Snowflake Inc (SNOW) 2024 Q2 法說會逐字稿

Good afternoon, and thank you for joining the Snowflake Q2 Fiscal Year 2024 Earnings Conference Call. My name is Kate, and I will be the moderator for today's call. (Operator Instructions)

下午好，感謝您參加 Snowflake 2024 財年第二季度收益電話會議。我叫凱特，我將擔任今天電話會議的主持人。 （操作員說明）

I would now like to pass the call over to your host, Jimmy Sexton, Head of Investor Relations at Snowflake. You may proceed.

現在我想將電話轉給主持人、Snowflake 投資者關係主管吉米·塞克斯頓 (Jimmy Sexton)。您可以繼續。

Good afternoon, and thank you for joining us on Snowflake's Q2 Fiscal 2020 Earnings Call. With me in Bozeman, Montana are Frank Slootman, our Chairman and Chief Executive Officer; Mike Scarpelli, our Chief Financial Officer; and Christian Kleinerman, our Senior Vice President of Product, who will join us for the Q&A...

下午好，感謝您參加 Snowflake 的 2020 年第二季度財報電話會議。和我一起在蒙大拿州博茲曼的是我們的董事長兼首席執行官 Frank Slootman； Mike Scarpelli，我們的首席財務官；以及我們的產品高級副總裁 Christian Kleinerman，他將與我們一起參加問答……

(technical difficulty)

（技術難度）

Excuse me, ladies and gentlemen, it does look like the speakers have disconnected. One moment while we get them reconnected.

對不起，女士們先生們，揚聲器似乎確實已斷開連接。一會兒我們讓他們重新連接。

Ladies and gentlemen, we have the speakers back on the call. You may proceed with the presentation.

女士們、先生們，我們請發言人重新進行通話。您可以繼續演示。

Good afternoon again, and thank you for joining us on Snowflake's Q2 Fiscal 2024 Earnings Call. With me in Bozeman, Montana are Frank Slootman, our Chairman and Chief Executive Officer; Mike Scarpelli, our Chief Financial Officer; and Christian Kleinerman, our Senior Vice President of Product, who will join us for the Q&A session.

再次下午好，感謝您參加 Snowflake 的 2024 年第二季度財報電話會議。和我一起在蒙大拿州博茲曼的是我們的董事長兼首席執行官 Frank Slootman； Mike Scarpelli，我們的首席財務官；我們的產品高級副總裁 Christian Kleinerman 將參加我們的問答環節。

During today's call, we will review our financial results for the second quarter of fiscal 2024 and discuss our guidance for the third quarter and full year fiscal 2024.

在今天的電話會議中，我們將回顧 2024 財年第二季度的財務業績，並討論我們對 2024 財年第三季度和全年的指導。

During today's call, we will make forward-looking statements, including statements related to the expected performance of our business, future financial results, strategy, products and features, long-term growth our stock repurchase program and overall future prospects. These statements are subject to risks and uncertainties, which could cause them to differ materially from actual results. Information concerning these risks is available in our earnings press release distributed after market closed today and in our SEC filings, including our most recent filed for 10-Q for the fiscal quarter ended April 30, 2023, and the Form 10-Q for the quarter ended July 31, 2023, that we will file with the SEC. We caution you to not place undue reliance on forward-looking statements and undertake no duty or obligation to update any forward-looking statements as a result of new information, future events or changes in our expectations.

在今天的電話會議中，我們將做出前瞻性陳述，包括與我們的業務預期業績、未來財務業績、戰略、產品和功能、股票回購計劃的長期增長以及總體未來前景相關的陳述。這些陳述存在風險和不確定性，可能導致它們與實際結果存在重大差異。有關這些風險的信息可在我們今天收盤後發布的收益新聞稿和我們向SEC 提交的文件中找到，包括我們最近提交的截至2023 年4 月30 日的財政季度的10-Q 表格以及該季度的10-Q 表格截至 2023 年 7 月 31 日，我們將向 SEC 備案。我們提醒您不要過度依賴前瞻性陳述，並且不承擔因新信息、未來事件或我們的預期變化而更新任何前瞻性陳述的責任或義務。

We'd also like to point out that on today's call, we will report both GAAP and non-GAAP results. We use these non-GAAP financial measures internally for financial and operational decision-making purposes and as a means to evaluate period-to-period comparisons. Non-GAAP financial measures are presented in addition to and not as a substitute for financial measures calculated in accordance with GAAP. To see the reconciliation of these non-GAAP financial measures, please refer to our earnings press release distributed earlier today in our investor presentation, which are posted at investors.snowflake.com. A replay of today's call will also be posted on the website.

我們還想指出，在今天的電話會議上，我們將報告 GAAP 和非 GAAP 業績。我們在內部使用這些非公認會計原則財務指標用於財務和運營決策目的，並作為評估期間比較的手段。非公認會計原則財務指標是根據公認會計原則計算的財務指標的補充，而不是替代。要查看這些非公認會計準則財務指標的調節表，請參閱今天早些時候在投資者介紹中發布的收益新聞稿，該新聞稿發佈在 Investors.snowflake.com 上。今天電話會議的重播也將發佈在網站上。

With that, I would now like to turn the call over to Frank.

有了這個，我現在想把電話轉給弗蘭克。

Thanks, Jimmy. Welcome and good afternoon. Q2 product revenue grew 37% year-over-year to reach $640 million. Non-GAAP product gross margin expanded to 78%, and non-GAAP adjusted free cash flow was $88 million, representing 50% year-over-year growth. In Q2, we continued to execute in an unsettled macro environment but with incremental improvement in general sentiment and engagement. Generative AI is at the forefront of customer conversations. However, enterprises are also realizing that they cannot have an AI strategy without a data strategy to base it on.

謝謝，吉米。歡迎，下午好。第二季度產品收入同比增長 37%，達到 6.4 億美元。非 GAAP 產品毛利率擴大至 78%，非 GAAP 調整後自由現金流為 8800 萬美元，同比增長 50%。第二季度，我們繼續在不穩定的宏觀環境中執行任務，但總體情緒和參與度逐步改善。生成式人工智能處於客戶對話的最前沿。然而，企業也意識到，如果沒有數據戰略作為基礎，他們就無法制定人工智能戰略。

We have a head start in this race as the epicenter of highly curated, optimized and trusted enterprise data. We now have a presence with 639 Global 2000 customers. AI reaches beyond enterprise boundaries. Models need external data to answer challenging questions. Data sharing makes Snowflake uniquely positioned to enable AI workloads. As of Q2, 26% of Snowflake customers are data sharing, up from 20% in the same period last year. Approximately 70% of customers with more than $1 million in trailing 12-month product revenue are data sharing with an average of 6 stable edges.

作為高度策劃、優化和值得信賴的企業數據的中心，我們在這場競賽中處於領先地位。我們現在擁有 639 家全球 2000 強客戶。人工智能超越了企業邊界。模型需要外部數據來回答具有挑戰性的問題。數據共享使 Snowflake 在支持 AI 工作負載方面具有獨特的優勢。截至第二季度，26% 的 Snowflake 客戶進行數據共享，高於去年同期的 20%。過去 12 個月產品收入超過 100 萬美元的客戶中，大約 70% 的數據共享平均有 6 個穩定邊緣。

For years, we focused on the programmability of our platform via Snowpark. We are seeing momentum. In Q2, we added more than 400 Snowpark customers, and our consumption grew approximately 70% quarter-over-quarter. 63% of our Global 2000 customers are using Snowpark on a weekly basis.

多年來，我們通過 Snowpark 專注於平台的可編程性。我們看到了勢頭。第二季度，我們增加了 400 多家 Snowpark 客戶，我們的消費量環比增長了約 70%。我們的 Global 2000 客戶中有 63% 每週使用 Snowpark。

Document AI is now on private preview. With Document AI, customers can use natural language to ask questions of unstructured data. Legal contracts or invoices are now available for inquiry and analytics. This is an early example of how language models are expanding our opportunity. With Snowflake Container Services, we are bringing LLM models like Reka and NVIDIA's NeMo into Snowflake. You heard in my conversation with Jensen, Snowflake is sitting on a gold mine of data. Together, we can help customers turn that gold mine into intelligence.

Document AI 現已提供私人預覽版。借助 Document AI，客戶可以使用自然語言來詢問非結構化數據的問題。現在可以查詢和分析法律合同或發票。這是語言模型如何擴大我們機會的早期例子。通過 Snowflake 容器服務，我們將 Reka 和 NVIDIA 的 NeMo 等 LLM 模型引入 Snowflake。您在我與 Jensen 的談話中聽說，Snowflake 正坐在數據金礦上。我們可以共同幫助客戶將金礦轉化為智能。

We announced Snowpark Container Services 2 months ago. Since then, hundreds of customers have requested access to the private preview. With our support of Iceberg Tables, we are expanding our data lake scope. Many customers already use Snowflake as a data link. Large financial services customer consolidates data in Snowflake to eliminate useless extracts in transfers of data. This means new use cases are deployed 80% faster. Iceberg Tables will bring additional scope and open file formats to Snowflake. We expect to unlock more data lake opportunities with these capabilities.

我們在 2 個月前推出了 Snowpark 集裝箱服務。從那時起，數百名客戶請求訪問私人預覽版。在 Iceberg Tables 的支持下，我們正在擴大數據湖範圍。許多客戶已經使用 Snowflake 作為數據鏈路。大型金融服務客戶將數據整合到 Snowflake 中，以消除數據傳輸中無用的提取。這意味著新用例的部署速度提高了 80%。 Iceberg Tables 將為 Snowflake 帶來額外的範圍和開放文件格式。我們期望通過這些功能釋放更多數據湖機會。

We've also reached an inflection point on the application front. At Summit, we launched so-called native apps in public preview. We have over 25 native application providers today. Snowflake is a safe, certified and sanctioned place to deploy applications. Grassroot support is building. We now have more than 145,000 monthly active developers on Streamlit. This represents an increase of 160% year-on-year. Our start-up program allocates resources to developers planning to build on Snowflake. Approximately 20% of new customers landed in Q2, landed on Snowflake through our startup program.

我們在應用程序方面也達到了一個拐點。在峰會上，我們在公共預覽版中推出了所謂的本機應用程序。目前，我們擁有超過 25 家本機應用程序提供商。 Snowflake 是一個安全、經過認證和認可的應用程序部署場所。草根支持正在建立。目前 Streamlit 上的月活躍開發者數量已超過 145,000 名。這意味著同比增長160%。我們的啟動計劃為計劃在 Snowflake 上進行開發的開發人員分配資源。大約 20% 的新客戶在第二季度登陸，通過我們的啟動計劃登陸 Snowflake。

General sentiment appears to be incrementally getting better. The Snowflake Summit in June was a highlight of energy and excitement about what is becoming possible in the world of data. We hosted over 20,000 on-site and virtual attendees. This was up over 85% from last year. Next up is our Data Cloud World Tour. The World Tour brings Summit messaging to a wider audience. We expect to double the attendance of Summit. This is in 26 cities worldwide.

總體情緒似乎正在逐漸好轉。六月的雪花峰會是一次充滿活力和興奮的亮點，人們對數據世界的可能性充滿了活力和興奮。我們接待了超過 20,000 名現場和虛擬與會者。這比去年增長了 85% 以上。接下來是我們的數據云世界之旅。世界巡演將峰會信息傳達給更廣泛的受眾。我們預計峰會的出席人數將增加一倍。該項目遍布全球 26 個城市。

With that, I will turn the call over to Mike.

這樣，我會將電話轉給邁克。

Thank you, Frank. Consumption came in line with our expectations for the quarter. In May, we saw a return to growth with strength continuing into June and July. From a booking standpoint, we saw promising signs of stabilization with new bookings outperforming our expectations. However, we believe productivity has room for further improvement. Q2 remaining performance obligations grew 30% year-over-year totaling $3.5 billion. Of the $3.5 billion in RPO, we expect approximately 57% to be recognized as revenue in the next 12 months. This represents a 32% increase compared to our estimate as of the same quarter last year.

謝謝你，弗蘭克。本季度的消費符合我們的預期。 5 月份，我們看到經濟恢復增長，且勢頭持續到 6 月和 7 月。從預訂角度來看，我們看到了有希望的穩定跡象，新預訂量超出了我們的預期。然而，我們認為生產率還有進一步提高的空間。第二季度剩餘履約義務同比增長 30%，總計 35 億美元。在 35 億美元的 RPO 中，我們預計大約 57% 將在未來 12 個月內確認為收入。與我們去年同期的估計相比，這意味著增長了 32%。

Our net revenue retention rate of 142% includes 6 new customers with $1 million in trailing 12-month product revenue. We now have 402 customers with trailing 12-month product revenue greater than $1 million. We continue to focus on growth and efficiency. We generated $88 million in non-GAAP adjusted free cash flow, outperforming our Q2 target. Q2 represented another quarter of continued progress on profitability. Our non-GAAP product gross margin was 77.9%, benefiting from a onetime credit from one of our cloud service providers.

我們的淨收入保留率為 142%，其中包括 6 個新客戶，過去 12 個月的產品收入為 100 萬美元。我們現在有 402 名客戶過去 12 個月的產品收入超過 100 萬美元。我們繼續關注增長和效率。我們產生了 8800 萬美元的非 GAAP 調整後自由現金流，超出了我們第二季度的目標。第二季度是盈利能力持續進步的又一個季度。我們的非 GAAP 產品毛利率為 77.9%，受益於我們的一家云服務提供商的一次性信貸。

Non-GAAP operating margin was 8%, benefiting from tight controls on headcount additions and the overachievement in product gross margin. Our non-GAAP adjusted free cash flow margin was 13%. We continue to have a strong cash position, with $4.9 billion in cash, cash equivalents and short-term and long-term investments. We did not repurchase any shares in the quarter but plan to opportunistically repurchase shares using our free cash flow.

非公認會計準則營業利潤率為 8%，得益於對員工人數增加的嚴格控制以及產品毛利率的超額完成。我們的非 GAAP 調整後自由現金流率為 13%。我們仍然擁有強勁的現金狀況，擁有 49 億美元的現金、現金等價物以及短期和長期投資。我們在本季度沒有回購任何股票，但計劃利用我們的自由現金流機會性地回購股票。

Now let's turn to guidance. Our forecast assumes that our largest customers will continue to be a growth headwind. We are seeing encouraging signs of stabilization but not recovery. Our forecast calls for these customers to more closely align their consumption with their annual contract value. For the third quarter, we expect product revenues between $670 million and $675 million, representing year-over-year growth between 28% and 29%.

現在讓我們轉向指導。我們的預測假設我們最大的客戶將繼續成為增長的阻力。我們看到了令人鼓舞的穩定跡象，但沒有看到復蘇。我們的預測要求這些客戶將他們的消費與其年度合同價值更加緊密地結合起來。對於第三季度，我們預計產品收入在 6.7 億美元至 6.75 億美元之間，同比增長 28% 至 29%。

Turning to margins. We expect on a non-GAAP basis 4% operating margin. And we expect 364 million diluted weighted average shares outstanding. For the full year fiscal 2024, we expect product revenues of approximately $2.6 billion, representing year-over-year growth of approximately 34%. Turning to profitability for the full year fiscal 2024, we expect on a non-GAAP basis approximately 76% product gross margin, 5% operating margin and 26% adjusted free cash flow margin, and we expect 362 million diluted weighted average shares outstanding.

轉向邊緣。我們預計按非公認會計準則計算的營業利潤率為 4%。我們預計攤薄加權平均流通股數量為 3.64 億股。 2024 財年全年，我們預計產品收入約為 26 億美元，同比增長約 34%。談到2024 財年全年的盈利能力，我們預計在非GAAP 基礎上，產品毛利率約為76%，營業利潤率為5%，調整後自由現金流利潤率為26%，我們預計稀釋後加權平均流通股數量為3.62 億股。

We will continue to prioritize hiring in product and engineering. We still expect to add approximately 1,000 employees in fiscal 2024, inclusive of M&A.

我們將繼續優先招聘產品和工程人員。我們仍預計在 2024 財年增加約 1,000 名員工，其中包括併購。

With that, operator, you can now open up the line for questions.

接線員，您現在可以撥打電話提問了。

(Operator Instructions) The first question will be from the line of Keith Weiss with Morgan Stanley.

（操作員說明）第一個問題將來自摩根士丹利的 Keith Weiss。

Mike, I wanted to dig into the comment about kind of a large customer activity. You talked to us about sort of consumption coming more in line with their committed contracts. Can you give us any visibility what's happening on the contract renewals? Because as you go through these big contract renewals, are you seeing any change in their behavior of what the large customers are willing to commit to? And any impacts that we should be thinking about on how that's going to impact bookings and sort of RPO on a go-forward basis?

邁克，我想深入研究有關大型客戶活動的評論。您與我們談到了某種消費更符合他們承諾的合同。您能否向我們介紹一下合同續籤的情況？因為當你續簽這些大合同時，你是否發現大客戶願意承諾的行為發生了任何變化？我們應該考慮這將如何影響未來的預訂和恢復點目標（RPO）？

Yes. No, we're seeing customers renew. This quarter was a good renewal quarter. We had our largest customer, they renewed under existing terms. They did $100 million 3-year renewal, even though their revenue run rate is at a higher amount than that. I think we did 9 or 10 $10 million-plus TCV deals this quarter, and most of those were renewals. And so customers are doing that. But remember, that doesn't necessarily equate to consumption. And we do know some of our largest customers are trying to consume at their contract rate rather than going above that.

是的。不，我們看到客戶續訂。本季度是一個很好的更新季度。我們有我們最大的客戶，他們根據現有條款續約。他們進行了 1 億美元的 3 年續約，儘管他們的收入運行率比這個數字更高。我認為本季度我們完成了 9 或 10 筆價值超過 1000 萬美元的 TCV 交易，其中大部分是續約。所以客戶正在這樣做。但請記住，這並不一定等於消費。我們確實知道我們的一些最大的客戶正試圖按照合同價格消費，而不是高於這個價格。

Got it. So the dynamic is really on consumption. It is not on contracting as of yet?

知道了。因此，動態實際上取決於消費。現在還沒有簽約嗎？

No, it's on consumption. As I said, the contracting actually feels -- the sentiment really seemed to change in July with customers really reengaging with us. And so -- and I think we'll have good bookings. But that doesn't equate to consumption. It takes time for the consumption to come in.

不，是消費。正如我所說，合同確實讓人感覺——隨著客戶真正重新與我們互動，7 月份的情緒似乎確實發生了變化。所以——我想我們會有很好的預訂。但這並不等於消費。消費的到來需要時間。

Got it. Got it. And if I could sneak one in for Frank as well. You talked about you need a good data strategy to have a good AI strategy, and that's something we hear a lot when we're talking to customers and people out there in the field. So that really resonates. When it comes to kind of go-to-market and sort of the selling motion, does having to have the gen AI conversation while a long-term positive, does that disrupt the -- that sort of typical kind of data cloud discussion that you guys have been having for the past like 5 years with these customers? Does it have that risk or has it been elongating the sales cycles in any real way?

知道了。知道了。如果我也能為弗蘭克偷偷送一份的話。您談到需要良好的數據策略才能擁有良好的人工智能策略，這是我們在與客戶和該領域的人員交談時經常聽到的說法。所以這確實引起了共鳴。當談到進入市場和銷售活動時，必須進行一代人工智能對話，同時具有長期的積極意義，這是否會擾亂您所討論的那種典型的數據云討論？過去五年來，大家一直在與這些客戶打交道嗎？它是否存在這種風險，或者它是否以任何實際方式延長了銷售週期？

No, I wouldn't say so, Keith. We were actually saying that having highly organized optimized trusted sanction data is incredibly important for deploying large language models. If you think, you can just drop a model on top of a data lake and just see what happens. That's not going to end well. And that's what people are realizing. So they really got to get super serious about their foundations before we -- if you don't have a good foundation, there's not much you can build on top of that.

不，我不會這麼說，基思。我們實際上是說，擁有高度組織的優化的可信制裁數據對於部署大型語言模型非常重要。如果您願意，您可以將模型放在數據湖上，然後看看會發生什麼。這不會有好的結局。這就是人們正在意識到的。因此，在我們之前，他們確實必須非常認真地對待他們的基礎——如果你沒有良好的基礎，那麼你就無法在此基礎上進行任何構建。

There's tons of governance issues involved as well. We spent literally decades as an industry making data highly governed. In other words, who can have access to what. So that now needs to translate into the world of large language models as well. So there's tons of questions that are coming up that are really important for the enablement of language models and AI generally. So being extremely organized on your data is going to become a premium thing. And we're obviously that's -- we've been on that, but it's become more important as a function of this.

還涉及大量治理問題。作為一個行業，我們花了幾十年的時間對數據進行嚴格管理。換句話說，誰可以訪問什麼。因此，現在也需要將其轉化為大型語言模型的世界。因此，出現了大量對於語言模型和人工智能的總體實現非常重要的問題。因此，對數據進行高度組織將成為一件很重要的事情。顯然，我們一直在這樣做，但它變得更加重要。

The next question will be from the line of Mark Murphy with JPMorgan.

下一個問題將來自摩根大通的馬克·墨菲。

Frank, I'm wondering if you can speak to the expanded Microsoft relationship. I believe you referred to it as nearly a doubling of the commitment. And I'm wondering if you're optimistic on seeing that alignment in the field coming together and perhaps unlocking some new opportunities relating to Azure or even some of the OpenAI workloads that you wouldn't have seen previously? Then I have a quick follow-up.

Frank，我想知道您是否可以談談擴大的 Microsoft 關係。我相信您所說的承諾幾乎增加了一倍。我想知道您是否對看到該領域的整合持樂觀態度，並可能釋放一些與 Azure 相關的新機會，甚至是一些您以前從未見過的 OpenAI 工作負載？然後我會進行快速跟進。

Yes. Look, the bottom line of working with the cloud vendors is not what gets set at the top levels, the good intentions. What matters is what are the incentives. When you get 14 layers down at street level, how do we get people get paid. That determines whether they are either going to fight you and double and triple down on that or they're going to partner with you. And we see models where we've worked out really, really well. Of course, with AWS. But with Microsoft, we were not in the place that we wanted to be at street level in terms of the incentives.

是的。看，與雲供應商合作的底線並不是高層設定的良好意願。重要的是激勵因素是什麼。當你在街道上往下走 14 層時，我們如何讓人們獲得報酬。這決定了他們是否會與你戰鬥並加倍或三倍地投入，或者他們是否會與你合作。我們看到了一些模型，我們的效果非常非常好。當然，使用 AWS。但對於微軟來說，我們在激勵措施方面並沒有達到我們想要的水平。

So we really took this opportunity when we were renegotiating our relationship with Microsoft to say, "Hey, we have to tackle this, right?" And Microsoft very much wanted to be a bigger percentage of our business because they're not punching at their weight at all. They're not as big a percentage of our business as they should be based on our market share. And they want to be. And this is the way to do it. It's really you need to bring alignment to the field organizations. Then you're going to get partnerships, and then you're going to get joint selling, and then you're going to get your fair share.

因此，當我們與微軟重新談判關係時，我們確實抓住了這個機會說：“嘿，我們必須解決這個問題，對嗎？”微軟非常希望在我們的業務中佔據更大的份額，因為他們根本沒有發揮自己的作用。根據我們的市場份額，它們在我們業務中所佔的比例並不大。他們也想成為這樣的人。這就是做到這一點的方法。您確實需要使現場組織保持一致。然後你將獲得合作夥伴關係，然後你將獲得聯合銷售，然後你將獲得公平的份額。

Okay. And Mike, as a follow-up, in Q2, did you observe any customers adjusting their indexing or reducing data retention time lines? It looks like the quarter went well overall. But I'm curious if you sense any more or less of that optimization activity heading into the second half here.

好的。邁克，作為後續行動，在第二季度，您是否觀察到任何客戶調整索引或減少數據保留時間線？看起來本季度總體進展順利。但我很好奇您是否或多或少感覺到進入下半年的優化活動。

In terms of customers changing their retention policy, we really didn't notice any of that. If anything, we saw growth in the amount of storage in the quarter. And I just want to call it, that was really one large customer that changed their retention from 5 to 3 years. That was pretty unusual.

在客戶改變保留政策方面，我們確實沒有註意到這一點。如果說有什麼不同的話，那就是本季度的存儲量有所增長。我只想說，這確實是一個大客戶將他們的保留期從 5 年改為 3 年。這很不尋常。

And in terms of optimizations, we continually work with customers on their own optimizations, but we're continuing to optimize our software as well, too, because we are really dedicated to delivering price performance for our customers. And I want to stress, we've been talking about optimizations with investors since we went public. These will always continue, optimizations, whether it's customer or us doing the optimizations because when -- history has shown, when we improved price performance, more workloads come to us.

在優化方面，我們不斷與客戶合作進行優化，但我們也在繼續優化我們的軟件，因為我們真正致力於為客戶提供性價比。我想強調的是，自從我們上市以來，我們一直在與投資者討論優化問題。這些優化將永遠持續下去，無論是客戶還是我們進行優化，因為歷史表明，當我們提高性價比時，更多的工作負載就會降臨到我們身上。

The next question will be from the line of Kirk Materne with Evercore ISI.

下一個問題將來自 Evercore ISI 的 Kirk Materne。

Mike, I was wondering if you could just talk a little bit more about the comment you're seeing stabilization in consumption, but not recovery. I'm just kind of curious, when you say that, is that the pressure from the top-down on practitioners starting to ease a little bit and they're starting to feel better about what they can consume, and they're just waiting for budgets to kind of get refreshed to sort of get going on the recovery side. I'm just kind of curious if it's sort of the top-down easing or it's more that they're taking a little while to sort of ramp back up on projects that perhaps they slowed down 6 months ago or 3 months ago?

邁克，我想知道您是否可以多談談您看到消費穩定但沒有復甦的評論。我只是有點好奇，當你這麼說時，從業者自上而下的壓力開始減輕一點，他們開始對自己可以消費的東西感覺更好，他們只是在等待讓預算得到更新，以便在復蘇方面有所進展。我只是有點好奇這是否是一種自上而下的寬鬆政策，或者更多的是他們需要一段時間來恢復可能在 6 個月前或 3 個月前放緩的項目？

I actually think it's both, but I definitely think the fact that we kind of saw customers more reengaging with us in July on contracts, and that continues into this quarter. I think it is easing a little bit at the top level in terms of approvals for customers and they're willing to commit. But it takes time to convert that to consumption. With that said, consumption is good. It was really good today as an example but it's only one data point. It's -- we want to see more days of that before we think the -- we're into a real recovery. I think stabilization is the right term. We're not seeing customers reduce their consumption right now.

我實際上認為兩者都是，但我絕對認為我們看到客戶在 7 月份更多地與我們簽訂合同，並且這種情況持續到本季度。我認為高層在客戶審批方面有所放鬆，他們也願意做出承諾。但將其轉化為消費需要時間。話雖如此，消費還是不錯的。今天作為一個例子確實很好，但它只是一個數據點。我們希望在我們認為我們已經進入真正的複甦之前看到更多的日子。我認為穩定是正確的術語。我們目前沒有看到客戶減少消費。

Okay. And then, Frank, just as you spoke to a lot of executives at Summit, do they recognize the fact that the road to AI does require perhaps a heavier level of investment than they were thinking 12 months ago? How do you think that factors into sort of their thinking on budgets as we go into '24?

好的。然後，弗蘭克，正如您在峰會上與許多高管交談一樣，他們是否認識到人工智能之路確實需要比他們 12 個月前想像的更多的投資？當我們進入 24 世紀時，您認為這會如何影響他們對預算的看法？

The reality is they don't really know yet in any real definitive terms what this is going to take. I mean I think a lot of people, and I think this correct, they have characterized their foray into language models as experimental, exploratory and they're sort of trying to get their arms around how big a bread box is this. So it's going to take a while before we get a real read on what the level of investment is that people are going to stomach to do this.

現實情況是，他們還不真正確切地知道這將帶來什麼。我的意思是，我認為很多人，我認為這是正確的，他們將他們對語言模型的嘗試描述為實驗性的、探索性的，他們試圖了解這個麵包盒有多大。因此，我們需要一段時間才能真正了解人們願意為此進行的投資水平。

I mean one of the challenges, one of the great things about search historically has been -- that search also had a very potent business model to go with us to pay for it. And we cannot sort of unleash AI and have no business model to pay for it, and people will get tired of that really, really quick. So these are -- these GPUs from NVIDIA, they aren't cheap as powerful as they are. So we all have to bring that into alignment and into focus and have a sensible go-forward strategy. So a lot of the use cases will focus on, what are we getting for this, right? This is not just fun and games and planning your next trip to Yellowstone. I mean people are going to be asking very, very hard-hitting questions, "What is it doing for us?"

我的意思是，搜索歷史上的挑戰之一、最偉大的事情之一是——搜索也有一個非常有效的商業模式來幫助我們支付費用。我們無法釋放人工智能，也沒有商業模式來為其付費，人們很快就會對此感到厭倦。所以這些是——這些來自 NVIDIA 的 GPU，它們並不便宜，但功能卻很強大。因此，我們所有人都必須將其納入重點，並製定明智的前進戰略。所以很多用例都會關注，我們能得到什麼，對嗎？這不僅僅是樂趣和遊戲以及計劃您的下一次黃石之旅。我的意思是，人們會問非常非常尖銳的問題，“它對我們有什麼作用？”

The next question will be from the line of Raimo Lenschow with Barclays.

下一個問題將來自巴克萊銀行的 Raimo Lenschow。

Great to see the stabilization as well from my end. Congrats there. The -- Frank, one for you more. As you -- as we all realize that data is kind of the new fuel, we do see more [ventriloquising] like, "Oh, no, I have the data, I have the data." Which kind of in a way like kind of puts them on a slightly different course than they used to be compared to your partnership. How do you see that competitive landscape evolving for you since you, in theory, are the natural kind of holding source for data that is used in AI, how do you see this playing out for you guys?

很高興看到我這邊也穩定下來。恭喜你。弗蘭克，再給你一個。正如我們都意識到數據是一種新的燃料一樣，我們確實看到了更多的[口技]，比如，“哦，不，我有數據，我有數據。”與您的合作夥伴關係相比，這在某種程度上使他們的路線與以前略有不同。您如何看待競爭格局的演變，因為從理論上講，您是人工智能中使用的數據的自然持有來源，您如何看待這對你們來說會發生什麼？

Well, we see it play out really well. I think we agree with you. I mean we think that data is becoming infinitely valuable, and that's for all kinds of reasons because we can no longer run enterprises and institutions based on what we call anecdotal observation because the world is too dynamic, too disruptive, and we have this massive disintermediation happening. We no longer have all these intermediaries between us and the end customer. You can only run direct-to-consumer businesses with data.

嗯，我們看到它的效果非常好。我想我們同意你的看法。我的意思是，我們認為數據正變得無限有價值，這是出於各種原因，因為我們不能再根據所謂的軼事觀察來運營企業和機構，因為世界太動態、太具有破壞性，而且我們有這種大規模的脫媒現象。正在發生。我們和最終客戶之間不再有所有這些中介。您只能利用數據來運營直接面向消費者的業務。

I mean you see that, for example, in insurance like auto industry, companies like GEICO and Progressive and Liberty Mutual. I mean you can only run these businesses on top of data, telemetry data being really important. And we think every business, every institutions is going to develop a complete and total dependency, not just on data, but the ability to harness that data. So this is a full-on transformation really of how industry and institutions have operated. And we're at the leading edge, but we're also very much at the beginning of it.

我的意思是，例如，在汽車行業等保險領域，GEICO、Progressive 和 Liberty Mutual 等公司。我的意思是，你只能在數據之上運行這些業務，遙測數據非常重要。我們認為每個企業、每個機構都會產生完全的依賴，不僅依賴於數據，而且依賴於利用這些數據的能力。因此，這實際上是行業和機構運作方式的全面轉變。我們處於領先地位，但我們也還處於起步階段。

The next question will be from the line of Karl Keirstead with UBS.

下一個問題將來自瑞銀集團的 Karl Keirstead。

So Mike, just one -- maybe two quickies. You mentioned there was a onetime credit from one of the CSPs. Are you able to define how material that was? It sounds like it might have certainly hit gross margins, but just wanted to clarify.

邁克，只有一個——也許是兩個快手。您提到其中一個 CSP 提供了一次性信用。你能定義它有多重要嗎？聽起來它肯定會影響毛利率，但只是想澄清一下。

That was about $4 million that hit in the quarter.

本季度的收入約為 400 萬美元。

Okay. Got it. And then maybe just a follow-up, Mike, what I'm hearing you describe is an effort -- you mentioned certainly your largest customer, but perhaps others taking their usage a little bit closer to their ACV. So one would think that, that would result in a decent amount of headwind. Yet your guidance for the third quarter, in my judgment anyway, is relatively solid. So what's the offset? What customer segment might be ramping nicely to offset to some extent the headwinds from your large customers?

好的。知道了。然後也許只是後續行動，邁克，我聽到你描述的是一種努力——你提到的當然是你最大的客戶，但也許其他人的使用方式更接近他們的 ACV。所以人們會認為，這會帶來相當大的阻力。然而，無論如何，根據我的判斷，你們對第三季度的指導是相對可靠的。那麼偏移量是多少呢？哪些客戶群可能會很好地增長，以在一定程度上抵消來自大客戶的阻力？

Well, we're now at 639 Global 2000 that are only consuming on average on a trailing 12 months at around $1.5 million, $1.6 million. A lot of those are still doing their migrations, and we don't see that stopping. It's the larger customers, they just are not forecast to grow as quickly. They're still growing but at a slower pace. Once again, this is a consumption model. That could turn around tomorrow.

好吧，我們現在在全球 2000 強中排名第 639 位，過去 12 個月的平均消費額約為 150 萬美元、160 萬美元。其中許多人仍在進行遷移，而且我們認為這種情況不會停止。這是較大的客戶，只是預計他們的增長速度不會那麼快。它們仍在增長，但速度較慢。再次強調，這是一種消費模式。明天情況可能會扭轉。

The next question will be from the line of Kash Rangan with Goldman Sachs.

下一個問題將來自高盛的卡什·蘭根 (Kash Rangan)。

Frank, I think your fireside chat with Jensen was absolutely illuminating. He was looking at the opportunity set with structured data in the Snowflake ecosystem and almost salivating. And yet you seemed a little moderated and -- that you need to have a business case. So when are we likely to reach the point where generative AI, coupled with Snowpark could really lead to a tangible increase in consumption outside of the core data cloud business.

弗蘭克，我認為你與詹森的爐邊談話絕對具有啟發性。他正在尋找 Snowflake 生態系統中結構化數據的機會集，幾乎垂涎欲滴。然而你似乎有點溫和——你需要有一個商業案例。那麼，我們什麼時候才能達到這樣的程度：生成式人工智能與 Snowpark 相結合才能真正帶來核心數據云業務之外的消費的切實增長。

And one for you, Mike. If we were to read your comments, 3 months of stability, 4 months of stability it looks like including August, does that mean that net expansion rates reach a bottom and could potentially start to stabilize and rebound as we head into the later part of the year.

還有一份給你，邁克。如果我們讀到你的評論，3 個月的穩定，4 個月的穩定（看起來包括8 月），這是否意味著淨擴張率已觸底，並可能在我們進入下半年時開始穩定和反彈。年。

This is Frank, Kash. In the really short term, I'm really talking days, weeks and months here where you're going to see language model begin to immediately impact the business is that SQL generation. I mean, in other words, the analyst job is going to be up-leveled so much, I mean people are going to be able to drive queries into the data much better, much faster with far less skill requirement than they ever have before. And we're showing that off every day. So I mean, these days, you don't even have to be literate in order to be able to have interactions with your data. So that really is an expansion vector that is just enormously, and it's very close to home. Because it's really how you use data and how you use a platform like Snowflake.

這是弗蘭克，卡什。從短期來看，我真的在這裡談論幾天、幾周和幾個月，您將看到語言模型開始立即影響業務，即 SQL 生成。我的意思是，換句話說，分析師的工作將得到大幅提升，我的意思是，人們將能夠更好、更快地對數據進行查詢，而對技能的要求比以前要低得多。我們每天都在展示這一點。所以我的意思是，如今，您甚至不需要識字就可以與數據進行交互。所以這確實是一個巨大的擴展向量，而且它非常接近我們的目標。因為這實際上關係到您如何使用數據以及如何使用 Snowflake 這樣的平台。

The other area where you're going to see drivers of workload is that people get to search for data related to what their general angle of inquiry is in a much more effective manner than they have been able before. And this is also where it's very important that you can search beyond enterprise boundaries because the context of data is not limited by your enterprise boundaries. We can go on and on and on about use case. There is a million of them. If you get further down, you can start asking really, really hard questions that in prior periods, prior eras, we really needed to launch whole analyst teams to go research and investigate topics where now the data, we'll be able -- or the systems will be able to generate queries in the type of data that it will immediately very, very quickly begin to generate insights. And that's -- by the way, that's going to become the leading edge for structured proprietary data, which is, of course, the center of our universe.

您將看到工作量驅動因素的另一個領域是，人們可以比以前更有效的方式搜索與他們的一般調查角度相關的數據。這也是您可以超越企業邊界進行搜索非常重要的地方，因為數據上下文不受企業邊界的限制。我們可以繼續討論用例。他們有一百萬。如果你再往下看，你可以開始問非常非常困難的問題，在以前的時期、以前的時代，我們確實需要啟動整個分析師團隊來研究和調查現在數據我們能夠的主題——或者系統將能夠生成數據類型的查詢，並立即非常非常快地開始生成見解。順便說一句，這將成為結構化專有數據的前沿，這當然是我們宇宙的中心。

And on your question on net revenue retention, I just want to remind you, I'm not going to guide to a net revenue retention. But I do think over time, it is going to continue to converge closer to our growth rate. I do think it will stabilize, but I do expect it's going to come down slightly from where it's at right now just on what we're seeing today.

關於你關於淨收入保留的問題，我只是想提醒你，我不會指導淨收入保留。但我確實認為，隨著時間的推移，它將繼續向我們的增長率靠攏。我確實認為它會穩定下來，但我確實預計它會比我們今天看到的情況略有下降。

The next question will be from the line of Brad Zelnick with Deutsche Bank.

下一個問題將來自德意志銀行的 Brad Zelnick。

I've got one for Frank and one for Mike. Frank, stable edges continue to tick up, which is great to see, and I know it's an important metric for the company and its strategic vision. Anything else perhaps qualitative that you can share in terms of how data sharing is progressing.

我有一本給弗蘭克，一本給邁克。坦率地說，穩定的邊緣繼續上升，這是很高興看到的，我知道這是公司及其戰略願景的一個重要指標。您可以分享有關數據共享進展情況的任何其他定性信息。

And for you, Mike, great to see the margin upside, everybody's happy about it. But with such a huge opportunity, how can you be sure you're striking the right balance of investment, especially when you're up against such well-capitalized competitors?

對於你來說，邁克，很高興看到利潤率上升，每個人都對此感到高興。但面對如此巨大的機會，您如何確定自己在投資上取得了正確的平衡，尤其是當您面對資本充足的競爭對手時？

On the topic of data sharing, we instrument that whole side of the business very, very carefully, and we drive it on a quarterly basis. But sometimes data edges are very enterprise-specific. In other words, they do just have things -- the use cases, that just pertain to their business, and it's -- these are bilateral relationships between Snowflake accounts and different institutions.

在數據共享這一主題上，我們非常非常仔細地對整個業務進行監測，並按季度推動它。但有時數據邊緣是非常特定於企業的。換句話說，他們確實擁有與他們的業務相關的用例——這些是 Snowflake 帳戶和不同機構之間的雙邊關係。

But what gets really interesting, where you get real network effect kicking in is when you have industries or subindustries where data sharing just makes sense. And obviously, in financial institutions, because financial institutions inherently have been pumping data around in massive, massive volumes for -- literally for generations. This is an absolute no-brainer. And we do the vast majority. Historically, data edges have been in the financial services sector. It's become almost the standard. This is how we move data from A to B to C. Asset management, particularly has a really big need for that.

但真正有趣的是，當您擁有數據共享有意義的行業或子行業時，您就會產生真正的網絡效應。顯然，在金融機構中，因為金融機構本質上一直在大量地輸送數據——實際上是幾代人的數據。這絕對是理所當然的。絕大多數都是我們做的。從歷史上看，數據優勢一直存在於金融服務領域。這幾乎已經成為標準。這就是我們將數據從 A 轉移到 B 再到 C 的方式。資產管理尤其對此有很大的需求。

But the other area -- and again, this is an industry, in supply chain management. In the supply chain, there are multiple entities. To the degree that they all have Snowflake accounts, it's very easy to get visibility in the supply chain across entities and being able to flag supply chain events much earlier and get visibility to that. So once you're in a supply chain, the need to be on Snowflake share data with your supply chain partners is going to become very, very compelling.

但另一個領域——再說一遍，這是一個供應鏈管理行業。在供應鏈中，存在多個實體。就他們都擁有 Snowflake 帳戶而言，很容易獲得跨實體的供應鏈的可見性，並且能夠更早地標記供應鏈事件並獲得可見性。因此，一旦您進入供應鏈，與供應鏈合作夥伴在 Snowflake 上共享數據的需求將變得非常非常引人注目。

And we announced at Summit and even earlier our relationship with Blue Yonder’ for example, who is really the largest software company in the world of supply chain management, that they are replatforming on Snowflake. So we think that's another sort of industry/subindustry where every manufacturer, every retailer it's going to become an opportunity for us. So just a little bit of color on how these things develop from our perspective.

我們在峰會上甚至更早的時候宣布了我們與 Blue Yonder 的關係——例如，他們是供應鏈管理領域最大的軟件公司，他們正在 Snowflake 上進行平台重組。因此，我們認為這是另一種行業/子行業，每個製造商、每個零售商都將成為我們的機會。因此，從我們的角度來解釋一下這些事情是如何發展的。

On your question on investments in the business, Brad, given the opportunity, we are investing as fast as we think we need to invest. You did see the guided margins to 4% for operating margin where we just did 5% this quarter. And that's because we're investing. As an example, we have 1,000 H100 GPUs reserved. That's an extra $1 million a month as we're working on AI. I'm not getting requests that people need more headcount in the engineering organization and the sales and marketing until we see an increase in productivity. We're going to be very methodical about how we add resources into those areas. So we're definitely not under-investing in the business. I'm not getting the feedback from any of the executive team with regards to that.

關於你關於業務投資的問題，布拉德，如果有機會，我們將按照我們認為需要的速度進行投資。您確實看到營業利潤率的指導利潤率為 4%，而本季度我們的指導利潤率僅為 5%。那是因為我們正在投資。例如，我們預留了 1,000 個 H100 GPU。由於我們正在研究人工智能，因此每月會額外增加 100 萬美元。在我們看到生產力提高之前，我沒有收到工程組織以及銷售和營銷人員需要更多人員的請求。我們將非常有條理地向這些領域添加資源。所以我們對這項業務的投資絕對不會不足。我沒有收到任何執行團隊的反饋。

Your next question will be from the line of Tyler Radke with Citi.

您的下一個問題將來自花旗銀行的泰勒·拉德克 (Tyler Radke)。

First question, just on the commentary around some of the projects, you're starting to see better momentum there, particularly in July. I was wondering if you could just comment on the nature of those projects. Are they larger deals than you typically see or maybe they include more generative AI or data science, given all the new products that you released. If you could just kind of contrast the pickup and kind of where that's coming from?

第一個問題，就一些項目的評論而言，你開始看到那裡更好的勢頭，特別是在 7 月份。我想知道您是否可以評論一下這些項目的性質。考慮到您發布的所有新產品，這些交易是否比您通常看到的規模更大，或者可能包含更多的生成人工智能或數據科學。您能否將拾音器與拾音器的來源進行對比？

Do you want to go...

你想去嗎...

You go ahead, Frank.

你繼續吧，弗蘭克。

You shouldn't equate projects with deals, okay, because there's tons and tons of projects going on, and projects relate to use cases and workloads and applications. What we said in the prepared remarks is we've really seen a sort of a sentiment change from the early quarters where people were sort of trying to cut off their limbs to fit within budgetary constraints and all this kind of stuff and where do -- and that's where you see unnatural acts to save money.

您不應該將項目與交易等同起來，因為有大量的項目正在進行，而項目與用例、工作負載和應用程序相關。我們在準備好的發言中所說的是，我們確實看到了從早期幾個季度開始的一種情緒變化，當時人們試圖切斷自己的四肢以適應預算限制和所有這些東西，但在哪裡——這就是你看到為了省錢而採取的不自然行為的地方。

That has really subsided considerably. And the conversation has really gone back to where it historically has been is we want to do these applications, these workloads, these migrations. And of course, we're pushing the boundaries on the much more sophisticated use cases in machine learning. And obviously, people want to understand, how do I deploy large language models on a Snowflake platform? And we have we have outlined that in excruciating detail and demonstrated, showcased how we are doing that, and we're super excited about how that's unfolding for us and our customers.

這確實已經大大減弱了。談話確實回到了歷史上的狀態，即我們想要執行這些應用程序、這些工作負載、這些遷移。當然，我們正在突破機器學習中更複雜用例的界限。顯然，人們想了解，如何在 Snowflake 平台上部署大型語言模型？我們已經以令人難以忍受的細節概述了這一點，並演示、展示了我們是如何做到這一點的，我們對我們和我們的客戶的進展感到非常興奮。

Great. And then a follow-up, just in terms of the Snowpark revenue, any update on kind of where you're expecting that to track as you exit this year? And then these related services, whether it's the Native Apps store or Container Services, would that all fall under Snowpark theoretically when that goes GA next year?

偉大的。然後是後續行動，就 Snowpark 的收入而言，您預計今年退出時會跟踪哪些方面的最新情況？然後這些相關的服務，無論是原生應用商店還是容器服務，理論上明年通用時這些服務都會屬於 Snowpark 的範圍嗎？

In terms of Snowpark, as we said, whereas Frank talked about seeing 70% growth in Snowpark consumption, is still relatively small but meaningful. We have a number of customers that are in the process of doing their migrations, a few quite large ones. I do think next year, it will be more meaningful to revenue. But on $2.6 billion in revenue, it's a couple percent of our revenue this year.

就 Snowpark 而言，正如我們所說，雖然 Frank 談到看到 Snowpark 消費增長 70%，但仍然相對較小，但很有意義。我們有許多客戶正在進行遷移，其中一些規模相當大。我確實認為明年，這對收入將更有意義。但就 26 億美元的收入而言，這只是我們今年收入的幾個百分點。

Once the Container Services becomes prime time, and that is part of Snowpark, obviously, that means any workload that comes for our game to be deployed on Snowflake, this is obviously running close to the data inside our governance perimeter, it's essentially virtualization for the cloud, so we would think there is enormous upside for us once those services become generally available across all our cloud platforms.

一旦容器服務成為黃金時間，這顯然是 Snowpark 的一部分，這意味著我們的遊戲的任何工作負載都將部署在 Snowflake 上，這顯然靠近我們治理範圍內的數據運行，它本質上是虛擬化雲，因此我們認為，一旦這些服務在我們所有的雲平台上普遍可用，我們就會有巨大的優勢。

Which will be next year.

這將是明年。

The next question will be from the line of Brent Thill with Jefferies.

下一個問題將由布倫特·希爾 (Brent Thill) 和杰弗里斯 (Jefferies) 提出。

Mike, you mentioned at the Analyst Day, you were idling back quota-carrying sales capacity on the new hire front. Have you seen any different to lean back into hiring quota reps in '23?

邁克，您在分析師日提到，您在新員工方面閒置了配額銷售能力。您是否發現 23 年重新招聘配額代表有什麼不同？

It really depends upon the territory and the opportunity. There are some territories where kind of the regions were shrinking, where it's overcapacity and we're reshifting those heads to other more productive territories. As we're -- I'm not planning on adding net a lot of new ones for the balance of this year. But as we're starting to plan for next year, there is additional headcount going into the quota-carrying rep area.

這實際上取決於領土和機會。在某些地區，有些地區正在萎縮，產能過剩，我們正在將這些負責人轉移到其他生產力更高的地區。因為我們——我不打算在今年餘下的時間裡淨增加很多新的。但當我們開始為明年做計劃時，有額外的人員進入配額代表領域。

Great. And for Frank, on the vertical side, any verticals that are showing more excitement that perhaps weren't online that weren't firing up? Are you seeing anything change here?

偉大的。對於弗蘭克來說，在垂直領域，是否有任何垂直領域表現出更多的興奮，而這些興奮可能不是在線的，也沒有被激發？你看到這裡有什麼變化嗎？

I mean actually that's a great question because we had massive outperformance by our healthcare vertical this quarter and healthcare usually runs fourth or fifth in the lineup of verticals, and they massively outperformed. Really excited about -- it feels like that healthcare is really getting a move on, if you will. They have not traditionally been an aggressive adopter of technology. But in the world of data, they are moving. They're moving hard, and you see it on the provider side, you see it on the payer side, you see it on the pharma side. So I think that's going to become a great contributing segment for us.

我的意思是，實際上這是一個很好的問題，因為本季度我們的醫療保健垂直領域表現出色，而醫療保健通常在垂直領域中排名第四或第五，而且它們的表現大幅優於其他領域。真的很興奮——如果你願意的話，感覺醫療保健真的在進步。傳統上，他們並不是技術的積極採用者。但在數據世界中，它們正在發生變化。他們正在努力行動，你可以在提供商方面看到它，你可以在付款人方面看到它，你可以在製藥方面看到它。所以我認為這將成為我們做出巨大貢獻的部分。

Yes. Healthcare and Life Sciences grew 61% year-over-year in revenue for us.

是的。我們的醫療保健和生命科學收入同比增長 61%。

The next question will be from the line of Michael Turrin with Wells Fargo.

下一個問題將來自富國銀行的邁克爾·特林 (Michael Turrin)。

Great. I think one of the comments mentioned new bookings outperformed expectations. I appreciate you're still seeing room for improvement, but anything you can add around what drove the improvement versus last quarter? It sounded like healthcare from the prior commentary, but wondering if some other or certain product releases maybe also contributed there.

偉大的。我認為其中一條評論提到新預訂超出了預期。我很高興您仍然看到了改進的空間，但是您可以補充一下與上季度相比推動改進的因素嗎？從之前的評論來看，這聽起來像是醫療保健，但想知道其他或某些產品的發布是否也可能對此做出貢獻。

Yes. Hard to say whether it was -- I don't think it was a product release. I would say we saw some nice renewals from customers with growth. We also saw two very large cap ones. One large one in Europe, which was -- a cap one as an initial deal was a $22 million TCV deal in an insurance industry, and we saw a large gaming company in Korea commit to $9.5 million as a cap one. So clearly, our message is getting across to these customers, and they see what we're doing, and a lot of these want to do more in the area of AI. But first, they need to get their data into Snowflake, and it's going to be a journey for these people. It's not going to happen overnight, AI, for our customers.

是的。很難說它是否是——我不認為這是一個產品發布。我想說，我們看到客戶的一些增長帶來了很好的續約。我們還看到了兩個非常大的帽子。歐洲的一項大型交易是——初始交易的上限是保險行業一筆 2200 萬美元的 TCV 交易，我們看到韓國的一家大型遊戲公司承諾支付 950 萬美元作為上限。很明顯，我們的信息正在傳達給這些客戶，他們看到我們正在做的事情，其中​​很多人希望在人工智能領域做更多的事情。但首先，他們需要將數據導入 Snowflake，這對這些人來說將是一個旅程。人工智能，對於我們的客戶來說，這不會在一夜之間發生。

I like and appreciate those large deal stats. Maybe just quickly on the back half. If you can just help level set what's embedded in the rest of the year outlook. You've seen multiple comments around stabilization. Is that fairly consistent with what informs the outlook? And maybe just any refresh on second half seasonality as expected?

我喜歡並欣賞這些重要的統計數據。也許只是在後半段很快。如果您能幫助確定今年剩餘時間展望中的內容。您已經看到了許多關於穩定性的評論。這與預測的情況相當一致嗎？也許下半年季節性的更新如預期的那樣？

Well, Q4 is usually one of our largest bookings quarter, and it's shaping up, but that's not necessarily consumption. And -- but the sentiment within our sales team has definitely shifted from where it was in the first half of the year.

嗯，第四季度通常是我們最大的預訂季度之一，而且它正在形成，但這不一定是消費。而且——但我們銷售團隊內部的情緒與上半年相比確實發生了變化。

The next question will be from the line of Patrick Colville with Scotiabank.

下一個問題將由豐業銀行的帕特里克·科爾維爾 (Patrick Colville) 提出。

I just want to double-click on your comments. I think you said customers are reengaging in July. You said that in the prepared remarks. I mean, I guess, what do you mean by that? Is that new customers, existing customers, consumption, contract negotiations? And then any color you can give us on, thus far, in August would be helpful.

我只想雙擊你的評論。我想你說過客戶會在七月份重新參與。你在準備好的發言中這麼說過。我的意思是，我想，你這是什麼意思？是新客戶、老客戶、消費、合同談判？到目前為止，八月份您可以給我們提供的任何顏色都會有所幫助。

All 3 of the above. We have new customers. I just pointed out, those two large cap ones. We usually don't do cap ones that big, and it was a very good quarter with some large new customers. We're seeing our existing customers. We saw some nice early renewals with customers where they were consuming faster. And we're seeing customers willing to do larger deals rather than just do a co-term to bridge them through to another period. So that's what I mean by the sentiment is changing with our customers.

以上 3 項全部。我們有新客戶。我剛剛指出，那兩個大盤。我們通常不會做那麼大的上限，這是一個非常好的季度，有一些大的新客戶。我們正在見到我們現有的客戶。我們看到了一些不錯的早期續訂，客戶的消費速度更快。我們看到客戶願意做更大的交易，而不是僅僅做一個聯合術語來將他們過渡到另一個時期。這就是我所說的客戶情緒正在發生變化的意思。

All right. And I guess my kind of follow-up is NVIDIA reported results tonight. I think looking at their numbers, their data center revenue rose about 150% sequentially. But clearly, like AI spend is hitting the silicon layer. I mean a question I get from investors is when will AI spend more clearly hit the software layer. I mean, do you have any thoughts on that?

好的。我想我的後續行動是 NVIDIA 今晚報告的結果。我認為從他們的數據來看，他們的數據中心收入比上一季度增長了約 150%。但顯然，人工智能支出正在觸及矽層。我的意思是，我從投資者那裡得到的一個問題是，人工智能支出何時會更明確地影響到軟件層。我的意思是，你對此有什麼想法嗎？

I think it's going to be next year. As I said, it's going to take some time for AI. And people are still struggling to get the GPUs, and there is a time lag between when a chip manufacturer sells their chips to it gets built into the hardware that actually gets deployed in a rack in a data center and it gets deployed to customers.

我想那會是明年。正如我所說，人工智能的發展還需要一些時間。人們仍在努力獲得 GPU，從芯片製造商將其芯片出售給芯片、將其內置到實際部署在數據中心機架中的硬件到部署給客戶之間存在一個時間差。

I think you will see the leading act to that happening in months to come with the material impact. I think most analysts out there are seeing it as a 2024 thing, and we tend to agree with that.

我認為您將在未來幾個月內看到這一重大舉措的發生，並產生重大影響。我認為大多數分析師都將其視為 2024 年的事情，我們也傾向於同意這一點。

And I would say in my prior life, when we were buying racks of servers, there's a 6-month delay between when we bought them and when they were actually going into production. And I don't see that any different with GPUs.

我想說的是，在我的前世，當我們購買服務器機架時，我們購買它們和它們實際投入生產之間有 6 個月的延遲。我認為 GPU 沒有什麼不同。

The next question will be from the line of Alex Zukin with Wolfe Research.

下一個問題將來自 Wolfe Research 的 Alex Zukin。

This is Ethan Bruck on for Alex Zukin. I just had a quick kind of numbers question. So if we calculate the product revenue of cRPO bookings growth, it decelerated 13% from about 30% last quarter. So just curious, how should we think about this as an indicator of future consumption for future product growth?

我是伊森·布魯克 (Ethan Bruck) 替補亞歷克斯·祖金 (Alex Zukin)。我只是有一個快速的數字問題。因此，如果我們計算 cRPO 預訂增長的產品收入，它比上季度的 30% 左右下降了 13%。所以只是好奇，我們應該如何將此視為未來產品增長的未來消費指標？

So I guided full year revenue at $2.6 billion, and we'll guide next year and next year.

因此，我預計全年收入為 26 億美元，我們將在明年和明年進行指導。

Okay. And then I guess to ask the August trends question in a different way. I guess, has some of the stability you called out trended into August kind of first month for the quarter? And then just on the seasonality, is there any kind of month to note as you think about the rest of the year that seems that either is seasonally strong or seasonally weak.

好的。然後我想以不同的方式問八月趨勢問題。我想，您所說的一些穩定性是否會延續到本季度第一個月的八月？然後就季節性而言，當您考慮今年剩餘時間時，是否有任何月份需要注意，該月份似乎季節性強或季節性弱。

Well, as I said, August is shaping up very good. I called out yesterday was actually a very good consumption, but 1 day doesn't make a trend. Q4 is definitely seasonality with the holidays, with Thanksgiving in the U.S. and the Christmas holidays, that does impact daily consumption. From a bookings perspective, Q4 though is clearly our largest bookings.

嗯，正如我所說，八月的情況非常好。我昨天喊的其實是很好的消費，但是1天並沒有形成趨勢。第四季度肯定是季節性假期，美國的感恩節和聖誕節假期確實影響了日常消費。從預訂量的角度來看，第四季度顯然是我們最大的預訂量。

Your next question will be from the line of Brent Bracelin with Piper Sandler.

您的下一個問題將來自 Brent Bracelin 和 Piper Sandler。

I wanted to go back to the discussion around the increase we're seeing in model training capacity. Clearly, billions of incremental dollars going into NVIDIA GPUs here, you need data to train the models. I appreciate there's going to be a lag relative to when the spend hits the data layer. But are there any technical hurdles that need to be overcome? Or you do think this cycle is different in that there are other considerations as well? Just thinking through that -- the investment we're seeing right now in infrastructure and thinking through what are the other factors we need to think about before it starts to impact the data layer.

我想回到關於我們所看到的模型訓練能力增長的討論。顯然，NVIDIA GPU 投入了數十億美元，您需要數據來訓練模型。我知道相對於支出到達數據層的時間而言，將會存在滯後。但有什麼技術障礙需要克服嗎？或者您確實認為這個週期有所不同，因為還有其他考慮因素？仔細想想——我們現在在基礎設施方面看到的投資，以及在它開始影響數據層之前我們需要考慮的其他因素。

I'll start, and then maybe Christian can follow. He'll thinking about the question while I'm talking. You can't characterize as AI as one thing, right, because you see the things that people are doing with unstructured data and the whole notion of copilots and the systems and tutors and all that. It's very much focused on contextual data. And we see action with support call records, contact centers and so on. But then you look at Snowflake who sits on mountains of structured proprietary enterprise data, that's a different realm for AI than the very text model-oriented type of inquiry.

我先開始，然後克里斯蒂安也許可以跟上。我說話的時候他會思考這個問題。你不能將人工智能描述為一件事，對吧，因為你看到人們用非結構化數據所做的事情以及副駕駛、系統和導師等等的整個概念。它非常關注上下文數據。我們看到了支持通話記錄、聯絡中心等方面的行動。但是，你看看 Snowflake，它擁有堆積如山的結構化專有企業數據，這對於人工智能來說是一個與面向文本模型的查詢類型不同的領域。

And I have to say that just from all my conversation with customers, I mean people are behind, I would say, the textual side with how -- with proprietary data how they're going to approach that. We view that as our business, and we're driving that very hard, and hence, the emphasis on getting your data house in order because you just cannot unleash large language model and hope for the best because of all the issues that we've mentioned before around governance and just understanding of what kind of data we are generating out in the process.

我不得不說，從我與客戶的所有對話來看，我想說的是，人們在文本方面落後於他們將如何利用專有數據來實現這一目標。我們認為這是我們的業務，我們正在非常努力地推動這一點，因此，重點是讓您的數據室井然有序，因為由於我們遇到的所有問題，您無法釋放大型語言模型並希望得到最好的結果之前提到過治理以及了解我們在此過程中生成的數據類型。

That's why I said the early going, you're going to see a lot of upside from AI, that analysts are going to be able to generate data far quicker, far better than they ever have been before. And we're really massively reducing the skill sophistication and requirements to be able to do that. That, in and of itself is going to be a big driver for us.

這就是為什麼我一開始就說，你會看到人工智能的很多好處，分析師將能夠比以前更快、更好地生成數據。為了做到這一點，我們確實大幅降低了技能的複雜性和要求。這本身將成為我們的一大推動力。

Yes. Christian here, I would add maybe 2 areas in addition to what Frank mentioned. The first one is around having the right data to be fed into these models. Frank started the call with no AI strategy without data strategy. And it is very visible that the results of traditional ML or gen AI is a function of having the right data, the right data quality, the right metrics. The technology will be as good as the data that is fed. And so all of the investments that we make on data quality and cleansing and pipelines, all of that is very important.

是的。克里斯蒂安，除了弗蘭克提到的內容之外，我可能還想補充兩個方面。第一個是將正確的數據輸入到這些模型中。 Frank 開始通話時沒有人工智能策略，也沒有數據策略。很明顯，傳統機器學習或新一代人工智能的結果取決於擁有正確的數據、正確的數據質量和正確的指標。該技術將與所提供的數據一樣好。因此，我們在數據質量、清理和管道方面進行的所有投資都非常重要。

The other piece that I think will be a technical imperative for everyone doing AI and gen AI is around the measurement and feedback, how good are the solutions? How do I know if there are potential biases into this data or there are gaps in their understanding and performance of the model? Those 2 are inherent parts of the life cycle and interestingly enough, they all run to having a great data foundation enabled service.

我認為對於每個從事人工智能和新一代人工智能的人來說，技術上的當務之急是測量和反饋，解決方案有多好？我如何知道這些數據是否存在潛在偏差，或者他們對模型的理解和性能是否存在差距？這兩個是生命週期的固有部分，有趣的是，它們都運行以擁有出色的數據基礎支持服務。

Helpful color. And then lastly for Mike on consumption, one follow-up. The implied Q4 product growth is, I think, 26% at the midpoint. I know there's a delta between signing growth and consumption. Exiting this year, do you think product growth stabilizes in the mid-20% range, maybe starts to reaccelerate next year? Or is it just too early to tell?

有用的顏色。最後是邁克關於消費的一個後續行動。我認為，第四季度產品隱含增長率的中點為 26%。我知道簽約增長和消費之間存在差異。今年退出後，您認為產品增長是否穩定在 20% 左右，也許明年會開始重新加速？或者現在說還為時過早？

Let us finish Q3, and then we'll guide to, and I'll see how next year is looking. But I do anticipate -- there's a lot of new things coming out next year that we think are going to have a very positive impact on our consumption from -- remember, we have Streamlit goes into GA, Unistore towards the end of this year on our public preview. We have Containerized Services next year. There's a number of things that are happening that are all going to have a positive impact on our revenue growth rate next year. So stay tuned for that.

讓我們完成第三季度，然後我們將進行指導，我將看看明年的情況如何。但我確實預計 - 明年將會出現很多新事物，我們認為這將對我們的消費產生非常積極的影響 - 請記住，我們有 Streamlit 進入通用版，Unistore 將於今年年底進入通用版我們的公開預覽。明年我們將提供容器化服務。正在發生的許多事情都將對我們明年的收入增長率產生積極影響。所以請繼續關注。

The next question will be from the line of Brad Reback with Stifel.

下一個問題將來自 Brad Reback 和 Stifel 的線路。

Gentlemen, you've talked about changing sentiment a couple of times during the call. How much of that is your sales team being better able to engage with the customers selling value, just looking at the problem from a different perspective given the macro trends versus customers just feeling better about their businesses and the macro unleashing the demand.

先生們，你們在電話會議中多次談到了改變情緒的問題。其中有多少是你的銷售團隊能夠更好地與銷售價值的客戶互動，只是從不同的角度看待問題，考慮到宏觀趨勢，而不是客戶對他們的業務和釋放需求的宏觀感覺更好。

It's Frank. Look, it's not the sales team. There's really a change in how customers engaged a couple of quarters ago. And like I said earlier, people were doing unnatural acts to force that themselves into the spending envelope, and they were doing it almost regardless of consequence. The expectation on that reset, we obviously felt that. The change in sentiment is that, that has passed. We are now sort of, "Okay, we're comfortable with the path that we're on. Now we're talking again of our projects and migrations and use cases. We're trying to basically get a grip on deploying large language models. What do we have to do with the data, with the infrastructure, answering governance questions and so on?" So in other words, the sentiment is very constructive and engaging on core data strategy. That's a big change from where we were a couple of quarters ago. Obviously, the sales people are perceiving that as very positive. That's where you want to be.

是弗蘭克。看，這不是銷售團隊。幾個季度前，客戶的參與方式確實發生了變化。正如我之前所說，人們正在採取不自然的行為來迫使自己進入支出範圍，而且他們幾乎不顧後果地這樣做。我們顯然感受到了對重置的期望。情緒的變化就是，那已經過去了。我們現在有點，“好吧，我們對我們所走的道路感到滿意。現在我們再次談論我們的項目、遷移和用例。我們正在努力基本上掌握大型語言的部署模型。我們與數據、基礎設施、回答治理問題等有什麼關係？”換句話說，這種情緒非常有建設性，並且對核心數據策略很有吸引力。與幾個季度前相比，這是一個巨大的變化。顯然，銷售人員認為這是非常積極的。那就是你想去的地方。

The next question will be from the line of Derrick Wood with TD Cowen.

下一個問題將來自 Derrick Wood 和 TD Cowen。

Mike, this is the strongest sequential growth quarter you've had in 3 quarters. I think you were at 6%, then 6% and this was 8.5%. Your guidance for Q3 is kind of 5.5% sequential. I'm sure there's some level of conservatism in there. But just in terms of the Q2, were there any kind of onetime consumption dynamics to call out? Or does that just really kind of inform us that the optimization headwinds that you saw in Q4 and Q1 kind of dissipated in Q2?

邁克，這是您三個季度以來最強勁的連續增長季度。我想你當時是 6%，然後是 6%，現在是 8.5%。你們對第三季度的指導是 5.5% 的連續增長。我確信其中存在一定程度的保守主義。但僅就第二季度而言，是否存在任何一次性消費動態值得關注？或者這是否真的告訴我們，您在第四季度和第一季度看到的優化逆風在第二季度消散了？

Yes. Q2 has more days in it. Remember where consumption model was actually quarter-over-quarter. Working days adjusted was the growth rate. And actually, Q3, it goes up when you look at the working days in Q3. Remember, on a consumption model, where really kind of 70% is the scheduled work, there is a big piece that is tied to workdays. That does have an impact. So there is growth next quarter in that guide on a days adjusted basis, working days.

是的。第二季度還有更多天。請記住消費模式實際上季度環比的情況。調整後的工作日為增長率。實際上，第三季度，當你查看第三季度的工作日時，它會上升。請記住，在消費模型中，實際上 70% 是計劃工作，其中很大一部分與工作日相關。這確實有影響。因此，該指南下季度的調整天數（工作日）將會有所增長。

Okay. Frank or Christian, I just -- I was hoping to double click on the new Container Services. It seems like it enables you to deploy different types of third-party engines, apps, code bases directly in the platform, you can streamline a lot more workflows, which I think when it comes to building AI model seems pretty interesting. So I was just curious, what are you most excited about in terms of this new capability and opening up new consumption, especially when it comes to AI?

好的。弗蘭克或克里斯蒂安，我只是 - 我希望雙擊新的容器服務。看起來它使您能夠直接在平台中部署不同類型的第三方引擎、應用程序、代碼庫，您可以簡化更多工作流程，我認為在構建人工智能模型時這似乎非常有趣。所以我很好奇，在這種新能力和開放新消費方面，特別是在人工智能方面，您最興奮的是什麼？

Yes. Container Services was the absolute hit star performer at our Snowflake Summit Conference. I mean, customers were just mesmerized by the possibilities that a platform capability has because we have essentially eliminated any limitation on deployment on Snowflake. And why do you care? I mean, the thing is, first of all, you want to deploy close to the data for all the reasons that we talk about. This enabled us -- you get a fully trusted sanctioned platform where you can deploy applications without any further questions. I mean one of the challenges that you have in cloud computing is who's managing this, right? I mean what is the safe space to deploy into and who is really guaranteeing the high-trust enterprise-grade capabilities of that platform.

是的。集裝箱服務公司絕對是雪花峰會上的熱門明星。我的意思是，客戶只是對平台功能的可能性著迷，因為我們基本上消除了 Snowflake 上部署的任何限制。你為什麼關心？我的意思是，首先，出於我們討論的所有原因，您希望部署靠近數據。這使我們能夠獲得一個完全值得信賴的認可平台，您可以在其中部署應用程序而無需任何進一步的問題。我的意思是，雲計算面臨的挑戰之一是誰來管理它，對嗎？我的意思是部署的安全空間是什麼以及誰真正保證該平台的高信任企業級功能。

So we are bringing that. So we're going to see a lot of services. A lot of them could be on-premise legacy engines that are going to be containerized and reserviced as a cloud service, right? So a lot of things that were old will be new again. So the it's virtualization for the cloud and having secure, safe, high performance, very, very efficient spaces to run services and applications in. So the sky is the limit on this capability, and we and our customers and our partners could not be more excited about the potentials and the possibilities here.

所以我們帶來了這個。所以我們會看到很多服務。其中很多可能是本地遺留引擎，將被容器化並作為雲服務重新提供服務，對吧？所以很多舊的東西都會重新煥然一新。因此，它是雲的虛擬化，具有安全、可靠、高性能、非常非常高效的空間來運行服務和應用程序。因此，這種能力的極限是無限的，我們、我們的客戶和我們的合作夥伴不能再多了。對這裡的潛力和可能性感到興奮。

But specific to AI, and this matters a whole lot. Because the containers are vehicle, are a vessel, if you will, to deploy large line models, there's no limit on which models and how many models and for which segments of the business we can deploy. And we can shift gears very, very quickly, and we have incredible flexibility in terms of deploying these capabilities because you're going to see a lot of change and a lot of movement. We've already seen an enormous amount. That's going to continue. So we're very, very well positioned architecturally platform-wise to enable the AI revolution with container services.

但具體到人工智能，這非常重要。因為集裝箱是車輛，是船隻，如果您願意的話，要部署大型生產線模型，我們可以部署哪些模型、多少模型以及針對哪些業務部分，沒有限制。我們可以非常非常快地轉變方向，並且在部署這些功能方面我們擁有令人難以置信的靈活性，因為你會看到很多變化和很多變動。我們已經看到了大量的情況。這種情況將會繼續下去。因此，我們在架構平台方面處於非常非常有利的位置，可以通過容器服務實現人工智能革命。

And the final question will be from the line of Sterling Auty with Moffat Nathanson.

最後一個問題將由斯特林·奧蒂（Sterling Auty）和莫法特·內桑森（Moffat Nathanson）提出。

Just one question from my side. You mentioned sales productivity a couple of times. I'm curious, how would you grade your go-to-market sales execution in the quarter? And are there any specific changes that you're making to further optimize, given the environment for the back half?

我這邊只有一個問題。您多次提到銷售效率。我很好奇，您如何評價本季度的上市銷售執行情況？考慮到後半部分的環境，您是否正在做出任何具體的改變來進一步優化？

I think in general, our execution in Q2 on the sales side was actually quite good. It improved. But there's still pockets though, where there is room for improvement when you look in certain territories or geos. We have a new leader in EMEA. EMEA -- certain markets in EMEA are doing good. others, there's a lot of room for improvement. We have a new leader in EMEA. There are certain pockets in Asia that are doing good, but there's others that have a lot of room for improvement. But in general, overall, as I talked about, you can see through our bookings, it was a good execution from a bookings perspective in the aggregate last quarter.

我認為總的來說，我們第二季度在銷售方面的執行實際上相當不錯。它有所改善。但當你在某些地區或地理區域觀察時，仍然存在一些改進的空間。我們在歐洲、中東和非洲地區有了一位新的領導者。歐洲、中東和非洲——歐洲、中東和非洲的某些市場表現良好。其他方面，還有很大的改進空間。我們在歐洲、中東和非洲地區有了一位新的領導者。亞洲的某些地區表現良好，但也有一些地區還有很大的改進空間。但總的來說，正如我所說，您可以通過我們的預訂看到，從上季度總體預訂的角度來看，這是一個很好的執行。

Thank you. That will conclude today's Q&A session and today's conference call. Thank you all for your participation, and you may now disconnect your lines.

謝謝。今天的問答環節和電話會議到此結束。感謝大家的參與，現在您可以斷開線路了。