Thank you for standing by, and welcome to the DigitalOcean Second Quarter 2024 Earnings Conference Call. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question-and-answer session. Thank you. I'd now like to turn the call over to Melanie Strate, Head of Investor Relations. You may begin.

Thank you, and good afternoon. Thank you all for joining us today to review DigitalOcean's second quarter 2024 financial results. Joining me on the call today are Paddy Srinivasan, our Chief Executive Officer, and Matt Steinfort, our Chief Financial Officer. After our prepared remarks, we will open the call to a question-and-answer session. Before we begin, let me remind you that the statements made on the call today may be considered forward-looking statements, which reflect management's best judgment based on currently available information. I refer specifically to the discussion of our expectations and beliefs regarding our financial outlook for the third quarter and full year 2024, as well as our business goals and outlook. Our actual results may differ materially from those projected in these forward-looking statements. I direct your attention to the risk factors contained in our filings with the Securities and Exchange Commission and those referenced in today's press release that is posted on our website. DigitalOcean expressly disclaims any obligation or undertaking to release publicly any updates or revisions to any forward-looking statements made today. Additionally, known GAAP financial measures will be discussed on this conference call, and reconciliations to the most directly comparable GAAP financial measures are also available in today's press release, as well as in our investor presentation that outlines the financial discussion on today's call. A webcast of today's call is also available on the IRS section of our website. And with that, I'll turn the call over to Paddy.

Thank you, Melanie. Good afternoon, everyone, and thank you for joining us today as we review our second quarter results. DigitalOcean delivered a strong second quarter, building on the momentum from the first quarter and continuing to execute on all key metrics. In my remarks today, I will briefly highlight our second quarter results, provide an update on the leaders we hired recently, share tangible examples of our increasing product velocity, and discuss how we are capitalizing on our AI growth opportunity. First, I would like to briefly recap our second quarter 2024 financial results. Revenue growth has continued to reaccelerate in the second quarter to 13% year-over-year, reflecting the growing signs of success we're seeing from both a product and go-to-market standpoint and the continued acceleration of our AI and machine learning products, where ARR has grown over 200% year-over-year from the Paperspace ARR we acquired last year. In Q2, we also saw the largest step-up in incremental total company ARR in nearly two years, excluding the quarter in which we acquired the AI/ML business. We also delivered strong adjusted EBITDA margins at 42% and adjusted free cash flow margins at 19%, exemplifying our ability to demonstrate ongoing cost discipline and optimization while continuing to accelerate product innovation. Our second quarter financial results highlight the progress we are making and our ability to execute on the plans we laid out at the beginning of the year. We're also encouraged by the signs of improvement within both our growth profile and our key fundamentals. Net dollar retention was flat versus the previous quarter at 97%, as expansion within our customer base continues to be lower than historical levels, given that we're still navigating a challenging macro environment, which is muting the positive impact of our increased product velocity and the stability we have seen in churn and contraction from our solid execution on the various customer success motions. In addition to the increased momentum from our AI/ML products, we received healthy revenue contributions from both our managed hosting products and new customers. Matt will walk you through more details on our financial results and guidance later in the call. In addition to our solid financial performance and accelerating product innovation traction, I'm also excited about the advancements we've made in building out the team. We added three critical new leaders to our executive team over the past several weeks. First is Bratin Saha, our Chief Product and Technology Officer, who will lead product strategy, product engineering, infrastructure, and security. Most recently, Bratin built AWS's multi-billion dollar AI, machine learning, and data platforms, which together represented one of its fastest growing business segments. Previously, Bratin worked at NVIDIA and Intel, running many of their software infrastructure platforms. We also announced Wade Wegner as our Chief Ecosystem and Growth Officer, which is a unique role that is highly appropriate for DigitalOcean, as we are a very unique company. Our cost-efficient self-service customer acquisition model is one of the most efficient in the industry. As I have said many times, one of DigitalOcean's strengths and a key driver of our customer acquisition model is our passionate community of developers, many of whom have grown or are growing up learning to code on our platform. Wade and his organization will be responsible for supercharging our engagement with this community and for driving our very distinct product-led growth motion. Finally, we recently announced Larry D'Angelo as our Chief Revenue Officer, who will bring his years of experience building and scaling high velocity go-to-market teams to drive direct sales and partner sales to augment our product-led growth engine, and also to build scalable customer success and support functions to help our customers be successful and expand their footprint on our platform. DigitalOcean's strong fundamentals and future potential drew these three world-class executives to join us in our journey. Their arrivals have also created further hiring momentum, as having top talents such as these three new executives tends to attract additional world class talent. We're already seeing this dynamic play out as they fill out their respective teams. I'm very confident that we now have the right executive team in place to fuel growth, increase product velocity, help our customers be successful, and continue executing on our mission of making cloud and AI simple and accessible for developers. Now let me give you an update on our products. As we continue to listen to our customers and incorporate their feedback, enabling them to grow and scale on our platform, we released 24 new product features throughout Q2, doubling our product velocity from the prior six months. We also revived Deploy, our virtual developer conference, which was held on July 9. I'm thrilled about the success of this event and look forward to continuing to engage with our developer community as we intend to increase the frequency of our deploy events and do them on a regular cadence going forward. During our July event, we announced a number of material product announcements in response to customer feedback. First, we announced GPU Droplets in early availability mode, and this launch democratizes on-demand access to Nvidia H100 GPU instances for our customers, enabling them to leverage one, eight, or more GPUs at a time, providing flexible deployment options tailored to various use cases and budgets. A lot more on this later. During Deploy, we also announced our global load balancer product, which we refer to as GLB, which is currently in public beta. This is engineered to bolster application resiliency, eliminate single points of failure, and significantly minimize end-user latency while securing GLB traffic from denial-of-service attacks. It offers global traffic distribution based on geographic proximity to the end-user, dynamic multi-regional traffic failover, data center prioritization, edge caching, and automatic scaling of the GLBs. It is intuitive, predictably priced, and tailored to the essential needs of growing technology companies for enhancing their global resiliency. We also recently announced that select DigitalOcean products can now be used to host electronic protected health information. This allows companies such as telehealth providers, healthcare software applications, and health tech organizations to build and scale sensitive workloads regulated under HIPAA on our developer cloud, leveraging select DigitalOcean covered products. During the quarter, we also launched Managed OpenSearch, a comprehensive solution designed for in-depth log analysis, simplifying troubleshooting and optimizing application performance. With Managed OpenSearch, customers can now pinpoint and analyze log data with ease, customize log retention, enhance security of their applications, scale to fit capacity needs, and forward these logs from multiple sources. During Q1, we announced that we offer premium memory optimized droplets and premium storage optimized droplets, and in Q2, we finished rolling this out to all our data centers, a huge milestone for us. We also announced improvements to our app platform, including auto scaling, dedicated egress, and an expanded lineup with entry-level dedicated instances, higher data transfer allowances, and reduced bandwidth overage fees. Dedicated egress provides application developers with fixed IP addresses, enabling them to meet the security needs of their customers or run applications that require whitelisting for authentication purposes. Additionally, with the new expanded lineup, customers can now start small and grow on the platform with auto-scaling. Reduced bandwidth overage fees help customers deploying bandwidth-intensive applications. These updates allow customers more flexibility and features to deploy their production applications. Now turning to our managed hosting cloud-based offering, we launched Malware Protection, which detects malware and protects our customers from cyberattacks. This add-on includes critical capabilities such as phishing protection, files protection, database protection for WordPress and Joomla, automated malware cleanup, proactive defense, and cron malware cleanup. These are just a few highlights as we continue to add new capabilities and features to achieve our objective of simplifying cloud and AI infrastructure for our customers. We will continue to listen closely to our customers and accelerate our product velocity so that customers continue to scale and grow on our platform, which is our primary focus as we work to drive up expansion and improve net dollar retention. And now I'll pivot to a part of the business that is seeing a lot of momentum, our AI/ML offerings. We continue to see very strong demand for our AI platform. To support that growing demand and to take the first step of our long-term data center optimization strategy, I'm very excited to announce that we will be opening a new state-of-the-art data center in Atlanta in Q1 of 2025. This not only expands our geographic footprint, providing us cost-effective additional coverage across the U.S. for our core workloads but also gives us near-term incremental space and power to support our AI strategy and growth. This new data center is also a key part of our medium-term strategy to reshape our data center footprint, including consolidating workloads from DCs that are currently in expensive locations, including New York City, San Francisco, and Toronto, enabling us to improve our gross margin profile over time. As a reminder, opening a new data center gives us ample runway to grow into the additional capacity, and we only add equipment and spend capital as needed to meet demand. As such, the financial impact of this long-term investment will appear steadily over time as we ramp capacity and leverage it for consolidation of our core workloads, and also for AI training and inferencing as that demand evolves over time. We will share additional details over the next few calls as we start building our new data center out and make further progress on our data center optimization strategy. Given this DC expansion and with the modest increase in AI-related capital that Matt will detail in his remarks, it is worthwhile for me to spend a little bit of time providing some context on our AI strategy and how we view this market opportunity. Today, the majority of AI action across the industry is in the foundational infrastructure layer, with a handful of companies providing GPU infrastructure to a relatively concentrated set of customers that require GPU compute for foundational model training. But over time, we expect generative AI and AI overall to follow a similar progression that the market has seen with other technology evolutions, with the action shifting up stack from infrastructure to platforms to eventually applications in the coming years to deliver actual business value to customers. The heavy users of the infrastructure layer today are those building foundational generative AI models or those extending those foundational models by injecting their own data. This requires a lot of deep expertise in machine learning, data, and foundational models. This restricts AI and associated innovation to well-funded startups and large enterprise companies with very skilled staff given the limited talent pool and high costs associated with this emerging technology, leaving behind the vast majority of companies who don't have access to these capabilities. Our mission at DigitalOcean is to change this paradigm by democratizing access to generative AI and AI infrastructure for all customers, just like we did with core cloud computing services, using simple-to-use software platform components rather than expensive CapEx-heavy hardware infrastructure. As a significant step in this direction, we announced the launch of GPU droplets, allowing customers to seamlessly leverage AI technology into their workflows and applications using as few as one or eight GPUs in an on-demand mode. GPU droplets remove the burden of managing the full lifecycle of GPUs and the orchestration associated with its usage. This type of fractional on-demand access to GPUs is not widely available in the market today. We have seen very robust demand for this capability, which is still in early availability mode. Additionally, applications that consume AI also need the usual cloud primitives like compute, storage, databases, security, and so on to be deployed in the real world and deliver real business value. Unlike applications that are built on pure GPU farms, software that consumes AI through GPU droplets can seamlessly take advantage of DigitalOcean's core cloud computing platform, making it easy for customers to transition from R&D mode to production very seamlessly rather than having to go through redeployment. Let me give you some specific examples of customers that are building on our AI platform. First example is an advanced stage startup building a lightweight but very fast AI code completion tool for developers with a very large context window using native neural network architecture on our platform. Another example is an AI infrastructure management company that offers a middleware layer to enable rapid training and inferencing for generative AI models on the DigitalOcean platform. To recap, our longer-term AI vision is more software-centric, with the mission of making it easy for our approximately 638,000 current customers and other companies that look like them to leverage AI in their application stack without needing super deep AI and machine learning expertise. Now, with Bratin Saha, one of the most accomplished AI leaders in the industry, leading the charge for us, we will build on this momentum we have generated over the last couple of quarters and fulfill our mission to democratize AI and make it accessible to all companies. In conclusion, I'm very pleased with the team's performance in the first half of the year. We have seen growing signs of success in our AI machine learning business, growth in our core business is reaccelerating, and I'm excited about our near and long-term growth potential across all areas of our business. We have the right leadership team in place and are focused on accelerating our product roadmap and delivering new capabilities that we announced this year at Deploy and enhancing our go-to-market motion in the second half of the year. I will now turn the call over to Matt to provide additional details on our financial results and for our outlook for Q3 and the remainder of the year. Over to you, Matt.

Thanks, Paddy. Good afternoon, everyone, and thanks for joining us today. In Q2, we continued to execute on the plans we laid out at the beginning of the year. We made progress on key metrics. We continued to see revenue growth reaccelerate, and we delivered favorable adjusted EBITDA and adjusted free cash flow margins. Revenue in the second quarter was $192.5 million, up 13% year-over-year and up 4% quarter-over-quarter. We added $32 million of annual run rate revenue, or ARR, in the quarter, which was 158% higher than the incremental ARR we generated in Q2 of 2023 and was also the highest step-up in nearly two years, excluding the quarter in which we acquired our AI/ML business. Contributing to this growth was healthy incremental revenue from new customers, increased momentum from our AI/ML platform, which saw significant growth quarter-over-quarter, and contributions from our managed hosting platform, which continues to be one of our faster growing platforms, all of these together offsetting a flat quarter-over-quarter net dollar retention rate from our existing installed base. Our Q2 net dollar retention rate was 97%. As we saw last quarter, we continued to see stable performance in net expansion, which is defined as expansion net of contraction on our core DigitalOcean platform. Contributing to this stability was our increased product velocity that drove an increase in ARPU, helping to offset the broader macro pressures on net expansion in our customer base. Our churn levels have also remained very stable for over a year across the business. We are encouraged by the stability in NDR and the modest sequential improvements we are seeing, despite the challenging macro environment, which is muting the pace of improvement in NDR, and despite a positive but lower contribution to NDR from our managed hosting platform, now that we have fully lapped last April's price increase. We continue to expect stable NDR and expansion levels through the end of the year, despite these ongoing headwinds. To further improve our net dollar retention rate, we will continue our solid execution, accelerating our product roadmap, refining our pricing and packaging models, and enhancing our customer success motions. Beyond NDR, we continue to see acceleration within our AI/ML platform. Q2 AI-ARR has grown over 200% year-over-year from the ARR we acquired last year. We have also successfully navigated much of the initial supply chain and implementation risk that we had identified earlier in the year and are now working aggressively to keep up with demand. We anticipate this momentum to continue for the balance of the year, given the demand for our AI solutions. Turning to the P&L. Gross margin was 61%, which was consistent with the prior quarter and up 100 basis points from the prior year. The 100 basis point year-over-year improvement is primarily a result of the success of our ongoing cost optimization efforts, which to date have more than offset our continued investment in AI infrastructure. Adjusted EBITDA margin was 42% in the second quarter, which was ahead of guidance and approximately 200 basis points higher than the prior quarter. This beat was primarily driven by strength in gross margin and our ongoing operating cost discipline. Diluted net income per share was $0.20, and non-GAAP diluted net income per share was $0.48. GAAP and non-GAAP diluted earnings per share increased by $0.19 and $0.04, respectively, on a year-over-year basis. This is a result of our ability to increase our per share profitability levels by driving both operating leverage and reducing our share count. Finally, Q2 adjusted free cash flow was $37 million, or 19% of revenue. Turning to our customer metrics. Our total Q2 customer count was approximately 638,000, representing an increase from 637,000 customers in Q1. The number of builders and scalers on our platform, those that spend more than $50 per month was approximately 161,000, an increase of 7% year-over-year. The revenue growth associated with builders and scalers was 15% year-over-year, ahead of our overall revenue growth rate of 13%. The number of builders and scalers on our platform, which represent 87% of our total revenue, increased by approximately 3,000 quarter-over-quarter. The continued growth of our largest spending cohorts is a direct result of our focusing our product development and customer success investments on these builders and scalers. The increase in our higher spend and higher growth customers also resulted in our total average revenue per user, or ARPU, increasing 9% year-over-year to $99.45. With our substantial pre-cash flow generation, our balance sheet remained very strong as we ended the quarter with $443 million of cash and cash equivalents. We also continued to execute against our ongoing share repurchase program and completed $10 million of repurchases in the quarter. Moving on to guidance. We expect Q3 revenue to be in the range of $196 million to $197 million, representing approximately 11% year-over-year growth at the midpoint of our guidance range. For the third quarter, we expect adjusted EBITDA margins to be in the range of 37% to 38%, and non-GAAP diluted earnings per share to be $0.39 to $0.41, based on approximately $102 million to $103 million in weighted average fully diluted shares outstanding. As a result of the steady performance in our core platform and strong demand we are seeing for our AI platform, we are increasing the bottom end of our full year revenue guide by $10 million, projecting revenue to be in the range of $770 million to $775 million, a $5 million increase in the midpoint of our guidance range, and representing year-over-year growth of approximately 11% to 12%. As demonstrated through the first half of 2024, we remain committed to driving continued operating leverage in our core DigitalOcean platform. Given our solid performance in the first half of the year, we are raising our adjusted EBITDA margins guidance for the full year to be in the range of 37% to 39%. Turning to adjusted free cash flow. We anticipate making appropriate incremental investments through the second half of the year, as we continue to capitalize on the AI opportunity to fuel future growth, although, we anticipate these investments having only a modest impact on our cash flow margins. We expect adjusted free cash flow margins for the full year to be in the range of 15% to 17%. As a reminder, adjusted free cash flow can vary quarter-to-quarter, given the variability of our capital spend and our working capital timing. From an overall strategic capital allocation perspective, we will continue to be good stewards of our capital and will evaluate opportunities to maximize shareholder return, maintaining financial flexibility while continuing to evaluate investments across share repurchases, incremental capacity, and balance sheet management. We are also raising the top end of our prior non-GAAP diluted earnings per share guidance and now expect this to be in the range of $1.60 to $1.70. That concludes our prepared remarks and we'll now open it up for Q&A.

Thank you. We will now begin the question-and-answer session. Your first question comes from the line of Gabriela Borges from Goldman Sachs. Your line is open.

Good afternoon. Thank you, and great to see the step-up in ARR. Paddy, I wanted to follow up on your comments on AI strategy. Help us understand, are you saying that other GPU providers don't offer the same level of fractional access that you do with droplets? And then, broadly speaking, comment on the competitive environment. There are a number of well-funded GPU providers that are spending orders of magnitude more CapEx than you are. How do you think about the sustainability of your differentiation when you're up against competitors that have that kind of CapEx fund? Thank you.

Thank you, Gabriela. Great question. So as I started the prepared remarks by talking about our AI strategy and how it is differentiated, I want to remind everyone we are also a very strong player in the GPU infrastructure as a service today. So we have competitive offerings all the way from bare metal to virtualized environments. What we announced with the GPU droplets is a strategy which enables access to fractional GPU and the ability to orchestrate the lifecycle through the virtualized environment that we provide with GPU droplets. So it just takes the overhead of having to manage the whole infrastructure, making it super easy for those workloads which we believe are going to be more important for customers that are digital natives as well as the types of customers that typically reside on our platform. They are looking to build applications that consume AI and leverage different parts of the AI infrastructure in a way that extends existing AI models. They're not model builders per se in the sense that they're not trying to build another foundational model. So we feel our AI strategy, which includes the GPU infrastructure, is tailor-made for customers that are looking to consume AI, not necessarily build foundational models. When I talked about the GPU droplets, that's an abstracted version of the core GPU as a service. We also announced and showed an early preview of what we call the next generation of our platform as a service offerings at the Deploy conference. These are endpoint APIs for well-known open-source models like Llama 3.1 and so forth, which we will be releasing in early Q4, enabling another layer of abstraction to build applications on top of LLM platforms. We believe our strategy is going more up stack and enabling applications that derive business value from AI rather than focusing on model builders that are building and training foundational models. There will be different needs for customers that are looking to derive business value and build applications and platforms on top of our infrastructure, which is our focus. There will be room and space for core infrastructure and GPU farms, and there is, as I mentioned, a very concentrated group of model training companies that are building foundational models, whether it is for question-and-answer, like a large language model for question-and-answer bots, or for text-to-image generation. But our hypothesis is that market is very concentrated, and over time, there will be a few but very big model companies in the world, and the rest will move towards building applications on top of this infrastructure, and that's where our focus is. That's historically been our strength, to simplify access to complex infrastructure by providing essential building blocks of platform and application building, which is the strategy we are following.

Thank you for the detail. Matt, the follow-up is for you on the updated outlook. How did you think about the pace of net new ARR into 3Q and 4Q? Back at the end of it, Matt, I think it implies a moderation from the result that you just put out this quarter. So help us understand how you thought about handicapping that, call it, $7 million, $8 million that you were able to deliver in 2Q into 3Q and 4Q as well.

As we consider growth, especially in the AI segment, it's somewhat variable, and we mentioned earlier this year that we had some risks factored in. This was part of our decision to provide a wide revenue range in our initial guidance due to supply chain concerns. We have successfully managed through those risks, allowing us to significantly increase our capacity, which we had ordered last year and was realized in the second quarter, contributing to an increase in annual recurring revenue. Looking ahead for the rest of the year as we continue to add capacity, we expect less volatility, so our guidance will be lower than what we experienced in the second quarter. The implied growth rates for the third and fourth quarters are impacted by two overlapping factors that may make it appear as though growth is slowing, despite clearly accelerating performance in our core and showing growth in other areas of the business. The third quarter will be the first quarter where we will see the Paperspace AI revenue compared to the previous period, which results in a 200 basis point reduction in growth rate. Additionally, we are comparing against a price increase and nominal customer growth in the Cloudways business, which adds about 100 basis points to the year-over-year growth apparent slowdown from the second to the third quarter.

Your next question comes from the line of Raimo Lenschow from Barclays. Your line is open.

Thank you. Can I go back to the first question that was asked? Just given that we are kind of early on that Gen AI journey, we don't fully understand how this is going to play out. So is the message from you guys that as I train a model, I need hundreds, thousands, tens of thousands of GPUs, but as I run the model, then if I have an application, I can do that with a much, much smaller GPU number, like the one to eight that you were suggesting today? It feels a bit like, because you're early in the journey, it feels a little too good to be true. Can you speak to that a little bit more?

Yeah, Raimo. Thank you for the question. There are three types of AI users today. First are the true foundational model builders, like an Anthropic Mistral, who are building the foundational models. The second one is AI extenders, who take these models, say, Llama 3.1, an open-source model, and inject or enhance it using their custom data. It could be a company that has geospatial data and enhances an existing LLM with their own custom data source, creating a modified version of an existing foundational model. For this class of customers, they actually don't need hundreds of thousands of GPUs. The third type of user is an AI consumer. For example, let's say you're creating a new AI native CRM application or a supply chain application relying on a robust foundational model, again, like Llama 3, but bulk of your application is to deliver a supply chain forecasting algorithm, heavily relying on AI. But you don't need the same physical raw compute power as the foundational model builders or extenders. So we are addressing the needs of the second and third categories that I talked about. As with any technology wave, you have infrastructure providers, in this case, NVIDIA and the foundational model builders. They lay the infrastructure, but true business value is built when platforms are created, like operating systems based on x86 architecture. As this AI wave goes up stack from one layer to the other, we feel there is tremendous need to democratize access to these GPUs and also provide other software frameworks on the platform and infrastructure layer.

Okay. Perfect. That makes total sense. Thanks for the clarification. And then one for Matt, if I think about the improvement in gross margins you mentioned from just the optimization work you do around the data centers, what sort of magnitude are you thinking? I don't want guidance, but is this like a few basis points? Is this kind of more meaningful? How should we think about that? Thank you, and congrats from me, as well.

Thanks, Raimo. That's a great question. I've touched upon this on prior calls. When I first joined the company, I observed that the data center footprint was interesting in that we were in tier 1 markets and tier 1 buildings. That was the way the company had architected its data center network, a relatively expensive way of building. Over time, we would want to move to more tier two locations and be more of a wholesale than retail kind of data center footprint. The Atlanta data center we announced is a great example of that, where we'll be able to take down incremental capacity. It will give us room for the expansion of our AI capabilities, but we'll also be able to take workloads out of our more expensive locations, reducing or consolidating those footprints and shifting less latency-sensitive workloads into different locations. This will be a multi-year effort; it's not something we can do quickly, but I think there's a meaningful improvement we can drive in gross margin over time. The goal is to continue to drive gross margins up, offsetting the lower gross margin mix that comes as our AI business grows.

Your next question comes from a line of Kingsley Crane from Canaccord Genuity. Your line is open.

Great. So first question would be, you cited customer success efforts as helping to create stability in churn and contraction. How much could customer success be used not just to prevent churn but also drive usage in these high potential accounts?

Hello, Kingsley. Good to hear from you. It's a great question. Our customer success efforts are currently nascent, but ramping up. They are tasked not only with managing relations with our top customers but also driving product usage and showing them the breadth of our platform. After I joined six months ago, we've put considerable resources into closing the gap in what our top end scalers need and adding new capabilities. Our customer success reps are trained and take this message to help expand the footprint of large customers in leveraging our platform. This effort is just beginning, but with Larry on board, there will be a push for us. The purpose of getting someone like Larry is to leverage his skills in building a high-velocity motion to augment our phenomenal product-led growth and ensure these customers know the breadth of our platform, leveraging the capabilities to drive up expansion.

That makes a lot of sense. Then I want to return to the GPU topic. A few comments on the availability. It's great to see that out in the market. The fractional demand consumption is relatively unique. We've spoken in the past about software differentiating you from other providers. So I just want to check in on Gradient specifically. How important is that to the vision and how is it acting as an on-ramp for customers?

Yes, I completely agree that software as a differentiator will be significant for us. So right now, let me refresh everyone's memory regarding our offerings. At the bottom layer, we have a bare metal GPU platform. We just announced the GPU Droplets, which provide virtualized, fractional, and on-demand GPU access. Above that is our platform as a service offering, Gradient. We are working on a lot of enhancements to the platform as a service layer and will announce several capabilities over the next 90 days. We believe this will be an important entry point for customers adopting AI capabilities. The abstractions currently allow customers to organize their AI/ML workspaces to share within teams and publish to different repos. There are many reasons for them to come there, and when they test or deploy to production, they leverage our GPU infrastructure, which includes various hardware options, not just H100s, to help customers with different use cases. Hence, software is extraordinarily important as we move up the stack, and Gradient is a differentiated offering for us.

Our next question comes from a line of Patrick Walravens from JMP Securities. Patrick. Your line is open.

Great. Thank you. First, I want to commend you on your recruiting. The background, like Bratin in particular, is just fabulous. Someday I'd like to hear how you did that. My more specific question is, what are the bottlenecks to growing your AI footprint in terms of the data centers? Data centers that were built for CPUs don't work very well for GPUs, right? How are you thinking through that?

Thank you, Pat, for the kind words. You're right. There are constraints, but just to be clear, the new data center I just mentioned is not live yet, so all AI workloads are currently in existing data centers, whether in New York or Amsterdam. Is it easy? No, but we've figured out a way to do that. There are multiple constraints. Power and cooling are critical, but we also have to consider the network stack, which differs from our CPU offerings. We need to establish parallel network stacks for GPUs, and the density of our GPU racks is quite different from traditional CPU racks due to power and heat. Finding more data center space in existing facilities is expensive and challenging. Therefore, we decided to go to Atlanta. Also, this helps us consolidate and balance workloads across our expensive data center footprint in the short term.

16 data centers across nine regions. Did you have your slide deck? Does that include the Paperspace, whatever Paperspace has?

Yes, it does. And Paperspace coincidentally happens to be in our New York City data center, so yes, it includes that.

Your next question comes from a line of Mike Cikos from Needham & Company. Your line is open.

Hey, guys. Thanks for taking the questions here. I want to continue the thought on the data center optimization strategy you're working through. With the Atlanta data center that you announced here, should we think about that as really taking on most of the AI workloads or is it freeing up capacity for you to build out those GPUs in your existing footprint? I know when you went through those three types, the foundational model builders, the extenders, the consumers, I would think that at least the extenders would probably be more concerned about security, needing to be in a tier three, maybe even a tier four data center given security restrictions. Can you help us think about the ability to re-architect your capacity for these GPUs? I know we're announcing Atlanta today, but should we expect a more significant build-out even beyond Atlanta?

Yes, thanks, Mike. I'll start with the last question. Should we expect a more significant build-out? I don't think so. We have enough capacity now for the foreseeable future. But as we go up the stack, workload goes from training to what's called inferencing. Inferencing requires minimizing latency. If you ask me whether all of Atlanta will be consumed by AI, no, because that’s why Matt and I have talked about needing to consolidate and load balance across our footprint. AI capacity will also be in the West Coast, Canada, Europe, or Asia because our inferencing demand grows. Some of our customers already ask for GPUs in inferencing mode closest to their customer density. Thankfully, we have a distributed data center footprint globally. We'll use this opportunity to optimize, load balance, and consolidate data centers to improve core business margins, get out of costly locations, and effectively manage AI workloads.

To answer the first part of your question, Mike, Atlanta won't be filled with GPUs, but it will take a big chunk of our capacity. Much of that comes from migrating GPU capacity out of expensive locations, providing a more cost-effective solution for relevant workloads. This opportunity combines extending our GPU runway with load balancing and cost optimization in our existing footprint.

Just to add, we will not rush to fill out the capacity to the brim. This gives us land, space, and power, but we will fill it up as part of our consolidation strategy and as AI demand grows in the next few years.

Got it. And then just a quick follow-up for Matt. Coming back to the prepared remarks on the net dollar retention. So the takeaway is that we're stable at 97%, and based on the guide you have, the team's assuming that 97% remains for the rest of the year. The second clarification here, I know last quarter you mentioned $19 million in AI-related ARR. Can you give us an update on what that is this quarter?

We're not going to guide or provide details at the platform level at this point. With the 200% growth rate from what we acquired, you should ascertain the ARR for the AI business. The first part of your question was about the 97%. Our aim is to drive NDR above 100 and have a tailwind. We're making significant progress on various initiatives contributing to new products and increased product velocity. Still, expansion is stubborn and not increasing as fast as we'd like. Customers aren't growing as quickly as before, and their recovery isn't as fast. So, we think it will remain stable. However, we're working actively to improve it and are optimistic about movement in it. It won't stay completely flat, but for guidance, we're saying it will remain stable.

Your next question comes from a line of Pinjalim Bora from JP Morgan. Your line is open.

Hi, great. Thanks for taking the questions and congrats on the results. Paddy, you have really strengthened the leadership team, obviously with the number of hires that you highlighted. As these leaders settle into their roles, how should we think about some of the changes you're driving? Are these changes focused more on next year or the second half of this year, and any expenses related to those changes we should consider in the model?

Yes. Thanks, Pinjalim, for the question. In terms of changes, I don't expect to see a different expense profile. We will remain disciplined. Nothing fundamentally will change that calls for a different-looking P&L or expense profile. In terms of impact, let’s take three key functions: product and technology, AI strategy, and community engagement. Bratin brings substantial AI/ML experience, so you should expect to hear from us regarding core product innovation and our AI strategy. We need to pick up the pace; we have doubled product velocity but aim to accelerate further. Our intent is to ramp up innovation and ensure our customers' success on the platform. The AI strategy must cater to our customers' needs. We intend to be focused and disciplined about which customers to pursue and their requirements with Bratin's expertise guiding us to navigate this strategy efficiently and responsively. Wade's team has done exceptional work on the recent deploy event, and we plan to sustain interest and showcase DigitalOcean as a destination for developers to enhance their capabilities. Larry's mission focuses on integrating high-velocity sales to complement our product-led growth, ensuring incoming clients understand the vast capabilities we offer while driving expansion.

And that concludes our question-and-answer session. I will now turn the call back over to CEO Paddy Srinivasan for closing remarks.

Again, thank you, everyone, for spending your hour with us. As I mentioned, we are super excited with the progress. We have a lot of work to do, and we are looking forward to staying engaged. Have a great rest of the day.

This concludes today's conference call. Thank you for your participation. You may now disconnect.