind of integrated into their own development environment, so we're part of their infrastructure. W here we lose opportunity, because we do think we have a big wallet share opportunity with those larger customers, is, they may not have all their cloud. W e may, t hey're not single cloud on us. They have multi-cloud deployments, and we can get a bigger share of that multi-cloud opportunity if we continue to innovate and address some of the issues that Paddy had described.
Great. Paddy, just, you're the new guy here, so I'd love to hear, in your, I don't know, 150 days now, what is it?
Hundred days.
100 days, okay. W hat has surprised you the most versus your kind of expectations coming into the job?
Yeah, my biggest surprise has been the vibrancy of our developer community, despite how we have actually engaged with them over the last couple of years. Like, yes, we have a momentum that is taking us forward, but just the resiliency of our community has been really something that really caught my attention, and you see that every week. There'll be a question on a Reddit forum or on Hacker News or something like that, and even before our support staff can jump on it, community would have addressed it and resolved it in a few minutes. It's just amazing to see, and that's why we are actually going to super double down on the community and do all of the things that I just talked about, right?
It's getting back on the road in terms of the developer conferences, really engaging with the startup ecosystem, getting active, which we have started doing in the open source community. There's a lot of little things that we should be doing that we kind of got distracted a little bit over the last couple of years, but we are going to get back to that in a big way.
We've seen some acceleration at some of your larger competitors on the cloud services side. Do you think that your exposure to kind of the tech vertical and kind of digital natives and startups, do you think that's created kind of disproportionate pressure on you guys in your business, just given some of the challenges in that startup world over the last couple of years?
That's a good question. T here are two ways to think about it. One, do we have exposure to small companies? Yes, and small companies across the world are not immune to some of the macro concerns that we read in the newspapers every day. A lso on the flip side, these are tech-heavy, tech natives, digital native companies, right? T hey need to spend on essential infrastructure to grow their business. I feel that is actually an opportunity for us. The macro notwithstanding, I feel like if any company was to be in the business of serving small companies, I would rather serve small tech companies rather than any other type of small company as a cloud provider, because it is just mission-critical, revenue-generating infrastructure that we offer them.
W e are well aware of how the enterprise-oriented hyperscalers are performing, and I feel like it is just a matter of time because any tech company, regardless of their size, has to innovate or die. I think the next wave of innovation is right around the corner, and we want to be ready with our infrastructure and with our innovation, so when the small companies do in fact start spending and scaling, we are going to be the partner they can trust.
That's a great segue into AI. We've got about five minutes left, so we'll talk about that for the rest of the time. I wanna talk about it both from the perspective of your customers and what their needs are, what your actual offerings are today, and then Matt, from your perspective on the financial impact, CapEx in particular. A question I get a lot, which is, like, what's the return on that CapEx investment?
Yeah, I can start with, what do we offer, and why do we offer what do we offer, right? That segues into the customer thing, and then Matt can close us out with, the financials. W hat we offer right now, as I was explaining, is we have two things. One, we acquired Paperspace, and I call it the Platform as a Service for AI. W hat does that mean? It means any developer that wants to incorporate AI into their application, so it could be a inventory management system or a digital order processing system, or AdTech is a real customer that I talked to a few days ago. All of them need several AI models.
LLMs get all the hype and the press coverage, but in fact, there are dozens of AI models in the world. Some of them are LLM, some of them are not. A ny tech company or any tech solution going forward will need to incorporate multiple AI models into their system. O ur platform as a service, what it does is it gives developers the full lifecycle support of discover the right models that you want to use, design your application, incorporate these models, test it, validate it, build data pipelines to inject your custom data into these AI models, and then finally run these models on our GPU infrastructure.
We provide a collaborative environment where they can do it and also start pulling from GitHub repos and things like that to build a full-fledged machine learning and AIO ps pipeline. T hat's Platform as a Service. We also announced in January that we are now offering Infrastructure as a Service, which is access to GPUs directly. T hese are typically used by companies that are developing a model themselves, right? Or taking a model and substantially customizing it by changing the weights or injecting their own data to augment the underlying model. What do they need? They need access to raw GPU horsepower.
Not only that, but they also get the advantage from us of having an orchestration layer because a lot of these model companies are. I use the reference of MIT kids or MIT scientists. They know machine learning and AI really well. What they don't know is how network-attached storage works, or how to move data back and forth between these highly demanding GPUs with low latency networks and things like that. F or that, you need an orchestration layer that can do all of these things so that they can focus on what they know best, which is creating world-changing machine learning models. T he numbers that Matt was referring to, overall AI for us, is showing a lot of promise.
Our Infrastructure as a Service is actually doing really, really well, and just shows how GPU and compute-hungry this world is when it comes to AI. W e have both those offerings, and I'll just finish by saying, what do these customers look like? F or both the Platform as a Service and Infrastructure as a Service, these customers look very remarkably similar to the ones we already have and know how to acquire. These are smaller companies that are starting their journey, startups and Series A, Series B type of companies. A lso for the Platform as a Service, these are established technology companies, and as I said, they are in the business of consuming AI. The Infrastructure as a Service customers are in the business of creating AI models, and the former are in the business of consuming AI models to make their software smarter. Matt, you wanna comment on the financials?
As Paddy said, you have to think about the economics in those two different categories. Platform as a Service, which is a more fulsome solution, the margins are not too dissimilar from the core cloud. They're a bit lower, from a gross margin standpoint. The Infrastructure as a Service, which is really just the leasing of hardware and the access to that, with not as much software wrapper around it, the margins on that are definitely lower than the gross margins, lower than the core cloud business. T he paybacks are still attractive, even though the market is very competitive. You're looking at 2.5-3-year cash-on-cash payback on just the incremental capital that you deploy for that.
It's still good margin, and I think that as the industry matures, as we get more than a single supplier of the predominant bulk of the equipment, I think you'll see the cost curve bend, and would expect that that comes down to better margins over time.
Great! Well, that's perfect timing, and we'll wrap it up there. We're gonna go upstairs. Thank you, everybody, for joining us, and thanks to you guys.
Thank you.
Thanks.