DigitalOcean Holdings Inc (DOCN) 2025 Q1 法說會逐字稿

Good day, and welcome to the DigitalOcean Q1 '25 earnings conference call.

大家好，歡迎參加 DigitalOcean 2025 年第一季財報電話會議。

(Operator Instructions) And finally, I would like to advise all participants that this call is being recorded. Thank you.

（操作員指示）最後，我想告知所有參與者，本次通話正在錄音。謝謝。

I'd now like to welcome Melanie Strate, Head of Investor Relations to begin the conference.

現在，我歡迎投資者關係主管梅蘭妮·斯特拉特 (Melanie Strate) 開始會議。

Melanie, over to you.

梅蘭妮，交給你了。

Thank you, and good morning. Thank you all for joining us today to review DigitalOcean's first-quarter 2025 financial results. Joining me on the call today are Paddy Srinivasan, our Chief Executive Officer; and Matt Steinfort, our Chief Financial Officer.

謝謝，早安。感謝大家今天加入我們，共同回顧 DigitalOcean 2025 年第一季的財務表現。今天與我一起參加電話會議的還有我們的執行長 Paddy Srinivasan；以及我們的財務長 Matt Steinfort。

Before we begin, let me remind you that certain statements made on the call today may be considered forward-looking statements, which reflect management's best judgment based on currently available information. Our actual results may differ materially from those projected in these forward-looking statements, including our financial outlook. I direct your attention to the risk factors contained in our filings with the SEC as well as those referenced in today's press release that is posted on our website.

在我們開始之前，請允許我提醒您，今天電話會議上發表的某些聲明可能被視為前瞻性聲明，這些聲明反映了管理層根據當前可用資訊做出的最佳判斷。我們的實際結果可能與這些前瞻性聲明中的預測有重大差異，包括我們的財務前景。我請您注意我們向美國證券交易委員會提交的文件中的風險因素以及我們網站上發布的今天的新聞稿中提到的風險因素。

DigitalOcean expressly disclaims any obligation or undertaking to release publicly any updates or revisions to any forward-looking statements made today. Additionally, non-GAAP financial measures will be discussed on the conference call and reconciliations to the most directly comparable GAAP financial measures can be found in today's earnings 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 in the IR section of our website.

DigitalOcean 明確表示不承擔任何義務或承諾公開發布今天所做的任何前瞻性聲明的任何更新或修訂。此外，電話會議將討論非公認會計準則 (non-GAAP) 財務指標，而與最直接可比較的公認會計準則 (GAAP) 財務指標的對帳可在今天的收益新聞稿以及概述今天電話會議財務討論的投資者介紹中找到。我們網站的 IR 部分也提供今天電話會議的網路直播。

And with that, I will turn the call over to Paddy.

說完這些，我將把電話轉給帕迪。

Thank you, Melanie. Good morning, everyone, and thank you for joining us today as we review our solid first quarter 2025 results. The top line growth momentum we generated in 2024 continued into Q1, and the results provide further evidence of our continued execution of our strategy. This strategy, as outlined on slide 4 of our earnings presentation includes scaling with our digital native enterprise customers by helping them run large and complex workloads on our cloud platform and continuing to democratize the access of AI for both new generation AI-native startups as well as for our existing 600,000-plus customers.

謝謝你，梅蘭妮。大家早安，感謝您今天加入我們，一起回顧我們 2025 年第一季的穩健業績。我們在 2024 年創造的營收成長動能延續到了第一季度，業績進一步證明了我們將繼續執行策略。正如我們的收益報告第 4 張投影片中所概述的那樣，這項策略包括與我們的數位原生企業客戶一起擴展，幫助他們在我們的雲端平台上運行大型和複雜的工作負載，並繼續為新一代人工智慧原生新創公司以及我們現有的 60 萬多名客戶提供人工智慧的民主化存取。

I'm very pleased to share today the excellent progress we are making on both of these strategic priorities. My comments today will include a quick recap of our first quarter results, details on the progress we are making for our customers as we invest in product innovation and go-to-market for both core cloud and AI platforms. And a quick summary of the progress we are making on our financing strategy that Matt will detail later in the call.

今天我很高興與大家分享我們在這兩個策略重點上的出色進展。我今天的評論將包括對我們第一季業績的簡要回顧，以及我們在投資產品創新和核心雲和人工智慧平台上市過程中為客戶取得的進展的詳細資訊。馬特將在稍後的電話會議中詳細介紹我們的融資策略進度。

Let me start with the first quarter financial results and turning to slide 5 of our earnings tab. We had another solid quarter as revenue growth increased in the first quarter to 14% year-over-year to $211 million with our AI ARR continuing to grow north of 160% year-over-year.

讓我先介紹一下第一季的財務業績，然後翻到我們的損益表的第 5 張投影片。我們又度過了一個穩健的季度，第一季營收年增 14%，達到 2.11 億美元，我們的 AI ARR 繼續年增 160% 以上。

Q1 net dollar retention rate, or NDR, improved 100% for the first time since Q2 of 2023, and we expect NDR to remain in a similar range as the last two quarters for the remainder of the year.

第一季淨美元留存率（NDR）自 2023 年第二季以來首次提高 100%，我們預計今年剩餘時間內 NDR 將保持與前兩季相似的水平。

In addition, we made further progress with our continued focus on our higher spending digital native enterprise customers, increasing revenue from customers who were at $100,000 plus annual run rate up 41% year-over-year and to 23% of total revenue. This is available market growth rate and was driven by a 27% year-over-year increase in customer count, and 11% increase in their average spend with DigitalOcean.

此外，我們繼續關注支出較高的數位原生企業客戶，取得了進一步的進展，來自年運行率 100,000 美元以上的客戶的收入同比增長 41%，佔總收入的 23%。這是可用的市場成長率，其推動力是客戶數量年增 27%，以及在 DigitalOcean 上的平均支出成長 11%。

This improved growth was not confined to just the segment as our higher spend customers as a whole, which includes our builders, scalers and scalers plus customers grew to over 170,000 in numbers and their revenue grew 16% year-over-year, making up 88% of our total revenue.

這種成長的改善並不局限於該細分市場，因為我們整體上支出較高的客戶（包括我們的建築商、擴展商和擴展商加客戶）的數量增長到超過 170,000 人，他們的收入同比增長 16%，占我們總收入的 88%。

The growth of these customers is a clear sign that our product innovation efforts and investments in strategic and targeted go-to-market motions are helping Digital Native Enterprises scale rapidly on our platform.

這些客戶的成長清楚地表明，我們的產品創新努力以及對策略性和有針對性的上市舉措的投資正在幫助數位原生企業在我們的平台上迅速擴大規模。

While we continued investments and drove higher growth, we did so while delivering healthy profitability metrics, including 61% gross margin and 41% EBITDA margins in Q1. Our 61% gross margins in Q1 is 200 bps higher than the prior year, an improvement that came from cost optimization that mitigated the near-term increase in cost of revenue that comes with the incremental capacity that we brought online in our new Atlanta data center.

在我們繼續投資並推動更高成長的同時，我們也實現了健康的獲利指標，包括第一季 61% 的毛利率和 41% 的 EBITDA 利潤率。我們第一季的毛利率為 61%，比去年同期高出 200 個基點，這項改善得益於成本優化，它緩解了我們在新亞特蘭大資料中心上線的增量容量所帶來的短期收入成本增加。

This new data center hosts multiple AI inferencing fabrics, some of which are already live for customers, powering real-world customer inferencing workloads and is the first step in our long-term data center optimization strategy, which includes both our cloud and AI workloads, giving us a path to further gross margin optimization.

這個新的資料中心承載著多個人工智慧推理結構，其中一些已經為客戶投入使用，為現實世界的客戶推理工作負載提供支持，這是我們長期資料中心優化戰略的第一步，該戰略包括我們的雲端和人工智慧工作負載，為我們進一步優化毛利率提供了途徑。

Our 2025 capital program was heavily front-loaded in Q1, which drove the decline in quarterly adjusted free cash flow margin. While lumpy, this spend was contemplated in our full year 2025 plan, and we remain on track for our full year free cash flow margin guidance.

我們的 2025 年資本計畫在第一季投入了大量資金，這導致季度調整後的自由現金流利潤率下降。儘管金額不大，但這些支出已納入我們的 2025 年全年計劃，而且我們仍有望實現全年自由現金流利潤率指引。

Having this new AI inferencing infrastructure in place in the new Atlanta data center enables us to go even faster on our AI initiatives, helping us observe and learn from these customer workloads, ultimately enabling us to innovate faster in this rapidly evolving space. It also enables us to win even larger inferencing workloads like the $20 million-plus multiyear inferencing commitment, we closed with a strategic customer and partner early in Q2.

在新的亞特蘭大資料中心建立這種新的人工智慧推理基礎設施使我們能夠更快地推進我們的人工智慧計劃，幫助我們觀察和學習這些客戶工作負載，最終使我們能夠在這個快速發展的領域更快地創新。它還使我們能夠贏得更大的推理工作量，例如我們在第二季度初與戰略客戶和合作夥伴達成的價值 2000 多萬美元的多年推理承諾。

With our focus now on large digital native enterprises, our push to win large workload migrations from other clouds, and our funnel of larger AI-native inferencing workloads, we are now seeing and winning larger and even multimillion dollar deals than we have in the past. While this is consistent with our strategy, and we are excited about its near- and long-term growth potential, these deals require larger amounts of capacity to be available for the customer.

我們現在專注於大型數位原生企業，努力贏得從其他雲端平台進行的大量工作負載遷移，以及更大的人工智慧原生推理工作負載管道，我們現在看到並贏得了比過去更大甚至數百萬美元的交易。雖然這符合我們的策略，我們對其近期和長期的成長潛力感到興奮，但這些交易需要為客戶提供更大的產能。

To support these types of scaled global workloads while maintaining our strong free cash flow generation, we are exploring additional funding strategies that Matt will explain in greater detail in his commentary, that would increase our ability to more rapidly deploy new capacity to drive growth while minimizing the near-term impact on free cash flow.

為了支持這些類型的規模化全球工作負載，同時保持我們強勁的自由現金流生成，我們正在探索額外的融資策略，馬特將在他的評論中更詳細地解釋，這將提高我們更快地部署新產能以推動增長的能力，同時最大限度地減少對自由現金流的短期影響。

Let me now give you some updates on product innovation that we are delivering for our digital native enterprise customers, starting with the enhancements to our core cloud computing platform outlined on page 6 of our earnings deck, if you're following along.

現在，讓我向您介紹一些我們為數位原生企業客戶提供的產品創新的最新情況，首先是我們在收益報告第 6 頁概述的核心雲端運算平台的增強功能，如果您一直在關注的話。

In Q1, we continued our torrid pace of innovation and released more than 50 new products and features, which is more than 5 times what we released in Q1 of the prior year. We accomplished this with no appreciable increase in R&D spend as a percentage of revenue, which is a testament to DO's engineering talent and our increasing use of AI for accelerating development and improving our operational posture. AI help developers improving coding output by up to 40%, and we also saw gains in our fleet health through AI-based predictive maintenance.

第一季度，我們持續保持強勁的創新步伐，發布了 50 多種新產品和新功能，是去年同期的 5 倍多。我們實現了這一目標，而研發支出佔收入的百分比並沒有明顯增加，這證明了 DO 的工程才能以及我們越來越多地使用人工智慧來加速開發和改善我們的營運狀況。人工智慧幫助開發人員將編碼輸出提高高達 40%，我們還透過基於人工智慧的預測性維護看到了車隊健康狀況的改善。

The new cloud capabilities we introduced in the quarter remain hyper focused on addressing the needs of our larger footprint customers to help them scale on the DigitalOcean platform.

我們在本季推出的新雲端功能仍然高度專注於滿足更大規模客戶的需求，以幫助他們在 DigitalOcean 平台上擴展規模。

Let me highlight a few of those. We announced the GA of DigitalOcean Partner Network Connect, a secure, high-performance connectivity solution that simplifies multi-cloud and hybrid cloud networking. This service enables our customers to establish private connections between DigitalOcean servers and other cloud providers or even on-premise data centers by bypassing the public Internet, it enhances security, reduces latency and optimizes multi-cloud and hybrid cloud networking.

讓我重點介紹其中的幾個。我們宣布推出 DigitalOcean Partner Network Connect，這是一個安全、高效能的連接解決方案，可簡化多雲和混合雲端網路。這項服務使我們的客戶能夠繞過公共互聯網，在 DigitalOcean 伺服器與其他雲端供應商甚至內部資料中心之間建立私人連接，從而增強安全性、減少延遲並優化多雲和混合雲網路。

Our recent product updates over the last three quarters and enterprise-grade availability are now attracting more complex workloads migrating from hyperscalers and this new capability of Partner Network Connect simplifies these by enabling a stage migration process by establishing secure connection between DigitalOcean and other cloud providers, allowing customers to run workloads across multiple clouds.

我們最近三個季度的產品更新和企業級可用性現在吸引了從超大規模遷移的更複雜的工作負載，而合作夥伴網路連接的這項新功能透過在 DigitalOcean 和其他雲端供應商之間建立安全連接來實現階段遷移過程，從而簡化了這些工作負載，允許客戶跨多個雲端運行工作負載。

DigitalOcean's Kubernetes service, or DOKS for short, now scales up to 1,000 nodes, supporting complex workloads with optimized network routing for enhanced speed and reduced latency, developed with heavy customer inputs, this enhancement allows for infrastructure scaling to handle unexpected traffic surges with minimal manual effort.

DigitalOcean 的 Kubernetes 服務（簡稱 DOKS）現在可擴展至 1,000 個節點，透過優化的網路路由支援複雜的工作負載，從而提高速度並減少延遲，該服務是在大量客戶輸入的基礎上開發的，這項增強功能允許擴展基礎設施以處理意外的流量激增，同時只需最少的手動工作。

DOKS provides robust cluster creation with optimized default settings, several CPU and GPU droplet choices and integrated cloud resources enabling customers to manage large, fluctuating workloads while maintaining high availability and performance as their DigitalOcean usage grows.

DOKS 提供強大的叢集建立功能，具有最佳化的預設設定、多個 CPU 和 GPU droplet 選擇以及整合的雲端資源，使客戶能夠管理大量波動的工作負載，同時在 DigitalOcean 使用量成長時保持高可用性和效能。

Our managed database offerings for MySQL and PostgreSQL now feature significantly expanded scalable storage options. Effective in Q1, we have doubled the storage for managed MySQL plans to 20 terabytes, managed PostgreSQL now supports up to 30 terabytes also doubling from its previous capacity. These enhancements enable our customers to seamlessly scale their database storage as their businesses grow, accommodating both increasing demands of existing workloads and also migrations from self-managed databases with greater flexibility.

我們針對 MySQL 和 PostgreSQL 提供的託管資料庫產品現在具有顯著擴展的可擴充儲存選項。從第一季開始，我們將託管 MySQL 計畫的儲存空間增加了一倍，達到 20 TB，託管 PostgreSQL 現在支援高達 30 TB 的儲存空間，也是以前容量的兩倍。這些增強功能使我們的客戶能夠隨著業務的成長無縫擴展其資料庫存儲，既能滿足現有工作負載日益增長的需求，又能更靈活地從自我管理資料庫進行遷移。

Another product release that I want to highlight is network load balancing, which was developed in response to strong demand from our larger digital native enterprise customers. It enables companies to distribute traffic efficiently across multiple servers. This feature automatically scales resources increasing capacity during peak loads and decreasing it during slower periods.

我想要強調的另一個產品發布是網路負載平衡，它是為了滿足我們較大的數位原生企業客戶的強烈需求而開發的。它使公司能夠在多個伺服器之間有效地分配流量。此功能會自動擴展資源，在高峰負載期間增加容量，在較慢的時段減少容量。

In Q1, we augmented these product innovations with increased support and engagement of our top customers with complementary go-to-market motions. We expanded named account coverage to our top 3,000 customers by revenue, which is double the coverage we had in 2024.

在第一季度，我們透過增加對頂級客戶的支持和參與以及互補的上市措施增強了這些產品創新。我們將指定帳戶覆蓋範圍擴大到收入最高的 3,000 名客戶，這是我們 2024 年覆蓋範圍的兩倍。

We now have accounting coverage for our top 3,000 accounts that includes a technical account manager who ensures that these customers leverage our platform to the fullest, a solution architect who helps with the actual adoption and a growth account manager who manages the overall relationships and looks for new workloads to run on DO.

現在，我們為前 3,000 名客戶提供會計服務，其中包括一名技術客戶經理，負責確保這些客戶充分利用我們的平台，一名解決方案架構師，負責幫助實際採用，以及一名增長客戶經理，負責管理整體關係並尋找在 DO 上運行的新工作負載。

This new named account engagement model and all the new product innovations enabled us to accelerate the traction with rapidly scaling and larger digital native enterprise customers.

這種新的指定帳戶參與模式和所有新的產品創新使我們能夠透過快速擴展和更大的數位原生企業客戶來加速發展。

To raise the visibility of these advanced features, we also launched a new webinar series called sail to success to drive product awareness and option featuring real-world case studies and topics, including the best practices in migrating workloads to deal and building real-world applications using our GenAI platform using advanced capabilities like guardrails, AI knowledge bases and so forth.

為了提高這些高級功能的知名度，我們還推出了一個名為“揚帆成功”的新網絡研討會系列，以提高產品知名度和選擇，其中包含真實案例研究和主題，包括將工作負載遷移到交易的最佳實踐以及使用我們的 GenAI 平台構建真實應用程序，使用護欄、AI 知識庫等高級功能。

We're also publishing numerous case studies and are averaging about a case study a week to provide reference materials on how our digital native enterprise customers are taking advantage of the expanding breadth of our platform.

我們也發布了大量案例研究，平均每週發布一個案例研究，提供參考資料，說明我們的數位原生企業客戶如何利用我們平台不斷擴大的廣度。

On the last call, I talked about a new migrations program and a small dedicated team to support it, and I'm very pleased to share that in Q1, we facilitated 79 migrations to the DigitalOcean platform. While these workloads will continue to grow in footprint, the average annual run rate from these migrations is already in the tens of thousands of dollars.

在上次電話會議上，我談到了一個新的遷移計劃和一個支持它的小型專門團隊，我很高興地告訴大家，在第一季度，我們促成了 79 次向 DigitalOcean 平台的遷移。雖然這些工作負載的佔用空間將持續成長，但這些遷移的平均年運行成本已達到數萬美元。

Our migrations team, along with specialized partners help these customers with planning, architecture and the actual technical work to make these migrations smooth and ensure that the most appropriate architecture is deployed to get the best ROI from our platform.

我們的遷移團隊與專業合作夥伴一起幫助這些客戶進行規劃、架構和實際技術工作，以使這些遷移順利進行，並確保部署最合適的架構以從我們的平台獲得最佳投資回報率。

In one such example, we won all workloads for ProMobi, a SaaS company known for its solutions in enterprise mobility management that holds electronic protected health information or EPHI workloads on this solution. Appwrite, an open-source platform that helps developers quickly build and scale back-end applications, signed a two-year commitment with us in Q1.

在一個這樣的例子中，我們贏得了 ProMobi 的所有工作負載，ProMobi 是一家 SaaS 公司，以其企業移動性管理解決方案而聞名，該解決方案上包含受保護的電子健康資訊或 EPHI 工作負載。Appwrite 是一個幫助開發人員快速建立和擴展後端應用程式的開源平台，它在第一季與我們簽署了為期兩年的承諾。

Their platform supports core back-end services like authentication, databases, storage and messaging across multiple frameworks and languages. Appwrite chose DigitalOcean over hyperscalers due to our superior price performance in our ability to rapidly deploy spaces in their preferred geographic regions.

他們的平台支援跨多個框架和語言的核心後端服務，如身份驗證、資料庫、儲存和訊息傳遞。Appwrite 之所以選擇 DigitalOcean 而不是超大規模伺服器，是因為我們能夠在其首選的地理區域快速部署空間，具有卓越的性價比。

Another example of a customer that has rapidly adopted our new enterprise-grade features is Centra, a leading e-commerce platform that leverages our role-based access control, droplet autoscaling, global load balancers, VPC peering and more.

另一個快速採用我們新的企業級功能的客戶的例子是 Centra，這是一個領先的電子商務平台，利用我們基於角色的存取控制、Droplet 自動擴展、全域負載平衡器、VPC 對等連接等。

Now turning to our AI/ML initiatives, which is covered on page 7 of our earnings deck, we are progressing on our strategy of democratizing access to AI. I strongly believe that we are entering an era of AI everywhere in which every software application is going to have AI as a core component of how it functions.

現在談談我們的 AI/ML 計劃，這在我們的收益報告第 7 頁有所介紹，我們正在推進 AI 民主化訪問策略。我堅信，我們正在進入一個無所不在的人工智慧時代，每個軟體應用程式都將以人工智慧作為其運作的核心組件。

For our target customers, the digital native enterprises, this translates into a heavy use of AI in inferencing model. The strong growth of AI in Q1, the majority of our customers AI workloads are now in inferencing mode giving us an early glimpse into an AI world dominated by use cases that are serving the needs of end customers and solving real-world problems.

對於我們的目標客戶，即數位原生企業，這意味著在推理模型中大量使用人工智慧。第一季人工智慧成長強勁，我們大多數客戶的人工智慧工作負載現在都處於推理模式，讓我們可以提前一睹以滿足最終客戶需求和解決實際問題的用例為主導的人工智慧世界。

To take advantage of the growing traction that we are seeing with inferencing workloads, we are optimizing our AI stack to serve inferencing by continuing to deploy capabilities across the infrastructure, platform and AI agent layers of our stack.

為了利用推理工作負載日益增長的吸引力，我們正在優化我們的 AI 堆疊，透過繼續在我們的堆疊的基礎設施、平台和 AI 代理層部署功能來服務推理。

Let me break out our advancements down a little bit more, starting with the infrastructure layer. In Q1, we announced that DigitalOcean customers now have access to NVIDIA HGX H200s GPUs, our deployment of these servers allows our customers to use them as stand-alone machines or multi-node clusters.

讓我更詳細地介紹我們的進步，從基礎設施層開始。在第一季度，我們宣布 DigitalOcean 客戶現在可以使用 NVIDIA HGX H200s GPU，我們部署的這些伺服器允許我們的客戶將它們用作獨立機器或多節點叢集。

Deepening our AMD partnership, in April, we announced that our customers now have access to a highly performing and cost-effective solution for AI inferencing workloads with the AMD Instinct MI300X GPUs with ROCm software.

為了深化我們與 AMD 的合作夥伴關係，我們在 4 月宣布，我們的客戶現在可以使用配備 ROCm 軟體的 AMD Instinct MI300X GPU 獲得高效能且經濟高效的 AI 推理工作負載解決方案。

These leading edge GPUs are now available in single-tenant bare metal configurations for customers seeking control and raw computing power to power AI inferencing. The demand is outpacing the supply for our AI products that leverage these leading edge GPU types, NVIDIA HGX H200s and AMD Instinct MI300X GPUs, which is further validation of our decision to invest in this growth capital and add material incremental capacity through our Atlanta data center.

這些領先的 GPU 現已以單一租戶裸機配置提供，可供尋求控制和原始運算能力以支援 AI 推理的客戶使用。對於利用這些尖端 GPU 類型、NVIDIA HGX H200s 和 AMD Instinct MI300X GPU 的 AI 產品來說，需求超過了供應，這進一步驗證了我們投資這一成長資本並透過亞特蘭大資料中心增加材料增量容量的決定。

One real world example of a customer already leveraging our NVIDIA HGX H200s service offering is an emerging AI start-up focused on developing next-generation search and recommendation solutions for e-commerce. This customer runs AI searches that scans billions of items and websites in real time, delivering curated recommendations tailored to their customers' fashion needs and transforms search results into stunning visual recommendations. All of this is powered by digital oceans NVIDIA HGX H200s GPUs.

現實世界中已經利用我們的 NVIDIA HGX H200s 服務產品的客戶的一個真實示例是一家新興的 AI 新創公司，專注於開發用於電子商務的下一代搜尋和推薦解決方案。該客戶運行人工智慧搜索，即時掃描數十億件商品和網站，提供根據客戶時尚需求量身定制的推薦，並將搜尋結果轉化為令人驚嘆的視覺推薦。所有這些都由數位海洋 NVIDIA HGX H200s GPU 提供支援。

Another example on the AI infrastructure side is WindBorne, a company whose mission is to help mitigate and manage the destructive aspects of climate change, extreme weather and weather uncertainty. They achieved this by instrumenting the environment and advancing weather forecast. WindBorne Leverages DigitalOcean's GPU droplets to build a record-breakingly accurate, deep learning-based global medium-range forecast model called WeatherMesh, which predicts wind speed, temperature, dew point, cloud cover, precipitation, geo potential height and more.

人工智慧基礎設施的另一個例子是 WindBorne，該公司的使命是幫助減輕和管理氣候變遷、極端天氣和天氣不確定性的破壞性影響。他們透過監測環境和推進天氣預報實現了這一目標。WindBorne 利用 DigitalOcean 的 GPU droplets 建立了一個破紀錄的、基於深度學習的全球中期預測模型 WeatherMesh，該模型可以預測風速、溫度、露點、雲量、降水、地理位勢高度等。

Moving up the stack to the platform layer, I'm very excited about the progress we are making and the adoption we are driving on our GenAI platform, which is our fully managed solution designed to help digital native enterprises build and scale generative AI applications with ease.

向上移動到平台層，我對我們正在取得的進展以及我們在 GenAI 平台上推動的採用感到非常興奮，這是我們完全託管的解決方案，旨在幫助數位原生企業輕鬆構建和擴展生成式 AI 應用程式。

It now supports state-of-the-art models from providers like Anthropic, Meta, Mistral, DeepSeek, OpenAI and others and offers advanced capabilities such as retrieval augmented generation, function calling, secure guardrails and more to ensure reliable and context of our outputs.

現在它支援來自 Anthropic、Meta、Mistral、DeepSeek、OpenAI 等提供者的最先進的模型，並提供檢索增強生成、函數呼叫、安全護欄等高級功能，以確保我們的輸出的可靠性和上下文。

Developers can now integrate our GenAI into their workflows using APIs or embed chat interfaces directly into their applications. The platform also includes serverless model end points, vector database integration and token-based billing, allowing companies to go from prototype to production without managing complex infrastructure. While this platform is still in beta, we already have over 5,000 customers leveraging the platform and over 8,000 agents have been created on it since announcing it in January.

開發人員現在可以使用 API 將我們的 GenAI 整合到他們的工作流程中，或將聊天介面直接嵌入到他們的應用程式中。該平台還包括無伺服器模型端點、向量資料庫整合和基於令牌的計費，使公司無需管理複雜的基礎設施即可從原型轉向生產。雖然該平台仍處於測試階段，但自 1 月宣布推出以來，我們已經有超過 5,000 名客戶利用該平台，並且已在該平台上創建了超過 8,000 名代理商。

More than 80% of our users for this platform are existing DigitalOcean customers, clearly showing the potential that we have within our existing customer base. One such customer example is Phoenix Secure, a leading India-based GPS tracking and fleet management provider. They leverage the digital ocean GenAI platform to create and integrate sales AI agents into their GPS tracking platform to enhance lead management.

該平台 80% 以上的用戶都是現有的 DigitalOcean 客戶，這清楚地表明了我們在現有客戶群中的潛力。其中一個客戶範例是 Phoenix Secure，一家領先的印度 GPS 追蹤和車隊管理提供商。他們利用數位海洋 GenAI 平台創建銷售 AI 代理並將其整合到他們的 GPS 追蹤平台中，以加強潛在客戶管理。

Now let me also cover the AI Agentic layer, one of my favorite case studies on impactful AI agent applications is our own Cloudways Copilot offering, which is currently in public preview. Cloudways Copilot is an AI-powered assistant integrated into our managed hosting platform that is designed to streamline server management and optimize website performance for growing small and medium businesses.

現在讓我也介紹一下 AI Agentic 層，我最喜歡的有影響力的 AI 代理應用案例研究之一是我們自己的 Cloudways Copilot 產品，該產品目前處於公開預覽階段。Cloudways Copilot 是一款整合到我們的託管平台的人工智慧助手，旨在簡化伺服器管理並優化成長型中小型企業的網站效能。

This helps our customers automate tasks, monitor performance and provide them with insight to help keep their websites up and running smoothly.

這可以幫助我們的客戶自動執行任務、監控效能並為他們提供洞察力，以幫助他們的網站保持平穩運行。

Cloudways Copilot currently has over 250 customers leveraging the tool with over 90% accuracy rate. One specific customer leveraging Cloudways Copilot is create a web design agency that specializes in creating and maintaining websites to effectively represent a client's brand.

Cloudways Copilot 目前已有超過 250 位客戶使用該工具，準確率超過 90%。一位利用 Cloudways Copilot 的特定客戶創建了一個網頁設計機構，專門創建和維護網站以有效地代表客戶的品牌。

Create has been able to improve productivity by leveraging the insight feature to quickly and accurately identify issues that otherwise would have been only found if found at all through manual processes and review of logs.

Create 利用洞察功能快速且準確地識別問題，從而提高了生產力，否則這些問題只能透過手動流程和日誌審查才能發現。

Another example is CurtisE, a digital marketing agency based in Australia that specializes in helping small businesses enhance their online presence through personalized and strategic digital solutions.

另一個例子是 CurtisE，位於澳洲的數位行銷機構，專門透過個人化和策略性的數位解決方案幫助小型企業提升其線上形象。

Cloudways Copilot provides precise insights that empower CurtisE to quickly pinpoint and resolve e-commerce store issues cutting problem-solving time from hours to just minutes.

Cloudways Copilot 提供精確的洞察，使 CurtisE 能夠快速找出並解決電子商務商店問題，將解決問題的時間從數小時縮短到幾分鐘。

Let me now give you an update on the financing matters. In addition to the great progress we have made on top line growth initiatives, we've also made material progress on our balance sheet and capital structure priorities.

現在讓我向你們通報融資事宜的最新進展。除了我們在營收成長計畫方面取得巨大進展之外，我們在資產負債表和資本結構優先事項方面也取得了實質進展。

We announced this morning that we have taken the first step in addressing our outstanding 2026 convertible debt, having entered into a new secured five-year credit facility agreement of $800 million, with a $500 million Term Loan A that we will leverage to refinance a portion of our existing convertible notes.

我們今天上午宣布，我們已邁出解決未償還的 2026 年可轉換債務的第一步，已達成一項新的 8 億美元五年期擔保信貸協議，其中包括 5 億美元的定期貸款 A，我們將利用這筆貸款為部分現有可轉換票據進行再融資。

Matt will walk you through more details on our financial results, refinancing actions and guidance later in the call.

馬特將在稍後的電話會議中向您詳細介紹我們的財務表現、再融資行動和指導。

In closing, DigitalOcean's Q1 2025 demonstrates significant progress in executing the strategy we outlined at our recent Investor Day. We achieved accelerated revenue growth, consistent improvement in net dollar retention and substantial advancements in our strategy to scale with the digital native enterprises.

最後，DigitalOcean 的 2025 年第一季在執行我們在最近的投資者日概述的策略方面取得了重大進展。我們實現了收入加速成長、淨美元留存率持續提高以及與數位原生企業一起擴大規模的策略取得了實質進展。

Our product innovation remains strong with over 50 new releases specifically designed for the sophisticated needs of our growing digital native enterprise customers. Notably, we're also observing a clear increase in inferencing, signifying the evolution of customer workloads towards practical real-world AI applications and strong adoption of our GenAI platform.

我們的產品創新保持強勁，有超過 50 個新產品專為滿足我們不斷成長的數位原生企業客戶的複雜需求而設計。值得注意的是，我們還觀察到推理能力明顯增強，這表明客戶工作負載正在向實際的現實世界 AI 應用發展，並且我們的 GenAI 平台得到了廣泛採用。

Strategic go-to-market investments and product innovation drove a 41% year-over-year increase in revenue from customers spending over $100,000 annually. These results demonstrate clear progress towards our strategic goals and bolster our confidence in achieving our ambitious long-term objectives.

策略性市場投資和產品創新推動年消費超過 10 萬美元的客戶收入年增 41%。這些結果表明我們在實現戰略目標方面取得了明顯進展，並增強了我們實現雄心勃勃的長期目標的信心。

Additionally, we have proactively started addressing our financing strategy by initiating debt refinancing, positioning us for sustained growth with the growing needs of our digital native enterprises.

此外，我們已開始積極解決融資策略問題，啟動債務再融資，以滿足我們數位原生企業日益增長的需求，從而實現持續成長。

The momentum generated in Q1 provides a solid base for the remainder of the year, and underscores our mission of simplifying cloud and AI so that digital native enterprise customers can focus on creating software that changes the world.

第一季的進展為今年剩餘時間奠定了堅實的基礎，並強調了我們簡化雲端和人工智慧的使命，以便數位原生企業客戶可以專注於創造改變世界的軟體。

Thank you. And now over to Matt.

謝謝。現在輪到馬特了。

Thanks, Paddy. Good morning, everyone, and thanks for joining us today. As Paddy discussed, we are very pleased with our Q1 2025 performance as we continue to execute against our key strategic priorities and deliver strong financial results. In my comments, I'll walk through our Q1 results, provide an update on our balance sheet and financing strategy and share our second quarter and full year 2025 financial outlook.

謝謝，帕迪。大家早安，感謝大家今天加入我們。正如帕迪所討論的，我們對 2025 年第一季的表現非常滿意，因為我們繼續執行我們的關鍵策略重點並取得強勁的財務業績。在我的評論中，我將介紹我們的第一季度業績，提供我們的資產負債表和融資策略的最新信息，並分享我們第二季度和 2025 年全年的財務展望。

Starting with the top line. Revenue in the first quarter was $211 million, and annual run rate revenue, or ARR, was $843 million, both up 14% year-over-year. We added $23 million of incremental ARR in the quarter, which is 50% higher compared to incremental ARR in the same quarter last year. We continue to focus our product innovation and go-to-market investments on our digital native enterprise customers, and the momentum we have generated with this target customer base continues to grow.

從第一行開始。第一季營收為 2.11 億美元，年運行率營收（ARR）為 8.43 億美元，均比去年同期成長 14%。本季我們增加了 2,300 萬美元的增量 ARR，與去年同期的增量 ARR 相比高出 50%。我們持續將產品創新和市場投入重點放在數位原生企業客戶上，並且我們透過這個目標客戶群所創造的勢頭持續成長。

In Q1, revenue from our customers whose annualized run rate revenue in the quarter was greater than $100,000, who we refer to as scalers plus and who represent 23% of overall revenue grew 41% year-over-year with a 27% year-over-year increase in customer count, as our largest customers continued their strong adoption of our core cloud and AI products.

在第一季度，來自我們本季年化運行率收入超過 100,000 美元的客戶（我們稱之為 scalers plus，他們佔總收入的 23%）的收入同比增長 41%，客戶數量同比增長 27%，因為我們最大的客戶繼續大力採用我們的核心雲和 AI 產品。

Q1 revenue growth was driven by improvements in both customer acquisition and customer expansion. The enhancements we've made to our product-led growth engine are paying off in improved customer acquisition with revenue growth from customers in their first 12 months on our platform continuing to accelerate in Q1.

第一季營收的成長得益於客戶獲取和客戶擴展的改善。我們對產品主導型成長引擎所做的改進正在帶來回報，客戶獲取率的提高，客戶在我們平台上使用後的前 12 個月的收入成長在第一季繼續加速。

Our product innovation and go-to-market efforts are also having an impact on customer expansion as our Q1 net dollar retention ticked up to 100%, up from 97% in the same quarter last year, which is a major accomplishment as we have largely eliminated the net expansion headwind we faced last year.

我們的產品創新和市場推廣努力也對客戶擴張產生了影響，因為我們第一季的淨美元保留率從去年同期的 97% 上升至 100%，這是一項重大成就，因為我們基本上消除了去年面臨的淨擴張逆風。

We expect NDR to remain in a similar range as the last two quarters for the remainder of the year. Increasing customer expansion was also evident in increased average spend as average revenue per user, or ARPU, grew at 14% year-over-year.

我們預計，今年剩餘時間內 NDR 將保持與前兩個季度相似的範圍。客戶數量的不斷擴大也體現在平均支出的增加上，每用戶平均收入 (ARPU) 年比增加 14%。

Turning to the P&L. Gross margin for the first quarter was 61%, which was 200 basis points higher than the prior year. The improved gross margin is driven by our ongoing cost optimization efforts with revenue growth at 14% year-over-year, outpacing cost of revenue growth of 8% and by the improved utilization and extended useful life of our servers from five to six years.

轉向損益表。第一季毛利率為61%，較上年同期高出200個基點。毛利率的提高得益於我們持續的成本優化努力，營收年增 14%，超過 8% 的收入成本成長，以及伺服器利用率的提高和使用壽命從五年延長至六年。

Adjusted EBITDA was $86 million, an increase of 16% year-over-year. Adjusted EBITDA margin was 41% in the first quarter, approximately 100 basis points higher than the prior year. As shown in our current results and historical adjusted EBITDA margins, we continue to balance disciplined growth investments with ongoing gains in operational efficiency.

調整後EBITDA為8,600萬美元，較去年成長16%。第一季調整後的 EBITDA 利潤率為 41%，比上年高出約 100 個基點。正如我們目前的業績和歷史調整後的 EBITDA 利潤率所示，我們繼續在嚴格的成長投資和持續提高的營運效率之間取得平衡。

Non-GAAP diluted net income per share was $0.56, a 30% increase year-over-year. This increase is a direct result of expanding per share profitability by driving ongoing operating leverage and offsetting dilution through consistent share repurchases.

非公認會計準則每股攤薄淨收益為 0.56 美元，較去年同期成長 30%。這一增長是透過推動持續的營運槓桿和透過持續的股票回購來抵消稀釋而擴大每股獲利能力的直接結果。

GAAP diluted net income per share was $0.39, a 160% increase year-over-year as we continue to both drive operating leverage and prudently manage stock-based compensation. Q1 adjusted free cash flow was effectively breakeven. As we discussed last quarter, our 2025 capital plan is heavily front-end loaded, and when coupled with normal Q1 specific cash outflows, caused Q1 adjusted free cash flow to be significantly lower compared to Q4. While the lower Q1 adjusted free cash flow level was expected and is fully contemplated in our projected 16% to 18% free cash flow margins for 2025, it does warrant an explanation of how we view growth investments and how they have been and can be funded in the future.

由於我們繼續推動營運槓桿並審慎管理股票薪酬，每股 GAAP 稀釋淨收益為 0.39 美元，較去年同期成長 160%。第一季調整後的自由現金流實際上實現了損益兩平。正如我們上個季度所討論的那樣，我們的 2025 年資本計畫前端負荷很大，再加上第一季度正常的特定現金流出，導致第一季度調整後的自由現金流與第四季度相比大幅下降。雖然第一季調整後的自由現金流水準較低是預料之中的，並且已充分考慮到我們預計的 2025 年 16% 至 18% 的自由現金流利潤率，但它確實需要解釋我們如何看待成長投資以及這些投資過去和未來的融資方式。

As a public cloud company, we invest each year in assets such as servers and network switches that underpin our simple and scalable cloud and AI solutions and the software that our customers consume for our infrastructure and Platform as a service offerings.

作為一家公有雲公司，我們每年都會投資伺服器和網路交換器等資產，這些資產支撐著我們簡單、可擴展的雲端和人工智慧解決方案，以及我們的客戶為我們的基礎設施和平台即服務產品所消費的軟體。

These assets have long useful lives and not only support our current customer usage and revenue, but also will support future customer usage and revenue for years to come.

這些資產使用壽命長，不僅支援我們目前客戶的使用和收入，還將支援未來幾年的未來客戶使用和收入。

The amount of capital that we invest each year in these assets is a function of two drivers: the first driver of capital spend is the need to replace existing infrastructure that has reached the end of its useful life.

我們每年在這些資產上投資的資本金額取決於兩個驅動因素：資本支出的第一個驅動因素是需要更換現有已達到使用壽命的基礎設施。

Each year, we buy new equipment to replace aged servers and networking gear that has been supporting our existing customers and their usage. This capital spend is considered maintenance capital, as it is required to maintain the health and stability of current platform and revenue.

每年，我們都會購買新設備來替換一直支援現有客戶及其使用的老舊伺服器和網路設備。這筆資本支出被視為維護資本，因為它需要維持當前平台和收入的健康和穩定。

We have continued to improve our maintenance capital efficiency over time by extending the useful life of our infrastructure and by leveraging the natural improvements in price performance of compute and storage infrastructure over time.

透過延長基礎設施的使用壽命以及利用運算和儲存基礎設施價格效能的自然改善，我們不斷提高維護資本效率。

Maintenance capital is generally less than 20% to 30% of our annual capital program or approximately 5% of revenue in a typical year. This is an important point for us to emphasize as it implies that our steady-state adjusted free cash flow margin can be approximated by including only our maintenance capital and our adjusted free cash flow calculations.

維護資本通常不到我們年度資本計畫的 20% 到 30%，或典型年份收入的約 5%。這是我們要強調的一點，因為這意味著我們的穩態調整自由現金流利潤率可以透過僅包括維護資本和調整後的自由現金流計算來近似。

For reference, in 2024, this would have reflected a steady state adjusted free cash flow margin in the mid-30s. This maintenance capital dynamic, where a small portion of annual capital is required to support the existing business is a sign of strong economics and is similar to what you see in data center and tower companies, where very little capital is required to maintain their profitable platforms.

作為參考，到 2024 年，這將反映出 30 年代中期穩定狀態調整後的自由現金流利潤率。這種維護資本動態，即每年只需一小部分資本來支持現有業務，是經濟強勁的標誌，類似於數據中心和塔式公司的情況，這些公司只需要很少的資本來維護其盈利平台。

This second and primary driver of capital is the need to provide incremental capacity for both existing customers as they expand their usage and for the new customers that we acquire. This capital spend is considered growth capital and it enables growth not only in the current year, but also in years to come.

資本的第二個也是主要的驅動力是需要為擴大使用範圍的現有客戶和我們獲得的新客戶提供增量容量。這筆資本支出被視為成長資本，它不僅能促進當年的成長，還能促進未來幾年的成長。

Paybacks on our growth capital are attractive, with core cloud growth capital paying back in less than two years, and AI growth capital having paybacks around three years. The amount that we spend on growth capital will vary from quarter-to-quarter based on our demand outlook and deployment timing, but it's typically at least 70% to 80% of our capital program and approximately 15% of revenue.

我們的成長資本報酬率很有吸引力，核心雲端成長資本的回報期不到兩年，而人工智慧成長資本的回報期約為三年。根據我們的需求前景和部署時間，我們在成長資本上花費的金額每季都會有所不同，但通常至少占我們資本計畫的 70% 到 80%，約佔營收的 15%。

As the company grew from the IPO to $840 million in ARR, we elected to fund all of our maintenance and growth capital ourselves, using our balance sheet and strong cash flow generation to buy and own the service and other equipment that we use to run our business. In fact, all of the company's capital expenditures since IPO have been funded by organic cash generation alone, as the $1.5 billion in convertible debt and the $723 million net proceeds from the IPO were all effectively used to fund $1.6 billion in share repurchases, over $450 million in M&A and repayment of approximately $260 million of debt.

隨著公司從 IPO 成長到 ARR 達到 8.4 億美元，我們選擇自行資助所有維護和成長資本，利用我們的資產負債表和強勁的現金流來購買和擁有用於營運業務的服務和其他設備。事實上，自 IPO 以來，該公司的所有資本支出均由有機現金產生提供資金，因為 15 億美元的可轉換債務和 7.23 億美元的 IPO 淨收益均被有效用於資助 16 億美元的股票回購、超過 4.5 億美元的併購以及償還約 2.6 億美元的債務。

Looking forward, as we work to reach and exceed $1 billion in revenue, it is even more important to understand the dynamics of our growth capital investments. Our strategy is squarely focused on both serving larger digital native enterprises and pursuing our differentiated AI opportunity.

展望未來，當我們努力達到並超過 10 億美元的收入時，了解我們的成長資本投資動態就變得更加重要。我們的策略重點是服務大型數位原生企業和追求差異化的人工智慧機會。

Both of these initiatives create opportunities for us to accelerate our growth, as we are seeing and winning larger workloads, individual customers and partnership opportunities than we have seen in the past as evidenced by the $20 million plus two-year commitment with a strategic customer and partner that Paddy had referenced earlier in his remarks.

這兩項舉措都為我們創造了加速成長的機會，因為我們看到並贏得了比過去更大的工作量、個人客戶和合作機會，正如帕迪早些時候在演講中提到的與戰略客戶和合作夥伴達成的 2000 萬美元以上的兩年期承諾所證明的那樣。

With these larger growth opportunities come more and longer customer commitments, but also larger capacity and capital requirements. It also changes the dynamics of how we plan for and provision for growth as we can no longer rely solely on broad utilization trends to trigger growth capital investments and instead have to have capacity available or be quickly able to deploy it to meet the needs of rapidly growing enterprises across both our core cloud and AI platforms.

隨著這些更大的成長機會而來的是更多和更長期的客戶承諾，但同時也帶來了更大的產能和資本需求。它也改變了我們規劃和準備成長的方式，因為我們不能再僅僅依靠廣泛的利用趨勢來觸發成長資本投資，而是必須擁有可用容量或能夠快速部署它以滿足我們核心雲端和人工智慧平台上快速成長的企業的需求。

Said differently, when we have an opportunity to win a large workload, customers are willing to make long-term commitments, but they expect that the capacity is immediately available, which requires us to have capacity in front of demand more so than we have needed in the past.

換句話說，當我們有機會贏得大量工作量時，客戶願意做出長期承諾，但他們期望容量能夠立即可用，這就要求我們比過去更多地滿足需求。

The investments we made in Q1 to add capacity in Atlanta are a great example of this upfront growth capital investment. With Atlanta, we are bringing on a larger amount of incremental capacity to serve our growing AI pipeline and we're doing so in a lower-cost facility that will improve our long-run gross margin profile.

我們在第一季為增加亞特蘭大的產能所做的投資就是這種前期成長資本投資的一個很好的例子。借助亞特蘭大，我們將帶來更大的增量產能來服務我們不斷增長的人工智慧管道，我們正在以低成本的設施來實現這一目標，這將改善我們的長期毛利率狀況。

As we look ahead at our growth potential in the second half of 2025 and beyond, we anticipate an increasing portion of our growth to come from our larger digital native enterprises and from larger (inaudible), both of which will come with more committed contracts. To accelerate growth beyond our current outlook while maintaining healthy free cash flow, we are actively evaluating adding additional funding tools for growth capital. They're similar to those currently deployed by our public hyperscaler peers.

展望 2025 年下半年及以後的成長潛力，我們預計，我們成長的越來越大的部分將來自更大的數位原生企業和更大的（聽不清楚），這兩者都將帶來更多的承諾合約。為了加速超越我們當前前景的成長，同時保持健康的自由現金流，我們正在積極評估增加成長資本的額外融資工具。它們與我們公共超大規模同行目前部署的類似。

With our plan to have fully addressed our balance sheet by the end of this year and with our ongoing adjusted free cash flow generation, we certainly can and will continue directly funding growth capital ourselves. However, we are also exploring adding additional financing options to our funding toolkit where we could leverage the strong appetite from capital partners to support cloud and AI growth capital investment through leasing arrangements.

由於我們計劃在今年年底前全面解決資產負債表問題，並且隨著我們不斷調整自由現金流的產生，我們當然可以並將繼續直接為成長資本提供資金。然而，我們也在探索在我們的融資工具包中增加額外的融資選擇，我們可以利用資本合作夥伴的強烈需求，透過租賃安排來支援雲端和人工智慧成長資本投資。

Pursuing such a strategy would enable us to maintain or improve our strong cash flow generation while also reducing the upfront capital required to take advantage of the tremendous growth potential we see in both serving larger digital native enterprises and growing AI inferencing workloads.

實施這樣的策略將使我們能夠保持或提高強勁的現金流產生能力，同時減少所需的前期資本，以利用我們在服務更大的數位原生企業和不斷增長的人工智慧推理工作負載方面看到的巨大成長潛力。

We are early in this evaluation and would only leverage this financing tool if it enabled us to accelerate growth beyond our current outlook. We will share more on this potential strategy as we advance our thinking over the coming months and quarters.

我們尚處於評估的早期階段，只有當這種融資工具能讓我們加速成長並超越目前的預期時，我們才會利用它。隨著我們在未來幾個月和幾個季度中進一步思考，我們將分享更多有關這項潛在策略的資訊。

Shifting back to Q1 results. We continue to maintain material cash and cash equivalents and ended the quarter with $360 million in cash. We continue to execute our share repurchase program in the quarter with $59 million of repurchases in Q1, buying back approximately 1.6 million shares. This brings our cumulative share repurchases since IPO to $1.6 billion and 34.1 million shares through March 31, 2025.

回到第一季的結果。我們繼續保持大量現金和現金等價物，本季末現金餘額為 3.6 億美元。我們在本季繼續執行股票回購計劃，第一季回購了 5,900 萬美元，回購了約 160 萬股。截至 2025 年 3 月 31 日，我們自 IPO 以來的累計股票回購額已達 16 億美元，累計股票回購量達 3,410 萬股。

At the end of Q1, we had $23 million remaining on our share repurchase authorization. We remain well-positioned to continue investment in both organic growth and share repurchases while we actively address our existing debt. On the debt front, we announced this morning that we have entered into a secured five-year credit facility agreement, consisting of a $500 million nine-month delayed draw term loan A as well as a $300 million revolving credit facility. This new loan bears interest on the drawn amount in the range of SOFR plus 1.75% or just over 6% based on the starting net leverage.

截至第一季末，我們的股票回購授權餘額還剩 2,300 萬美元。我們仍處於有利地位，可以在積極解決現有債務的同時繼續投資於有機成長和股票回購。在債務方面，我們今天上午宣布，我們已經達成了一項有擔保的五年期信貸協議，其中包括一筆 5 億美元的九個月延期提取定期貸款 A 以及一筆 3 億美元的循環信貸。這筆新貸款的利息在 SOFR 加 1.75% 的範圍內，或根據起始淨槓桿率計算略高於 6%。

We will use the proceeds from this facility to refinance a portion of our existing convertible notes that are maturing in December 2026. This loan is the first step in 2025 to fully address the 2026 convert before the existing facility goes current. With our strong balance sheet, growing adjusted EBITDA and cash flow generation, we have multiple attractive options available to us that we are considering, including convertible debt, Term Loan B and high-yield options. We will continue to evaluate these options over the balance of this year to fully address the 2026 convert and optimize our long-term cost of debt.

我們將利用該筆貸款的收益為 2026 年 12 月到期的部分現有可轉換票據進行再融資。這筆貸款是 2025 年的第一步，旨在現有設施投入使用之前全面解決 2026 年的轉換問題。憑藉我們強勁的資產負債表、不斷增長的調整後 EBITDA 和現金流生成，我們有多個有吸引力的選擇可供考慮，包括可轉換債務、定期貸款 B 和高收益選擇權。我們將在今年剩餘時間內繼續評估這些選擇，以全面解決 2026 年的轉換問題並優化我們的長期債務成本。

Moving on to guidance. I will now share our financial outlook for the second quarter of 2025 and for the full year. For the second quarter of 2025, we expect revenue to be in the range of $215.5 million to $217.5 million, representing approximately 12.5% year-over-year growth at the midpoint. For the full year 2025, we are maintaining our revenue guidance in the range of $870 million to $890 million, representing approximately 13% year-over-year growth at the midpoint.

繼續指導。我現在將分享我們對 2025 年第二季和全年的財務展望。對於 2025 年第二季度，我們預計營收將在 2.155 億美元至 2.175 億美元之間，中間值年增約 12.5%。對於 2025 年全年，我們將營收預期維持在 8.7 億美元至 8.9 億美元之間，中間值年增約 13%。

Given our strong Q1 performance and our visibility into our customers' usage trends, we remain confident in our full year guidance despite the uncertainty in the current economic and geopolitical environments.

鑑於我們第一季的強勁表現以及我們對客戶使用趨勢的了解，儘管當前經濟和地緣政治環境存在不確定性，我們仍然對全年業績指引充滿信心。

For the second quarter of 2025, we expect our adjusted EBITDA margins to be in the range of 38% to 40%. For the full year, we maintained our adjusted EBITDA margin guide, which is in the range of 37% to 40%.

對於 2025 年第二季度，我們預計調整後的 EBITDA 利潤率將在 38% 至 40% 之間。全年而言，我們維持調整後的 EBITDA 利潤率指引，即在 37% 至 40% 的範圍內。

For the second quarter of 2025, we expect non-GAAP diluted earnings per share to be $0.42 to $0.47 based on approximately 103 million to 104 million in weighted average fully diluted shares outstanding. For the full year 2025, we expect non-GAAP diluted earnings per share to be $1.85 to $1.95 based on approximately 104 million to 105 million in weighted average fully diluted shares outstanding.

對於 2025 年第二季度，我們預計非 GAAP 稀釋每股收益為 0.42 美元至 0.47 美元，基於約 1.03 億至 1.04 億股加權平均完全稀釋流通股。對於 2025 年全年，我們預計非 GAAP 稀釋每股收益為 1.85 美元至 1.95 美元，基於約 1.04 億至 1.05 億股加權平均完全稀釋流通股。

Turning to adjusted free cash flow. We maintain our guided adjusted free cash flow margins for the full year at 16% to 18%. We intend to take advantage of the nine months delayed drop feature of our new $500 million Term Loan A and do not anticipate drawing that facility until late in 2025. And as such, do not anticipate material interest expense on that facility in 2025.

轉向調整後的自由現金流。我們將全年調整後的自由現金流利潤率維持在 16% 至 18% 之間。我們打算利用我們新的 5 億美元定期貸款 A 的九個月延遲提取功能，預計要到 2025 年底才會提取該貸款。因此，預計 2025 年該設施不會產生重大利息支出。

Consistent with our historical guidance practice, we are not providing adjusted free cash flow guidance on a quarter-by-quarter basis given it is heavily influenced by working capital timing, as you saw in our Q1 results. That concludes our prepared remarks, and we will now open up the call to Q&A.

與我們的歷史指導實踐一致，我們不會按季度提供調整後的自由現金流指導，因為它受到營運資本時機的嚴重影響，正如您在我們的第一季業績中看到的那樣。我們的準備好的演講到此結束，現在我們將開始問答環節。

(Operator Instructions)

（操作員指示）

Jason Ader, William Blair.

傑森·阿德、威廉·布萊爾。

Paddy, just on the GenAI platform, can you give us a sense of when it's expected to be generally available? I know it's still essentially like the beta situation now. And then -- and then just related to that, you talked about sort of the differentiated AI opportunity for DigitalOcean. Can you just walk us through how you think about DigitalOcean's differentiation within that AI landscape? Thank you.

帕迪，關於 GenAI 平台，您能告訴我們預計什麼時候可以全面推出嗎？我知道現在基本上還處於測試階段。然後 — — 與此相關，您談到了 DigitalOcean 的差異化 AI 機會。您能否向我們介紹您如何看待 DigitalOcean 在 AI 領域的差異化？謝謝。

Thank you, Jason. On your first question, yes, the GenAI platform, we announced in January, and we have been progressively adding a lot of new features while still in private Beta and public Beta now, we expect to go live sometime by end of Q2, beginning of Q3, it's just a function of how many additional capabilities we can add to the platform, making it super comprehensive for customers to be able to deploy real-world agentic workloads on it.

謝謝你，傑森。關於您的第一個問題，是的，我們在一月份宣布了 GenAI 平台，並且一直在逐步添加許多新功能，目前仍處於私人 Beta 版和公共 Beta 版階段，我們預計將在第二季度末或第三季度初的某個時候上線，這只是我們可以向平台添加多少額外功能的一個功能，使客戶能夠在平台上部署真實世界的代理工作負載，使其變得非常全面。

I gave several examples of companies that are leveraging it in production environments today, and we're getting excellent feedback from customers, both our existing customers as well as new AI native start-ups and companies that are giving us excellent feedback.

我給了幾個目前在生產環境中利用它的公司的例子，我們從客戶那裡得到了非常好的回饋，包括我們現有的客戶以及新的人工智慧原生新創公司和給我們非常好反饋的公司。

So -- and I gave the -- some count as well in terms of 5,000 customers, 8,000 agencies are tremendous numbers for a product that is not even in GA currently. So we're super bullish about the future of the GenAI platform, and we feel our full -- the second part of your question was what is our strategy and differentiation from an AI perspective.

所以 — — 我也給了 — — 一些統計數據，就 5,000 名客戶而言，8,000 個代理商對於目前甚至尚未進入 GA 階段的產品來說是一個驚人的數字。因此，我們對 GenAI 平台的未來非常樂觀，我們感到我們完全——您的問題的第二部分是從人工智慧的角度來看我們的策略和差異化是什麼。

So I've been talking about our IPA stack, infrastructure platform and applications. And as I outlined in today's call, we are making good progress on all three layers. The best part about our infrastructure layer is that a majority of the workloads that we are seeing now are on the inferencing side. And what is important to understand on the inferencing layer is that it is powering real-world use cases rather than just proof-of-concept or experimentation, which means it is serving the needs of end users, either in a B2B scenario or B2C scenario.

我一直在談論我們的 IPA 堆疊、基礎設施平台和應用程式。正如我在今天的電話會議中所述，我們在所有三個層面都取得了良好進展。我們的基礎設施層最好的部分是，我們現在看到的大部分工作負載都在推理方面。在推理層上需要理解的重要一點是，它支援現實世界的用例，而不僅僅是概念驗證或實驗，這意味著它滿足最終用戶的需求，無論是在 B2B 場景還是 B2C 場景中。

So I strongly believe that with our infrastructure capabilities today, both on AMD and NVIDIA, we are differentiating ourselves in our ability to run a full stack cloud that powers inferencing.

因此，我堅信，憑藉我們目前在 AMD 和 NVIDIA 上的基礎設施能力，我們在運行支援推理的全端雲端方面具有獨特的優勢。

As I explained in the Investor Day, inferencing needs a lot more than just bringing GPUs online. Inferencing typically requires a full stack general-purpose cloud, which is where we distinguish ourselves from a number of the other [Neo] GPU clouds that have emerged over the last few quarters.

正如我在投資者日所解釋的那樣，推理需要的不僅僅是將 GPU 上線。推理通常需要全端通用雲，這也是我們與過去幾季出現的其他一些 [Neo] GPU 雲的區別所在。

On the platform layer, we just talked about the GenAI platform. Our strategy there is to democratize the access and make it super, super easy for non-AI native companies to consume GenAI in their application stack. And the proof is in the numbers I talked about like -- over 80% of the traction today we are seeing is from existing DigitalOcean Digital Native Enterprise customers who are not AI native, but they recognize the importance of adding GenAI into their application. And we are making it super easy for them to build sophisticated GenAI features into their platform.

在平台層，我們剛剛談到了GenAI平台。我們的策略是使訪問民主化，並使非 AI 原生公司能夠非常輕鬆地在其應用程式堆疊中使用 GenAI。證據就在我談到的數字中——我們今天看到的超過 80% 的吸引力來自現有的 DigitalOcean Digital Native Enterprise 客戶，他們不是 AI 原生的，但他們認識到將 GenAI 添加到他們的應用程式的重要性。我們正在讓他們能夠非常輕鬆地在他們的平台中建立複雜的 GenAI 功能。

And finally, the Agentic layer, our Cloudways Copilot is doing really well. We are getting real customer feedback, and that should go into GA sometime in the next couple of months as well. So overall, that is our AI strategy to democratize access of AI starting from inferencing on the infrastructure layer to GenAI platform and agents that will change the cloud experience. So hopefully, that gives you a perspective on our infrastructure platform and application layer.

最後，在 Agentic 層，我們的 Cloudways Copilot 表現得非常好。我們正在獲得真實的客戶回饋，這些回饋也應該會在未來幾個月內進入 GA 階段。總的來說，我們的人工智慧策略是從基礎設施層的推理到 GenAI 平台和代理，實現人工智慧存取的民主化，從而改變雲端體驗。希望這能讓您對我們的基礎架構平台和應用層有大致的了解。

Pinja Bora, JP Morgan.

摩根大通 (JP Morgan) 的 Pinja Bora。

Oh great, thank you, congrats on quarter. Paddy, maybe talk about just what are you seeing from a macro standpoint? Obviously, there's a lot of uncertainty over there. But so far in April in Q2, are you seeing any change in customer buying behavior or any change in the top of the funnel? Are you seeing people deploy any cloud optimization efforts back again, both domestically and internationally?

哦，太好了，謝謝，恭喜這個季度。帕迪，也許你可以從宏觀角度談談你看到了什麼？顯然，那裡存在著許多不確定性。但是到第二季的 4 月為止，您是否看到客戶購買行為有任何變化或漏斗頂部有任何變化？您是否看到人們在國內和國際上再次部署雲端優化工作？

Great. Pinjalim, thank you for the question. I'll give you the answer first. Whatever we are observing is reflected in our outlook that Matt talked about. So let me take a step back and remind everyone that our customers are Digital Native Enterprises, right?

偉大的。Pinjalim，謝謝你的提問。我先給你答案。正如馬特所談到的，我們所觀察到的一切都會反映在我們的觀點中。所以讓我退一步提醒大家，我們的客戶是數位原生企業，對嗎？

And we had really solid evidence in Q1 that our strategy around the digital natives around builders, scalers, scalers plus, it's growing really robustly at 16%.

我們在第一季獲得了確鑿的證據表明，我們圍繞數位原生代（建造者、擴展者、擴展者+）的策略正在以 16% 的強勁速度成長。

And for digital natives, cloud is a top line driver, right? And we have very clear visibility into customer usage on a daily basis, we monitor several utilization trends across the world. And we are seeing some pockets of customers that have unique impacts on their business. So for example, Ad tech is a vertical that we have seen that is a little bit more in a heightened sense of cautiousness.

對數位原住民來說，雲端運算是最重要的驅動力，對嗎？我們對客戶的日常使用有非常清晰的了解，我們監控著世界各地的幾個使用趨勢。我們看到一些客戶對他們的業務產生了獨特的影響。例如，我們發現廣告技術是一個垂直領域，需要更加謹慎。

But as you know, we don't have any vertical, geographic or even customer concentration. So whatever we are seeing is baked into our outlook. And to the second part of your question, Pinjalim. We have several leading indicators we look at. We look at the usage patterns of our large customers.

但如您所知，我們沒有任何垂直、地理甚至客戶集中度。因此，我們所看到的一切都已融入我們的觀點中。關於你問題的第二部分，Pinjalim。我們關注幾個領先指標。我們關注大客戶的使用模式。

We look at the usage patterns of the long tail of our customer base. We look at the top of the funnel in terms of new business acquisition through our self-service funnel.

我們關注客戶群長尾的使用模式。我們透過自助服務漏斗來專注於新業務取得的漏斗頂部。

We look at the pipeline that is being generated by our sales team, but also from our partners. We look at our AI demand. And from what we can see, we have taken a very appropriately cautious approach to projecting the outlook for the rest of the year because we don't know exactly how things are going to unfold. But given our -- the lack of concentration of our customers and geography and vertical segments, we feel fairly confident in our full year guidance.

我們不僅關注我們的銷售團隊，也關注我們的合作夥伴所產生的管道。我們關注我們的人工智慧需求。從我們所看到的情況來看，我們對今年剩餘時間的前景預測採取了非常謹慎的態度，因為我們不知道事情究竟會如何發展。但考慮到我們的客戶、地理和垂直細分市場缺乏集中度，我們對全年業績指引相當有信心。

Gabriela Borges, Goldman Sachs.

加布里埃拉·博爾赫斯，高盛。

;Hi,, good morning. Thanks for the details. I'm observing a shift in how you're talking about the business from being mostly reactive to being a little bit more predictable and larger deals. So maybe talk to us a little bit about the $20 million multiyear deal that you signed in particular.

嗨，早安。謝謝你的詳細資料。我觀察到，你們談論業務的方式發生了變化，從主要是被動反應，到更可預測和更大的交易。那麼，能否請您與我們具體談談您簽署的價值 2000 萬美元的多年期合約。

How do you think about the potential for more deals like that? And what does the conversation look like given how quickly AI is changing and inference in particular, is evolving, how do you get to the point where our customer is comfortable committing to you long term? What sort of road map discussion is that? Thank you.

您認為會有更多類似交易嗎？鑑於人工智慧的變化速度，尤其是推理的發展速度，對話是什麼樣的，您如何讓客戶願意長期與您合作？那是什麼樣的路線圖討論？謝謝。

Yeah. Thank you, Gabriela. Great question. So yes, as I mentioned in my prepared remarks, we are having a lot of [ads] with these larger digital native enterprise customers. And the example that I quoted, we'll have more details as we start working on the implementation of that customer commitment.

是的。謝謝你，加布里埃拉。好問題。是的，正如我在準備好的演講中提到的那樣，我們與這些大型數位原生企業客戶投放了大量廣告。至於我所引用的例子，當我們開始履行對客戶的承諾時，我們會提供更多的細節。

But taking a big step back and looking at the overall landscape, these conversations that we are having with customers for multiyear large commitment contracts is both on AI inferencing as well as on core cloud.

但退一步來看整體情況，我們與客戶就多年期大承諾合約進行的對話既涉及人工智慧推理，也涉及核心雲。

I think a lot of the functionalities that we have released and the one specifically that I highlighted this earnings call, are all lending themselves to having these kinds of conversations and giving our customers confidence that they can now move larger workloads to our platform without having to be all or nothing.

我認為我們發布的許多功能以及我在本次財報電話會議上特別強調的功能，都有助於進行此類對話，並讓我們的客戶相信他們現在可以將更大的工作負載轉移到我們的平台上，而不必全有或全無。

So the Partner Connect functionality, specifically mentioned how it enables stage migration, where you can run parts of the workload in multi-cloud scenario and move some of the heavier workloads to DigitalOcean in a staged manner.

因此，Partner Connect 功能特別提到了它如何實現分階段遷移，您可以在多雲場景中運行部分工作負載，並以分階段的方式將一些較重的工作負載遷移到 DigitalOcean。

There are so many other things we are doing both on the core infrastructure side, but also in our Platform as a Service with DOKS, now scaling to 1,000 nodes, larger storage and database capabilities with our MySQL and Postgres. So it is a collection of all of these things, plus the continuing maturity of our go-to-market motion.

我們在核心基礎設施方面以及使用 DOKS 的平台即服務方面正在做很多其他事情，現在擴展到 1,000 個節點，使用我們的 MySQL 和 Postgres 提供更大的儲存和資料庫功能。所以它是所有這些事情的集合，再加上我們進入市場的動力的持續成熟。

That is why I specifically called out the webinar case studies even though they may be -- they may sound very obvious and natural for a company like us to be doing that. These are some new muscles that we are exercising and also the enterprise sales portion that we now have, putting our arms around our biggest customers and really give them true onboarding experience.

這就是為什麼我特別提到網路研討會案例研究，儘管對於像我們這樣的公司來說，它們可能聽起來非常明顯和自然。這些是我們正在鍛鍊的一些新能力，也是我們現在擁有的企業銷售部分，我們將擁抱我們最大的客戶並真正為他們提供真正的入職體驗。

So it's a collection of all of these things that are giving us better at [back]. And of course, as the quarters go by, we are getting better and better at converting a lot of these things. And I just wanted to bring attention to it that this is not a blue bird by any stretch. We are actually starting to see these kinds of deals.

所以，所有這些事情都讓我們能夠更好地[後退]。當然，隨著時間的推移，我們在轉換這些東西方面做得越來越好。我只是想提醒大家注意，這絕對不是一隻藍鳥。我們實際上開始看到這類交易。

And hence, we are also evolving our thinking in terms of what we need to build, how we need to support it, how we should sell these capabilities. And Matt talked about how we should be financing these things going forward. So it is a whole company kind of conscious push towards getting more predictability into our platform, especially as we scale with the larger digital natives.

因此，我們也在不斷改進我們的思維，考慮我們需要建立什麼、如何支援它、如何銷售這些功能。馬特談到了我們今後應該如何為這些項目提供資金。因此，這是整個公司有意識地推動我們平台更加可預測的舉措，特別是當我們隨著更大的數位原生世代而擴展時。

James Fish, Piper Saber.

詹姆斯·菲什、派珀·塞伯。

Hey guys, Just wanted to build off of Gabriel's question a little bit there. But with the increase -- with the need to increase capacity for new customers coming on, Matt, how are you thinking about CapEx investments for the year? And how much ARR is the pipeline for these large strategic deals for the year that requires this level of CapEx investments that have you kind of having to weigh all these options beyond just refinancing the $1.5 billion really?

嘿夥計們，我只是想稍微回答一下 Gabriel 的問題。但是隨著新客戶增加，需要增加容量，馬特，您如何考慮今年的資本支出投資？那麼，今年這些大型策略交易的 ARR 是多少，需要這種程度的資本支出投資，以至於您必須權衡所有這些選擇，而不僅僅是再融資 15 億美元？

Good question, Fish. The CapEx that we accelerated into the first quarter and that we front-loaded into the first quarter, was primarily around getting Atlanta data center up and -- which gave us a tremendous amount of capacity. That capacity is what underpins the growth projection and the revenue outlook and the incremental ARR for 2025.

問得好，Fish。我們在第一季加速並預先投入的資本支出主要用於建立亞特蘭大資料中心——這為我們提供了巨大的容量。這一產能是 2025 年成長預測、收入前景和增量 ARR 的基礎。

And as Paddy indicated with that one large customer, we have the opportunity to fill it in chunks as well as using kind of the more on-demand AI capabilities with our GPU droplets and other capabilities around the GenAI platform. So as we think about the CapEx requirements going forward, again, we're very comfortable with the estimates that we had in the capital we need for this year and the revenue projections that we have for this year.

正如 Paddy 向那位大客戶指出的那樣，我們有機會分塊填充它，並使用我們的 GPU droplet 和 GenAI 平台周圍的其他功能來提供更多按需 AI 功能。因此，當我們考慮未來的資本支出需求時，我們對今年所需資本的估計以及今年的收入預測感到非常滿意。

But as we think about it over a longer period of time, you got to come back to the fact that we are very, very focused on driving revenue growth and doing so while generating strong free cash flow margins. And as we think about some of the lumpier potential needs for the -- for capital around that when we win some of these bigger customers, that's why we're thinking about tools like, hey, well, we could -- we can lease some of the incremental gear if it accelerated our growth beyond what we're already contemplating because we can fund what we're already contemplating very cleanly within the free cash flow margins that we've articulated.

但當我們從更長遠的角度考慮這個問題時，我們必須認識到，我們非常非常專注於推動收入成長，並在實現強勁自由現金流利潤率的同時實現這一目標。當我們考慮一些更不確定的潛在需求時——當我們贏得一些大客戶時，我們會考慮一些工具，比如，我們可以——如果它能加速我們的增長，超出我們已經考慮的範圍，我們可以租賃一些增量設備，因為我們可以在我們已經闡明的自由現金流利潤率內，非常乾淨地為我們已經考慮的項目提供資金。

But as we see bigger and bigger opportunities, we want to make sure that we have the flexibility to pursue those. And so we'll add some financing tools to our toolkit to be able to accommodate that so that we can both accelerate growth and not only maintain, but even potentially improve the free cash flow generation of the business.

但隨著我們看到越來越大的機會，我們希望確保我們有足夠的彈性去追求這些機會。因此，我們將在工具包中添加一些融資工具來適應這種情況，以便我們既可以加速成長，又不僅可以維持，甚至可能改善業務的自由現金流產生。

Thomas Blakey, Cantor.

湯瑪斯·布萊基，領唱者。

Thanks for taking my question and, congratulations on the uptick in demand here, maybe. Just building off of Mr. Fish questions here. What kind of changed from the April 4, Analyst Day in terms of what you're seeing in the market, timing, I think, from an element would be helpful here. What type of like uptick in accelerated growth in the out year, like are you kind of seeing from this uptick?

感謝您回答我的問題，也許祝賀這裡的需求上升。這只是根據 Fish 先生的問題進行的補充。就您所看到的市場情況而言，與 4 月 4 日分析師日相比，時機發生了哪些變化？我認為，從某種程度上來說，時機的選擇會有所幫助。您從中看到了今年加速成長的哪些類型的上升趨勢？

And what would the makeup be? Would it be more recurring revenue? Or would it be more consumption based from the GPU base. Just a lot more color there just in terms of this. What are you seeing in terms of an uptick of these large deals and potential needs for increasing this CapEx?

那麼妝容是什麼樣的呢？這會帶來更多的經常性收入嗎？或者說從 GPU 基礎來看，它的消耗會更大。就這一點而言，那裡的色彩更加豐富。您認為這些大型交易的數量增加，並且有增加資本支出的潛在需求，這種情況如何？

That would be helpful. Thank you.

那將會很有幫助。謝謝。

What I'd say is -- and I'll tie back to what Gabriela has asked earlier, which is we're now at a point -- we've got through Investor Day, we've got a lot of the AI capabilities in place. We're observing the rapid changes in the market and the shift towards inferencing. Majority of our AI revenue right now is inferencing. And again, we feel very good about the near-term outlook of business despite the economic kind of uncertainty.

我想說的是——我將回到加布里埃拉之前提出的問題，那就是我們現在處於一個階段——我們已經度過了投資者日，我們已經擁有了許多人工智慧能力。我們正在觀察市場的快速變化和向推理的轉變。我們目前的人工智慧收入大部分來自推理。儘管存在一些經濟不確定性，但我們對近期的業務前景仍然非常樂觀。

But as we think about '26 and '27 and beyond, and we think about, okay, how do we position ourselves to take advantage of even accelerating the growth beyond what we articulated at Investor Day. We're just thinking like, what are the physics of that? And how would we do that? And how would we both take advantage of some of these larger, lumpier deals while maintaining our free cash flow or even improving it. And that's why we're thinking about the -- some of these other structures.

但是，當我們考慮 26 年、27 年及以後時，我們會思考，好吧，我們如何定位自己，以利用甚至加速成長，超越我們在投資者日所闡明的成長。我們只是在想，這其中的物理原理是什麼？那我們該怎麼做呢？我們該如何利用這些規模更大、更不穩定的交易，同時維持甚至改善我們的自由現金流呢？這就是我們考慮其他一些結構的原因。

And so the demand, as one, we didn't have the Atlanta data center until just recently, literally in the last month or so. And so a lot of the bigger opportunities, we didn't have the ability to win those because we didn't have capacity available.

因此，就需求而言，我們直到最近，也就是上個月左右，才建立了亞特蘭大資料中心。因此，我們沒有能力贏得許多更大的機會，因為我們沒有可用的能力。

And so for us, as we've turned that capacity up, we've been able to get a big win already. That's what's really accelerating the thinking about, okay, well, we got to -- we feel good about 2025, but we got to start planning and thinking about the ways in which we can position ourselves to grow beyond what we had communicated at Investor Day in 2026 and beyond. And that's why we're considering the things that we're considering.

因此對我們來說，隨著我們提高產能，我們已經能夠獲得巨大的勝利。這才是真正加速思考的原因，好吧，我們必須——我們對 2025 年充滿信心，但我們必須開始規劃和思考如何定位自己，以超越我們在 2026 年及以後的投資者日上所傳達的目標。這就是我們考慮這些事情的原因。

Kingsley Crane, Canaccord Genuity.

金斯利·克蘭（Kingsley Crane），Canaccord Genuity。

Hi, thank you. On the named account engagement model, how many accounts have you targeted today? And then how are you evaluating expanding that program to more prospects as you continue to see nice expansion results with those customers?

你好，謝謝。在指定帳戶參與模型中，您今天定位了多少個帳戶？那麼，當您繼續看到這些客戶取得良好的擴展成果時，您如何評估將該計劃擴展到更多潛在客戶？

Hi, Kingsley, so right now, we are targeting our top 3,000 spenders, and that list has expanded, as I mentioned, from last year, we covered 1500 exiting the year. So we just expanded coverage. So we are looking at several leading indicators to inform us when we start scaling it further.

你好，金斯利，現在，我們的目標是消費金額最高的 3,000 名客戶，而且這個名單已經擴大，正如我提到的，從去年開始，我們在今年年底就覆蓋了 1,500 名客戶。所以我們只是擴大了覆蓋範圍。因此，我們正在關注幾個領先指標，以便在我們開始進一步擴大規模時獲得資訊。

My philosophy is always to nail before scaling. So we are still going to be in the face of maximizing the ability to engage with these companies. We also have a different propensity based engagement model with another 5,000-or-so accounts based on their propensity to spend based on their demographic, their usage of certain aspects of our platform and so forth.

我的理念始終是先釘釘子，再縮放。因此，我們仍將面臨最大限度地發揮與這些公司合作的能力。我們還擁有一個不同的基於傾向的參與模型，該模型基於另外 5,000 個左右的帳戶的人口統計、他們對我們平台某些方面的使用情況等的消費傾向。

So think of them as 5,000 customers that look like our top 3,000 customers who have the propensity or the potential to be scalers plus or spend $100,000 or more on our platform. We are proactively engaging with them and bringing the goodness of our platform to these customers. So between these two, I expect us to be busy for the remainder of the year.

因此，可以把他們想像成 5,000 名客戶，他們看起來像我們的前 3,000 名客戶，他們有傾向或潛力成為擴展者，或者在我們的平台上花費 100,000 美元或更多。我們正在積極與他們合作，並將我們平台的優勢帶給這些客戶。因此，我預計我們今年剩餘的時間都會很忙碌。

We are still trying to ensure that we have the right sales enablement model, the right sales rep productivity metrics and things like that before we plan to expand that, because we are getting to a point where we can predict exactly what the productivity of single rep can be and making sure that we are able to scale that to a reasonable degree will give us the confidence to invest more in the future. But for this year, we want to ensure that we are nailing these two motions.

在計劃擴展之前，我們仍在努力確保我們擁有正確的銷售支援模型、正確的銷售代表生產力指標等，因為我們已經能夠準確預測單一銷售代表的生產力，並確保我們能夠將其擴展到合理的程度，這將使我們有信心在未來進行更多投資。但今年，我們要確保這兩項動議能夠順利通過。

Josh Baer, Morgan Stanley.

摩根士丹利的喬希貝爾。

Great, thanks for the question. Really interesting dynamic to explore the alternative financing, if it would enable growth to accelerate beyond your current outlook. Just wondering what you're seeing from the supply constrained side, if you could maybe touch on GPUs, but also more broadly on the leasing market, how tight it is and how that might come into things?

太好了，謝謝你的提問。如果替代融資能夠使成長加速超出目前的預期，那麼探索替代融資確實是一個有趣的動態。我只是想知道您從供應受限方面看到了什麼，如果您可以談談 GPU，但更廣泛地談談租賃市場，它的緊張程度如何以及它會如何影響事情？

Yeah, Josh, on the supply side, we remain kind of, I'd say, in a real healthy position from a supply standpoint. We don't order GPUs or CPUs in such massive quantities that we're having to get in line behind super large orders. And in general, the market, as I'm sure you guys have heard, it's a lot less supply constrained from a GPU standpoint than it was last year. So we've not had any issues with getting new capacity. And we have a variety of sources.

是的，喬希，就供應方面而言，我想說，從供應的角度來看，我們仍然處於非常健康的狀態。我們不會訂購如此大量的 GPU 或 CPU，因此我們必須排隊等待超大訂單。總的來說，我相信你們已經聽說了，從 GPU 的角度來看，今年的市場供應限制比去年少了很多。因此我們在獲取新產能方面沒有遇到任何問題。我們有各種各樣的消息來源。

We've got four different global OEMs that we can access both NVIDIA and AMD, GPUs from and clearly, a lot of sources for CPU. So we don't really worry too much about the supply chain side on the infrastructure side.

我們有四家不同的全球 OEM，可以從他們那裡獲得 NVIDIA 和 AMD 的 GPU，顯然還有很多 CPU 來源。因此，我們並不太擔心基礎設施方面的供應鏈面向。

In terms of the leasing alternatives, there's a tremendous amount of interest from whether it's PEs or its other kind of capital providers to enter into kind of leasing or similar arrangements. There's a lot of money that -- around the globe that is wanting to put to work in support of AI infrastructure in particular.

就租賃替代方案而言，無論是 PE 或其他類型的資本提供者，都對租賃或類似安排表現出濃厚的興趣。全球有大量資金希望投入人工智慧基礎設施建設。

But certainly, in our case, it would be AI and GPU and CPU. So there's a lot of different alternatives. And part of the reason that we're bringing it up now proactively is that we're engaging in some of these calls, and it gives us the ability to talk to a lot of different people and explore options.

但當然，在我們的例子中，它將是 AI、GPU 和 CPU。因此有很多不同的選擇。我們現在主動提出這個問題的部分原因是我們正在參與一些這樣的電話會議，這讓我們有能力與許多不同的人交談並探索各種選擇。

I just want to reiterate, though, we don't need to do this to deliver the growth rate that we're delivering now. We generate a ton of cash. We've got a really good free cash flow profile. We're going to take care of our balance sheet and convert by the end of this year, we're in a very, very good position.

但我只是想重申，我們不需要這樣做就能實現現在的成長率。我們賺取了大量的現金。我們的自由現金流狀況非常好。我們將在今年年底前處理好我們的資產負債表並實現轉型，我們目前處於非常非常好的地位。

This would enable us to grow faster than that without having to compromise the near-term free cash flow. So that's why we're interested in and we think there's a lot of counterparties out there that would be supportive of that kind of incremental addition to our funding strategy.

這將使我們能夠實現更快的成長，而無需損害近期的自由現金流。這就是我們感興趣的原因，我們認為有很多交易對手會支持我們這種增量融資策略。

Mark Zhang, Citi.

花旗銀行的 Mark Zhang。

Hey, great, thanks guys for taking the question. So obviously, nice to see the NDR improvements and traction with the enterprise cohort. But really, you're holding, I guess, like S&M or sales and marketing spend pretty steady. So can you, number one, give a sense of how you're really able to achieve this incremental level of leverage especially given the new programs you're already running, I guess, like customer behavior changes relative to your expectations. And number two, how should we really think about the spend trajectory from here? And when should we maybe see an acceleration of the sales and go-to-market investments? Thanks.

嘿，太好了，謝謝大家回答這個問題。顯然，很高興看到 NDR 的改進以及對企業群體的吸引力。但實際上，我猜你的 S&M 或銷售和行銷支出相當穩定。那麼，首先，您能否介紹一下您是如何真正實現這種增量槓桿水平的，特別是考慮到您已經在運行的新程序，例如客戶行為相對於您的期望的變化。第二，我們該如何思考未來的支出軌跡？我們什麼時候可以看到銷售和市場投資的加速？謝謝。

Yeah. I can start Mark, thank you for the question. So I just answered the question from Kingsley on sales and marketing and our philosophy to investment there. I think we're still early in the cycle in terms of understanding the best ways to serve our customers. Largely, our customer acquisition engine has been from self-service funnel.

是的。我可以開始了，馬克，謝謝你的提問。所以我剛剛回答了金斯利關於銷售和行銷以及我們在那裡投資的理念的問題。我認為，在了解為客戶提供最佳服務的方式方面，我們還處於早期階段。很大程度上，我們的客戶獲取引擎來自自助服務管道。

And now we are just expanding it to additional motions like adding a channel partner program, getting a small, but mighty team in terms of AI outbound sales and also expanding a couple of other partnership funnels.

現在，我們只是將其擴展到其他舉措，例如添加通路合作夥伴計劃，在 AI 外向銷售方面組建一支規模雖小但強大的團隊，並擴展其他幾個合作夥伴管道。

But we are learning very, very rapidly what it takes to, a, bring in new business through these channels, but also leverage this named account model, and farming to expand the footprint of our existing customers.

但我們正在非常迅速地學習如何透過這些管道帶來新業務，同時利用這種指定帳戶模式，並擴大現有客戶的覆蓋範圍。

So there's a lot of learning ahead of us in the spirit of nailing some of these unit economics so that we understand exactly what the points of leverage are and feel conviction to invest behind this to drive expansion even further. And -- we also want to be appropriately cautious given what our customers are dealing with in today's economic climate.

因此，我們還有很多工作要做，本著搞清楚單位經濟學的精神，以便我們準確地理解槓桿點，並有信心為此進行投資，以進一步推動擴張。而且—考慮到我們的客戶在當今的經濟環境下所面臨的情況，我們也希望採取適當的謹慎態度。

I believe there are two major factors that are going on. One is the economic dislocation that is happening, but also more importantly, a technology dislocation, which is largely positive for digital native customers and they are going to be more technology hungry as part of this. So we are very cautiously observing these things, and we will invest at the right time behind the right unit economics to get more leverage and drive expansion and new business appropriately.

我認為有兩個主要因素正在發生。一是正在發生的經濟錯位，但更重要的是技術錯位，這對數位原生代客戶來說基本上是有利的，他們將因此更加渴望科技。因此，我們非常謹慎地觀察這些事情，並將在正確的時間、正確的單位經濟條件下進行投資，以獲得更多槓桿並適當地推動擴張和新業務。

As we are at the top of the hour, I would like to thank our speakers for today's presentation and thank you all for joining us. This now concludes today's conference. You may now disconnect.

現在正值整點，我要感謝今天的演講者，也感謝大家的參與。今天的會議到此結束。您現在可以斷開連線。