Okay, great. So why don't we go ahead and get started? Look, we're just delighted to have DigitalOcean with us at the Citizens JMP Technology Conference in San Francisco, at the Ritz-Carlton Hotel. It's our final session, and I'm delighted that Matt, who is the CFO, could join us. There's a lot of interesting things going on at DigitalOcean, right?
Yeah.
So I was looking at your background. It was more operational and consultative than I realized. So I actually do wanna talk about that a little bit and why you took this role. Then I want to talk about your new CEO, and you can put your consulting hat on for a second-
Yeah
If you like. And then, we'll talk about GPU clouds and what that's all about, and that'll definitely spend most of our session. We'll leave some time open for questions. So let me start with the question I've been asking everyone, which is: How's business, Matt? What would you say?
I'd say I'm pretty optimistic. It's, you know, coming off a challenging year last year, my first year at the company.
Yeah.
You know, growth is, was definitely declining. We bottomed there kind of midway through the year, and we've seen some pretty positive kind of early indications of a return to growth. Still being, I think, appropriately, cautious as we think about 2024. There's still market headwinds. But you know, usage was up in our core platform in January and February. That's encouraging. As you said, Patty joining the company, which we can talk about in a little bit, is incredibly exciting. I think he's going to be a great fit. If you heard him on the earnings call, he's exactly what we were-
I did, yeah.
... we were looking for. And again, you know, talking to our customers, our customers are also cautiously optimistic, you know, about returning to growth and seeing kind of improvements in their core business, which is definitely a good sign. So-
Yeah.
I'd say all in all, it's going pretty well.
Yeah. Good. I feel like you have good instrumentation in your business and, like, good metrics and good visibility. So when you say that, you feel like it bottomed in the middle of the year and it's coming back, what do you base that off of?
Yeah. So, it's hard, you know, externally to see. Like you said, we're pretty well instrumented.
Yeah.
If you pull out the price increase, the impact of the price increase that we did in Core DO in July 2022, and you pull out the Cloudways acquisition, which was September 2022, you pull out the acquisition of Paperspace here, you know, back in July, you pull out the price increase that Cloudways did in April 2023, and you just look at the core, kind of unadulterated cloud business that we have. Like the hyperscaler, if we saw the bottom, it was. I can very finely point to, it was June.
Hmm.
That was the low point of our growth. We've seen a steady improvement in the growth in that core business. Again, normalizing for all those extraneous events, we've seen an improvement in the core business.
Uh-huh.
And so when I looked at January, I saw an improvement. When I looked at February, I saw an improvement over January. And if you look at some of the even further back leading indicators, we added, you know, a healthy number of new learners, that's, you know, customers new to our platform that are only spending less than $50. But we also grew and graduated a bunch of builders and scalers on the platform. So the customer accounts, you know, look good. And when you look at, like, the... we still have a headwind with NDR being below 100, which is, again-
Yeah
... the amount of money someone spends a year later. It's 96%, which is, you know, you'd rather have that be north of 100%. But if you look at both the trends we're seeing in there, and you look at the number of customers that fall into each of the categories of are they growing, are they shrinking, are they staying flat? The numbers are working in our favor, too. We've got fewer customers and the fewer dollars in the declining categories and, you know, more in the expanding categories. So that's, that's all positive.
Yeah. Okay, awesome. So, you went to Princeton, Engineering, graduated in 1992, and then did you go to Bain? Right? Is that what you did there?
No, I was at Cambridge Technology Partners.
Oh, you were?
Software development, IT consulting for a couple of years, and then I went to business school, then I went to Bain.
Oh, so Bain was post-business school?
Yeah.
Okay, so you did... And what kind of stuff were you doing at Cambridge Technology Partners?
Um-
Was it ERP kind of-
No, no.
No?
It was client-server software development, doing rapid application prototyping and screen scraping mainframes for making, you know, cool GUI interfaces. I did a little bit of database work, so I started as a developer.
Yeah, so you-
Federal law.
... legit. So this product resonates with you?
It does.
Yeah.
It does, and it definitely resonates—for me because if you think of me as a developer, you're like: "Okay, it must have been really easy what you were doing.
No.
That's what we make for our customers. We focus on making it simple and easy for developers to get on, to build a business, to try an idea, and not have to deal with the complexity of, one, owning their own infrastructure, but two, dealing with some of the larger providers that are a little bit more complicated. So it makes it practical. And if you think about AI is doing the same thing, there's going to be a lot more developers in the world because we're making it easier for people to code and be more efficient. And so more people likely kind of join that rank, and that's really good for us 'cause that's our target market.
We're gonna come back to that. We're the two sides of that argument. Yeah, we'll come back to that. Did you graduate Princeton in three years?
No.
No? Okay. And then at Bain, what were you doing?
Did a lot of-
Bain was late 1990s, right?
Yeah, late, late 1990s, right. Internet kind of bubble time.
Yeah.
A lot of software, a lot of telecom.
Mm-hmm.
Spent almost a year in Australia working for Optus, in telecom, but I did a lot with like BMC Software-
Mm-hmm
... and Motorola and a handful of other things. Those were the most interesting.
Yeah. Okay, and then from there, there's a brief stop. We'll skip the brief stop, unless ICG was... So what was Envysion?
Envysion was a, again, kind of coming back. That was a small software company that I co-founded and ran as the CEO for 10 years. Sold business intelligence software to restaurants and retailers.
Come on! Like who in Denver?
Chipotle.
Really?
We're in every Chipotle in the world. Yeah. We're in all the Qdoba-
What does, what does your software do? Is it still in there now, or do you-
Yeah.
Yeah?
It marries-
What does it do?
- the video that's in the stores with the point-of-sale data and other transactional data. So think about artificial intelligence. If I could tell you how long it took somebody to get through the line and what they bought, and tell you-
Oh, the video of the people in the store.
Of the store.
Wow, yeah.
Married with transaction data. So anyway, I, I resonate with people who want to try to build applications and-
That's so interesting. Okay, and so when did the CFO part start?
That was late in my career.
Yeah.
It was 2016 at Zayo, when I
What was Zayo Group?
Zayo is one of the largest communication infrastructure companies in the world, founded in 2006 by former Level 3 executives. Built basically a business by acquiring all the fiber properties in the U.S. and some in Europe.
Uh-huh.
So we own, they own a material amount of the underlying fiber infrastructure in the U.S. and across the globe.
Okay, and you're CFO of that until the end of December 2022, and then, I mean, you were geographically in the right place, but how did you get connected with-
So-
-with DigitalOcean?
Yancey was on the board at Zayo.
Ah!
When we took the company private, he and I worked very closely on that transaction.
Ah, I remember this now.
So I know Yancey through that. And when I heard the DigitalOcean story, again, contrast it with telecom. You know, telecom growing 2-3%, spending $1 billion in capital, burning a couple hundred million dollars of cash every year, levered at, like, 6x.
Yeah.
And then I heard the DigitalOcean story. I'm like: Wait a minute, you guys are complaining on your earnings call that you went from 30%-20% growth, and you're generating free cash flow already, and you got-
Yeah
... this massive customer base, and you've got a debt facility with zero coupon. And, I mean, I was like, you know,
Sign me up.
I tried to negotiate a little bit, but I was pretty, pretty happy to, to be included in that conversation.
Yeah. Love it. Is there an office? Where's the office?
We're basically fully remote.
Is it fully remote?
Yeah. We've got a reasonable presence in Denver. We've got a reasonable presence in New York-
Yeah
... San Francisco, Seattle. A lot of people in Pakistan-
Yeah
India, Mexico.
So in the Yancey era, when he would want to have people together, where would it be? Would it be in Boulder?
Yeah.
It would be, right? Yeah. Okay, so tell us a little bit, and then we'll come to Patty in a second, but tell us a little bit... Like, so you start basically January 2023. What was 2023 like?
It was interesting because this business that I stepped into that has so much going for it, you know, hit some headwinds that-
Mm-hmm
... that, you know, I clearly, not having been in the space, hadn't been kind of familiar with. But I think that the, you know, it was the broad market had benefited from COVID. You know, software, the spend was up. Maybe the paying attention to the level of spend was not as big a deal in that timeframe 'cause everything seemed to be going up and to the right. And as that started to kind of, you know, the winds started to lessen, you realized that: Oh, shoot, there's a lot of things we need to do to improve, you know, the business, and we need to get ourselves set up for a lower growth profile. So we did that.
We reduced the cost structure quite a bit, got to the long-term margins, made some, you know, hard decisions. We shifted our cost model, so we've got more resources offshore, and set ourselves up to be—you know, generate really good cash flow regardless of the economic environment. And that, that was, I think, a really positive, a positive thing. And, you know, I hadn't anticipated that was gonna be the journey-
Yeah
... that was my first year, but that's okay.
Yeah.
I feel like we're in a really good position. I also hadn't anticipated that we'd make a CEO transition during that year. And, you know, but I, you know, I understood it. I understood Yancey and the board's perspective, and I think, you know, again, you want to talk about Paddy, well-
I think we'll get there. We'll get there
... he's the perfect fit for what we need to do in the next-
We'll get there. I want to do one thing in between before you do that.
Yeah, for the next-
Staying chronological helps me keep my thoughts organized.
Yeah, sure.
You weren't there for Cloudways, right?
No, I started in January. They had done that acquisition in September.
Yeah, but you were there for Paperspace.
I was there for Paperspace.
Yeah, yeah. So tell us how Paperspace came about. And, and tell, and for the audience, tell them what the, what Paperspace is.
Yeah, so Paperspace is a, an AI/ML platform that very much like DigitalOcean, caters to the developer community, making it super easy for people to get on and experiment with AI/ML capabilities and build it into your application. So it's not like the hyperscalers or some of these pure plays, where it's all about how many GPU do you have and what kind of data centers. This is about the software and making it easy for developers to build it into their application. So we thought that was a tremendous overlap with our strategy and a great fit, and a way for us to get kind of into the AI/ML space very quickly. So it was a small acquisition.
We paid $110 million for it, and it didn't come with a ton of revenue. It was, you know, a modest amount of revenue, but a lot of growth potential-
Yeah
... and a very good fit. And so we've been, you know, working to integrate that into our business. And that by itself, that business that really didn't contribute much in 2023, we think contributes 3% growth for the overall business in 2024.
Yeah, so dollar amount is, what are we talking about?
It's in the 20s.
In the twenties, yeah. And then you didn't give a milestone for it on this last earnings call, right? Or did I miss it?
We said it'd be 3%, 3% growth in,
For the year.
For the year.
But you gave us, like, it crossed $1 million in ARR the quarter before.
Yeah, we-
Right?
No, we didn't give an update-
You didn't update that.
... on the, on the run rate. And, and part of the, the reluctance to be precise on that is it's a supply chain constrained environment, so it's not a demand constrained environment. Like, there's lots of demand.
Yeah.
And there's lots of people chasing that demand, so it's not like it's not, you know, super price competitive, to win that demand. But our ability to hit and exceed our plan is a function of can we get the gear when they said the, you know, the distributors said they would deliver it, can we get it turned up on time? And then, you know, I don't worry about us filling it.
We had another CEO of a private company who was here, who was saying that. He's like, "Even once you get the GPUs," he's like, so he had $20 million of stuff delivered to his new data center, and no cabling.
Yeah, no optics. The optics are also-
And he's like, and he's like: "Where's, where's, where's, you know, where's the, the cabling?" And they're like, "Oh." He's like: "When, when can I get it?" "We can't even give you a date on that.
Oh, yeah.
Right.
Yeah.
And so he's like: "Take it back," right?
Yeah.
He made them take it back. Then he's trying to open a data center in an Asian country, and he's talking to the number one, you know, telecom and electricity provider in that country, and they're like: "We can get you one megawatt in Q2 of 2025.
Yeah.
He's like: "1 megawatt is like $10-12 million of ARR.
Yeah.
My ability, my ability to actually grow in this country is totally constrained by power.
Yeah.
Right? So for Paperspace, tell us a little bit more about what the current constraints are.
Yeah, our-
Is it just GPUs, or is it-
No, it's the optics, too.
It's the optics.
But we're not buying at the quantity and scale-
No
... those guys are buying, 'cause it's not like, we're not gonna differentiate on the hardware. We're gonna differentiate on the software, and the ability for developers to take AI, make it simple, and build it into their apps. And that's a different customer base, and our customers don't need like hundreds or thousands of-
Right, it's not the next ChatGPT.
No, it's not.
Yeah.
You know, some customers are using our infrastructure to train, you know, language models, some of them are doing it for inference, some of them are doing the inference on our, you know, the Core DO platform. They're not even using GPUs.
Really?
So it's a different size customer, different problems. They're still solving really cool problems and building, you know, interesting apps, and-
Yeah
... just like Core DO, some of those will be really successful, and they'll grow on our platform, and some of them will be hobbies or ideas, and they won't grow. But that's the different market that we participate in than the hyperscalers looking for these enterprise workloads.
Yeah. Okay, super interesting. All right, now let's get to the big development, right? Which was, were you involved in the search? You, you probably interviewed a ton of the candidates, right?
Yeah, I did.
Without obviously telling us anything that you shouldn't tell us, what was the pool of candidates like, and why did Patty rise to the top?
Patty rose to the top... So the pool was very strong. I was very impressed by the candidates that we saw. It's a great platform. It's a great opportunity.
Yeah. Well, what were you looking for? Let's start with that.
Well-
What were the characteristics that...
So-
- the board was looking for?
... what we said was we wanted someone with more of a product and engineering background-
Yeah
... that had also been an operator, that had a deep understanding of cloud technology, and that had built kind of, you know, applications at scale.
So product and engineering background, that was number one. What was number two?
Cloud tech. Deep, deep understanding-
Cloud, right
... of the cloud market.
Right.
You know, ideally developer, you know, the understanding the needs of developers, and then had some level of operational experience, as well. And we got a lot of really, really interesting candidates. And Patty stood out because, one, from a product standpoint, and engineering, you know, he's developed his kind of craft from Microsoft and Amazon and Oracle and GoTo, and he's got a tremendous amount of experience there on both the product and the technology side. But he's also been a kind of a, you know, a business leader. He's the CEO of GoTo for the last several years as a private company.
The passion that he has, and again, you would've heard it in the earnings call, and when you speak to him, you know, you'll hear this, is he's just got a maniacal passion for us understanding the developer experience on our platform. And, you know, the pushing us to say: "Okay, you guys got to increase the pace of innovation. You got to get more new products out, more capabilities. Keep up with the developers. We've got to be thinking two and three years ahead. What is the developer gonna need in three or four years?
Mm.
And that's what we should be working on." And AI, ML fits into that-
It definitely does.
... but it's not it, right?
No.
That's not the only thing. So it's, I think, you know, the kind of stages of a company as it grows, the last stage was, hey, you got to bring some discipline, you got to get the company public.
When did you meet him first?
Oh, shoot, I don't know.
It was a Zoom, basically, interview sort of thing?
Yeah, Zoom interviews, probably in, well, at some point in the fourth quarter.
Yeah. Then when did he actually start? What was the first day?
He started on the twelfth, I think. It was, I think it was the twelfth. It was a Monday, three weeks ago.
Yeah. And what do you do on the first day if you're Patty and you just landed?
Well, we had a whole day, a whole week set up for him where I had been-
Okay, let's hear what week one was like.
I had been... Well, we had, we had earnings that-
You did, yeah.
... first week, so that was, that was one thing. Or is that - that was the second week. We had a whole week lined up for him, and I had been meeting with him regularly ahead of that. But we basically said, and this is what we communicated on the earnings: What are the two biggest levers for us in terms of hitting our plan this year and, and near-term growth? It's getting NDR back above 100, and it's delivering the Paperspace plan.... So that's what we spent our first week on, is, "Hey, what are we doing to do those two things? And, and here's the team that's doing it, and here's our initiatives, and, and our milestones and our metrics, and how we're gonna, we're gonna hold ourselves accountable for that." And we. It was a deep dive in our top priorities.
And again, that's what we communicated on the earnings call and what we talked to the board about. You know, those are, you know, the Paperspace business is going from effectively zero to some meaningful number-
Yeah.
there's supply chain execution risk. What are we doing?
Yeah.
The second one is, okay, you know, you're seeing an improvement in NDR, the, that bottomed last year, but what are you gonna do to make sure that you get to 100% and above? And some of that is product, you know, releases to drive ARPU up. Some of that is investing in customer success and customer care to try to minimize contraction. You know, those are the initiatives you would think we'd be doing, and then that's what we took Patty through.
Does he primarily listen?
Oh, he's a great. I've found him to be. He didn't come in with a, "Hey, here's my agenda, it's gonna be different than what you guys are doing.
Yeah.
So he's listening to what we're doing. What he's coming in with is more of the thematics, which is for us to be successful-
Mm.
We have to listen to the customers.
I like that, yeah.
We have to embrace and obsess over the customer journey. We have to make decisions more quickly. We have to be more nimble. Those are kind of, you know, operating model constructs, and not like: "Well, I don't like what you're doing here.
Right.
I don't... You know, I've done it a different way there. It's like, let's align on what it's gonna take for us to be successful, and then now let's talk about what are we doing and what fits with that and what doesn't, so.
Oh yeah, listen to the customer, be more nimble. What was the third one? Do you remember? You just reeled them off. I know, like, this happens to my guests all the time. They reel them off and-
Yeah, I don't know what the third one is. It's probably in the transcript.
You know, Snowflake has a new CEO.
Mm. I saw that.
Yeah. And so, the Snowflake CFO was sitting in that chair, and we had a very similar conversation. And, you know, one difference between, you know, new and old. Well, one has a PhD in computer science, and the other one doesn't.
Yeah.
So that's one big difference, you know. One was, is in the San Mateo office every day, the other one lives in Montana, right? One is, like a force of nature from go-to-market kind of perspective. The other one delivers products really fast. So one of the things they were talking about was how, you know, if you look at, like, the three big products that Snowflake's been working on, two of them took way too long. Like, one took, like, two years, and the other one, over two years. But the one that Sridhar was working on went from idea to generally available in a year, right? And they really felt a need to get more products to the salespeople faster, right? So I just- I'm giving you that detail-
Yeah
... just again, just to give you some sense of what would be helpful for us to hear in terms of, you know, differences.
Yeah. Well, yeah, the differences are, I think, fairly stark in our case, in that, you know, Yancey was a former CFO-
Yeah
... an incredibly bright and sophisticated guy, and definitely understood the capital markets and brought a lot of discipline and structure to the business.
Mm-hmm.
But just, you know, that's not his wheelhouse around the product and the engineering side.
Yeah.
Patty lives and breathes that. You, again, just sitting there on the earnings call and listening to him answer the questions, in his seventh day at the company-
He knows what it's called.
... he knows the space.
Yeah.
He knows the space, he understands developers, he understands what it's like to work in a company that's large but try to be innovative. So he's worked in, you know, he ran a startup that he sold to Microsoft. He's worked at Microsoft, he's worked at, you know, Amazon, some of these bigger companies.
Yeah.
So I think he has a good balance for, you know, how do you, how do you be nimble and scrappy in the context of a bigger business that's trying to scale? So I again, I think he's a perfect, perfect fit, and very, very happy in the first handful of weeks-
Yeah
... within the company. Employees, we're a development company. Like, we sell to developers, and we're a company that is largely the product we have is developers that build-
Right
... technology and make it available to other people. He resonates with them.
Yeah.
They've been very, very happy to just hear him talk and provide the vision for the company.
Yeah, that's great. All right, any questions from our audience? How important is innovation in AI and ML to your growth story?
Well, in the near term-
I wish you repeated the question.
Oh, I'm sorry.
Yes.
Yeah, so the question was, how important is AI ML to our growth story? In the near term, it's, you know, it's pretty clear it's gonna be three points of growth in 2024, we think. But I think over the longer term, I think it's incredibly important for the business. Not just for the revenue that we're gonna derive from that, but as you build in AI ML capabilities into the rest of our platform, into our own business, I think it's gonna have a huge, huge impact.
And as we were talking earlier, it sounds like you might have a different opinion. I believe that it's going to enable more developers and more technology people to kinda enter the, you know, that part of the economy, because it's gonna make things, you know. Again, it's not easier per se, but you can be more effective with, you know, a little bit less in terms of the specific technical knowledge out of the gate. That doesn't mean that super technical people are at risk of having their jobs removed.
It just means that you can have more people dabble, and you can have more people experiment, and maybe more people will go out on a limb and say, "Hey, I'm gonna work on this at night while I'm working another job, and I'm gonna see if it works.
Mm.
I would hope that it would lead to more experimentation in the market. Again, that new business formation is very, very good for us in terms of long-term growth.
... Please?
He was first.
Okay.
You talked a little bit about constraints before. To what extent could constraints impact your growth rate going forward? So I think you mentioned 3% as a number for this year. If you had upside, would you go to 5%? If you had, what kind of, what numbers around that might those constraints be?
Sure. So the question was, given the constraints we have, like, what is the potential impact on the growth rate? I'd say, the magnitude of investment that we're making relative to some of the others is relatively tiny. You know, $50 million of incremental CapEx this year on AI/ML-related gear. The good news is we've got a lot of dry powder from the cash we have on hand. We've dialed back the share repurchases to give us some additional flexibility. We could ramp that up if we saw the demand. For us, it's a little bit of, you know, we're just getting started, we're integrating it, we've gotta get it up and running, and we've gotta demonstrate that we have a competitive advantage and differentiated solution.
It's not just, "Hey, go buy a ton of GPUs and put them in a data center and try to turn them on." That's not our business model, and so I don't feel constrained. I'm constrained in that if we wanna go faster, we gotta kinda decide that now, because it's like six-9 months of supply chain stuff you gotta worry about. And so you're having to make decisions a little bit earlier than you would want to, but that's okay. We'll see how things go, and we'll ramp the spend as appropriate.
Last one.
Yeah, thank you. My question is more towards like, how have you guys navigated the uncertainties in the tech space as of right now? 'Cause we're seeing a lot of hiring freezes here and there. But more towards like, how has that impacted your, or shifted the way your business plan or your long-term outlook?
Yeah, it's an interesting question. The question was the uncertainty in the tech space with all the layoffs and everything, has that changed our business plan? And I'd say we're... And this is part of Paddy's mandate, is: How do we get more resource into our development organization, and, you know, keep driving efficiencies in the rest of the business, so we can focus on increasing the throughput? So I'd expect that we're kind of net hirers on the technology front, not worried about the risk around technology, you know, as some of the other folks are. Now, again, we took really aggressive moves last year to get the cost structure right. So it may be that other companies are just catching up to adjusting for the right growth rate.
But I feel like from a technology standpoint, we're gonna be a net grower, not a shrinker in development resources.
Awesome. Well, Matt, thank you for closing out our tech conference.
Yeah.
It was great to have you.
Well, thanks for hanging around.
Thanks a lot. Great to see you.
Good to see you, Pat.