GitLab Inc. (GTLB) 2027 Q2 法說會逐字稿

Good day, everyone, and welcome to today's GitLab's second quarter fiscal year 2027 conference call. (Operator Instructions) Please note, this call is being recorded.

各位好，歡迎參加 GitLab 2027 會計年度第二季財報電話會議。(接線員指示) 請注意，本通話將被錄音。

It is now my pleasure to turn the conference over to Nic Edwards.

現在我很榮幸將會議交給 Nic Edwards。

Good afternoon, and welcome to GitLab's second quarter 2027 financial results earnings call. I'm Nic Edwards, VP of Finance, Strategy and Operations. And with me are Bill Staples, our CEO; and Jessica Ross, our CFO.

各位下午好，歡迎參加 GitLab 2027 年第二季財務業績電話會議。我是 Nic Edwards，財務、策略與營運副總裁。與我一同出席的有我們的執行長 Bill Staples，以及財務長 Jessica Ross。

During this afternoon's call, we will provide an overview of the business, commentary on our second quarter and full year results and guidance for the second quarter and fiscal year 2027. In addition to our prepared remarks on this call, additional information can be found in our shareholder letter, press release and investor presentation and SEC filings on our Investor Relations website at ir.gitlab.com.

在今天下午的電話會議中，我們將提供業務概覽，並就第二季與全年業績發表評論，同時提供第二季及 2027 會計年度的展望指引。除本次電話會議的事先準備講稿外，更多資訊可於我們的股東信、新聞稿、投資人簡報，以及位於 ir.gitlab.com 的投資人關係網站上的 SEC 申報文件中查閱。

Before we begin, I'll cover the safe harbor statement. I would like to direct you to the cautionary statement regarding forward-looking statements on page 2 of our presentation and in our earnings release, issued earlier today, both of which are available under the Investor Relations section of our website. The presentation and earnings release include a discussion of certain risks, uncertainties, assumptions and other factors that could cause our results to differ from those expressed in any forward-looking statements within the meaning of the Private Securities Litigation Reform Act. As is customary, the content of today's call and presentation will be governed by this language.

在開始之前，我將先說明安全港聲明。請各位參閱我們簡報第 2 頁以及今日稍早發布的財報新聞稿中，關於前瞻性陳述的警示性聲明；兩者皆可於我們網站的投資人關係專區取得。該簡報與財報新聞稿包含對若干風險、不確定性、假設及其他因素的討論，這些因素可能導致我們的實際結果與任何前瞻性陳述所表達者不同；該等前瞻性陳述係依《私人證券訴訟改革法》之定義。依慣例，今日電話會議與簡報內容將受上述措辭所規範。

In addition, during today's call, we will be discussing certain non-GAAP financial measures. These non-GAAP financial measures exclude certain unusual or nonrecurring items that management believes impact the comparability of the periods referenced. Please refer to our earnings release and presentation materials for additional information regarding these non-GAAP financial measures and the reconciliations to the most directly comparable GAAP measure.

此外，在今天的電話會議中，我們將討論若干非 GAAP 財務衡量指標。這些非 GAAP 財務衡量指標排除某些不尋常或非經常性項目；管理層認為這些項目會影響所提及期間之可比性。關於這些非 GAAP 財務衡量指標的更多資訊，以及與最直接可比之 GAAP 指標的調節表，請參閱我們的財報新聞稿與簡報資料。

I will now turn the call over to Bill. Bill?

接下來我將把電話會議交給 Bill。Bill？

Thanks, Nic, and good afternoon, everyone. Q2 was an exceptional quarter. Let me start with just some of the highlights. Revenue was $286.3 million, up 21% year-over-year, and non-GAAP operating income was $42.6 million, representing a 15% operating margin. Our sales team delivered the largest gross bookings quarter in company history, and net ARR grew more than 40% year-over-year, our highest in several years. We also beat all of our targets on first orders, which grew more than 100% year-over-year.

謝謝你，Nic，各位下午好。第二季是非常出色的一季。我先從幾個重點開始。營收為 2.863 億美元，年增 21%；非 GAAP 營業利益為 4,260 萬美元，營業利益率為 15%。我們的銷售團隊創下公司史上最高的總訂單（gross bookings）季度表現，淨 ARR 年增超過 40%，為數年來最高。我們也在所有首購（first orders）目標上全面超標，首購年增超過 100%。

Underneath those highlights, I'd like to share a brief update on each of the five growth initiatives that we laid out for this year. First, we told it was important that we accelerate first orders. Q2 was our largest first order quarter in three years. First order count grew more than 100% year-over-year and first order net ARR increased nearly 40%.

在上述亮點之下，我想就我們今年提出的五項成長倡議逐一做個簡短更新。首先，我們表示加速首購非常重要。第二季是我們三年來首購表現最佳的一季。首購筆數年增超過 100%，首購淨 ARR 也增加近 40%。

Our dedicated first order organization is building more repeatable sales motions, while our product-led investments are widening the top of the funnel. The combination is bringing significantly more new customers into GitLab. It's important to win customers of all sizes, including AI start-ups and others with small orders to begin with because more than half of our current $1 billion plus in run rate revenue comes from customers whose first order was less than $5,000. Our strategy is to land customers wherever they're ready to start and grow with them over time.

我們專責的首購組織正在建立更可複製的銷售動作（sales motions），同時我們以產品為主導（product-led）的投資也在擴大漏斗頂端。兩者結合，正為 GitLab 帶來顯著更多的新客戶。贏得各種規模的客戶都很重要，包括 AI 新創以及其他一開始下單金額較小的客戶，因為我們目前超過 10 億美元的年化營收（run rate revenue）中，有一半以上來自首購金額低於 5,000 美元的客戶。我們的策略是在客戶準備開始的任何起點先落地（land），並隨時間與他們一同成長。

Second, we said increased predictive sales capacity would be an important driver of reacceleration. And in Q2, we saw that thesis begin to pay off. Account executive capacity increased approximately 30% year-over-year, while productivity per rep improved approximately 10%. Attrition also improved year-over-year for the second consecutive quarter. Adding capacity while simultaneously increasing productivity is difficult, and we believe it demonstrates that the investments that we've made into our sales organization are beginning to compound. The result was the largest gross bookings quarter in GitLab history and net ARR growth of more than 40%.

第二，我們表示提升可預測的銷售產能將是再加速的重要驅動因素。而在第二季，我們看到這個論點開始奏效。客戶經理（Account Executive）產能年增約 30%，同時每位業務代表的生產力提升約 10%。離職率也連續第二季較去年同期改善。在增加產能的同時又提升生產力並不容易，我們認為這顯示我們對銷售組織的投資正開始產生複利效應。其結果是 GitLab 史上最高的總訂單季度表現，以及淨 ARR 年增超過 40%。

Third, we told you it was important for us to expand our monetization vectors. Our seat-based business continues to grow and AI is creating an additional opportunity for GitLab to monetize an increasing amount of work happening across the software life cycle. Flex is an important enabler of that strategy. After only six weeks in market, more than 130 customers committed more than $20 million to Flex. That early response reinforces our belief that customers want a simpler way to allocate their GitLab spend across seats and new consumption products as their needs evolve.

第三，我們告訴各位，擴大我們的變現向量（monetization vectors）對我們很重要。我們以席位（seat）為基礎的業務持續成長，而 AI 也為 GitLab 帶來額外機會，可針對軟體生命週期中愈來愈多的工作進行變現。Flex 是推動該策略的重要促成因素。上市僅六週，就有超過 130 家客戶對 Flex 承諾超過 2,000 萬美元。這樣的早期回饋強化了我們的信念：客戶希望以更簡單的方式，隨著需求演進，在席位與新的用量型（consumption）產品之間分配其 GitLab 支出。

Prior to Flex, customers committed dollars to fixed quantities of seats for premium or ultimate access to our platform. Flex changes that. Customers now make a dollar commitment and each month, they could decide how to put it to work across seats and all of our consumption products. The commitment is fixed and can be used for any product in the platform. With work accelerating due to agents and more builders coming into the picture, we have a tremendous opportunity ahead to create and capture value. Anything we can do to shorten the time between customer interest and realized value is good for customers and it's good for GitLab. Flex was designed to do this in three important ways.

在 Flex 之前，客戶會針對固定數量的席位承諾金額，以取得我們平台的 Premium 或 Ultimate 存取權。Flex 改變了這一點。客戶現在先做出金額承諾，並且每個月都可以決定要如何在席位與我們所有用量型產品之間分配使用。承諾金額是固定的，且可用於平台中的任何產品。隨著代理（agents）加速工作進程、更多建置者（builders）加入，我們面前有巨大的機會去創造並擷取價值。任何能縮短客戶興趣到實現價值之間時間的作法，對客戶有利，也對 GitLab 有利。Flex 的設計正是要以三個重要方式達成這點。

First, it helps customers get more value from every dollar they commit to GitLab. Instead of leaving unused capacity stranded as shelfware, customers can redirect dollars toward products creating the most value for them. We believe better utilization should strengthen our already healthy retention profile by reducing the potential for churn and contraction driven by unused spend.

第一，它幫助客戶從其對 GitLab 承諾的每一美元中獲得更多價值。客戶不必讓未使用的容量閒置成為「架上軟體」（shelfware），而是可以把資金重新導向到對他們最能創造價值的產品。我們相信，更好的使用率應能強化我們原本就健康的留存表現，藉由降低因未使用支出而導致的流失（churn）與縮減（contraction）風險。

Flex also reduces sales and procurement friction because once the financial commitment is in place, customers can allocate dollars across existing and new products on their own timing throughout the year without requiring a new commercial agreement each time. We've already increased sales capacity and productivity. Flex should help us make that capacity even more productive by allowing our teams to spend less time on incremental contracting and more time helping customers realize value.

Flex 也能降低銷售與採購摩擦，因為一旦財務承諾到位，客戶就能在一年中依其時程自行在既有與新產品間分配金額，而不需要每次都重新簽訂新的商務合約。我們已經提升了銷售產能與生產力。Flex 應能讓這些產能更有效率，讓團隊把更少時間花在增量簽約上，把更多時間用於協助客戶實現價值。

And third, it lets usage expand seamlessly as customer needs grow. Customers can consume any eligible product on demand above reserve capacities, including premium and ultimate seats and any usage above their commitment or reservation is billed in the month it occurs. That removes many of the true-up and contracting cycles that are required today to capture that incremental usage and gives customers more flexibility. It creates another potential growth tailwind for GitLab as usage expands.

第三，當客戶需求成長時，它讓使用量能無縫擴張。客戶可在保留容量（reserve capacities）之上，按需使用任何符合資格的產品，包括 Premium 與 Ultimate 席位；任何超出其承諾或預留的使用量，將在發生當月計費。這消除了目前為了擷取該增量使用量所需的許多補差額（true-up）與簽約週期，並給客戶更大的彈性。隨著使用量擴張，這也為 GitLab 帶來另一個潛在的成長順風。

Paid consumption run rate ended the quarter above $40 million, up from $15 million existing in Q1, thanks to the introduction of Flex. As a reminder, paid CRR is a point-in-time annualized measure that includes GitLab credit commitments, Flex commitments and paid on-demand consumption. It excludes trials and promotional credits.

付費用量年化執行率（paid consumption run rate）在本季末超過 4,000 萬美元，高於第一季既有的 1,500 萬美元，主要受 Flex 推出帶動。提醒一下，付費 CRR 是一個特定時點的年化衡量指標，包含 GitLab 點數（credit）承諾、Flex 承諾，以及付費的隨用隨付（on-demand）用量。它不包含試用與促銷點數。

We believe paid CRR gives investors a useful way to track the dollars moving through our consumption model as it scales. Our objective is to exceed $100 million of paid CRR by the end of this fiscal year. We entered this year with a business monetized entirely through seat subscriptions and we now expect to exit it with seats continuing to grow alongside a meaningful and scaling consumption business.

我們相信，付費 CRR 能為投資人提供一個有用的方法，用以追蹤我們用量模式在擴張過程中流經的金額。我們的目標是在本會計年度結束前，付費 CRR 超過 1 億美元。我們在今年初的業務完全透過席位訂閱變現，而我們現在預期在今年結束時，席位將持續成長，同時也會有一個具規模且正在擴張的用量型業務並行發展。

Fourth, we need to improve our performance in our price-sensitive customer segments. And I'm pleased to report that we now have seen stabilization in SMB and mid-market this quarter with performance ahead of our targets for both of those segments in both expansion and first orders. Our thesis was increased focus and investment in this segment could improve performance, and this quarter's results are encouraging. We want to see that performance sustained over the next few quarters, but we're increasingly confident that we're on the right path.

第四，我們需要提升在價格敏感型客戶細分市場的表現。我很高興地報告，本季我們已看到 SMB 與中端市場趨於穩定，且在這兩個細分市場的擴張與首次訂單表現皆優於我們的目標。我們的論點是：加大對該細分市場的聚焦與投資可以改善表現，而本季的結果令人鼓舞。我們希望在接下來幾季能維持這樣的表現，但我們也愈來愈有信心，我們正走在正確的道路上。

Fifth, we're executing a bold AI strategy. This quarter, Duo Agent Platform CRR grew roughly 50% quarter-over-quarter, inclusive of credit commitments, paid on-demand credits and flex reservations. We also broadened adoption geographically after a more US-centric first quarter. And one top 20 US commercial bank expanded its AI credit pool nearly tenfold this quarter.

第五，我們正在執行一項大膽的 AI 策略。本季，Duo Agent Platform 的 CRR 約較上季成長 50%，包含信用承諾、按需付費的點數以及彈性預留。在第一季較以美國為中心之後，我們也在地理區域上擴大了採用。此外，本季有一家美國前 20 大商業銀行將其 AI 點數池擴大了近十倍。

But I think the broader AI story is even more important. For more than a decade, GitLab has brought together the context of how software is built, secured and shipped across source code, issues, merge requests, pipelines, vulnerabilities, policies, approvals and deployments. As AI becomes more capable, we believe that connected context becomes even more valuable, and we're seeing early evidence of that with GitLab Orbit, our context graph for the software life cycle.

但我認為更廣泛的 AI 故事更為重要。十多年來，GitLab 將軟體如何被建置、保護並交付的脈絡整合在一起，涵蓋原始碼、議題、合併請求、管線、弱點、政策、核准與部署。隨著 AI 能力愈來愈強，我們相信這種相互連結的脈絡將變得更有價值；我們也在 GitLab Orbit（我們用於軟體生命週期的脈絡圖譜）上看到了早期證據。

Since opening the beta in June, more than 2,200 organizations have enabled orbit indexing, an increase of 70% in four weeks. Customers have generated more than 170,000 queries and roughly 80% of customer query volume comes from customers connecting Orbit to external agents such as Claude Code and Codex.

自 6 月開放 Beta 版以來，已有超過 2,200 家組織啟用 Orbit 索引，四週內增加了 70%。客戶已產生超過 170,000 次查詢，且約 80% 的客戶查詢量來自於將 Orbit 連接到外部代理（例如 Claude Code 與 Codex）的客戶。

We're also beginning to see evidence that better context improves agent performance. Compare the market tested Orbit against traditional RAG with 79 real merger requests and saw accuracy improve from 58% to 70% and the same dynamic is visible across our platform. Year-over-year, secure repositories grew 60%, code pushes grew 50%, CI/CD pipelines grew 40%. Among some customers moving aggressively into AI-assisted development, we've seen code bases grow as much as 500%.

我們也開始看到更佳脈絡能提升代理表現的證據。我們以 79 個真實的合併請求，將經市場測試的 Orbit 與傳統 RAG 進行比較，準確率從 58% 提升至 70%；同樣的動態也在我們的平台各處可見。以年對年來看，安全儲存庫成長 60%，程式碼推送成長 50%，CI/CD 管線成長 40%。在一些積極導入 AI 輔助開發的客戶中，我們看到程式碼庫規模最高成長達 500%。

The pattern is increasingly clear. As enterprises adopt more AI development tools, they use more of GitLab. That matters because our core business remains very strong. Ultimate now represents 59% of ARR and 8 of our top 10 deals this quarter. Premium and Ultimate are the foundation of a business now exceeding $1 billion in trailing four-quarter revenue, and we continue to see significant opportunity ahead. AI gives us the opportunity to build on top of that foundation with new products, new consumption and entirely new kinds of work flowing through GitLab.

這個模式愈來愈清楚。當企業採用更多 AI 開發工具時，他們會使用更多 GitLab。這很重要，因為我們的核心業務仍然非常強勁。Ultimate 目前占 ARR 的 59%，並占本季前 10 大交易中的 8 筆。Premium 與 Ultimate 是一項業務的基礎，該業務在過去四季合計（TTM）營收已超過 10 億美元，我們仍持續看到前方有顯著機會。AI 讓我們有機會在此基礎之上打造新產品、新的消費模式，以及讓全新類型的工作流經 GitLab。

Before I turn it over to Jessica, I want to recognize two important groups. First, Ian Steward, our CRO, and the entire sales organization. Ian has now been with GitLab for five quarters. The results we're seeing today reflect many quarters of disciplined work, building the fundamentals, sharpening our strategy and creating a team capable of executing against it.

在我把時間交給 Jessica 之前，我想表揚兩個重要團體。第一個是我們的 CRO Ian Steward，以及整個銷售組織。Ian 加入 GitLab 至今已五個季度。我們今天看到的成果，反映了多個季度以來的紀律性工作：建立基本功、精煉策略，並打造一支能夠依策略執行的團隊。

I especially also want to recognize the entire GitLab team because at the beginning of this quarter, we made the difficult decision to restructure the company. Many people chose to stay and help build GitLab's next chapter. Together, we delivered one of the strongest quarters in our company's history. To everyone at GitLab, thank you for your hard work, your resilience and your trust. I'm so incredibly proud of you and what we've accomplished.

我也特別想表揚整個 GitLab 團隊，因為在本季一開始，我們做出了重組公司的艱難決定。許多人選擇留下來，協助打造 GitLab 的下一個篇章。我們一起交出公司史上最強勁的季度之一。致 GitLab 的每一位同仁，感謝你們的努力、韌性與信任。我為你們以及我們所達成的成果感到無比自豪。

We're entering Act 2 from a position of strength, a healthy core business, accelerating customer growth, new products gaining traction and a consumption model designed for a world where humans and agents increasingly build software together.

我們正以強勢姿態進入第二幕：核心業務健康、客戶成長加速、新產品逐步獲得動能，以及一個為人類與代理愈來愈共同建置軟體的世界所設計的消費模式。

With that, I'll turn it over to Jessica.

接下來，我把時間交給 Jessica。

Thanks, Bill, and thanks to everyone joining us today. Q2 was an exceptional quarter, and we believe an important inflection point for this business. Revenue was $286.3 million, up 21% year-over-year and roughly 5 points ahead of our guidance. We delivered the strongest gross bookings quarter in company history. Net ARR grew 42% year-over-year, our second highest growth quarter in the last four years. Dollar-based net retention accelerated sequentially for the first time since 2024, and we did all of this while also beating our profitability expectations.

謝謝 Bill，也謝謝今天加入我們的各位。第二季是非常出色的一季，我們相信這是這項業務的重要拐點。營收為 2.863 億美元，年增 21%，並較我們的指引高出約 5 個百分點。我們創下公司史上最強的總預訂額季度。淨 ARR 年增 42%，為過去四年中第二高的成長季度。以美元計的淨留存率自 2024 年以來首次連續季度加速，而我們在達成這些成果的同時，也超越了我們對獲利能力的預期。

The strength in our business was broad-based across new and existing customers and multiple geographies, every customer size and across our core platform. That breadth is why we're increasingly confident in the underlying trajectory of this business.

我們業務的強勁表現具有廣泛性，涵蓋新舊客戶與多個地理區域、各種客戶規模，以及我們的核心平台。正因為這種廣度，我們對這項業務的基本走勢愈來愈有信心。

Our go-to-market team executed at an extremely high level. We closed significantly more large deals than we forecasted. We saw better-than-expected linearity and sales attainment improved. Deals of $500,000 or more grew more than 150% year-over-year, a clear signal that our largest customers are deepening their commitment to GitLab.

我們的市場推進（go-to-market）團隊以極高水準執行。我們完成的重大交易數量顯著高於預測。我們看到線性度優於預期，銷售達成率也有所改善。金額達 50 萬美元以上的交易年增超過 150%，清楚顯示我們最大的客戶正在加深對 GitLab 的承諾。

Public sector also rebounded meaningfully in the quarter, and we see room for that recovery to continue as buying patterns normalize through the back half of the year. We also saw a genuine step change in new customer formation. We had approximately 1,700 first orders in the quarter, more than double a year ago, and new logo net ARR grew 39% year-over-year.

公部門在本季也明顯回升；隨著採購模式在下半年趨於正常化，我們認為這波復甦仍有延續空間。我們也看到新客戶形成出現真正的階躍式變化。本季我們約有 1,700 筆首次訂單，較一年前增加逾一倍，而新客戶（new logo）淨 ARR 年增 39%。

The magnitude of this quarter's performance exceeded even our own ambitious expectations. We care about all the customers that choose GitLab, large and small. More than half of our $1 billion-plus run rate revenue base today came from customers whose very first order was under $5,000. Landing them early and growing with them is part of our competitive advantage. It allows us to acquire customers efficiently and grow with them over time.

本季表現的幅度甚至超出我們自身雄心勃勃的預期。我們重視所有選擇 GitLab 的客戶，不論規模大小。目前我們年化營收運行率超過 10 億美元的收入基礎中，超過一半來自於那些第一筆訂單低於 5,000 美元的客戶。及早落地並與他們一同成長，是我們競爭優勢的一部分。這讓我們能有效率地獲取客戶，並隨時間推進與他們共同成長。

GitLab Ultimate had an exceptional quarter as well. Ultimate ARR grew approximately 35% year-over-year and now represents 59% of our total ARR.

GitLab Ultimate 本季同樣表現出色。Ultimate ARR 約年增 35%，目前占我們總 ARR 的 59%。

Security, governance and compliance are becoming more central to buying decisions as artificial intelligence proliferates across the software development life cycle. We're starting to see demand extend beyond the traditional developer seat entirely as AI makes software creation accessible to a much broader set of builders across the enterprise. We think that expands who GitLab can ultimately serve.

隨著人工智慧在軟體開發生命週期中快速普及，安全、治理與合規正變得更為核心，並影響採購決策。我們開始看到需求完全延伸到傳統「開發者席位」之外，因為 AI 讓軟體創作對企業內更廣泛的建置者群體變得可及。我們認為這擴大了 GitLab 最終能服務的對象範圍。

Turning to customer metrics. Gross retention stayed well above 90%, consistent with our historical trends. Dollar-based net retention was 117%, up sequentially for the first time since 2024. Total RPO grew 16% year-over-year to $1.2 billion and current RPO grew 20% to $744.7 million. Calculated billings grew 24%, doubling the 12% growth rate we posted just last quarter.

接著看客戶指標。總留存率仍遠高於 90%，與我們的歷史趨勢一致。以美元計的淨留存率為 117%，自 2024 年以來首次較前一季上升。總 RPO 年增 16% 至 12 億美元，當期 RPO 年增 20% 至 7.447 億美元。計算帳單（calculated billings）成長 24%，是我們僅在上一季公布的 12% 成長率的兩倍。

Our emerging products are gaining real traction. Duo Agent Platforms paid CRR grew 50% sequentially and platform-wide paid consumption run rate exceeded $40 million exiting the quarter. Usage trends tell the same story. CI pipelines created and push actions were both up more than 40% year-over-year.

我們的新興產品正獲得實質動能。Duo Agent Platform 的付費 CRR 較前一季成長 50%，而全平台的付費消費運行率在季度末已超過 4,000 萬美元。使用趨勢也傳達相同訊息。CI 管線建立數與推送動作年增皆超過 40%。

Moving down the income statement. Non-GAAP gross margin was 86.5%. SaaS was 34% of total revenue and grew 36% year-over-year, powered by continued strength in GitLab Dedicated and Duo. Non-GAAP operating income was $42.6 million, up from $39.6 million a year ago for a non-GAAP operating margin of approximately 15%, ahead of our expectations, driven by improved sales and marketing productivity and the timing of certain investments. We incurred approximately $23.3 million in restructuring charges, in line with what we outlined last quarter.

接著往下看損益表。非 GAAP 毛利率為 86.5%。SaaS 佔總營收的 34%，年增 36%，主要受 GitLab Dedicated 與 Duo 持續強勁表現帶動。非 GAAP 營業利益為 4,260 萬美元，高於一年前的 3,960 萬美元，非 GAAP 營業利益率約 15%，優於我們的預期，主要受銷售與行銷生產力提升以及部分投資時點影響所帶動。我們認列約 2,330 萬美元的重組費用，與我們上季所說明的內容一致。

On JiHu, Q2 non-GAAP expenses were essentially flat year-over-year at $3.0 million. Our goal remains to deconsolidate JiHu, though we can't predict if or when that will happen. Adjusted free cash flow was $9.8 million, a 3% margin with the timing of collections weighing on the number this quarter.

關於 JiHu，第二季非 GAAP 費用年比基本持平，為 300 萬美元。我們的目標仍是將 JiHu 去合併（deconsolidate），但我們無法預測是否或何時會發生。調整後自由現金流為 980 萬美元，利潤率 3%，本季收款時點對該數字造成壓力。

We returned capital to shareholders by repurchasing approximately 3.5 million shares. We have about $245 million remaining under our current authorization. We ended Q2 with $1.3 billion in cash and investments.

我們透過回購約 350 萬股股票向股東返還資本。在目前的授權額度下，我們尚餘約 2.45 億美元可用。第二季末我們持有 13 億美元的現金與投資。

Before I get to guidance, I want to spend some time on Flex because it is central to where this business is going and because it changes how some of our results will show up in reported financials. This is the first quarter Flex shows up in our results, and the impact was therefore immaterial relative to the size of our existing revenue base. However, in the first six weeks on the market, over 130 customers committed more than $20 million to Flex. That is a fast emphatic signal of demand for a fundamentally better way to buy from us.

在我談指引之前，我想花一些時間談 Flex，因為它是這項業務未來走向的核心，也因為它會改變我們部分結果在財報中的呈現方式。這是 Flex 首次反映在我們的業績中，因此相對於我們既有的營收基礎規模，其影響並不重大。然而，在上市後的前六週，已有超過 130 家客戶對 Flex 承諾超過 2,000 萬美元。這是一個快速且明確的訊號，顯示市場對一種根本更好的向我們採購方式有需求。

At its core, Flex is a single annual or multiyear dollar commitment that customers draw down flexibly across premium and ultimate seats, GitLab credits and other usage-based capabilities without renegotiating a contract or restarting procurement every time their needs shift. We think this is simply a better model for customers, and we believe over time, it becomes the default way customers transact with GitLab.

Flex 的核心是一筆單一的年度或多年期美元承諾，客戶可在不重新談判合約、也不必在需求變動時每次都重新啟動採購流程的情況下，於 Premium 與 Ultimate 席位、GitLab 點數（credits）及其他按使用量計費的功能之間彈性扣抵使用。我們認為這對客戶而言就是更好的模式，並且相信隨著時間推進，它會成為客戶與 GitLab 交易的預設方式。

It's a better model for us, too. It means our growth is tied to how much value customers get from the whole platform, not just how many seats they buy. Given the strong signal we are seeing from our customers, we have become more convicted in the growth and adoption of Flex. It also means Flex may become materially impactful to our financial results as this commercial model becomes our customers' first choice for transacting with us.

這對我們而言也是更好的模式。這代表我們的成長將與客戶從整個平台獲得的價值連動，而不僅僅取決於他們購買了多少席位。鑑於我們從客戶端看到的強烈訊號，我們對 Flex 的成長與採用更具信心。同時也意味著，隨著這種商業模式成為客戶與我們交易的首選，Flex 可能會對我們的財務結果產生實質影響。

From that standpoint, there are two reporting implications investors will need to understand clearly when it comes to the impact of a more material adoption of Flex by our customer base, revenue recognition and RPO.

從這個角度來看，當客戶群更大幅度採用 Flex 時，投資人需要清楚理解兩項申報（報導）層面的影響：收入認列與 RPO。

First, revenue recognition. Today, approximately 15% of a traditional self-managed license is recognized upfront. Under Flex, the license fee is no longer recognized upfront. Since customers can shift product mix within their committed dollars, that revenue is instead recognized over the contract term.

第一，收入認列。目前，傳統自管（self-managed）授權約有 15% 會在前期一次性認列。在 Flex 之下，授權費不再於前期一次性認列。由於客戶可在其承諾金額內調整產品組合，該收入將改為在合約期間內分期認列。

To help you model this, for every $50 million of our self-managed available to renew that converts to Flex in FY27, we estimate approximately $5 million of revenue that would otherwise have landed in FY27 instead shifts to be recognized over future periods. Given the size of the available-to-renew pool in the back half of the year as well as our visibility into the H2 pipeline, we expect the maximum potential impact on our revenue to be approximately $13 million in FY27.

為協助你們建模，若 FY27 有每 5,000 萬美元原本可續約的自管業務轉換為 Flex，我們估計約有 500 萬美元原本會落在 FY27 的收入，將改為在未來期間認列。考量今年下半年可續約池的規模以及我們對 H2 管線的能見度，我們預期 FY27 對營收的最大潛在影響約為 1,300 萬美元。

To be clear, this is just a timing change in revenue recognition. It is not a change in the underlying customer commitment, and it is not a change in cash economics. Customer commitments are unchanged and cash billings continue to be billed annually upfront regardless of which model a customer is on. Because Flex is a contractual dollar commitment, the full committed amount is captured in total RPO. However, Flex is excluded from current RPO because of signing, the customer controls the timing, product mix and pace of future consumption, so we can't reliably say how much relates to the next 12 months. That means as Flex scales, you may see total RPO and revenue growth move at different paces, and you should expect that.

需要說明的是，這只是收入認列時點的變化。這不是底層客戶承諾的改變，也不是現金經濟（cash economics）的改變。客戶承諾不變，且不論客戶採用哪種模式，現金帳單仍會按年在前期一次性開立。由於 Flex 是合約性的美元承諾，完整承諾金額會納入總 RPO。然而，Flex 會被排除在當期 RPO（current RPO）之外，因為在簽約後，客戶可自行掌控未來消耗的時點、產品組合與消耗速度，因此我們無法可靠地判斷其中有多少與未來 12 個月相關。這表示隨著 Flex 規模擴大，你可能會看到總 RPO 與營收成長以不同步調變動，而你應該預期會出現這種情況。

This quarter, cRPO saw a 3-point headwind relative to RPO purely from the absence of Flex commitments in that metric. Every quarter, as Flex adoption grows, we will explicitly quantify its impact so you can cleanly separate the accounting effect of Flex from the underlying performance of the business.

本季 cRPO 相對於 RPO 出現 3 個百分點的逆風，純粹是因為該指標中未包含 Flex 承諾。未來每一季，隨著 Flex 採用增加，我們都會明確量化其影響，讓你們能清楚區分 Flex 的會計影響與業務的底層表現。

Now turning to our outlook. Our confidence in the underlying trajectory of this business has increased following our first half performance.

接著談我們的展望。在上半年表現之後，我們對這項業務的底層發展軌跡更具信心。

A few things to keep in mind as you think about the second half. First, we're taking a more constructive view of public sector given the improved buying patterns we saw this quarter, and we expect that recovery to continue gradually through the rest of the year.

在你們思考下半年時，有幾點需要留意。第一，鑑於本季我們看到公部門採購模式改善，我們對公部門採取更為正向的看法，並預期這項復甦將在今年剩餘時間逐步延續。

Second, Q2 benefited from exceptional execution, more large deals than forecasted, improved linearity and our best pipeline conversion ever. Our second half outlook assumes a more normalized pace of bookings from here, not a repeat of Q2's exceptional levels.

第二，第二季受惠於卓越的執行力、大型交易數量高於預期、線性度改善，以及我們史上最佳的管線轉換率。我們對下半年的展望假設從此處起訂單（bookings）回到較正常的節奏，而非重現第二季的異常高水準。

Third, on Duo Agent Platform, we assume limited contribution in FY27 relative to our large existing revenue base. Our focus this year remains on adoption, converting pilots into production deployments and building the foundation for monetization ahead.

第三，關於 Duo Agent Platform，我們假設 FY27 的貢獻相對於我們龐大的既有營收基礎仍然有限。我們今年的重點仍在於推動採用、將試點轉為正式上線部署，並為後續變現建立基礎。

And finally, to remain consistent with the guidance philosophy and forecasting methodology we have had all year, we have not incorporated the potential impact of Flex in our guide, but we do intend to quantify the revenue recognition impacts through the rest of the year.

最後，為了與我們全年一貫的指引理念與預測方法保持一致，我們尚未將 Flex 的潛在影響納入指引之中，但我們確實打算在今年剩餘時間量化其對收入認列的影響。

With these points in mind, I am pleased to share that we are raising our guidance to reflect the strong momentum we experienced in the first half of the year as well as the momentum we anticipate will continue into the back half.

基於以上幾點，我很高興分享：我們正在上調指引，以反映我們在上半年所經歷的強勁動能，以及我們預期將延續至下半年的動能。

For Q3 FY27, we expect total revenue of $281 million to $283 million, representing approximately 15% to 16% year-over-year growth. We expect non-GAAP operating income of $35 million to $37 million and non-GAAP net income per share of $0.19 to $0.20, assuming approximately 172 million weighted average diluted shares outstanding.

針對 FY27 第三季，我們預期總營收為 2.81 億至 2.83 億美元，約為年增 15% 至 16%。我們預期非 GAAP 營業利益為 3,500 萬至 3,700 萬美元，非 GAAP 每股淨利為 0.19 至 0.20 美元，假設加權平均稀釋流通股數約為 1.72 億股。

For the full year, we now expect total revenue of $1.129 billion to $1.133 billion, representing approximately 18% to 19% year-over-year growth. We expect non-GAAP operating income of $148 million to $152 million and non-GAAP net income per share of $0.85 to $0.87, assuming approximately 172 million weighted average diluted shares outstanding. There's no change to our underlying investment priorities or how we're balancing growth and profitability.

針對全年，我們目前預期總營收為 11.29 億至 11.33 億美元，約為年增 18% 至 19%。我們預期非 GAAP 營業利益為 1.48 億至 1.52 億美元，非 GAAP 每股淨利為 0.85 至 0.87 美元，假設加權平均稀釋流通股數約為 1.72 億股。我們的底層投資優先順序，以及我們在成長與獲利之間的平衡方式，沒有改變。

I'd also like to provide a few additional points for modeling purposes. First, we continue to expect full year gross margins to be between 85% to 87%. We expect approximately $15 million of JiHu-related expenses for the year compared with $13 million last year.

我也想提供幾點額外資訊，供你們建模之用。第一，我們仍預期全年毛利率介於 85% 至 87%。我們預期今年與 JiHu 相關的費用約為 1,500 萬美元，去年為 1,300 萬美元。

Stepping back, this was a genuinely exceptional quarter, and it gives a strong foundation as we begin the transition to Flex. We believe Flex is the right long-term model for our customers and for GitLab, even though the transition will introduce some near-term noise in revenue timing and reported metrics. Our job through that transition is to give you the transparency to see through those timing effects to the performance underneath.

回過頭來看，這確實是一個格外出色的季度，並在我們開始轉向 Flex 之際提供了堅實的基礎。我們相信 Flex 對客戶與 GitLab 而言都是正確的長期模式，儘管這項轉換在短期內會在收入時點與已報導指標上帶來一些雜訊。我們在這段轉換期間的工作，是提供透明度，讓你們能穿透這些時點效應，看見底層的業務表現。

During this quarter, we saw record bookings, accelerating net ARR growth, strong retention, broad-based strength across the business and profitability ahead of plan. This performance is what gives us real confidence in where the business is headed.

在本季度，我們看到創紀錄的預訂額、淨 ARR 成長加速、強勁的留存、業務各面向的廣泛強勢，以及超出計畫的獲利能力。正是這樣的表現，讓我們對公司未來的走向充滿真正的信心。

Thank you for joining us today. I will now turn it over to Nic to open it up for Q&A.

感謝各位今天加入我們。接下來我把時間交給 Nic，開始 Q&A。

(Operator Instructions) Kingsley Crane, Canaccord Genuity.

(接線員指示) Kingsley Crane，Canaccord Genuity。

Congrats on what is truly a monumental quarter. Bill, you've talked about how AI labs are building custom systems around their Git providers because Git wasn't designed for agent scale operations. And since then, you're now rebuilding source code management to better service agents. So I was wondering if you could just expand on that strategy. Is this for all customers? Does this allow you to do more with AI natives or labs that you otherwise may not have?

恭喜你們交出真正堪稱里程碑的一季。Bill，你曾談到 AI 實驗室正在圍繞其 Git 供應商打造客製化系統，因為 Git 並非為代理（agent）規模的運作而設計。而自那之後，你們現在正重新打造原始碼管理，以更好地服務代理。所以我想請你進一步談談這個策略。這是面向所有客戶嗎？這是否讓你們能在 AI 原生客戶或實驗室上做得更多，而這些原本可能做不到？

Thanks, Kingsley. This quarter's performance really demonstrated what we've always believed possible with GitLab. And now we're focused on systematically executing that opportunity to repeat that performance. In terms of the product road map, the demands of AI are really creating tailwinds for GitLab and some really exciting opportunities. As you mentioned, one of those is the next-generation Git product.

謝謝你，Kingsley。本季度的表現確實證明了我們一直相信 GitLab 能做到的事。而現在我們專注於系統性地執行這個機會，以便重複這樣的表現。就產品路線圖而言，AI 的需求確實正在為 GitLab 帶來順風，並創造一些非常令人振奮的機會。如你所提，其中之一就是下一代 Git 產品。

But let me put it into context because we started this year by introducing our first consumption product, Duo Agent Platform. It's been now in market for two quarters and been performing really well. Last month, in August, we also launched Secrets Manager and Dedicated Runners, our second and third new consumption products this year.

不過我先把它放在脈絡中說明，因為我們今年一開始就推出了第一個用量型（consumption）產品：Duo Agent Platform。它上市已兩個季度，表現非常好。上個月，也就是 8 月，我們也推出了 Secrets Manager 與 Dedicated Runners，這是我們今年的第二與第三個新用量型產品。

Adding to that, now we've also got GitLab Orbit, which entered public beta in Q2. That's our Knowledge Graph. It's really the secret sauce of GitLab as we connect code, issues, MRs, pipeline, security findings so that agents and humans get better answers to questions and higher quality Agentic outcomes at lower cost. We highlighted on the call, early beta adoption is really promising. Over 2,200 organizations enabled it, 70% increase in just four weeks and 80% of the queries come from external tools like Claude Code and Codex. They drive more consumption of GitLab platform.

此外，我們也有 GitLab Orbit，已在第二季進入公開 Beta。那是我們的知識圖譜（Knowledge Graph）。它其實是 GitLab 的祕密武器：我們把程式碼、議題（issues）、合併請求（MRs）、管線（pipeline）、安全發現（security findings）串連起來，讓代理與人類都能以更低成本，對問題得到更好的答案，並產出更高品質的代理式（Agentic）成果。我們在電話會議中也提到，早期 Beta 的採用情況非常令人鼓舞。已有超過 2,200 個組織啟用，僅四週就成長 70%，而且 80% 的查詢來自 Claude Code 與 Codex 等外部工具。它們也帶動 GitLab 平台更多的用量。

The other one that you mentioned is the next-generation Git product. As experts in Git, we're one of the leading contributors to that project. We're now in the middle of rearchitecting that Git infrastructure to achieve roughly 100x scale what humans have ever required. That's really important because agents operate at that machine scale level with one engineer or builder invoking dozens and sometimes hundreds of agents to accomplish a particular task.

你提到的另一項是下一代 Git 產品。作為 Git 的專家，我們是該專案的主要貢獻者之一。我們目前正在重新架構 Git 基礎設施，以達到約為人類過往需求 100 倍的規模。這非常重要，因為代理是在機器規模層級運作：一位工程師或建置者可能會呼叫數十個、甚至數百個代理來完成特定任務。

The other product road map item I want to highlight in addition to next-generation Git is our artifact management product, which is in private beta now, but we expect will be coming in public beta this quarter. It extends our platform to allow for our customers to store, version, govern and sign binaries in artifacts so that we can complete the software supply chain. We already have their source code. We already build and test that source code, and we help them deploy it.

除了下一代 Git 之外，我還想強調另一個產品路線圖項目：我們的成品（artifact）管理產品，目前處於私有 Beta，但我們預期本季度會進入公開 Beta。它擴展我們的平台，讓客戶能夠儲存、版本控管、治理並簽署成品中的二進位檔，從而補齊軟體供應鏈。我們已經擁有他們的原始碼。我們已經建置並測試那些原始碼，也協助他們部署。

And with artifact management, we'll also help them store those artifacts so that they can have secure end-to-end software supply chain all within GitLab. That's one of the top concerns of enterprises today is how do they trust their software supply chain in an Agentic era where agents are doing a lot of the work. So DAP is our agentic layer.

而透過成品管理，我們也將協助他們儲存這些成品，讓他們能在 GitLab 內部就完成安全的端到端軟體供應鏈。這也是當今企業最關切的議題之一：在代理式時代，當代理承擔大量工作時，他們要如何信任自己的軟體供應鏈。因此，DAP 是我們的代理式層。

And underneath that, we've got these new consumption-based products, including like you asked, next-generation Git, GitLab Orbit and Artifact Management, all coming into the market this coming year with capabilities that help software engineers and their agents operate at much higher scale. It's a pretty exciting time. And maybe it's a good time for me to also tee up our GitLab Transcend event coming in just about six weeks. We hope you tune in for more information on the road map ahead.

在其下方，我們有這些新的用量計費產品，包括你問到的下一代 Git、GitLab Orbit 與成品管理；它們都將在接下來一年陸續進入市場，提供能力，協助軟體工程師及其代理以更高的規模運作。這是一個相當令人振奮的時刻。也許這也是個好時機，先預告一下我們大約六週後即將舉辦的 GitLab Transcend 活動。希望你能收看，以獲得更多關於未來路線圖的資訊。

Matt Hedberg, RBC.

Matt Hedberg，RBC。

Can you hear me okay?

你們聽得到我嗎？

Yes.

聽得到。

Congrats on the results. Really, really good to see. I guess for Bill or Jessica, a lot of things to think about here. The seat growth is really, really exciting to see. And I think when we think about that plus the optimism around Flex, it feels like there could be some increased durability here. I guess I'm wondering on the seat side of it, obviously, developers are a focus. But could you talk about that non-developer opportunity? I mean I think we all think that that could be a pretty significant driver as well. But could you touch on that piece as well?

恭喜這次的成果。真的、真的很高興看到這樣的表現。我想不論是 Bill 或 Jessica，這裡有很多值得思考的點。席位（seat）成長真的非常令人振奮。而當我們把這點再加上對 Flex 的樂觀情緒時，感覺這裡的成長韌性可能會更強。我想問的是，在席位這一塊，顯然開發者是重點。但你們能否談談非開發者的機會？我的意思是，我想我們都認為那也可能是相當重要的驅動因素。也請你們談談那一塊。

Absolutely. AI is proving to be a real durable secular tailwind for GitLab and really not just the seats, but every component of our growth algorithm. Maybe this is also a good time for a quick plug. New to this quarter, we published an investor letter, which Jessica and I wrote to give you a look into how we think about the growth algorithm for GitLab going forward.

當然可以。AI 正在證明它對 GitLab 是一股真正具韌性的長期結構性順風，而且不只是席位，還包括我們成長演算法的每一個組成部分。也許這也是個適合快速補充一下的時機。從本季度開始，我們發布了一封投資人信，由 Jessica 和我共同撰寫，讓各位了解我們如何看待 GitLab 未來的成長演算法。

I'll give you a quick summary and cover as part of that, your question around the opportunity for seats. So the growth algorithm really touches on three components. First, more customers. And within more customers, more builders than ever. We believe GitLab's opportunity here is enormous. We're underpenetrated in the global market, and we've been building a stronger competitive position as evidenced by our first -- by this quarter's results. We added more than 1,700 first orders this quarter. That's more than double a year ago, and first order net ARR is up nearly 40%.

我先快速摘要一下，並在其中回應你關於席位機會的問題。我們的成長演算法主要涵蓋三個組成部分。第一，是更多客戶。而在更多客戶之中，是比以往更多的建置者（builders）。我們相信 GitLab 在這裡的機會非常巨大。我們在全球市場的滲透率仍偏低，而我們也在建立更強的競爭地位，這從本季度的成果就能看出。本季度我們新增了超過 1,700 筆首次訂單。這比一年前多出一倍以上，而首次訂單淨 ARR 也成長了將近 40%。

One of the things that's also inside of that more new customers number is really an increased demand for seats. AI has significantly enabled anyone to become a builder. I'm guessing many people on this call have experimented with AI tools and been able to create dashboards or applications with code entirely written by the agent.

而在「更多新客戶」這個數字裡，還包含一個重點：席位需求的提升。AI 大幅讓任何人都能成為建置者。我猜在座許多人都試過 AI 工具，並能建立儀表板或應用程式，而且程式碼完全由代理撰寫。

Well, that code has to be stored somewhere. It has to be governed by the organization if it belongs to an enterprise. It has to meet those organizational compliance and security standards, and that's what GitLab is best at. So AI is creating a secular tailwind for more customers and more builders within every one of our customer accounts.

而那些程式碼必須存放在某個地方。如果它屬於企業，就必須由組織進行治理。它必須符合組織的合規與安全標準，而這正是 GitLab 最擅長的。因此，AI 正在為「更多客戶」以及我們每個客戶帳戶內「更多建置者」帶來結構性順風。

Second part of our growth algorithm is around more products. And you've heard me talk in response to Kingsley's question about some of the exciting product road maps ahead. But this is an important part of our growth strategy because we're expanding beyond just the platform seats capabilities that we've had that led to a $1 billion run rate business that's still growing strong. The proof that this part of our growth algorithm is working is also evidenced in this quarter. We've shared, for example, that secure repositories are growing at 60%, code pushes 50% and CI/CD pipelines at over 40% year-over-year. That's our core platform in action.

我們成長演算法的第二部分，是更多產品。你也聽到我在回應 Kingsley 的問題時，談到一些令人振奮的產品路線圖。但這是我們成長策略的重要一環，因為我們正在擴展到不僅僅是平台席位能力；而這些席位能力曾帶動我們成為年化營收運行率達 10 億美元、且仍強勁成長的業務。本季度也顯示這部分的成長演算法正在奏效。例如，我們分享過：安全儲存庫（secure repositories）成長 60%、程式碼推送（code pushes）成長 50%，以及 CI/CD 管線年增超過 40%。這就是我們核心平台發揮作用的證明。

Also exciting beyond just the core platform is the introduction of our consumption model and credits. Evidence of this in the quarter, we can highlight is our Duo Agent Platform growth, where sequentially, we saw a 50% increase.

除了核心平台之外，同樣令人振奮的是我們推出了消費模式與點數（credits）。本季的佐證之一是我們的 Duo Agent Platform 成長：按季（sequentially）來看，我們看到成長了 50%。

Finally, the third part of our growth algorithm is all around unlocking consumption for every customer. That's where Flex comes in because it unlocks friction-free usage across both seats and credits with a single commitment that allows customers to choose every month how they want to consume the platform to solve their specific problems.

最後，我們成長演算法的第三個部分，核心在於為每一位客戶解鎖消費。這就是 Flex 的用武之地，因為它透過單一承諾，讓客戶能在席位（seats）與點數之間進行無摩擦的使用；客戶每個月都可以選擇要如何消費平台，以解決其特定問題。

Jessica highlighted in our last six weeks of Q2, more than 130 customers opted into Flex. They committed more than $20 million and our paid CRR went up, thanks to Flex from $15 million exiting Q1 to over $40 million exiting Q2. So stepping back, AI is dramatically lowering the bar to build software. Anyone can become a builder. Customers are needing more of GitLab than ever, and we've seen that pattern now for several quarters in a row. Thanks for the question.

Jessica 提到，在 Q2 的最後六週，超過 130 位客戶選擇加入 Flex。他們承諾超過 2,000 萬美元，而我們的付費 CRR 也因 Flex 而上升，從 Q1 結束時的 1,500 萬美元提升到 Q2 結束時超過 4,000 萬美元。退一步看，AI 正在大幅降低打造軟體的門檻。任何人都可以成為建置者（builder）。客戶對 GitLab 的需求比以往任何時候都更高，而這種趨勢我們已連續好幾季都看到了。謝謝你的提問。

Sanjit Singh, Morgan Stanley.

Morgan Stanley 的 Sanjit Singh。

Congrats on an excellent quarter. It's great to see. When I think about how the business model, the pricing model evolves, you guys have a lot of things that you're working on. Bigger picture, Bill, when we think about ultimate versus premium, is that still a relevant construct? I mean you guys have -- you're growing seats. It sounds like the non-developer population could be a secular driver for that. So does the model potentially just simplify to seats and then you have these other monetization vectors with these product add-ons, and I'm sure there's going to be a lot more of that over the next couple of years. So just wanted to think kind of bigger picture on how the business model pricing model progresses from here.

恭喜你們交出非常出色的一季。很高興看到這樣的表現。當我思考你們的商業模式、定價模式如何演進時，你們正在推進很多事情。從更宏觀的角度來看，Bill，當我們比較 Ultimate 與 Premium 時，這樣的區隔仍然重要嗎？我的意思是，你們的席位在成長。聽起來，非開發者族群可能會成為一個長期（secular）的驅動因素。那麼，這個模式是否可能簡化為以席位為主，然後再透過這些產品加購（add-ons）帶來其他變現途徑？我也相信未來幾年會有更多這類做法。所以我想從更大格局來理解，從現在開始你們的商業模式與定價模式會如何推進。

Yes. Thanks, Sanjit. Seats absolutely core to our business, continue to grow healthy, and we're excited about continuing to offer those premium and ultimate capabilities that have led to a $1 billion run rate business continuing to grow strong.

是的。謝謝你，Sanjit。席位絕對是我們業務的核心，並且持續健康成長；我們也很期待持續提供 Premium 與 Ultimate 的能力，這些能力已推動我們達到 10 億美元年化營收（run rate）的業務規模，並且仍在強勁成長。

But to up-level the question a little bit, I think the way that investors should think about our growth going forward is to understand the customer benefits and the GitLab benefits of Flex. So let me unpack the GitLab strategy here because that's core to understanding our business model going forward.

不過把問題提升一個層次，我認為投資人應該從客戶端的效益，以及 Flex 對 GitLab 的效益，來看待我們未來的成長。所以我來拆解一下 GitLab 的策略，因為這是理解我們未來商業模式的關鍵。

We introduced Flex in Q2, and I already highlighted some of the early wins, 130 customers, more than $20 million committed. And you might ask yourself like, why are we introducing that? What is driving customers to adopt at such a rapid early rate? The benefits are pretty clear. For customers, Flex represents more value for every dollar because every month, they get a chance to redirect unused capacity from shelfware towards whatever products that create the most value for them.

我們在 Q2 推出了 Flex，我也已經提到一些早期成果：130 位客戶、承諾超過 2,000 萬美元。你可能會問：我們為什麼要推出它？是什麼驅動客戶在早期就以這麼快的速度採用？其好處相當明確。對客戶而言，Flex 代表每一美元能獲得更高價值，因為每個月他們都有機會把未使用的容量，從「擱置未用的授權」（shelfware）重新導向到對他們最有價值的產品上。

So for example, with the emerging need to give GitLab to non-engineering users, they can flex capacity on to more premium or ultimate seats. In addition, those users may have a need for a Duo Agent Platform to provide agents that assist their work across the software life cycle, they can allocate a reservation on Duo Agent Platform.

例如，隨著把 GitLab 提供給非工程使用者的需求浮現，他們可以把容量彈性調整到更多 Premium 或 Ultimate 席位上。此外，這些使用者也可能需要 Duo Agent Platform，提供能在軟體生命週期中協助其工作的代理（agents）；他們就可以把預留（reservation）配置到 Duo Agent Platform。

Second, it also gives them the ability to adopt new products without going through a new contract cycle, without securing more new budget from procurement. This is one of the major benefits of the Flex model because right inside the Flex portal, new products show up, they can allocate reservations against them, and they can even set budgets so they have control over where they spend their commitment.

第二，它也讓客戶能在不進入新的合約週期、也不必向採購端爭取新增預算的情況下，採用新產品。這是 Flex 模式的一大優勢：在 Flex 入口網站（portal）內，新產品會直接出現，客戶可以對其配置預留，甚至可以設定預算，從而掌控其承諾金額的支出流向。

And finally, it also allows customers to seamlessly expand usage even beyond their commitment. So if they see the demand, they can provision more seats and more credits as needed going beyond their commitment, and those are then billed each month.

最後，它也讓客戶能在其承諾額度之外，無縫擴大使用量。因此一旦看到需求，他們就能視需要配置更多席位與更多點數，超出承諾的部分則按月計費。

For GitLab then, the benefits are pretty clear. First, we believe Flex will help us drive an already healthy profile around retention even better. It allows for us to capture any unused capacity into a Flex commitment and then allow customers to redirect that as needed increasing our retention that's already, as Jessica mentioned, gross retention already above 90%. This also stands to help us with improving our already improving productive sales capacity.

對 GitLab 而言，其好處也很清楚。第一，我們相信 Flex 能讓我們在原本就健康的留存表現上更上一層樓。它讓我們能把任何未使用的容量納入 Flex 承諾中，並讓客戶按需重新導向，進一步提升留存；如 Jessica 所提，我們的總留存率（gross retention）已高於 90%。這也有助於提升我們本來就正在改善的銷售產能生產力。

I mentioned that we've increased capacity 30% this quarter and a 10% increase in per rep productivity. Well, now reps can spend more time helping customers realize value and less time negotiating contracts and true-ups and working through procurement.

我提到過，本季我們把產能提高了 30%，且每位業務代表（per rep）生產力提升了 10%。而現在，業務代表可以把更多時間用在協助客戶實現價值，減少花在合約談判、補差額（true-ups）以及與採購流程周旋上的時間。

And finally, this also represents a new growth tailwind since customers can unlock usage on demand and we bill them immediately as that usage is incurred, giving us another monetization tailwind. So the strategy around the business model is pretty clear. We want to see customers shifting to Flex because it gives them more value, it reduces sales and contracting friction and allows for on-demand consumption.

最後，這也代表一個新的成長順風（tailwind），因為客戶可以按需解鎖使用量，而我們在使用發生時即可立即按月向其計費，為我們帶來另一個變現的順風。因此，圍繞商業模式的策略相當清楚。我們希望看到客戶轉向 Flex，因為它能帶來更高價值、降低銷售與簽約摩擦，並支援按需消費。

We started the year with 100% seat-based subscription business. And now we're focusing on ending the year, continuing to grow that core platform, but also want to focus on exiting the year with a meaningful consumption business scaling toward our second billion as we monetize the work that humans and agents are doing together.

我們在年初時，業務 100% 都是以席位為基礎的訂閱模式。而現在我們的重點是：在年底前持續成長這個核心平台，同時也希望在年底時，能有一個具規模且有意義的消費型業務，朝向我們的第二個 10 億美元邁進，並在「人類與代理共同工作」的過程中實現變現。

Ethan Weeks, Piper Sandler.

Piper Sandler 的 Ethan Weeks。

This is Ethan filling in for Rob tonight. Bill, I wanted to ask just how much of the strength in the quarter do you think was driven by internal execution improving versus kind of the overall demand environment just accelerating as all organizations really start to think about the tool chain that they're enabling their developers with?

今晚由我 Ethan 代替 Rob。Bill，我想問的是：你認為本季的強勁表現，有多少是來自內部執行力的提升，相較於整體需求環境的加速——因為各組織都開始思考要用哪些工具鏈來賦能其開發者？

Yes. Great question. As we mentioned, a record gross bookings quarter in company history. Net ARR growing more than 40% year-over-year. It's our second highest growth quarter in the last four years to give you a sense of that. And I think it's really a result of many quarters of investment and some tailwinds kicking in across multiple dimensions.

是的。好問題。如我們提到的，這是公司史上總訂單（gross bookings）創紀錄的一季。淨 ARR 年增超過 40%。為了讓你有個概念，這是我們過去四年中成長第二高的一季。我認為這確實是多季投資累積的結果，並且有一些順風在多個面向開始發酵。

As Jessica mentioned, PubSec rebounded meaningfully. This is a really strong signal after a few quarters of less-than-ideal results given government shutdowns and other things. We also see AI tailwinds, though, helping drive large deals. Jessica mentioned our $500,000 and above deals grew more than 150% year-over-year. Our largest customers deepening their commitment with GitLab. Those commitments span both seats and credits for our new consumption products.

如 Jessica 所說，公部門（PubSec）明顯回升。在經歷幾季因政府停擺等因素而導致表現不理想之後，這是一個非常強的訊號。我們也看到 AI 順風正在幫助推動大型交易。Jessica 提到，我們 50 萬美元以上的交易年增超過 150%。我們最大的客戶也在加深對 GitLab 的承諾。這些承諾同時涵蓋席位與我們新消費型產品的點數。

Underneath all of this, we really do see AI beginning to be a durable tailwind for GitLab. And it really impacts all three levers of our growth algorithm. It's leading to more customers. It's leading to the new products getting early adoption as we highlighted with GitLab Orbit and it's leading to more consumption. You should think of net ARR as that output metric for the growth algorithm because it takes into account all three of those dimensions compounding one another.

在這一切之下，我們確實看到 AI 正開始成為 GitLab 一個可持續的順風。而且它實際上影響我們成長演算法的三個槓桿。它帶來更多客戶。它也推動新產品的早期採用——正如我們在 GitLab Orbit 上所強調的——並且帶來更多消費。你可以把淨 ARR 視為成長演算法的輸出指標，因為它把這三個面向彼此疊加、相互增強的效果都納入考量。

Koji Ikeda, Bank of America.

Bank of America 的 Koji Ikeda。

Really nice job here with the commentary and the growth bookings. And so with a lot of the focus on Flex over the next, call it, a couple of years, there's a lot of moving pieces in the financial model with Flex. And so what would be the single best metric to gauge the performance of Flex?

你們在評論與訂單成長方面做得非常好。因此，隨著未來一兩年（大概這麼說）會有很多焦點放在 Flex 上，Flex 會讓財務模型裡出現許多變動因素。那麼，用來衡量 Flex 表現的單一最佳指標會是什麼？

And just one quick, I guess, housekeeping question on the guide, would the guide have been raised $13 million more if there was no Flex this year? Just around the question of is the guide actually understating what's happening in the business because of what's happening in the shifting with Flex?

另外再快速問一個，我想算是關於指引的例行問題：如果今年沒有 Flex，指引是不是會再提高 1,300 萬美元？主要是想問，因為 Flex 帶來的轉換移轉，指引是否其實低估了業務實際發生的狀況？

Thanks, Koji. Let me just step back and give you some thoughts about like how we thought about guidance this quarter. So first, again, Flex has only been in the market for 10 weeks. So it's just too new to bake into our guidance with precision. At the same time, we've been really pleasantly surprised by customer demand and how potentially impactful Flex could be in the back half of the year. So as we thought about our guide, we didn't want the accounting mechanics to distort the guidance conversation or investors' views of the underlying health of the business. So we've really applied I'd say two key principles to our approach, consistency and transparency.

謝謝你，Koji。我先退一步，分享一下我們這一季是如何思考指引的。首先，再次強調，Flex 上市才 10 週。所以它還太新，無法精準地納入我們的指引。同時，我們對客戶需求感到非常驚喜，也看到 Flex 在今年下半年可能帶來的重大影響。因此在制定指引時，我們不希望會計處理機制扭曲指引的討論，或投資人對業務基本健康狀況的看法。所以我們在方法上，我會說採用了兩個關鍵原則：一致性與透明度。

So first on consistency. We wanted to guide the same way all year. So we started FY27 guiding without the accounting timing impact of Flex, and we've stayed consistent with that methodology rather than changing course midyear. Second, on transparency. In line with our commitment to transparency, we are providing a couple of key data points to help you think about your models.

先談一致性。我們希望全年都用同一種方式提供指引。因此我們在 FY27 一開始提供指引時，就排除了 Flex 的會計認列時點影響，並且一直維持這個方法，而不是在年中改變作法。第二，透明度。依循我們對透明度的承諾，我們提供幾個關鍵數據點，協助你們思考模型。

First, the heuristic, which I mentioned in my prepared remarks. So for every $50 million of self-managed customers that convert to Flex, there is about a $5 million revenue recognition timing impact that shifts out of FY27 into the future period and then the maximum FY27 impact of $13 million.

第一個是經驗法則（heuristic），我在事先準備的發言中提到過。也就是：每有 5,000 萬美元的自管（self-managed）客戶轉換到 Flex，大約會有 500 萬美元的營收認列時點影響，從 FY27 延後到未來期間；而 FY27 的最大影響為 1,300 萬美元。

So going forward, we will continue to explicitly and transparently quantify the flex accounting timing impact each quarter so you all can separate the accounting effect from the underlying business performance. We are committed.

展望未來，我們會在每一季持續明確且透明地量化 Flex 的會計認列時點影響，讓各位能把會計效果與基本業務表現區分開來。我們會持續承諾。

And I think just in terms of that $13 million, the other thing I'd highlight is we've got a lot of visibility into our business. So in calculating that amount for us, we feel confident in the precision because, one, we understand our self-managed available-to-renew for the back half of the year, and we've also got very clear line of sight on pipeline.

另外我想就那 1,300 萬美元補充一點：我們對自身業務有很高的可視性。因此在計算這個金額時，我們對精準度很有信心，因為第一，我們了解下半年自管客戶可供續約（available-to-renew）的情況；第二，我們對銷售管線也有非常清楚的掌握。

And Koji, in terms of the other part of your question around what metric to look at, that paid CRR metric that we introduced last quarter and updated this quarter with more than $40 million of paid CRR is really the best way to understand how dollars are flowing through our consumption model. Paid CRR captures both Flex commitments, credit commitments as well as on-demand paid usage. So all three of those inside of that metric. It does not take into account promotional usage or trials or other unpaid usage. And as we mentioned on the prepared remarks, we are setting a bold target to achieve more than $100 million paid CRR by the end of this fiscal year.

Koji，至於你問題的另一部分，也就是該看哪個指標：我們上季推出、這季更新並顯示已超過 4,000 萬美元的 paid CRR 指標，確實是理解資金如何在我們的消費（consumption）模型中流動的最佳方式。Paid CRR 同時涵蓋 Flex 承諾、點數（credit）承諾，以及隨用隨付（on-demand）的付費使用量。所以這三者都包含在該指標內。它不包含促銷使用、試用或其他未付費使用。而且如同我們在事先準備的發言中提到的，我們設定了一個大膽目標：在本會計年度結束前，達成超過 1 億美元的 paid CRR。

Radi Sultan, UBS.

Radi Sultan，UBS。

Awesome. Now that you have some data points on conversion to Flex, I mean, in sizing those initial Flex deals you convert customers, like is there a higher propensity to upsize their deals because of the increased flexibility and sort of new product consumption? I'm trying to understand how much is just a form factor shift to Flex versus baking in anticipated new product consumption or seat expansion would be great.

太棒了。現在你們已經有一些轉換到 Flex 的數據點，我想問在為這些初始 Flex 交易定規模、把客戶轉過來時：因為彈性提高以及新的產品消費模式，客戶是否更傾向把合約規模做大（upsize）？我想了解的是，有多少只是形式上轉成 Flex， versus 把預期的新產品消費或席位擴張也一起納入，若能分享會很有幫助。

Yes. No, again, we're really excited about it. It's been about 10 weeks in market, so it's too early for us to tell. But I just -- I think really leaning into the strategy, Bill, is there anything you want to add there from that lens?

是的。不過再次強調，我們對此非常興奮。Flex 上市大約 10 週，所以現在要下結論還太早。但我想——從策略角度來看，Bill，你是否想從那個面向補充一些？

Yes. From a sales strategy perspective, it's an opportunity to do multiple things. First, some customers are early adopters and already testing our beta products. And Flex gives us a way not only to renew their seat-based subscription but shift that commitment into a Flex agreement where they can create headroom to start using those products as soon as they become generally available. And wherever we can do that, create headroom for customers and their commitment, obviously, we want to take advantage of that.

是的。從銷售策略的角度，這是一個可以同時做多件事的機會。第一，有些客戶是早期採用者，已經在測試我們的 beta 產品。Flex 讓我們不僅能續約他們以席位為基礎的訂閱，還能把該承諾轉到 Flex 協議中，讓他們預留空間（headroom），一旦這些產品正式全面上市（generally available）就能立刻開始使用。只要我們能做到、為客戶及其承諾創造 headroom，我們當然希望把握這個機會。

It also lets us look at customers who may have unused capacity because they were previously forced to commit to a seat forecast for the full year and rather than take some contraction on the account, keep their commitment the same or even higher and redeploy that budget onto new consumption-based products. So it really helps us on both the upside as well as reducing or mitigating potential downside, which we're really excited about deploying anywhere we can align with customers on the value of their GitLab commitment.

它也讓我們能檢視那些可能有未使用容量的客戶：過去他們被迫依全年席位預測做承諾；現在我們可以不必在帳戶上出現收縮（contraction），而是維持相同甚至更高的承諾，並把那筆預算重新部署到新的、以用量為基礎的產品上。因此它同時有助於上行成長，也能降低或緩解潛在下行風險；只要我們能與客戶在其 GitLab 承諾價值上達成一致，我們就非常期待在各處推動。

Derrick Wood, TD Cowen.

Derrick Wood，TD Cowen。

Congrats from me. Bill, could you touch on just what you're seeing across the competitive landscape when it comes to both new AI natives coming on to the scene as well as what's going on with your traditional competitors? And where you feel that you're competing better or gaining share right now?

我這邊也恭喜你們。Bill，你能談談你們在競爭格局上看到的情況嗎？包括新的 AI 原生（AI natives）競爭者進場，以及你們傳統競爭對手的動態。以及你認為目前你們在哪些方面競爭得更好或正在取得市占？

Yes. Great question. Competitive dynamics have been really consistent in this quarter. We still have one primary competitor. Our position versus that competitor is, I think, stronger than it's ever been. They've struggled with reliability and meeting the needs of customers in this new Agentic era with security and trust and other dimensions. And so we're seeing higher win rates. We're seeing both wins on the first order side as well as the expansion side as reflected in our numbers, 100% growth in first orders and more than 40% net ARR expansion.

是的。好問題。這一季的競爭態勢其實相當一致。我們仍然只有一個主要競爭對手。我認為我們相對於該競爭對手的定位，比以往任何時候都更強。他們在可靠性方面遇到困難，也難以在這個新的 Agentic 時代於安全、信任及其他面向滿足客戶需求。因此我們看到更高的勝率。我們在首單（first order）以及擴張（expansion）兩方面都在贏，這也反映在我們的數字上：首單成長 100%，以及淨 ARR 擴張超過 40%。

In terms of AI natives, what I would say there is really, this is an opportunity. They're creating tailwinds for us. They've dramatically simplified the ability for anyone to create code and all of that code needs GitLab. And so we're now seeing, as I've mentioned, our customers coming to us and saying they need more GitLab seats because of non-engineering people who need access to the platform. And those same users who are now needing seats will ultimately also need credits in order to use our consumption-based products. So we're happy to partner with any Agentic coding tools out there because we see it as creating tailwinds for the business that we have.

至於 AI 原生競爭者，我會說這其實是一個機會。他們正在為我們創造順風（tailwinds）。他們大幅簡化了任何人產生程式碼的能力，而所有那些程式碼都需要 GitLab。因此我們現在看到，如我提到的，客戶來找我們說他們需要更多 GitLab 席位，因為有非工程人員也需要存取平台。而這些現在需要席位的使用者，最終也會需要點數（credits）來使用我們以用量為基礎的產品。所以我們很樂意與任何 Agentic 程式撰寫工具合作，因為我們認為這會為我們的業務帶來順風。

Nick Altmann, US Bancorp.

Nick Altmann，US Bancorp。

Awesome. I wanted to follow up on Radi's earlier question. How much of the interest in Flex or the $100 million CRR target is being driven by some of the newer products like Duo Agent Platform, Orbit, Secrets Manager and the Dedicated Hosted Runners? Because I think you mentioned that not all of it will be incremental to revenue, but presumably some of those newer products should be incremental. So just any color on how much those newer products are driving some of the interest in Flex or that $100 million CRR target would be super helpful.

太好了。我想延續 Radi 先前的問題。對 Flex 的興趣或 1 億美元 CRR 目標，有多少是由一些較新的產品所驅動，例如 Duo Agent Platform、Orbit、Secrets Manager，以及 Dedicated Hosted Runners？因為我記得你們提到並非全部都會對營收帶來增量，但推測其中一些新產品應該會帶來增量。所以若能分享這些新產品在推動 Flex 興趣或 1 億美元 CRR 目標上大概占多少比重，會非常有幫助。

Just on the financial piece of it, it's really too early to tell, but I'll let Bill speak to a little bit more in terms of the interest that we're seeing so far.

就財務面的部分來說，現在真的還太早下定論；不過我先讓 Bill 多談一些我們目前看到的興趣狀況。

Yes, we only had six weeks in Q2, so the 130 customers and $20 million is what we've already captured. But I think it's important to step back and understand the strategy because even when customers are shifting to Flex and drawing down on that commitment with seats, it changes the nature of the relationship. And it's an important part of the strategy to understand because the needs today of software engineering teams are changing dramatically every week, every month, every quarter. New tools emerge, new techniques emerge, new models come into market and customers want the flexibility to allocate their spend as their needs evolve.

是的，我們在第二季只有六週，所以這 130 位客戶與 2,000 萬美元是我們已經拿到的成果。但我認為重要的是退一步理解策略，因為即使客戶正在轉向 Flex，並以席位來消耗該承諾，這也會改變關係的本質。而這也是策略中很重要的一部分需要理解，因為如今軟體工程團隊的需求每週、每月、每季都在劇烈變化。新工具出現、新技術出現、新模型進入市場，而客戶希望能隨著需求演進，彈性地分配其支出。

If the current seat-based subscription model was meeting their needs, there would be no need for Flex. There'd be no demand for it. We wouldn't be seeing the rapid uptake and adoption that we've already seen. So clearly, customers see the benefit of it, and they may start to fund their Flex agreement with their previous seat-based subscription dollars, but they see the value proposition and having the flexibility to adapt and evolve that as our new products come into market.

如果目前以席位為基礎的訂閱模式能滿足他們的需求，就不會需要 Flex。也不會有對它的需求。我們也不會看到目前已經出現的快速導入與採用。因此很明顯，客戶看到了它的好處；他們可能會用先前以席位訂閱的預算來為 Flex 協議提供資金，但他們看重的是其價值主張，以及能夠隨著我們新產品上市而具備調整與演進的彈性。

And certainly, some of that is driven by our current consumption product, Duo Agent Platform that launched in January and then the new emerging ones that just launched in August, the Secrets Manager and the Dedicated Hosted Runners. But we have meaningful usage of our beta products already going on. And I think customers are also looking ahead and appreciating the flexibility that we're extending to them via the Flex agreement.

當然，其中一部分是由我們目前的用量型產品所帶動，也就是 1 月推出的 Duo Agent Platform，以及 8 月剛推出的新興產品：Secrets Manager 與 Dedicated Hosted Runners。但我們的 beta 產品也已經有相當可觀的使用量在進行中。我認為客戶也在向前看，並且欣賞我們透過 Flex 協議提供給他們的彈性。

Shrenik Kothari, Baird.

Shrenik Kothari，Baird。

This is Zach on for Srini. I'll echo our congrats on the strong quarter. So my question on gross margins. It's obviously come down a bit as the business has begun supporting more AI and usage-based workloads, while credits still increasingly monetize both the human and nonhuman activity. So I guess the question is just how much of the recent gross margin movement is actually related to AI inference and infrastructure versus other factors?

我是 Zach，代替 Srini 發問。我也要呼應一下，恭喜你們這一季表現強勁。我想問的是毛利率。隨著業務開始支援更多 AI 與用量型工作負載，而點數（credits）也愈來愈多地同時對人類與非人類活動變現，毛利率顯然下滑了一些。所以我的問題是，近期毛利率的變動有多少其實與 AI 推論與基礎設施有關，相較於其他因素？

And then as agent usage scales, what gives you guys the confidence that pricing, model efficiency, orbit and just your ability to route workloads across models can drive attractive incremental economics maybe rather than creating a structurally lower margin revenue mix?

另外，隨著代理（agent）使用量擴大，你們憑什麼有信心：定價、模型效率、orbit，以及你們將工作負載在不同模型間路由的能力，能帶來具吸引力的增量經濟效益，而不是造成結構性更低毛利的營收組合？

Yes. No, thanks for the question. And just I think in terms of gross margins for the quarter, there's not really a big shift from what we've said previously. So our SaaS dedicated contribution is about 34%. That grew 36% year-over-year. And so that's really in line with what we've been messaging at the time of the IPO, we were at 22%. So we expect those margins to come down over time.

是的。不，謝謝你的問題。就本季毛利率而言，基本上沒有出現我們先前所說之外的重大變化。因此，我們的 SaaS 專屬貢獻（dedicated contribution）約為 34%。年增 36%。這與我們在 IPO 時所傳達的訊息一致，當時是 22%。所以我們預期這些毛利率會隨時間下降。

And then as we think about going forward, again, this is a year of investment and execution. We have intentionally invested in consumption products and pushing customers to focus on transitioning pilots to production. And so that's what you're seeing there.

接著展望未來，同樣地，今年是投資與執行的一年。我們刻意投資於用量型產品，並推動客戶把重點放在將試點（pilots）轉為正式上線（production）。所以你看到的就是這些因素的反映。

But Bill, I don't know if you have anything to add there.

不過 Bill，我不確定你是否要補充？

Yes. On the -- you sort of alluded to tokens and AI optimization potential within how GitLab looks at AI-native products. So let me talk just a little bit about that because I think we're in the very early days, and we actually have several structural advantages versus maybe typical AI native tools whose almost entire monetization is based on tokens.

是的。關於——你剛才提到 tokens 與 AI 最佳化的潛力，這涉及 GitLab 如何看待 AI 原生產品。我稍微談一下，因為我認為我們仍在非常早期的階段，而且相較於一些典型的 AI 原生工具（其幾乎全部的變現都建立在 tokens 上），我們其實有幾個結構性優勢。

One of the beautiful things about Duo Agent Platform is that we provide it in a cloud-agnostic model-agnostic way, meaning we support all of the models, including open-source open-weight models, and we allow customers to deploy it and use it in any cloud, including, for example, air-gapped data center environments that they run.

Duo Agent Platform 很棒的一點是，我們以雲端無關（cloud-agnostic）、模型無關（model-agnostic）的方式提供，也就是說我們支援所有模型，包括開源、開放權重（open-weight）模型，並允許客戶在任何雲端部署與使用它；例如也包括他們自行運行的隔離式（air-gapped）資料中心環境。

And what that means is for many of our customers, the token or inference cost is actually not embedded in the GitLab agreement. They pay us for the access to the platform, and they pay for the work done in the platform, the context, the harness, the governance and auditability that we provide, not the inference. Those are all very high-margin products.

這代表對我們許多客戶而言，token 或推論成本其實不包含在 GitLab 合約中。他們付費給我們是為了取得平台存取權，並為平台內完成的工作、情境（context）、框架（harness）、以及我們提供的治理與可稽核性付費，而不是為推論付費。這些都是非常高毛利的產品。

In addition to that, we're in the very early days of AI adoption. There's very simple things on the road map that we and others are approaching over time, which would allow for dynamic model routing, model optimization to deliver high-quality agentic outcomes in a more cost-efficient way. Those are also things that we want to provide to give customers the very best experience and price to value performance.

此外，我們仍處於 AI 採用的非常早期階段。在路線圖上，有一些我們與其他業者會隨時間推進的相對簡單事項，例如動態模型路由、模型最佳化，以更具成本效率的方式交付高品質的代理式（agentic）成果。這些也是我們希望提供的能力，讓客戶獲得最佳體驗，以及最佳的「價格對價值」表現。

So I encourage you to think about gross margins maybe a little bit differently than maybe typical AI natives. And as Jessica alluded to, a lot of the margin changes that we've seen in the business have been driven more by the mix shift to SaaS than the early AI adoption.

因此，我鼓勵你用與典型 AI 原生公司稍微不同的方式來思考毛利率。而如 Jessica 所提到，我們在業務中看到的許多毛利率變化，更多是由於組合轉向 SaaS 所驅動，而非早期的 AI 採用。

Lucky Schreiner, D.A. Davidson.

Lucky Schreiner，D.A. Davidson。

Okay. Great. This is Lucky with D.A. Davidson. Maybe just on the customer story who expanded their monthly commitments by 10x, was that primarily driven by an increase in usage and the [are] they got? And how much of that was maybe from vendor consolidation and replacement of other tools? And how applicable do you view that kind of expansion motion across your customer base?

好的。很好。我是 D.A. Davidson 的 Lucky。想問一下那位將每月承諾金額擴大 10 倍的客戶案例，這主要是由使用量增加以及他們獲得的 [are] 所驅動嗎？另外其中有多少可能來自供應商整併、替換其他工具？你們認為這種擴張動能在你們客戶群中有多大的適用性？

Yes. It's a top US bank, and they previously had started with a fairly modest commitment to Duo Agent Platform. And once they started rolling it out across their engineers and seeing the value that it provides, saw the need to increase their commitment 10x this quarter, which we would love to replicate across every single customer. We believe that the value of Duo Agent Platform is there, and it's complementary to other AI native tools that are more focused on the coding aspects of software.

是的。那是一家美國大型銀行，他們先前以相當適度的承諾額度開始使用 Duo Agent Platform。當他們開始在工程師之間推廣並看到它帶來的價值後，本季就看到了需要把承諾提高 10 倍——我們也很希望能在每一位客戶身上複製這種情況。我們相信 Duo Agent Platform 的價值確實存在，而且它能與其他更聚焦於軟體程式撰寫面向的 AI 原生工具互補。

So yes, we're really excited by that lighthouse customer. We've got several now that are spending multiples in excess of their premium or ultimate seat price now in terms of credits. And it's a pattern that we're trying to learn and optimize for and repeat everywhere we can.

所以是的，我們對這個指標性（lighthouse）客戶感到非常振奮。目前我們已有好幾個客戶，在點數（credits）方面的支出，已經是其 Premium 或 Ultimate 席位價格的數倍。而這是一種我們正在學習、最佳化並盡可能在各處複製的模式。

All right. That's the end of our call. Thank you, everyone, for joining our Q2 call this quarter. We're really excited by the results we're able to deliver this quarter, and thank you for joining us. We'll see you in the investor conferences to come.

好。本次電話會議到此結束。感謝各位參加我們本季的第二季電話會議。我們對本季能交付的成果感到非常振奮，也謝謝各位的參與。我們將在接下來的投資人會議上再見。