DigitalOcean Holdings, Inc. (DOCN) Q2 2023 Earnings Report, Transcript and Summary

Hello and thank you for standing by. My name is Regina and I will be your conference operator today. At this time, I would like to welcome everyone to the DigitalOcean Second Quarter 2023 Earnings Conference Call. [Operator Instructions] I would now like to turn the conference over to Rob Bradley, Vice President of Investor Relations. Please go ahead.

Thank you, Regina and welcome, everyone, to DigitalOcean's second quarter 2023 earnings call. Joining me today is Yancey Spruill, our Chief Executive Officer; and Matt Steinfort, our Chief Financial Officer. Before we begin, I would like to cover our safe harbor statement. During this conference call, we will be making forward-looking statements, including our financial outlook for the third quarter and full year as well as statements about goals and business outlook, industry trends, market opportunities and expectations for future financial performance and similar items. All of these statements are subject to risks, uncertainties and assumptions. You can review more information about these in the Risk Factors section of our filings with the SEC. We remind everyone that our actual results may differ and we undertake no obligation to revise or update any forward-looking statements. Finally, we will be discussing non-GAAP financial measures on our call today. Reconciliations between our non-GAAP and GAAP financial results can be found in our earnings press release which was issued earlier this afternoon and the investor presentation on our IR website. With that, let me turn the call over to our CEO, Yancey Spruill. Yancey?

Thanks, Rob. Good afternoon and thank you for joining us today. I'm pleased to share the results of another strong quarter for DigitalOcean. As discussed throughout this year, we made it a priority to position the company in 2023 to generate significantly higher free cash flow margins to create a durable platform that generates compelling shareholder returns regardless of the challenges of the macro growth environment. We have also been focused on leveraging our balance sheet to accelerate our business with the acquisition of Paperspace which positions us with immediate offerings in the rapidly growing market for artificial intelligence and machine learning application development. In Q2, we delivered a quarter with an attractive combination of growth and profitability. Revenue grew a solid 27% year-over-year. We delivered strong adjusted EBITDA margins of 43% and delivered healthy free cash flow margins of 27% on the strength of progress we are making in transforming our cost structure. From a top line growth perspective, we continue to see slower growth in the core DigitalOcean cloud business. We saw growth moderate across our diversified customer base with all regions and most industry verticals seeing slower growth which points to the broad weakness that the technology market faced during the quarter. We expect these ongoing headwinds to put further pressure on our second half growth. Although we saw positive signs of stabilization across churn and contraction, both of which have been stable during Q2. The third piece of the cohort performance puzzle that must stabilize is expansion which continued to moderate in Q2, although at a decelerating pace. Until we see stability in all 3 metrics, we can only say that we are bottoming but have not yet reached a bottom for growth deceleration. With respect to churn, it has been stable in the low double digits about where it was before the broader slowdown occurs which is a positive indicator of the strength of our value proposition. Contraction also stabilized in the second quarter, potentially indicating that optimizations are moving behind us, although it remains several hundred basis points higher than when the slowdown began. This reflects a change in our customers' practices. They have been much more disciplined about managing their cloud usage, a trend that we expect may be with us for a while. As for expansion, we have been pleased to see the sequential deceleration of the declines as we've moved through this year. In speaking with our customers, they remain optimistic. Even though they are growing more slowly than they were previously, they are still growing. This provides us with optimism that although we have not bottomed, we are certainly in the process of bottoming. Our lowered revenue outlook announced today reflects our expectation that the bottoming process continues in the second half. We continue to deliver on the transformation of our cost structure which we announced in Q1. There are several components to this transformation, including prudently managing our third-party spend, leveraging our strong procurement capabilities, operating our capital infrastructure more efficiently and shifting our talent mix to be more global. We made significant progress in the quarter driving towards our longer-term target free cash flow margins. Doing so provides us the flexibility to invest both organically and inorganically to accelerate future growth. We have taken advantage of this flexibility by investing on both the organic and inorganic fronts. Internal investment is focused on expanding our capabilities to better meet the needs of our customers as they grow their own businesses and expand on our platform, graduating from learners to builders to scalers. Previously launched product capabilities, such as the premium optimized droplet which are tailored to specific bandwidth-intensive use cases, enabled us to increase our ARPU at share While the expansion of our footprint, such physical footprint, such as the new Sydney data center, enable us to do more -- to more effectively serve the global market opportunity. Other investments such as evolving the customer onboarding process are enabling us to better identify higher-growth potential customers early and get them the support they need to accelerate their use of our cloud services as well as better match them from a product perspective earlier in their life cycle which yields higher ARPU. A good example of this is identifying candidates for a managed hosting solution through cloud rates instead of them self-serving directly onto the DigitalOcean platform with the mismatch of their needs, leading to under optimization of their experience. The newer capabilities that we've introduced over the past 10 months are growing significantly faster than is our overall company. We expect these offerings and the other capabilities on our second half road map such as enhancements to our storage capabilities to be more material drivers of our future growth. We also continue to invest to build out direct sales and partner channels to augment our proven self-serve go-to-market loan. While it is early innings in the establishment of these new channels and they are not yet material contributors to growth. We continue to see them as important growth levers next year and beyond. On the inorganic investment front, we are very happy with the results we are realizing from the strategic acquisitions that we have made to date. We are incredibly excited about the acquisition we reached when we announced. The Cloudways managed hosting business continues to be a very strong addition to our platform with performance that exceeds our initial expectations. We've seen net new Cloudways customers grow rapidly since the time of acquisition, benefiting from both our cross-selling activities and the strength of our highly efficient self-serve model. Cloudways revenue which grew 45% year-over-year in Q2 continues to be a strong contributor to overall growth -- company growth, aided by our focus on top of funnel and customer acquisition enhancements and is becoming a larger percentage of our revenue mix and therefore, an increasingly more meaningful contributor to driving a higher overall company growth rate in 2024 and beyond. Which brings me to the exciting announcement that we made about a month ago that we acquired Paperspace for $111 million. We couldn't be more excited about this highly strategic and synergistic acquisition in the dynamic and explosively growing AI/ML market, whose impacted opportunities projected to be transformative to how the technology sector is driving the global economy. We are very focused on building an AI platform that enables developers and SMBs to leverage the power of these technologies to build and run applications while maintaining our differentiation of simplicity, in how we deliver these capabilities. Paperspace has tremendous potential. We have been very familiar with the team, having tracked them carefully over the years as they have built their business. The addition of Paperspace to the DigitalOcean platform makes tremendous sense on 3 key dimensions. First, DigitalOcean and Paperspace are philosophically aligned both in the segment of the market that we target and in how we differentiate ourselves versus the larger enterprise-focused providers. We are both focused on enabling smaller companies from individual developers to startups to emerging growth SMB customers to leverage technology to grow their businesses. We both win customers in this segment of the market by providing simpler, more cost-effective solutions that can be leveraged on a consumption basis without the constraints of long-term contracts with a higher level of support than these customers get from the larger providers. We expect AI will be a force multiplier for small and medium-sized businesses, just as the advent of cloud computing eliminated many of the barriers to entry for developers and startups to create a digital business, allowing anyone in any geography to start the business. We believe that AI/ML will provide a similar accelerant to new SMB creation and their subsequent growth as it will enable them to scale more efficiently to create new and innovative solutions and to compete in novel and dynamic ways. Paperspace's software layer allows SMBs to execute AI/ML use cases simply rather than just providing access to physical GPUs on demand. This simplicity strategy fits very well with DigitalOcean's key differentiators. With Paperspace, we will be able to provide multiple products for AI and ML use cases, including compute storage and databases, allowing these customers to scale on DigitalOcean's integrated platform. The second dimension of importance is there are substantial synergies between DigitalOcean and Paperspace that will accelerate both of our growth rates. Paperspace serves the fast-growing segment of the market that brings together the combination of AI platforms and accelerated compute which allows customers to build and maintain AI applications while providing the computational capacity required to do this at scale. Together, IDC estimates that the compound growth of the SMB portion of this market will be 36% over the next 3 years. Today, Paperspace has established itself as a leading player serving this market with a proven differentiated GPU-based AI/ML product that serves over 12,000 paying customers today despite having invested very little in marketing of go-to-market. We will accelerate Paperspace growth by leveraging DigitalOcean's more scale, marketing and global go-to-market motions to increase the top of funnel and drive new customers for Paperspace's business. There is also a significant cross-selling opportunity to sell Paperspace capabilities to DigitalOcean's 616,000 plus learners, builders and scalers many of whom are already evaluating how AI/ML can be leveraged to accelerate their businesses. In addition, AI use cases clearly drive increased compute requirements but Paperspace had no capability to capture those additional production workloads requiring customers to leverage other cloud platforms. The combined platform of AI/ML software and high-performance and GPU-based compute will enable us to capture more paper space's current and future customers' broader infrastructure needs for database-as-a-service, high-throughput networking, spaces and Kubernetes, driving higher ARPU and creating more stickiness and loyalty on the DigitalOcean platform. The third dimension that has us so excited is how clearly it will integrate with our platform and helps dramatically accelerate our time to bring this capability to market. With Paperspace, we immediately have a proven AI/ML offering and have demonstrated a clear product market fit and it comes with a team with valuable experience operating in this dynamic market. The talented and entrepreneurial Paperspace team adds more than 3 dozen employees to the DigitalOcean team and brings significant experience providing AI solutions to thousands of SMB customers. Also as an active and important player in the AI ecosystem that is not tied up with the hyperscalers that have plans to build their own chips Paperspace on its own has strong industry relationships, such as having elite status with NVIDIA. These critical industry relationships will only be strengthened when combined with the larger scale relationships that DigitalOcean has with other players in the market. We are already leveraging these combined relationships to ramp up Paperspace's investment in GPU capacity. Clearly, we are very excited about the possibilities that paper space brings to us and you should expect to hear more from us on this potential over the coming quarters. While Matt will walk you through the financial implications of the Paperspace deal on our overall business later in our discussion, I can say unequivocally that after owning Paperspace for just a few short weeks, the potential is far greater than we had expected. We acquired a triple-digit growing business that we feel confident we can accelerate from here. And as such, we expect Paperspace to contribute at least 3 percentage points towards our total growth rate next year. We are excited to be on a path to deliver a truly differentiated set of capabilities with an aspiration to be the AI cloud for SMBs. With our acquisition of Paperspace and aggressive entry into the AI/ML market, many of you are likely interested in the implications on our projected cash flow margins and our current capital allocation strategy. While Matt will cover the specific projections in his commentary, I'd like to share my thoughts on how we view the economic impact on our business at a strategic level and how this translates to our go-forward capital allocation framework. We have demonstrated a clear focus on driving attractive returns on invested capital since going public. With that as our guiding principle over the past 2 years, we have repurchased more than 27 million shares for $1.3 billion and lowered our fully diluted shares outstanding well in excess of the shares granted to employees over the same time frame. When combined with the 6x increase in free cash flow since 2021, our first fiscal year as a public company. Through these actions, we have increased free cash flow per share by more than 590%. As another part of this capital allocation strategy, we have invested nearly $500 million on content, technology and capability acquisitions which have bolstered customer acquisitions, driven ARPU growth and meaningfully increased our addressable market. Although we continue to evaluate M&A opportunities that are consistent with our goals of enhancing our market position to drive profitable growth. In the near term, we are prioritizing the integration investments in Cloudways and Paperspace, both of which are growing substantially faster than the core DigitalOcean business. Our commitment to driving attractive shareholder returns under any market conditions and to delivering the long-term free cash flow margins required to do so, does not change as a result of our entry into the AI/ML market with Paperspace. Our growing free cash flow margin and the $551 million on the balance sheet afford us the flexibility to both invest appropriately in this exciting new market and, at the same time, maintain focus on delivering profitable growth across our entire business and delivering capital return to our investors. We have consistently said that investing in organic and inorganic growth as our first priority use for capital and that remains true. Given the significant opportunity we are looking at with key growth levers, we may moderate the magnitude of buybacks from 2023 levels. However, we will remain committed to managing the portfolio of balanced growth and capital return through share repurchases going forward. I'll close by saying we've made good progress through the first half of 2023. We've dramatically improved our financial profile by driving greater efficiency in our core business, while maintaining the flexibility to aggressively invest in several attractive growth opportunities. We will invest in 2 of our fastest-growing segments in Cloudways and Paperspace, while we continue to target investment in the highest revenue potential opportunities and optimize profitability as we work to return to higher growth in our core business. We continue to see a material opportunity in an expanding addressable market in the growing $100 billion-plus market for SMB cloud infrastructure and are increasingly well positioned to capture our fair share. Now, over to Matt to provide details on our financial results and our outlook for Q3 and for the balance of this year.

Thanks, Yancey. Good afternoon and thank you for joining us today to discuss another solid quarter that highlighted the material progress we have made towards achieving our long-term profitability targets, delivering solid revenue growth in the quarter, while meaningfully improving our margins. We continue to deliver durable and attractive free cash flow growth regardless of the macro growth environment. In my commentary today, I will review our second quarter financial results, provide additional insight into Paperspace's anticipated impact on our financials and share our updated financial outlook for both Q3 and through the full year. Before I begin that commentary, I'd like to provide some context on the material weakness and the tax expense error described in the 8-K we filed earlier today. Over the past several months, we have made meaningful upgrades to our tax capabilities and expertise. With new tax leadership recently in place and with the support of incremental third-party tax advisers, we identified an error in our tax expense calculation related to the treatment of R&D expense in the context of Section 174. The error had an immaterial impact on our full year 2022 financials but did have a material impact on our reported Q1 2023 financials and rose to the level of a material weakness in both periods. The correction of the roughly $18 million overstatement of Q1 tax expense will result in lower net loss and higher non-GAAP earnings per share in Q1 of 2023 and will result in a lower net operating loss balance as of December 31, 2022. We are confident that our new tax leadership, the incremental internal and external resources we have added and other changes we have made to the finance organization will enable us to remediate this issue and to provide more confidence in our tax estimates going forward. Turning to our Q2 performance. Revenue in Q2 was $169.8 million which was 27% year-over-year growth and 3% growth sequentially over the first quarter of 2023. Net dollar retention was 104% for the quarter. NDR declined 300 basis points from Q1 of 2023 which was a deceleration from the 500 basis point decline we saw from Q4 of 2022 to Q1 of 2023. NDR has 3 main components: expansion, contraction and churn. As discussed in our last call, churn has been stable since early in Q4 as customers remaining -- are remaining on our platform at near historical levels. We have also seen a deceleration of contraction in the past 3 months which is a positive signal that customers may be reaching the end of their optimization cycles. These 2 components, either flat or decelerating, we are looking to see a similar flattening or deceleration of the slowdown in expansion before we will be able to conclude that we've reached the bottom of the macro growth headwinds.. On the customer graduation front, we saw continued growth in our higher spend customers as our builders and scalers represented 86% of total revenue in Q2. We added more than 3,600 builders and scalers in Q2 versus the approximately 2,300 we added in Q1, bringing our total to more than 150,000 -- these customers are collectively growing revenue of 28% year-over-year and are a key driver of overall are which increased 14% year-over-year to $90.84. The addition of interest rates with its ability to cross-sell our new AMI -- AI/ML capabilities into our builder and scaler base gives us another to further increase ARPU Profitability was very strong in Q2 as a result of our disciplined execution and the good progress we have made on our cost savings initiatives. GAAP gross margin improved from 56% in Q1 to 60% in Q2 as we both grew into the co-location and bandwidth capacity increases that we had made in late 2022 and early 2023 and we drove efficiency into our cost structure. Adjusted EBITDA was 43% which was significantly above the 34% that we delivered in Q1. This 900 basis point increase was driven by continued cost management, the full quarter impact of the cost reductions we announced in February and the structural impact of shifting more of our employee base to our global capability centers in India, Pakistan and Mexico. Roughly 2/3 of that margin increase was driven by people-related cost savings, with the remaining 1/3 driven by efficiencies in cost of goods sold and other expense reductions. Free cash flow was very strong at $45 million, representing 27% of revenue. This 1,100 basis point sequential improvement from Q1 resulted from higher gross margin, higher adjusted EBITDA margin and the continued discipline we have shown in our capital investment program with capital expenditures remaining at 15% of revenue, consistent with Q1 levels. Given the approximately $18 million overstatement tax expense in Q1 of 2023 that I mentioned earlier on the call, we're still finalizing non-GAAP diluted net income per share for Q1 and Q2 2023. We do, however, anticipate that the correction will increase our previously reported Q1 non-GAAP diluted net income per share and that looking forward, Q2 2023 non-GAAP diluted net income per share will be above our previously guided $0.40 to $0.41 per share. Contributing to the anticipated growth in earnings per share for Q1 to Q2 was the repurchase of 2.8 million shares in Q2 for a total of $103 million at an average price of $37.08 per share. We ended Q2 with fully diluted shares outstanding of 105 million shares, down from 120 million in Q2 of 2022. On the operational front, we made substantial progress locking in the $60 million run rate cost savings opportunity that we had disclosed in Q1. And based on that progress, we expect to exceed our initial full year savings estimates which will give us cushion in the face of ongoing top line pressure. Our strong balance sheet and healthy profitability levels give us significant flexibility to invest -- to accelerate the faster-growing segments of our business without materially impacting our ability to achieve our long-term profitability target. Before providing our financial outlook for Q3 and the full year, I will provide some additional financial details on the exciting Paperspace acquisition that we announced in early July. As Yancey shared, we believe that the AI/ML market is a tremendous growth opportunity for DigitalOcean. Paperspace is a great fit as it is very highly aligned with our existing target customer profile and the acquisition expands our total addressable market within the large and growing SMB cloud market. Given DigitalOcean's proven and efficient go-to-market model and sizable customer base, we believe that we can accelerate Paperspace's already 100% plus growth rate and the Paperspace will begin to contribute meaningfully to overall ARPU, NDR and top line revenue growth as early as 2024. As previously disclosed, Paperspace will have an immaterial impact on DigitalOcean's revenue in 2023, attributing less than $5 million in the second half. Despite the modest 2023 revenue impact, we expect it to contribute at least 3 points of growth for overall DigitalOcean in 2024 based on the known and growing demand in our [indiscernible] To facilitate this growth and to keep up with the rapidly increasing demand we have seen in the short time that we have owned them, we have already accelerated planned investment in Paperspace's GPU capacity. Turning towards our financial guidance for Q3 of 2023; we expect revenue to be in the range of $172.5 million to $174 million which implies roughly 13% to 14% year-over-year growth for Q3. NDR for the third quarter will decline to the mid-90s as we lap the approximately 10% price increase that took effect in July of 2022. On our current trajectory, NDR will increase steadily over the back half of 2023 as we reach the 1-year anniversary of our Cloudways acquisition and Cloudways higher growth begins to contribute within our NDR comfort. For the third quarter, we expect adjusted EBITDA margins to be in the range of 38% to 39%. We estimate 105 million to 106 million weighted average fully diluted shares will be outstanding at the end of Q3. We will provide an estimate of Q3 non-GAAP diluted net income per share in the coming weeks after we finalize our directions and file our revised financials. Given the continued pressure we are seeing on net expansion, we are updating our full year 2023 revenue guidance to be between $680 million to $685 million. This 3% to 5% reduction in our full year projection is driven by a weaker-than-anticipated outlook for cohort growth in the second half of 2023. We've seen steady and durable performance on net new customer revenue from our self-serve funnel. Cloudways continues to grow faster than it was growing before we acquired it and we have made solid progress on the planned monetization efforts that were part of our original 2023 guidance. But as we look ahead, these contributions are not projected to offset the shortfall we expect from lower-than-anticipated growth in our [indiscernible]. Despite the lower projected revenue and the additional costs we're taking on to the Paperspace acquisition, we continue to expect adjusted EBITDA margins to be in the range of 38% to 39% for the full year. Thanks to our strong execution, driving operational efficiencies. We estimate that our investment in Paperspace will have a $15 million to $20 million impact on free cash flow in 2023, excluding the roughly $4 million to $6 million in onetime integration expenses that will be excluded from our non-GAAP metrics. Despite this investment, we remain confident in delivering our previously guided 21% to 22% full year free cash flow margin as we leverage our improving core DigitalOcean margins to fund our Paperspace growth investments. Despite an estimated 3 to 4 percentage point impact on longer-term free cash flow margins, from Paperspace and despite becoming a federal cash taxpayer in the U.S. beginning in 2023, we continue to target long-term free cash flow margins in the mid- to high 20s as we see additional opportunities to drive incremental operating leverage in our core business. We estimate that we will have between $105 million to $107 million weighted average fully diluted shares outstanding for the full year 2023. As I indicated for Q3, we will provide an estimate for full year non-GAAP diluted net income per share in the coming weeks after we finalize our calculations and file our revised financials. That concludes our prepared remarks and we will now open up the call to Q&A.

[Operator Instructions] Our first question will come from the line of Raimo Lenschow with Barclays.

Can I start with Paperspace. Obviously, it's a really exciting one. AI is coming into -- it's a topic for everyone. If you think about your installed base and what kinds are doing with them, like what do you see as kind of uptick use cases or where people are kind of doing more work with that? Is this going to be like a very broad-based adoption product? Or is this going to be some specialized side that I'm going to spend a lot with you? And then I had 1 follow-up for me, please.

Well, I think as we've said before, we've been participating in the portion of AI that's in the inference aspect. What Paperspace brings to us is large language models and other sort of complex machine learning algorithms and applications that require or utilizing GPU capacity versus the standard high-performance compute. What we've seen with our customers is that -- and really a rising tide of inquiries over the last year, related to AI and our potential for investment in AI capabilities. And so again, we serve a long tail of use cases industry verticals, etcetera. And so I think everybody across the SMB landscape and our customer base, like everywhere else in the world is trying to figure out how to get most leveraged, where to get most leveraged. And we're really excited that we now can bring this capability that's consistent with our historical process and capabilities around simplicity. And I could tell you the inquiries have accelerated since we've made the acquisition a month ago in terms of our sales folks, our customer success folks engaging with customers educating them on what Paperspace bring. So I would expect to see a broad adoption, not necessarily tied to any particular industry really as people start to really want to invest to resources to build out use cases to help them drive productivity, whatever the case may be and they can do that on our platform. And likewise, with the paper space customer base, paper space has had a very narrow, just sort of the ML ops and AI capabilities for application development. When their customers have historically wanted to leverage that on the broader compute platform once the applications were developed, they have to go elsewhere and now we can serve those customers. So we see a lot of synergy on top line synergies in both directions here. And again, we're really exciting about that.

Yes. Okay, perfect. That makes sense. And then, one for you, Matthew is, obviously changing full year guidance in the middle of the year is always like a tough decision for a management team. If you think about it like as you that decision, what was for you like the driving thing to say like, okay, I need to change it? And how did you come up with a new level? What are kind of the puts and takes in there?

Yes, it's always a tough decision. I think as everyone saw in our guidance, we had expected moderate growth, only moderate growth in the first two quarters which we've delivered pretty much as expected. But embedded in the second half increase was two things. One, we needed to see a flattening of the -- at least the flattening of the cohort and they couldn't get worse and we talked about that. And then we needed to see the ramp of the monetization initiatives and some of the other investments we made. And the good news is we are seeing the ramp around the monetization initiatives and some of the new products we launched, the city data center, etcetera, are growing kind of generally in line with what we thought. But what we missed was that the cohort is just continuing to see pressure longer than we had anticipated. And as we sat here in July, I mean, clearly, we would have liked to have beat the revenue guidance for the second quarter. And as we saw that coming in line with the expectations instead of being ahead of it and we saw the continued kind of deceleration or the continued slowdown of expansion. Even though we're seeing some green shoots around contraction is decelerating. We just didn't think it was appropriate to hold that out there still. So what we've done is we've given a range which is fairly tight and it's reflective of kind of if it doesn't change from where it is today, it continues at the current decline -- rate of decline that NDR have been about in the mid-90s here in the second half that's what you'll see. And so that doesn't assume any improvement in the cohort. It doesn't assume any dramatic improvement in the rates of growth around our monetization initiatives or on the go-to-market initiatives. It assumes a very modest impact from the Paperspace at less than $5 million. So we believe that we rotated to a, I'd say, a reasonably conservative view of what the potential is for the second half. And I think that's appropriate until we see a bottom because we can't keep hoping every month that hit a bottom. We have to accept the reality that it may not for a couple of more months or quarters.

Your next question will come from the line of Patrick Walravens with JMP Securities.

Maybe Paperspace, first. Just -- so, I saw that just this morning, Yancey, there's a company called [indiscernible] that I think did a $2.3 billion debt financing secured in part by the NVIDIA. It seems like a Paperspace you're entering that same general area. How are these businesses similar? And how are they different?

Well, as I mentioned on the prepared comments, we -- Paperspace has sort of a delivered product with an API that makes it simple and easy for people to build applications on top of the GPU capability for language models and machine learning applications, what have you. And it's a traditional or cloud service I think other folks who are sort of focusing on essentially renting GPU for higher for -- and don't have the other applications, don't have the other build-outs. And so for our customer base, we think they're going to need that because, again, our customers don't have big DevOps, IT development capabilities and whatever they have, they want to dedicate that to their end products. And so I think what we have is an offering tailored towards developers, start-ups and small businesses. And I think other folks in the space are catering to the large enterprise opportunity. We think our opportunity is large as well. So I don't know if it's a winner take all. I think it's -- a lot of people are going to win in this market. We think we'll be 1 of them. And we're sticking to our knitting which is focusing on the -- our end of the market where people need the core value differentiators we have which is support, our community investment and simplicity.

All right. Great. And then, Matt, I think you addressed it but just to be really explicit about it. Are you seeing higher churn?

No. The churn, as we talked about, term saw elevate over the balance of last year but then moderated in the beginning of the year to more historical levels and it's been relatively steady. So it's not customers leaving us. It's the combination of higher contraction which is really customers that are staying on the platform but optimizing that has been of the 3 drivers, probably the biggest headwind for us over the first 6 months of this year. And as Yancey said in his remarks, several hundred basis points higher than it has been historically. But it's moderating -- it's not getting worse. It's at an elevated level but it's not increasing. But what we've seen is that expansion still continues to come down a bit. So the customers just aren't growing as fast as they were a year ago and we've seen kind of month-over-month that, that has continued to get a little bit worse. And so when we say that we're looking to see deceleration, we've seen stability in churn which is in a good spot. We've seen stability recently in contraction but it's still elevated. And we've seen a continued kind of decline in the rate of expansion. And that last one, the decline in the rate of expansion, that's what you need to see flatten or turn that the direction before you're going to be able to say, okay, we're at the bottom and now we're going to start to pick growth up as contraction moderates more. And contraction is stable but it's at an elevated level and expansion is continuing to weaken. But churn has been fine this year.

Your next question comes from the line of Michael Turits with KeyBanc Capital Markets.

This is Billy [ph] on for Michael. You talked about how potentially there is some indication that optimizations are moving behind us. So when you speak about optimization in your customer base, kind of what does that look like? Is it more or less what we've heard from the hyperscalers? Or are there differences in how customers optimize spend on DigitalOcean? Just some more color on that would be great.

Well, there's two points to that. One, because we have a consumption-based model, as we saw last year, as weaker demand happened in the macro, we saw a much faster growth deceleration because as our customers business has slowed dramatically. They're slowing less today than they were a year ago; their spend corrects immediately; so there's that aspect. There's a second aspect about this -- so versus the longer-term contract model that others have, where volumes might be lower than committed payments and people need to just correct -- that's not what we saw because we have 30-day contracts. And so our customers correct on spending for what they're buying. What we've been seeing in sort of late last year, early this year, a pickup in customers calling saying, am I using this the right way? How do I use less but still satisfy my current demand? Did I activate too many compute instances, too many droplets. Am I in too many different locations. Can I use Kubernetes to potentially have a more efficient deploy from my cloud infrastructure. That is the optimization we've been seeing. And I would say the first couple of earnings calls this year, we talked about a pretty active customer conversations. That has really slowed in terms of our support and customer success engagements with customers. I think they're in a good spot and we're seeing that as we talked about, the contraction is flat over a quarter, flattens or stable, let's say, it's at a higher level which I -- as I mentioned, I think we're in a different time as people are in a lower-growth environment, they're going to be much more vigilant about any new instance or how they are executing operationally because they want to be as efficient as possible. Our customers tend to be bootstrap company. So they -- $1 saved is $1 earned for them. So I think that aspect is here to stay. And that plays really into a key strength of ours which is high support, high-touch we engage with our customers on how to best use the cloud, our community investment, our tutorials are also involved in that. And so I think the churn, that's why I'm so heartened by churn being flat relative to where this whole slowdown started because it reflects the fact that our value proposition is very high. Even in a challenging environment, we're able to help our customers through it and they're sticking with us which is going to be critical. We said we're seeing their slowdown in their growth, the deceleration in their growth has really slowed and it's not at a point where we're going to call a bottom, it's flat but it is clearly slowing at a much more dramatically slower rate than it was the beginning of this year or this time -- certainly this time last year. And having that stable churn for me is a leading indicator that when things bounce, we're going to see a significant bounce as the recovery ultimately happens. We don't know when that's going to happen which is why we've changed the outlook as we reflected today but we're positioned well for that because we've been helping our customers through this process. And the consumption-based model sort of immediately tricks their spend so they can focus on operational efficiency and we can help them do that.

Your next question comes from the line of Mike Cikos with Needham & Company.

I just wanted to circle up on some of the earlier comments but I guess, take it from a different angle. I know you guys have been talking about churn contraction and expansion. If I tie all those up and just think about the linearity that played out over the course of the June quarter and even into July now that the month is behind us, can you discuss how things trended as we went from April to May to June, just to give us a sense of what's going on in the quarter for some of those real-time data points that you're seeing?

Yes. So this is Mike, thanks. It's Yancey. When you look at each of those meters were looking at it on a monthly basis in addition to a quarterly basis, churn stayed relatively flat, that's in the kind of low double-digit, 10%, 11% range and it stayed there; hasn't gotten any worse. What we saw though was a continued kind of an increase in contraction which means people are taking money off of our platform and a decrease in expansion, meaning the amount people are adding year-over-year, each month, as I look at it, is getting smaller. But what we said is the rate of change of contraction has moderated. So it's decelerating. It's not getting worse as fast as it was in -- as it was earlier in this year. And in fact, in the last month that we've seen it actually flipped the other direction and we saw it go back on an absolute basis. It actually -- the amount of contraction we saw went down in the last month which is a really positive sign. So, multiple months in a row with deceleration and actual flip in the other direction but the expansion which is the other driver just continued to kind of get smaller. And so that's why we're saying when we're looking at this as we looked at that over the course of those months, we couldn't call the bottom and we can't say whether expansion is going to continue to get smaller in the coming months and is contraction stay at the level it is? Or is it going to continue the 1-month trend we have of moderate? And with that level of uncertainty, we just couldn't continue to go into the third quarter and fourth quarter without providing an update on the on the guidance because it's providing enough of a headwind that is offsetting the other good growth that we're seeing through our self-serve funnel and through the other initiatives that we've been pursuing.

I appreciate the color, Matt. And I really appreciate the transparency there as well. I just wanted to be crystal clear then when we turn to the outlook, again, it seems like you guys are juggling a number of different pieces all in real time. But as far as those assumptions that you have is -- I don't want to characterize but like between churn contraction and expansion, what are the assumptions that you're embedding in your guidance today which gives you confidence in your ability to achieve the numbers that we're communicating to the Street investors?

I'd say we're assuming that nothing gets better than what we're currently seeing. So the rates have declined and the level that we're landing on the things that stabilized. Stay the same. We don't see an improvement in any of those which would imply we're not forecasting a bottom in our guidance.

That's great. That's exactly 1 other 1 in to clear up with my understanding.

Your next question comes from the line of Pinjalim Bora with JPMorgan.

Yancey, can you talk a little bit more about Paperspace. Can you talk about maybe the customers that Paperspace brings to the table some of the use cases, those customers are working on. And as I look at the numbers, I think you're saying $5 million for the second half. So call it, $10 million, $10 million, $12 million for the year, 12,000 customers. So is it something like $1,000 ACV about. Help me understand through that. And then lastly, what portion of Paperspace's customers are on the ML Ops platform versus kind of the core infrastructure?

I think we'll provide more detail around some of that as we sort through some of the analytics on the business in terms of the mix, etcetera at a subsequent call. What I would say is their customers are very similar to ours. They're small emerging startups, many names you would have never heard of. Similar to our customer base from all over the place. Lots of different use cases. Obviously, language models is a big growth driver in the near term. Generative Media is a very interesting one, a very dynamic advertising, media, video etcetera, based upon whatever the vertical or use case. So that's a lot of people taking a lot of data and leveraging that for outcomes and is very exciting. What I learned about it is that the fact is you can look at the logos, you can look at the names in terms of the customers. And like our business, it's long tail. It's -- we enable people all over the planet to leverage their ideas in the cloud. And now we have the ability to do that in AI/ML. And so I think the ARPU is higher. Then I think our average customer will give more specifics later than that. And it's -- that's what we said earlier. It's a really hand in glove fit in so many dimensions. And -- and I think I'm excited to get this in the hands of our DigitalOcean customers and vice versa to give the Paperspace customers the same -- similar kind of experience they've had building their AI apps once they get those over the hump and are running a business around that application that can run on our platform and have the same experience in terms of ease-of-use, simplicity, the support model or community investment and the low cost.

Your next question comes from the line of Jim Fish with Piper Sandler.

This is Quinton on for Jim Fish. Yancey, maybe first for you. I'd like to touch a little bit on the time line side of coming to this Paperspace deal. Thinking back to Q3, Q4, we talked a little bit about how CPUs would be able to support AI workloads. And then kind of Q1, we transitioned to GPUs are probably an attractive expansion to the platform. And then obviously, now we have the Paperspace offering. Can you talk through what trends or maybe inputs changed from Q3, Q4 to Q1, Q2, where now you know you need this kind of GPU as a service offering? And then any color you can provide around the acquisition process of Paperspace, whether it was competitive bidding or anything would be helpful.

Yes. So I don't know that we never -- we ever said. I ever said that we didn't need GPUs. What I think we were clarifying is that there were a number of AI-based algorithmic type business models that we're running on our platform because high-performance compute was fine. I think that tended to be more in the inference market versus the language model market. So I just want to be clear. We have always said that we wanted to build a GPU capability and AI/ML platform capability, certainly since we've been public. We've known Paperspace for years. I've stayed very close to the founders since I became CEO in 2019. So we're excited to, let's say, finally have them part of the DigitalOcean family. But we certainly saw with late last year, early this year, a pickup in the threshold level of capability and market perception and interest in the language model opportunity which the market wants to use GPUs for those, certainly, to get the model up and running, maybe not to operate the model once it is running in customers' hands, certainly, the level of compute needs change and we'll see how that plays out. We're learning a lot very rapidly. But we're excited to have this new dimension of compute now on the platform. And I think whether that was accelerated or not, this is a capability we wanted to add and excited to be able to get this done here in the middle of this year. I'm not going to comment on the process that led to the conversations and what led to our announcement of the Paperspace a month ago.

Yes. Understand. And then, Matt, maybe for you. As I look at the customer segment, it looked like really the scalers is the segment of the business that saw some deceleration in the ARR growth. And I think you've been pretty clear that it's really not churn. So is it fair to think that the optimizations hit this segment of your business, the heaviest compared to maybe some of the builders or learners?

Yes. That's very, very accurate. The larger customer, the more opportunity for optimization. And we've got big examples, Yancey, gave a few but examples where a customer -- a large customer that's storing a fair bit of data on our platform, might change the policy and say, what we used to store 60 days of data and we've kind of -- we've rethought it and we think we can get away with 15 and we just don't need as much given the consumption model, they can make that happen. But that typically happens in the larger customer base. That's why you've seen the slowdown in the ARPU growth in the scalers versus some of the others. But it's not -- the scalers have the lowest churn of all of our segments. The customer gets on our platform and the more that they consume and the more products that they use, the lower their churn. So the churn in scalers is incredibly attractive churn level, just that we're seeing more contraction in that area, as you expect.

And our final question will come from the line of Tim Horan with Oppenheimer.

Yancey, can you talk about how you balance free cash flow generation with growth here? And I think you said high 20% on the free cash flow, maybe some timing there. I mean I only mentioned because it's basically insatiable demand for GPUs. If you doubled your CapEx spend you could probably double your revenue growth. So -- and it seems like now is the opportunity to do that. So, can you just walk us through how you're thinking about balancing that out?

Well, I think what we've reflected today is what we see. I will say this, in the month that we've acquired Paper space, the potential growth that we see is significantly higher than we saw a month ago. I'm not going to put a quantitive on it but it's pretty significant. And we've already stepped up capital purchases ahead of what we thought a month ago to get us through this year and to support acceleration of the growth rate which again, triple-digit growth rate, we already see our ability to accelerate that a month in. And we're getting very focused right now on what's the 2024 opportunity. I think 1 of the reasons we are sort of reframing expectations around capital allocation is this issue because to your point, if we can grow 500%, we will look to do that. We'll just see what the opportunity presents itself and we're very focused on that. And I know we talked a lot about free cash flow this year in a year where the growth was uncertain. And I think -- I hope people didn't misread that as we were allergic to grow. We've always said we want to invest our primary uses of capital or whether it's internal or externally generated growth and we're demonstrating that now we're excited. And I remember when we had the conversation with the Board that we're going to take margins down from what they could have been to invest more in paper space and there was all hands in the air. So we're excited about this. And we'll see where it plays out. Obviously, we'll have more of an update, a formal update at the next earnings call on what we're seeing. But I'll just say that the pace of this opportunity continues as it has in the first 30 days, we'll continue to see a lot of money going into paper space because it's a very exciting and opportunity we have.

Congratulations on developing in an AI strategy and expense reductions looks really impressive.

I'll now hand the conference back over to Yancey Spruill for closing remarks.

Thank you and thanks, everybody, for joining us today. As usual, we really appreciate your support. And we're looking forward to talking with you over the coming weeks and months about where we are and where we're heading. And we hope that you all have a good rest of the day. Thank you so much.

That will conclude today's conference call. We thank you all for joining and you may now disconnect.