(SNOW) 2024 Q4 法說會逐字稿

Hello, and welcome to the Q4 Fiscal Year '24 Snowflake Earnings Call. My name is Alex. I'll be coordinating the call today. (Operator Instructions)

您好，歡迎參加 '24 財年第 4 季雪花收益電話會議。我的名字叫亞歷克斯。我今天將協調通話。 （操作員說明）

I'll now hand over to your host, Katherine McCracken, Senior Manager, Investor Relations. Please go ahead.

我現在將會議交給東道主投資者關係高級經理凱瑟琳·麥克拉肯 (Katherine McCracken)。請繼續。

Good afternoon, and thank you for joining us on Snowflake's Q4 Fiscal 2024 Earnings Call. With me in Bozeman, Montana are Sridhar Ramaswamy, our Chief Executive Officer; Frank Slootman, our Chairman; Mike Scarpelli, our Chief Financial Officer; and Christian Kleinerman, our Executive Vice President of Product, who will join us for the Q&A session.

下午好，感謝您參加 Snowflake 的 2024 年第四季財報電話會議。和我一起在蒙大拿州博茲曼的是我們的執行長斯里達爾·拉馬斯瓦米 (Sridhar Ramaswamy)；法蘭克‧斯洛特曼，我們的主席； Mike Scarpelli，我們的財務長；我們的產品執行副總裁 Christian Kleinerman 將參加我們的問答環節。

During today's call, we will review our financial results for the fourth quarter of fiscal 2024 and discuss our guidance for the first quarter and full year fiscal 2025. During today's call, we will make forward-looking statements, including statements related to our business operations and financial performance. These statements are subject to risks and uncertainties, which could cause them to differ materially from actual results. Information concerning these risks and uncertainties is available in our earnings press release, our most recent Forms 10-K and 10-Q and our other SEC reports. All our statements are made as of today based on information currently available to us. Except as required by law, we assume no obligation to update any such statements.

在今天的電話會議中，我們將回顧2024 財年第四季的財務業績，並討論2025 財年第一季和全年的指導。在今天的電話會議中，我們將做出前瞻性陳述，包括與我們的業務營運相關的陳述和財務表現。這些陳述存在風險和不確定性，可能導致它們與實際結果有重大差異。有關這些風險和不確定性的資訊可在我們的收益新聞稿、最新的 10-K 和 10-Q 表格以及其他 SEC 報告中找到。截至今天，我們的所有聲明都是根據我們目前掌握的資訊做出的。除法律要求外，我們不承擔更新任何此類聲明的義務。

During today's call, we will also discuss certain non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures is included in today's earnings press release. The earnings press release and an accompanying investor presentation are available on our website at investors.snowflake.com. A replay of today's call will also be posted on the website.

在今天的電話會議中，我們還將討論某些非公認會計準則財務指標。今天的收益新聞稿中包含了 GAAP 與非 GAAP 指標的調整表。收益新聞稿和隨附的投資者簡報可在我們的網站 Investors.snowflake.com 上取得。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, Katherine. Welcome, and good afternoon. By now you've heard the great news about Sridhar becoming our next CEO. Before we get to that, I would like to highlight our fiscal 2024 results.

謝謝，凱瑟琳。歡迎光臨，下午好。到目前為止，您已經聽說了 Sridhar 成為我們下一任執行長的好消息。在此之前，我想先強調一下我們 2024 財年的表現。

FY '24 product revenue grew 38% year-over-year to reach $2.67 billion. Non-GAAP product gross margin expanded to 77.8%. Non-GAAP adjusted free cash flow was $810 million, representing 56% year-over-year growth. We continue to pay our high growth with efficiency. The year began against an unsettled macroeconomic backdrop. We witnessed lackluster sentiment and customer hesitation due to lack of visibility in their businesses. Customers prefer to wait and see posture versus leaning into longer-term contract expansions. This reversed in the second half of the year, and we started seeing larger multiyear commitments.

24 財年產品營收年增 38%，達到 26.7 億美元。非 GAAP 產品毛利率擴大至 77.8%。非 GAAP 調整後自由現金流為 8.1 億美元，較去年同期成長 56%。我們繼續以效率來回報我們的高成長。今年伊始，宏觀經濟背景不穩定。我們目睹了由於業務缺乏可見性而導致的低迷情緒和客戶猶豫。客戶更願意觀望，而不是傾向於長期合約擴張。這種情況在下半年發生了逆轉，我們開始看到更大的多年承諾。

Q4 was an exceptionally strong bookings quarter. We reported $5.2 billion of remaining performance obligations, representing accelerated year-on-year growth of 41%. Our international theaters outperformed the company as a whole. We continue to see success in our effort to campaign the largest enterprises globally. We added 14 Global 2000s in the quarter and 8 of our top 10 customers grew sequentially. Meanwhile, Snowflake has announced many new technologies that let customers mobilize AI. Streamlit and Snowflake -- Snowpark ML Modeling API and Cortex ML functions are all generally available. We also received FedRAMP High Authorization on the AWS GovCloud. This enables Snowflake to protect some of the federal government's most sensitive, unclassified data.

第四季的預訂量異常強勁。我們報告的剩餘履約義務為 52 億美元，較去年同期加速成長 41%。我們的國際影院的表現優於整個公司。我們在爭取全球最大企業的努力中不斷取得成功。本季我們新增了 14 家全球 2000 強客戶，前 10 名客戶中有 8 家連續成長。同時，Snowflake 宣布了許多讓客戶調動人工智慧的新技術。 Streamlit 和 Snowflake－Snowpark ML 建模 API 和 Cortex ML 功能都已普遍可用。我們也在 AWS GovCloud 上獲得了 FedRAMP 高授權。這使得 Snowflake 能夠保護聯邦政府的一些最敏感的非機密資料。

Now on the topic of CEO transition. I was brought to Snowflake 5 years ago to help the company break out and scale. I wanted to grow the business fast but not at all cost. It has to be efficient and establish a foundation for long-term growth. I believe the company succeeded on that mission. The Board has run a succession process that wasn't based on arbitrary time line, but instead look for an opportunity to advance the company's mission well into the future. The arrival of Sridhar Ramaswamy through the acquisition of Neeva last year represented that opportunity. Since joining us, Sridhar has been leading Snowflake's AI strategies, bringing new products and features to market at an incredible pace. He led the launch of Snowflake Cortex like new fully managed service that makes AI simple and secure.

現在談談執行長過渡的話題。五年前，我被帶到 Snowflake，幫助公司突破並擴大規模。我想快速發展業務，但不是不惜一切代價。它必須高效並為長期成長奠定基礎。我相信公司成功地完成了這個使命。董事會的繼任流程並非基於任意時間線，而是尋找機會將公司的使命推進到未來。去年，斯里達爾·拉馬斯瓦米 (Sridhar Ramaswamy) 透過收購 Neeva 的到來就代表了這個機會。自從加入我們以來，Sridhar 一直領導 Snowflake 的人工智慧策略，以令人難以置信的速度將新產品和功能推向市場。他領導推出了 Snowflake Cortex 之類的新型完全託管服務，讓人工智慧變得簡單又安全。

Prior to Neeva, Sridhar led all of Google's advertising products. During his 15-year tenure at Google, he helped grow AdWords in Google's advertising business from $1.5 billion to over $100 billion. With the onslaught of generative AI, Snowflake needs hard-driving technologists to navigate the challenges the new world represents. Sridhar's vision for the future and his proven ability to execute at scale made it clear to us as a Board that he is the right executive at the right time to lead Snowflake.

在加入 Neeva 之前，Sridhar 負責領導 Google 的所有廣告產品。在 Google 的 15 年任期中，他幫助 Google 廣告業務中的 AdWords 從 15 億美元成長到超過 1,000 億美元。隨著生成式人工智慧的衝擊，Snowflake 需要積極進取的技術人員來應對新世界帶來的挑戰。 Sridhar 對未來的願景以及他經過驗證的大規模執行能力讓我們作為董事會清楚地認識到，他是在正確的時間領導 Snowflake 的正確的高管。

This marks my retirement from an operating role, I will remain on duty as Chairman of the Board and look forward to working with Sridhar and the team going forward.

這標誌著我從營運職位上退休，我將繼續擔任董事會主席，並期待與 Sridhar 和團隊繼續合作。

With that, I will pass it over to Sridhar.

這樣，我將把它交給斯里達爾。

Thank you, Frank. I'm honored to have been chosen to lead this great company. The success Snowflake has achieved is a testament to the great customers, employees and partners who have contributed along the way. And of course, Frank has been a huge part of getting us here, which I gratefully acknowledge.

謝謝你，弗蘭克。我很榮幸被選來領導這家偉大的公司。 Snowflake 所取得的成功證明了一路以來做出貢獻的偉大客戶、員工和合作夥伴。當然，弗蘭克為我們走到這一步做出了很大的貢獻，對此我深表謝意。

Snowflake is a once-in-a-generation company that will revolutionize the world with its cloud data platform. This has become more true in the past year with the rapid technology innovations we have seen. Generative AI is at the forefront of my customer conversations. This drives renewed emphasis on data strategy in preparation of these new technologies. You heard the team say it many times, there is no AI strategy without a data strategy. And this has opened a massive opportunity for Snowflake to address. To deliver on the opportunity ahead, we must have clear focus and move even faster to bring innovation on the Snowflake platform to our customers and partners. This will be my focus. I look forward to working with the team, and I'm extremely excited for the opportunity.

Snowflake 是一家千載難逢的公司，將透過其雲端資料平台徹底改變世界。在過去的一年裡，隨著我們所看到的快速技術創新，這一點變得更加真實。生成式人工智慧是我與客戶對話的最前線。這促使人們重新重視準備這些新技術的數據策略。你多次聽到團隊說過，沒有數據策略就沒有人工智慧策略。這為 Snowflake 提供了一個巨大的解決機會。為了抓住未來的機遇，我們必須有明確的重點並更快採取行動，為我們的客戶和合作夥伴帶來 Snowflake 平台上的創新。這將是我的重點。我期待與團隊合作，對於這個機會我感到非常興奮。

With that, I'll pass it over to Mike.

有了這個，我會把它交給麥克。

Thank you, Sridhar. Q4 marked a strong finish to a challenging year. Product revenue was $738 million, growing 33% year-over-year. Similar to prior years, we experienced significant holiday impacts in December and January. Holidays make it difficult to discern meaningful consumption trends. In the quarter, younger customers led revenue growth. These accounts are adding new workloads and migrating from legacy vendors. Financial services and retail were our largest revenue contributors, and we are seeing emerging momentum from the EMEA region and technology vertical.

謝謝你，斯里達爾。第四季為充滿挑戰的一年畫上了圓滿的句點。產品收入為 7.38 億美元，年增 33%。與往年類似，我們在 12 月和 1 月經歷了重大的假期影響。假期讓我們很難辨別有意義的消費趨勢。本季度，年輕客戶引領營收成長。這些帳戶正在新增新的工作負載並從傳統供應商遷移。金融服務和零售是我們最大的收入貢獻者，我們看到歐洲、中東和非洲地區和技術垂直領域的新興勢頭。

Customer optimizations returned to a normal level, with 8 of our top 10 accounts growing sequentially. We proactively engaged with customers to help them optimize their Snowflake usage, and will continue to do so. History has shown that optimizations expand our long-term opportunity. We now have 83 customers with trailing 12-month product revenue greater than $5 million, up from $75 million in Q3.

客戶優化恢復到正常水平，前 10 個客戶中有 8 個連續成長。我們積極與客戶合作，幫助他們優化 Snowflake 的使用，並將繼續這樣做。歷史表明，優化可以擴大我們的長期機會。我們現在有 83 位客戶過去 12 個月的產品收入超過 500 萬美元，高於第三季的 7,500 萬美元。

Q4 was an exceptional booking quarter for us. Bookings are not a leading indicator of revenue; they do signal an improving macro environment. Remaining performance obligations grew 41% year-over-year to $5.2 billion. Of the $5.2 billion in RPO, we expect approximately 50% to be recognized as revenue in the next 12 months. We signed our largest deal ever in Q4, a 5-year $250 million contract with an existing customer. Our international territories returned to meaningful growth, outperforming expectations for the first time in a year. We made significant progress in delivering margin expansion.

第四季對我們來說是一個特別的預訂季度。預訂量並不是收入的領先指標；它們確實預示著宏觀環境正在改善。剩餘履約義務年增 41%，達到 52 億美元。在 52 億美元的 RPO 中，我們預計大約 50% 將在未來 12 個月內確認為收入。我們在第四季簽署了有史以來最大的一筆交易，與現有客戶簽訂了一份為期 5 年、價值 2.5 億美元的合約。我們的國際領土恢復了有意義的成長，一年來首次超越預期。我們在實現利潤率擴張方面取得了重大進展。

Non-GAAP product margin of 78% was up approximately 300 basis points year-over-year. Improved terms from the cloud service providers have contributed to margin expansion. Non-GAAP operating margin of 9% was ahead of expectations. Operating margin benefit from increased hiring scrutiny. Non-GAAP adjusted free cash flow margin was 42%. Bookings outperformance increased collections. We ended the quarter with $4.8 billion in cash, cash equivalents and short-term and long-term investments. We did not repurchase any shares in Q4. We have approximately $1.4 billion remaining under our original authorization of $2 billion.

非 GAAP 產品利潤率為 78%，較去年同期成長約 300 個基點。雲端服務提供者條款的改善有助於利潤率的擴大。非 GAAP 營業利益率為 9%，超出預期。營業利潤率受益於招聘審查的加強。非 GAAP 調整後自由現金流率為 42%。預訂表現優於收藏增加。截至本季末，我們的現金、現金等價物以及短期和長期投資為 48 億美元。我們在第四季沒有回購任何股票。我們最初授權的 20 億美元還剩下約 14 億美元。

Now let's turn to outlook. Consumption trends have improved since the ending of last year, but have not returned to pre-FY '24 patterns. We have evolved our forecasting process to be more receptive to recent trends. For that reason, our guidance assumes similar customer behavior to fiscal 2024. We are forecasting increased revenue headwinds associated with product efficiency gains, tiered storage pricing and the expectation that some of our customers will leverage Iceberg Tables for their storage. We are not including potential revenue benefits from these initiatives in our forecast.

現在讓我們轉向展望。自去年年底以來，消費趨勢有所改善，但尚未恢復到 24 財年之前的模式。我們改進了預測流程，以便更容易接受最新趨勢。因此，我們的指導假設客戶行為與 2024 財年類似。我們預測，與產品效率提升、分層儲存定價以及我們的一些客戶將利用 Iceberg Tables 進行儲存的預期相關的收入阻力將會增加。我們的預測中並未包括這些舉措所帶來的潛在收入收益。

These changes in our assumptions impact our long-term guidance. Internally, we continue to march towards $10 billion in product revenue. Externally, we will not manage expectations to our previous targets until we have more data. We are focused on executing in FY '24 to ensure long-term durable growth.

我們假設的這些變化影響了我們的長期指導。在內部，我們持續朝 100 億美元的產品收入邁進。從外部來看，在獲得更多數據之前，我們不會管理對先前目標的預期。我們的重點是在 24 財年執行，以確保長期持久成長。

Now turning to FY '25 guidance. For the first quarter, we expect product revenue between $745 million and $750 million, representing year-over-year between 26% and 27%. For the first quarter, we expect non-GAAP operating margin of 3% and 366 million diluted weighted average shares outstanding. For the full year, we expect product revenue of approximately $3.25 billion, representing 22% year-over-year growth.

現在轉向 25 財年指導。我們預計第一季產品營收在 7.45 億美元至 7.5 億美元之間，年增 26% 至 27%。我們預期第一季非 GAAP 營業利益率為 3%，稀釋加權平均流通股數為 3.66 億股。我們預計全年產品收入約為 32.5 億美元，年增 22%。

We expect Snowpark to contribute 3% of product revenue. We are not including any other new products in our forecast at this time. For the full year, we expect non-GAAP product gross margin of 76%, non-GAAP operating margin of 6%, non-GAAP adjusted free cash flow margin of 29% and diluted weighted average shares outstanding of 368 million.

我們預計 Snowpark 將貢獻產品收入的 3%。目前我們的預測中不包括任何其他新產品。對於全年，我們預期非 GAAP 產品毛利率為 76%，非 GAAP 營業利益率為 6%，非 GAAP 調整後自由現金流利潤率為 29%，稀釋加權平均已發行股數為 3.68 億股。

We plan to add approximately 1,000 employees this year, inclusive of M&A. For expenses, our forecast assumes meaningful investments in our AI initiatives. We expect approximately $50 million of GPU-related costs in fiscal year '25, approximately $10 million flowing through cost of product revenue. For the purpose of forecasting, we are not including any incremental revenue associated with these features.

我們計劃今年增加約 1,000 名員工，其中包括併購。對於費用，我們的預測假設我們對人工智慧計劃進行了有意義的投資。我們預計 25 財年 GPU 相關成本約為 5,000 萬美元，其中約 1,000 萬美元將流經產品收入成本。出於預測目的，我們不包括與這些功能相關的任何增量收入。

We have also evolved our go-to-market motion. As we compensate more reps on a consumption quota, we will see increased commission expense. Consumption-based commissions are expensed immediately rather than amortized over a 5-year period. This has no impact on cash flows, but is expected to have approximately $30 million impact to P&L.

我們也制定了進入市場的行動。當我們在消費配額上補償更多的代表時，我們會看到佣金費用增加。基於消費的佣金立即支出，而不是在 5 年內攤提。這對現金流沒有影響，但預計會對損益產生約 3000 萬美元的影響。

Lastly, I would like to acknowledge Frank's retirement. Working with Frank for the past 17 years has been an incredible learning experience, and I'm grateful for our time together. Frank's contributions to Snowflake has set the company up for the long-term success, and I look forward to being part of that journey. I've committed to Sridhar and the Board that I will be with Snowflake for at least the next 3 years.

最後，我要感謝弗蘭克的退休。過去 17 年與 Frank 一起工作是一次令人難以置信的學習經歷，我很感激我們在一起的時光。 Frank 對 Snowflake 的貢獻為公司的長期成功奠定了基礎，我期待成為這趟旅程的一部分。我已向 Sridhar 和董事會承諾，我將在 Snowflake 工作至少 3 年。

Before closing, we will host our Investor Day on June 4 in San Francisco, in conjunction with Data Cloud Summit, our annual users conference. If you're interested in attending, please e-mail ir@snowflake.com.

在交易結束之前，我們將於 6 月 4 日在舊金山舉辦投資者日活動，同時舉辦我們的年度用戶大會資料雲高峰會。如果您有興趣參加，請發送電子郵件至 ir@snowflake.com。

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

接線員，現在可以撥打電話提問了。

(Operator Instructions) Our first question for today comes from Mark Murphy of JPMorgan.

（操作員指令）我們今天的第一個問題來自摩根大通的馬克墨菲。

So Mike, I think with the understanding that bookings are not a perfect forward indicator of revenue, it's hard not to notice the total bookings performance is quite spectacular. And even the short-term backlog number looks good. I think we try to bridge from that, growing closer to 30%. And then to arrive at the revenue growth in the low 20s, I understand it's mechanically more complicated than that. But on the surface, it looks like you're guiding with a lot more conservatism than a year ago. So I'm just wondering, does it feel -- does that part of this feel a little more comfortable to you? Or do you think it could take a little longer to convert bookings to consumption this year?

麥克，我認為，在了解預訂量並不是完美的收入前瞻性指標的情況下，很難不注意到總預訂量的表現相當驚人。甚至短期積壓訂單數量看起來也不錯。我認為我們正在努力彌補這一點，將這一比例增長到接近 30%。然後要實現 20 多歲左右的收入成長，我知道這在機制上比這更複雜。但從表面上看，你的指導方針似乎比一年前更保守。所以我只是想知道，這部分對你來說是否感覺更舒服一些？或者您認為今年將預訂轉化為消費可能需要更長的時間？

I think we are definitely being more conservative this year, given the consumption patterns we saw in '24. And as we said at our Analyst Day last year, we needed to see consumption patterns more in line with what we saw pre-'24 to get to our longer-term goal. And as a result, we've decided to forecast this year based upon the consumption patterns we saw in '24. And as you know, we forecast based upon historical consumption of our products. There's a lot of new products coming into GA in public preview this year that we have taken into account in our forecast. And we will do so once we start to see that consumption. And so we will take this quarter-by-quarter for the year.

我認為，考慮到我們在 24 年看到的消費模式，今年我們肯定會更加保守。正如我們在去年的分析師日所說，我們需要看到消費模式與 24 世紀 24 年前的情況更加一致，才能實現我們的長期目標。因此，我們決定根據 24 年的消費模式對今年進行預測。如您所知，我們根據我們產品的歷史消費量進行預測。今年有許多新產品進入 GA 公開預覽版，我們在預測中已考慮到這些新產品。一旦我們開始看到這種消費，我們就會這樣做。因此，我們將逐季看待這一年的情況。

Okay. And then as a quick follow-up, Frank, I wanted to thank you for everything and wish you all the best. And Sridhar, just wanted to ask you because of your background in generative AI. Do you envision any changes in the technology road map perhaps relating to Snowpark or perhaps your role in LLMs or how heavily you'd see the go-to-market in those areas. I'm just wondering if there's -- if you foresee any change in philosophy or approach?

好的。然後，弗蘭克，我想快速跟進一下，我想感謝你所做的一切，並祝你一切順利。 Sridhar，只是想問你，因為你有產生人工智慧的背景。您是否預計技術路線圖會發生任何變化，這些變化可能與 Snowpark 相關，或者可能與您在法學碩士中的角色有關，或者您會在這些領域看到多大程度的市場推廣。我只是想知道您是否預見到哲學或方法會發生任何變化？

Yes. I've had 100 conversations with customers over the past year about generative AI, in particular. And the product announcements that we've already made, things in private preview including Snowflake Cortex, which is our managed AI and search layer combined with applications like Document AI or extracting structured data or our co-pilot. These have been very well received. Document AI, for example, has hundreds of customers waiting for it to hit GA that are on our waitlist. So I would say it is a matter of executing to the road map that we have already laid out. Cortex will hit public preview soon. To Mike's point, getting this to GA, getting this in the hands of our customers and having them realize value is the top priority. I don't think of this as needing a new strategy.

是的。在過去的一年裡，我與客戶進行了 100 次對話，尤其是關於生成式 AI。我們已經發布的產品公告，包括 Snowflake Cortex 在內的私人預覽版，它是我們的託管人工智慧和搜尋層，與文件人工智慧或提取結構化資料或我們的副駕駛等應用程式相結合。這些都非常受歡迎。例如，Document AI 在我們的候補名單上有數百名客戶正在等待正式發布。所以我想說，這是執行我們已經制定的路線圖的問題。 Cortex 即將推出公開預覽版。就麥克而言，將其交付給 GA、將其交到我們的客戶手中並讓他們實現價值是首要任務。我不認為這需要新的策略。

Our next question comes from Keith Weiss of Morgan Stanley.

我們的下一個問題來自摩根士丹利的基斯‧韋斯。

Excellent. This is Theo on for Keith. Maybe we just start off with one question, your consumption. I was wondering if you could give us a sense for where the consumption was softer as it is impacting your guidance. I think last quarter, you gave us a good sense for digital native seem to stabilize. Enterprise seems to get better. So can you just kind of double-click where the consumption is softening now versus your expectations and where may be stronger?

出色的。這是西奧替基思發言。也許我們從一個問題開始，即您的消費。我想知道您能否讓我們了解哪裡的消費疲軟，因為它影響了您的指導。我認為上個季度，您給了我們一種良好的感覺，即數位原生代似乎趨於穩定。企業似乎變得更好了。那麼，您是否可以雙擊當前消費與您的預期相比正在疲軟的地方以及可能會更強的地方？

Well, as I said, the strength was in financial services and retail. We are seeing the technology vertical do well. And I wouldn't say it's soft. We did beat the last quarter. But what I would say is it's improving, but it's not back to the pre-2024 levels, as I mentioned. And that's the -- we're basing. We've revised our model to look at more recent history rather than going back too far in history for forecasting consumption patterns.

嗯，正如我所說，優勢在於金融服務和零售。我們看到垂直技術表現良好。我不會說它很軟。我們確實擊敗了上個季度。但我想說的是，它正在改善，但並沒有回到 2024 年之前的水平，正如我所提到的。這就是我們的基地。我們修改了模型，著眼於更近期的歷史，而不是回溯太久的歷史來預測消費模式。

Got it. That's very helpful. And then one more follow-up on sort of the existing customer business and your net retention rate. That seems to be on top of the bookings and other KPI that is at least getting better on the rate of change basis. Is there any level that you would guide us to in terms of where that net retention rate could stabilize? Is that informing your guide aside from kind of the consumption you're seeing in any meaningful way? Any color you can give us on that would be helpful, too.

知道了。這非常有幫助。然後再跟進現有客戶業務和淨保留率。這似乎是在預訂量和其他 KPI 之上，至少在變化率的基礎上有所改善。您是否可以指導我們達到淨保留率穩定的水平？除了您所看到的消費類型之外，這是否以任何有意義的方式告知您的導遊？您可以給我們的任何顏色也會有幫助。

Well, as we've said before, over time, net revenue retention will converge with our revenue growth rate. And as I've said before, I'm not going to guide to net revenue retention.

好吧，正如我們之前所說，隨著時間的推移，淨收入保留率將與我們的收入成長率趨同。正如我之前所說，我不會指導淨收入保留。

Our next question comes from Raimo Lenschow of Barclays.

我們的下一個問題來自巴克萊銀行的 Raimo Lenschow。

Frank, all the best from me as well. A question for Sridhar. If you think about the customer conversations with AI and people will think differently about the data and the data platform. Can you speak a little bit about how do you see that playing out for you guys from a Snowflake perspective in terms of the one part that you have like the data warehouse, but also then more on the league side. Like what do you see -- what are you hearing customer conversations? And how are you positioned now? And what would be the push for you there? And then I have one follow-up for Mike.

法蘭克，我也祝你一切順利。問斯里達爾一個問題。如果你考慮一下客戶與人工智慧的對話，人們會對數據和數據平台有不同的看法。您能否從雪花的角度談談您如何看待這對您來說，不僅涉及資料倉儲等部分，而且還涉及聯盟方面。就像您所看到的—您聽到的客戶對話是什麼？你現在的定位如何？那裡對你的推動力是什麼？然後我有一個關於麥克的後續行動。

Yes. On Snowflake Cortex, we are implementing it as a core platform layer. It ships with every deployment and it makes AI readily accessible from SQL, so that even an analyst that's not an LLM expert or a Python expert can simply write SQL for things like summarization or sentiment detection data that is already in Snowflake.

是的。在 Snowflake Cortex 上，我們將其實作為核心平台層。它隨每次部署一起提供，使 AI 可以輕鬆地從 SQL 訪問，因此即使不是 LLM 專家或 Python 專家的分析師也可以簡單地為 Snowflake 中已有的摘要或情緒檢測資料等內容編寫 SQL。

Our overall aspiration here is to make AI really, really simple for our customers to use. And in some sense, the prototypical AI application is a chatbot over a specialized corpus using what's called RAG, retrieval-augmented generation. But the idea is basically you're able to, say, talk to your documentation or talk to the support cases that you have and get answers back in natural language, but with things like citation, so you can actually believe the answers that are coming back from the chatbot. So both of these are ready-made applications that our customers are excited about.

我們的整體願望是讓人工智慧對我們的客戶來說真的非常簡單。從某種意義上說，原型人工智慧應用程式是一個基於專門語料庫的聊天機器人，使用所謂的 RAG（檢索增強生成）。但這個想法基本上是，你可以與你的文件交談或與你擁有的支持案例交談，並以自然語言獲得答案，但通過引用之類的東西，這樣你就可以真正相信即將到來的答案從聊天機器人回來。因此，這兩個都是現成的應用程序，我們的客戶對此感到興奮。

But I would say the big, big unlock is being able to get at the structured data that is in Snowflake and have that be accessed by many, many more people. Today, people go through an elaborate process of getting the data ready using BI tools, going through a pretty slow cycle. And so the thing that we are driving towards is creating easy ways for people to be able to talk to subsets of data, like Mike to be able to talk to finance data or other things like that. I would say that, that is the thing that is truly, truly exciting for our customers.

但我想說，最大的解鎖是能夠取得 Snowflake 中的結構化數據，並讓更多人存取這些數據。如今，人們使用 BI 工具經歷了一個精心準備資料的過程，經歷了一個相當緩慢的週期。因此，我們正在努力創造簡單的方法，讓人們能夠與資料子集對話，例如麥克能夠與財務數據或其他類似的東西對話。我想說，這才是真正令我們的客戶興奮的事。

And with respect to where data is sitting, as you know, we support things like Iceberg formats is getting more and more popular. So data interoperability is very much a thing and our AI products can generally act on data that is sitting in cloud storage as well. So the solutions that we are offering as part of Snowflake are broadly applicable, both to data within Snowflake but also data that's sitting in cloud storage.

至於資料的位置，如您所知，我們支援像 Iceberg 格式這樣的東西變得越來越流行。因此，資料互通性非常重要，我們的人工智慧產品通常也可以對雲端儲存中的資料進行操作。因此，我們作為 Snowflake 的一部分提供的解決方案廣泛適用，既適用於 Snowflake 中的數據，也適用於雲端儲存中的數據。

Okay. Perfect. And then, Mike, a quick question for you. On the gross margin side, congrats on the amazing improvement there. Is that -- was that -- did I understand you correctly? Is that just one hyperscale area where the contract terms kind of got improved? And is that something that we could kind of hope for others? Or is the level now, the level?

好的。完美的。然後，麥克，問你一個簡單的問題。在毛利率方面，恭喜您取得了驚人的進步。我對你的理解正確嗎？這只是合約條款得到改善的超大規模領域嗎？這是我們對其他人的希望嗎？或者說現在的水平，水平？

So as a reminder, we did AWS and Azure before our Investor Day last year. We just did our Google contract last quarter as well. So all 3 of those contracts now, and those are where we are running in those 3 clouds, are current. And I don't expect any changes to those terms in any meaningful way in the short term.

提醒一下，我們在去年的投資者日之前就做了 AWS 和 Azure。我們上個季度剛完成了與Google的合約。因此，現在所有 3 個合約以及我們在這 3 個雲端中運行的合約都是最新的。我預計這些條款在短期內不會發生任何有意義的變化。

Our next question comes from Kirk Materne from Evercore.

我們的下一個問題來自 Evercore 的 Kirk Materne。

Frank, congrats on a great run at Snowflake and before. Mike, maybe for you, every year, you've told us that you'd always include technology advancements that you pass back to customers in your guidance. It seems like there's a higher level of conservatism around that, just given the lack of visibility on that. Can you try to, I guess, qualitatively give us some idea of how much bigger a headwind that is than prior years just to go along with some of the conservatism around consumption trends?

弗蘭克，恭喜雪花及之前的出色表現。麥克，也許對您來說，每年您都告訴我們，您總是將技術進步納入您的指導中，並將其回饋給客戶。考慮到這一點缺乏可見性，似乎存在更高程度的保守主義。我想，您能否試著從定性上讓我們了解，僅僅跟隨圍繞消費趨勢的一些保守主義，與前幾年相比，阻力有多大？

It's about a 6.2%, 6.3% impact this year. And -- but coupled with that, too, in the revenue, we rolled out in Q4 tiered storage pricing. So the amount of revenue associated with storage is coming down. But on top of that, we do expect a number of our large customers are going to adopt Iceberg formats and move their data out of Snowflake where we lose that storage revenue and also the compute revenue associated with moving that data into Snowflake. We do expect, though, there'll be more workloads that will move to us, but until we see that incremental revenue on workloads, we're not going to forecast that. I will say, last year, we saw a 62% increase in the number of jobs running on Snowflake year-over-year with a corresponding 33% increase in revenue. And that's because we continue to show our customers that we become cheaper and cheaper to them every year. And when we do that, it opens up new workload opportunities for us, and we'll continue to do that.

今年的影響約為6.2%、6.3%。而且，除此之外，在營收方面，我們也推出了第四季的分層儲存定價。因此，與儲存相關的收入正在下降。但最重要的是，我們確實預計我們的許多大客戶將採用 Iceberg 格式並將他們的資料從 Snowflake 中移出，這樣我們就會失去儲存收入以及與將資料移入 Snowflake 相關的計算收入。不過，我們確實預期會有更多的工作負載轉移給我們，但在我們看到工作負載收入增加之前，我們不會對此進行預測。我想說，去年，我們看到 Snowflake 上運行的作業數量年增了 62%，收入也相應增加了 33%。這是因為我們不斷向客戶展示我們的價格每年都變得越來越便宜。當我們這樣做時，就會為我們帶來新的工作機會，我們將繼續這樣做。

Our next question for today comes from Brent Thill of Jefferies.

我們今天的下一個問題來自傑富瑞 (Jefferies) 的布倫特希爾 (Brent Thill)。

Mike, I just want to reconcile a really good RPO close acceleration in Q4 and then how you marry that with the guide. I think there's a lot of questions about the acceleration you saw and perhaps, why that may not flow through into next year.

麥克，我只想在第四季度實現非常好的 RPO 接近加速，然後如何將其與指南結合。我認為對於你所看到的加速存在著許多疑問，也許還有為什麼這種加速可能不會延續到明年。

Well, I would say, we signed a lot of large multiyear deals in Q4. I mentioned the one that was a 5-year $250 million deal. There is -- and so with customers, what that really shows is the commitment customers are making to Snowflake. And from a customer's perspective, as long as they use that -- those credits up within their contract period, they're fine. There is a lot of pent-up demand for a lot of our new products that are coming out. Sridhar talked about Document AI. We have a lot of customers that want Cortex and Snowpark container services. So I think it's going to be more back-end loaded this year. And until we see that consumption by our customers, it's hard to forecast that. So stay tuned.

嗯，我想說，我們在第四季簽署了很多大型多年期協議。我提到了一份為期 5 年、價值 2.5 億美元的交易。對客戶而言，真正體現的是客戶對 Snowflake 的承諾。從客戶的角度來看，只要他們在合約期間內使用這些積分，就可以了。我們即將推出的許多新產品都有大量被壓抑的需求。 Sridhar 談到了文件人工智慧。我們有許多客戶需要 Cortex 和 Snowpark 容器服務。所以我認為今年後端負載會更多。在我們看到客戶的消費情況之前，很難預測這一點。所以請繼續關注。

Okay. And Mike, just on linearity in Q4, I know the holidays are the holidays, but did you see things bounce back similar to what you've seen historically in Q4? Or was there anything unique closing out the quarter?

好的。麥克，就第四季度的線性而言，我知道假期就是假期，但您是否看到情況反彈，類似於您在第四季度看到的歷史情況？或者本季結束時有什麼獨特之處嗎？

Are you talking from a revenue? Or are you talking about bookings? Bookings was very strong in January. I would say from a revenue consumption pattern, I think there was a little bit more of a prolonged holiday that went into mid-January. Coming out of January, consumption is good. But once again and through today, but it's more in line with 2024 consumption versus the pre-'24 consumption patterns.

你說的是收入嗎？或者你在談論預訂嗎？一月份的預訂量非常強勁。我想說，從收入消費模式來看，我認為一月中旬有一個更長的假期。進入一月份，消費情況良好。但今天再一次強調，與 24 年前的消費模式相比，它更符合 2024 年的消費模式。

Our next question comes from Alex Zukin of Wolfe Research.

我們的下一個問題來自沃爾夫研究中心的亞歷克斯祖金。

This is Allan on. Mike, just as a follow-up about consumption in the quarter. I'd love to maybe get a comparison on February. I know we got a day left but maybe how this month shaped out compared to February of last year. And then I've got a quick follow-up. And that's from a consumption perspective.

這是艾倫。麥克，作為本季消費的後續行動。我很想在二月進行比較。我知道我們還剩下一天，但也許這個月與去年二月相比​​情況如何。然後我會進行快速跟進。這是從消費的角度來看的。

The consumption is tracking where we are, and I just gave guidance for the quarter. So I don't know what else you want me to tell you.

消費正在追蹤我們的情況，我剛剛給出了本季度的指導。所以我不知道你還想讓我告訴你什麼。

Okay. And as a follow-up, I guess, you talked about some of the free cash flow impacts from the go-to-market change around pushing consumption more. Is there any way we should be thinking about maybe the top line impacts as a result of that with the guide?

好的。我想，作為後續行動，您談到了更多推動消費的市場化變化對自由現金流的影響。我們是否應該考慮該指南可能對營收產生的影響？

What I was talking about was cash flow was the switch in the comp plan from paying people on consumption versus the booking. And it doesn't really change the free -- the cash flow associated because we're still doing the payments the same. And it has a $30 million P&L impact, but no cash flow impact is what I was referring to.

我所說的是現金流是補償計劃中從支付人們消費費用到預訂費用的轉變。它並沒有真正改變相關的自由現金流，因為我們仍然以同樣的方式進行支付。它對損益產生了 3000 萬美元的影響，但我指的是對現金流沒有影響。

And obviously, one of the reasons why we switched to paying reps more on consumption because we want reps to be driving revenue more rather than bookings necessarily. And our reps are heavily compensated on -- 2 things this year. You have approximately 35% of our reps are focused just on initial new customers, 55% are just paid on existing customers with driving consumption, identifying new workloads within customers. We have about 10% that are in a hybrid, a mix of new customers and consumption. And those are more in emerging territories or territories where maybe the installed base of customers isn't as high.

顯然，這就是我們轉向向銷售代表支付更多消費費用的原因之一，因為我們希望銷售代表更多地推動收入而不是預訂。今年我們的代表在兩件事上獲得了豐厚的報酬。我們約 35% 的銷售代表只專注於最初的新客戶，55% 的銷售代表只向現有客戶支付報酬，以推動消費，並識別客戶中的新工作負載。我們大約有 10% 處於混合狀態，即新客戶和消費的混合體。這些更多地發生在新興地區或客戶群可能不那麼高的地區。

Our next question comes from Brent Bracelin of Piper Sandler.

我們的下一個問題來自 Piper Sandler 的 Brent Bracelin。

Frank, I'll -- well wishes here in your retirement. I'll forever remember you're working with you on the first IPO. And the last one here, it's been an amazing 17-year journey. Mike, glad to know you'll be with us for another 3 years. My question here is for Sridhar. As we think about driving more AI workloads to the Snowflake platform, what's the tip of the spear going forward? Is it Cortex? Is it Snowpark? Help us understand what you're focused on accelerating AI workloads to Snowflake.

法蘭克，我在這裡祝你退休。我將永遠記得您在第一次 IPO 中與您合作。這是最後一次，這是一段令人驚嘆的 17 年旅程。麥克，很高興知道你將與我們共事三年。我的問題是問斯里達爾的。當我們考慮將更多人工智慧工作負載轉移到 Snowflake 平台時，未來的矛尖是什麼？是皮質嗎？是雪園嗎？幫助我們了解您在將 AI 工作負載加速到 Snowflake 方面的重點。

Yes. When it comes to AI, as I was outlining, first of all, simplifying it so that it is easy for our customers to use. We have Cortex, is very first thing. And a lot of things that you and I, like analysts do with text now becomes so much easier to do with language models of different sizes. But I would say the applications that truly drive customer excitement that there is incredible demand for are on the Document AI side and on the co-pilot offering.

是的。當談到人工智慧時，正如我所概述的，首先，簡化它，以便我們的客戶易於使用。我們有皮質，這是第一件事。你和我，就像分析師對文字所做的很多事情，現在使用不同大小的語言模型變得更加容易。但我想說，真正讓客戶興奮的應用程式是在文件人工智慧方面和副駕駛產品上，這些應用程式有著令人難以置信的需求。

They're different, but they've used the same underlying technologies. Document AI is about extracting structured information from unstructured documents like PDFs that every enterprise has boatloads of. And then on the other side, co-pilot finally makes real the possibility of like being able to just talk to your data, ask questions in natural language and for that to get translated underneath. To SQL for that SQL to be run against Snowflake and for you to get back a tabular answer and then soon visualizations as well.

它們是不同的，但它們使用相同的底層技術。文件人工智慧是指從非結構化文件（例如每個企業都有的 PDF）中提取結構化資訊。另一方面，副駕駛最終實現了這樣的可能性：能夠與您的數據對話，用自然語言提出問題，並在下面進行翻譯。 SQL 以便針對 Snowflake 執行該 SQL，並讓您傳回表格答案，然後很快也可以獲得視覺化結果。

It's that kind of fluid access that customers are really, really excited by. And what you get from Snowflake is that this comes out of the box. This is very easy. You can build a stream let app with it. And so it avoids all of the complicated exfiltration of data using other tools, needing to stitch things together as an IT project in order to get something done.

這種流暢的訪問讓客戶非常非常興奮。您從 Snowflake 中得到的是，它是開箱即用的。這很容易。您可以用它來建立一個 Stream Let 應用程式。因此，它避免了使用其他工具進行的所有複雜的資料洩露，需要將所有內容拼接在一起作為一個 IT 專案才能完成某些工作。

And that's the power of our platform where we know the data. We know the schema. We know all the previous queries that have been run against it and are, therefore, able to create a very effective copilot solution. So hopefully, that gives you a flavor for the kinds of conversations that we've been having with customers and what they're looking forward.

這就是我們了解數據的平台的力量。我們知道這個模式。我們知道之前對其運行的所有查詢，因此能夠創建非常有效的副駕駛解決方案。希望這能讓您了解我們與客戶進行的各種對話以及他們的期望。

Very helpful there. And then, Mike, if I just look at the guide on a per customer basis, it does look like consumption per customer could be slower this year than last year. What's the delta there? Is the primary delta the assumption that tiered storage pricing and Iceberg will put more pressure on consumption growth? Is that the primary delta? Or is there something else I'm missing?

那裡非常有幫助。然後，麥克，如果我只看每個客戶的指南，看起來今年每個客戶的消費量可能會比去年慢。那裡的三角洲是什麼？主要增量是否是分層儲存定價和 Iceberg 會為消費成長帶來更大壓力的假設？這是主要的三角洲嗎？還是我還缺其他東西？

No. There's a lot of new performance enhancements being rolled out on our software this year that are going to have an impact. There's also -- well, I'll tell you, but I really don't want to because I know you're going to ask a lot of questions. We're also rolling out the arm chip in the Azure. It's not as big as AWS' impact. That will impact that as well, too. And clearly, we do expect customers will begin to adopt iceberg table format.

不會。今年我們的軟體推出了許多新的效能增強功能，這些功能將會產生影響。還有——好吧，我會告訴你，但我真的不想告訴你，因為我知道你會問很多問題。我們也在 Azure 中推出了 Arm 晶片。它的影響沒有AWS那麼大。這也會產生影響。顯然，我們確實預期客戶將開始採用冰山表格式。

Our next question comes from Kash Rangan from Goldman Sachs.

我們的下一個問題來自高盛的卡什·蘭根。

Mike, one for you and one for you, Sridhar. Mike, how much of the conservatism is due to the transition at the senior-most level? And one for you, Sridhar. Congratulations. What are the biggest challenges and opportunities facing Snowflake in your opinion?

麥克，一份給你，一份給你，斯里達爾。麥克，保守主義有多少是因為最高層的過渡所造成的？斯里達爾，還有一份給你。恭喜。您認為 Snowflake 面臨的最大挑戰和機會是什麼？

I don't know how to answer that question, Kash. But what I would say is I'd like to set the company up to be successful throughout the year as we progress with Sridhar coming on board. Sridhar?

我不知道該如何回答這個問題，卡什。但我想說的是，隨著斯里達爾的加入，我們將不斷進步，我希望公司能夠全年取得成功。斯里達爾？

Yes. And to answer your question, Kash. As you know, like Frank, our founders, have built Snowflake to be the trusted, efficient and cost-effective platform for enterprises. So I'm standing on the shoulder of giants to take us to the next chapter. And we already have a pretty ambitious plan, for example, to be able to write applications on top of Snowflake and those underlying technologies, native applications, container services or rolling out to GA this year.

是的。回答你的問題，卡什。如您所知，像我們的創辦人 Frank 一樣，Snowflake 已將 Snowflake 打造成值得企業信賴、高效且具有成本效益的平台。所以我站在巨人的肩膀上帶領我們進入下一個篇章。我們已經有了一個相當雄心勃勃的計劃，例如，能夠在 Snowflake 和那些底層技術、本機應用程式、容器服務之上編寫應用程序，或者今年推出到 GA。

And then I would add what is unique about this moment in technology, of course, is AI and its power to democratize access to enterprise data. I think this dramatically changes our understanding and notions of what an application is, all things can be stitched together. I think it is that a data platform, combined with applications, with AI powering things like interoperability that I think is the biggest opportunity in front of us. And we have a ton of investments, as I said, many of them are coming to GA. Getting them out quickly and driving adoption is easily my highest priority.

然後我想補充一下，當前技術領域的獨特之處當然是人工智慧及其使企業數據存取民主化的能力。我認為這極大地改變了我們對應用程式的理解和概念，所有的東西都可以縫合在一起。我認為，數據平台與應用程式相結合，以及人工智慧為互通性等提供動力，我認為這是我們面前最大的機會。正如我所說，我們有大量投資，其中許多都來到了 GA。快速將它們推出並推動採用是我的首要任務。

And in terms of challenges, some of these were deep platform surgeries. Unistore, as you folks know, is an incredibly ambitious project that has never been done before. And -- but it will roll out. So I think you're going to see us realize a lot of benefits of the investments that we've been making in the core technology platform over the last 2 years. But as you know, AI is also moving at lightning speed. We have an amazing team that's at work on it. But I think there's just lots more disruption to the software landscape to come. And that's why acting with speed and urgency is especially important for us.

就挑戰而言，其中一些是深度平台手術。如你們所知，Unistore 是一個前所未有的雄心勃勃的計畫。而且——但它將會推出。因此，我認為您將會看到我們認識到過去兩年我們在核心技術平台上進行的投資所帶來的許多好處。但如您所知，人工智慧也在以閃電般的速度發展。我們有一支出色的團隊正在致力於此。但我認為軟體領域未來還會面臨更多的顛覆。這就是為什麼快速而緊迫地採取行動對我們來說尤其重要。

Our next question comes from Patrick Colville of Scotiabank.

我們的下一個問題來自豐業銀行的派崔克·科爾維爾。

Under Frank's leadership, Snowflake's undoubtedly been a bit of a rocket ship. So Frank, I want to ask you, why leave now? And then Sridhar, can ask you on as well, looking forward to you being in charge for exciting next steps. How are you thinking about your strategy for Snowflake and changes you want to make now that you're in the CEO position? I've got a quick follow-up, if that's right.

在弗蘭克的領導下，雪花無疑有點像一艘火箭。那麼法蘭克，我想問你，為什麼現在離開？然後 Sridhar 也可以邀請您參與，期待您負責令人興奮的後續步驟。您如何看待 Snowflake 的策略以及您擔任執行長後想要做出的改變？如果是的話，我會快速跟進。

Yes. The important thing that I tried to highlight in the prepared remarks was that we haven't run succession as a time-based process. And I say, why ask now, it's not a timing issue. It is, do we have the person that we think is going to be an incredible win for the company going forward. And that's not -- you can't dictate that or mandate that. That is just based on opportunities that will or will not present themselves. So we feel incredibly fortunate that we crossed paths with Sridhar through the acquisition of Neeva.

是的。我在準備好的發言中試圖強調的重要一點是，我們並沒有將繼任過程作為一個基於時間的過程來運作。我說，為什麼現在要問，這不是時間問題。問題是，我們是否擁有一個我們認為將為公司未來帶來令人難以置信的勝利的人。那不是——你不能規定或強制這樣做。這只是基於機會是否會出現。因此，我們非常幸運能夠透過收購 Neeva 與 Sridhar 相遇。

If I think of myself not just as the former CEO of Snowflake, but also as an individual shareholder in Snowflake, this is the move I want to make at this time. And I cannot tell you as an investor how strongly I used to feel about the succession. This is not about -- just about changing of the guard. This is also about positioning the company really, really well for the challenges that are coming at us.

如果我認為自己不僅是Snowflake的前首席執行官，而且還是Snowflake的個人股東，這就是我此時想要採取的舉措。身為投資者，我無法告訴你我曾經對繼任者的感受有多強烈。這不是關於——只是關於換崗。這也關係到公司的定位是否真正能夠很好地應對我們面臨的挑戰。

Yes. And just adding on to what Frank said just now. Our short-term goals are very clear. As I said, we have a slew of product enhancements and everything from like transactional systems like Unistore, to interoperable storage to making applications on top of our data cloud possible with native applications and container services and of course, AI on top of that. So I think the short term and the need to react pretty quickly to a very quick silver AI landscape.

是的。補充一下弗蘭克剛才所說的話。我們的短期目標非常明確。正如我所說，我們有大量的產品增強功能，從像Unistore 這樣的交易系統，到可互通的存儲，再到透過本機應用程式和容器服務，當然還有人工智慧，使我們的資料雲之上的應用程式成為可能。所以我認為從短期來看，需要對非常快速的銀色人工智慧景觀做出快速反應。

This is what I'm going to be focused on. But I would almost say that, that actually translates pretty well into a longer-term strategy. Our belief is that a cloud that starts with data at the center combined with our product philosophy of creating a tightly integrated, easy-to-use product is the long-term winning strategy. Yes, there are going to be details that are different about which are the applications that we are going to be developing, what are ones that we are going to be doing in partnership, but that combination of the data cloud applications built on top of it with AI as an orchestrator is actually a pretty solid long-term strategy as well.

這就是我要關注的重點。但我幾乎想說，這實際上可以很好地轉化為長期策略。我們相信，以數據為中心的雲端結合我們創造緊密整合、易於使用的產品的產品理念，才是長期制勝策略。是的，對於我們將要開發的應用程式、我們將合作開發的應用程序，會有一些不同的細節，但構建在其之上的數據雲應用程序的組合讓人工智能作為協調者實際上也是一個相當可靠的長期戰略。

Of course, as I said, we have to be adaptive because the world of AI and its capabilities are changing by the month. And so we have to be receptive to that kind of change. But I feel very good about the path that we have set out for ourselves in how effective it's going to be both in the short term and the long term.

當然，正如我所說，我們必須適應，因為人工智慧的世界及其功能正在每月發生變化。因此，我們必須接受這種變化。但我對我們為自己設定的道路感到非常滿意，因為它在短期和長期都非常有效。

Mike, can I just -- one quick one for you. You gave the guidance about 3 points of product revenue will be from Snowpark in 1Q. Thanks for that color.

麥克，我可以給你一個快速的嗎？您給出了第一季 3 個百分點的產品收入來自 Snowpark 的指導。謝謝那個顏色。

For the year.

今年。

I guess my question is when -- for the year. Okay. Fiscal '25, okay. When might Snowpark hockey stick? It's been a product you guys have been talking about and invested pretty heavily in. It's seen terrific momentum. But can we expect the hockey stick at some point?

我想我的問題是何時——今年。好的。 25財年，好。 Snowpark 曲棍球棒什麼時候可以上市？這是你們一直在談論並投入大量資金的產品。它的勢頭非常好。但我們能在某個時候期待曲棍球棒嗎？

Anything is possible. What I would say is we did about mid-$30 million in revenue, I think $35 million, $36 million in revenue last year associated with Snowpark. Clearly, what I'm saying, 3% is going to be just under $100 million, $95 million or so this year. I think that is pretty phenomenal growth. And if we can get it to grow faster, we obviously will.

一切皆有可能。我想說的是，去年我們與 Snowpark 相關的收入約為 3,000 萬美元，我想是 3,500 萬美元、3,600 萬美元。顯然，我想說的是，今年 3% 的金額將略低於 1 億美元，約 9,500 萬美元。我認為這是相當驚人的成長。如果我們能讓它成長得更快，我們顯然會的。

Our next question comes from Brad Sills of Bank of America.

我們的下一個問題來自美國銀行的布拉德·希爾斯。

This is [Carly] on for Brad. I guess, first question, I just wanted to ask on your guide for the full year fiscal '25. I think you assumed a number of large customers going to adopt Iceberg Table. So some expectation on data moving out of Snowflake losing some storage revenue and some compute revenue there. Can you just double-click on that why some of the existing large customers are going to choose Iceberg Table rather than their original?

這是布拉德的[卡莉]。我想，第一個問題，我只是想問一下你們 25 財年全年的指南。我認為您假設許多大客戶將採用 Iceberg Table。因此，人們預計資料從 Snowflake 移出會損失一些儲存收入和一些計算收入。可以雙擊為什麼一些現有的大客戶會選擇 Iceberg Table 而不是他們原來的嗎？

A lot of big customers want to have open file formats to give them the options. And by the way, this is not necessarily customers moving all of their storage out of Snowflake, but a lot of the growth in their storage will be put into Iceberg Tables is what we think is going to happen. So you're just not going to see the growth associated with the storage and many of those customers. As a reminder, about 10% to 11% of our overall revenue is associated with storage.

許多大客戶希望擁有開放的文件格式來提供他們選擇。順便說一句，這不一定是客戶將所有儲存空間從 Snowflake 中移出，但我們認為將會發生的情況是，他們儲存中的大量成長將放入 Iceberg Tables 中。因此，您不會看到與儲存和許多客戶相關的成長。提醒一下，我們總收入的大約 10% 到 11% 與儲存相關。

Got it. Okay. And then just a follow-up, I guess, on the -- it's like encouraging to know 3% of the contribution will be from Snowpark. And then at the same time, you guys are expecting some headwinds from Iceberg and also the tier storage pricing. Can you just quantify for us what's like headwinds for the storage pricing in the Iceberg Table for the full year?

知道了。好的。然後，我想，這只是一個後續行動——知道 3% 的貢獻將來自 Snowpark，就像令人鼓舞一樣。同時，你們預期 Iceberg 和分層儲存定價會帶來一些阻力。您能否為我們量化一下冰山表中全年儲存定價的阻力？

Well, it's built into our guidance. I'm not going to break them out all separately. I would say the performance improvements, which have nothing to do with that, are around 6.2%, 6.3%.

嗯，它已納入我們的指南中。我不打算將它們全部分開。我想說的是，與此無關的效能改進約為 6.2%、6.3%。

Okay. And then I think you said the consumption is going to be like similar to fiscal year '24.

好的。然後我想你說過消費會跟 24 財年類似。

We're forecasting consumption patterns similar to what we saw in '24.

我們預測的消費模式與 24 年類似。

Our next question comes from Matt Hedberg of RBC.

我們的下一個問題來自加拿大皇家銀行的馬特·赫德伯格。

Great. And I guess following up on the Snowpark conversation. It feels like $35 million going to $100 million is, like you said, Mike, that is good growth. Could you talk a little bit more about sort of what the customer feedback has been? It feels like pipeline, it's like every check you do, Snowpark in the conversation. So maybe just double-click on what that customer feedback is and how encouraged you are with the pipeline growth there?

偉大的。我想是在跟進 Snowpark 的對話。就像你說的，麥克，感覺 3500 萬美元到 1 億美元是很好的成長。能多談談客戶的回饋嗎？感覺就像管道，就像你所做的每一次檢查，對話中的 Snowpark。因此，也許只需雙擊客戶回饋是什麼，以及您對那裡的管道成長有何鼓勵？

Christian here. The feedback has been very strongly positive. What we hear more often is better economics and better performance, but probably more important, more simplicity into how data doesn't have to be moved between systems and it's just an integrated solution. Some of the enhancements that you see apply only to data engineering, but also to traditional machine learning, which we see an increasing number of use cases also being deployed. So the sentiment is very positive.

基督徒在這裡。反饋非常正面。我們更常聽到的是更好的經濟性和更好的性能，但可能更重要、更簡單的是數據不必在系統之間移動，而只是一個整合的解決方案。您看到的一些增強功能僅適用於資料工程，也適用於傳統機器學習，我們看到越來越多的用例也在部署。所以情緒非常正面。

Got it. And Mike, maybe just one other -- just a little click on the guidance loss. You said consumption patterns from '24 is influencing your '25 guidance. I'm curious, and it feels like consumption has improved as '24 progressed. Are you sort of weighting it more towards like what you saw in the first half of the year or more sort of like some of the better trends you saw in the second half? Or just maybe double click a little bit on kind of what those fiscal '24 assumptions because it feels like things have gotten better as the year progressed?

知道了。麥克，也許只是另一個——只需點擊一下指導損失。您說 24 年的消費模式正在影響您的 25 年指引。我很好奇，感覺隨著 24 年的過去，消費有所改善。您是否更傾向於像上半年看到的那樣，或者更像是下半年看到的一些更好的趨勢？或者只是雙擊一下 24 財年的假設，因為隨著時間的推移，感覺事情已經變得更好了？

I would say it's more the average of '24 and which we saw stability happen in our customer base. I'm not forecasting any type of recovery inside there.

我想說這更像是 24 年的平均水平，我們看到我們的客戶群出現了穩定性。我並不預測內部會有任何類型的復甦。

Our next question comes from Tyler Radke of Citi.

我們的下一個問題來自花旗銀行的泰勒拉德克。

I want to direct this question at Sridhar, maybe Christian can jump in. But just as we think about the product road map for FY '25. You talked a little about Unistore, container services, Cortex. Can you just provide an update on when you expect these products to go GA, what you're seeing in terms of customer momentum, any customer statistics you can call out? And then how do you just kind of think about the maturity of these products, Sridhar, now that you've had time to look at the progress of these? Are you still expecting these to be contributors by year-end?

我想向 Sridhar 提出這個問題，也許 Christian 可以介入。但正如我們思考 25 財年的產品路線圖一樣。您談到了 Unistore、容器服務、Cortex。您能否提供有關您預計這些產品何時上市的最新資訊、您所看到的客戶動力以及您可以提供的任何客戶統計數據？那麼，Sridhar，既然您已經有時間了解這些產品的進展，那麼您如何看待這些產品的成熟度呢？您是否仍然期望這些人在年底前成為貢獻者？

I'll give a brief initial answer and then have Christian take over from that. One of the things that the Snowflake team is very, very good at doing is making sure that everything that we ship is tightly integrated with everything else, but there's just fluid interoperability between our various features and that is also rock solid. I think that, that culture of integrated and well-built features is a hallmark of Snowflake. Of course, in areas like AI, which is moving at lightning speed, we want a stable infrastructure but we also need to be flexible enough, whether it's new models that we put inside Cortex or other functionality that we need to develop on, on top of it. But a lot of the core investments that the team has made to allow for this kind of extensibility is what is coming in handy for us in terms of being able to ship things with speed.

我將給出一個簡短的初步答复，然後由克里斯蒂安接替。 Snowflake 團隊非常非常擅長做的一件事是確保我們發布的所有內容都與其他所有內容緊密整合，但我們的各種功能之間只有流暢的互通性，而且這也是堅如磐石的。我認為，整合和精心構建的功能文化是 Snowflake 的標誌。當然，在像人工智慧這樣以閃電般的速度發展的領域，我們需要一個穩定的基礎設施，但我們也需要足夠的靈活性，無論是我們放入Cortex 內部的新模型還是我們需要開發的其他功能。它的。但是團隊為實現這種可擴展性而進行的許多核心投資對於我們能夠快速交付東西來說是派上用場的。

I'll hand it off to Christian to talk about when he expects different things to hit GA. Christian?

我會把它交給 Christian，讓他談談他對 GA 的不同預期。基督教？

Yes. So as both Sridhar and Mike mentioned, we expect a number of meaningful GA milestones this fiscal year, starting with generative AI. Cortex will be in public preview very soon, and we expect it to be generally available on and around the Summit's time frame. And as Sridhar said, we expect all sorts of interesting use cases of generative AI coming to the data and preserving privacy and security.

是的。因此，正如 Sridhar 和 Mike 所提到的那樣，我們預計本財年將出現一些有意義的 GA 里程碑，首先是生成式人工智慧。 Cortex 很快就會推出公開預覽版，我們預計它將在高峰會期間及前後全面發布。正如 Sridhar 所說，我們預期生成式人工智慧的各種有趣用例都會出現在資料中並保護隱私和安全。

Snowpark container services is already in public preview in AWS, and we expect it to be generally available in that same time frame, give or take a couple of months from Summit, and is the ultimate extensibility capability for bringing computation into Snowflake. Iceberg Tables is already on public preview across all 3 and will be generally available again, also in the Data Cloud Summit time frame.

Snowpark 容器服務已經在 AWS 中公開預覽，我們預計它將在同一時間範圍內普遍可用，距離 Summit 大約有幾個月的時間，並且是將計算引入 Snowflake 的最終可擴展功能。 Iceberg Tables 已在所有 3 個平台上進行公開預覽，並將在資料雲高峰會時間範圍內再次全面推出。

Unistore, which enables combining transactional and analytical capabilities in single went very recently into public preview in AWS and will be generally available in the second half of the year. And native apps, which is how we accelerate time to value for both partners and customers, is clearly on AWS on Azure and we're continuing to round up the enhancement. So this is imminent, and we expect a strong showing of product capabilities at the Data Cloud Summit.

Unistore 能夠將事務和分析功能結合在一起，最近在 AWS 上推出了公開預覽版，並將於今年下半年全面上市。原生應用程式（我們加快為合作夥伴和客戶實現價值的方式）顯然已在 Azure 上的 AWS 上，我們將繼續完善這項增強功能。所以這已經迫在眉睫，我們期待在數據雲高峰會上強勢展示產品能力。

Great. And a follow-up for Mike, just in terms of what you've seen so far in February, and I know you're talking about FY '25 consumption -- assumptions consistent with a year ago. But have you seen February trends improve relative to January? Does February feel similar to February a year ago? Or is it closer to the strong growth that you saw in consumption towards the end of last year?

偉大的。麥克的後續行動，就您在 2 月到目前為止所看到的情況而言，我知道您正在談論 25 財年的消費 - 與一年前一致的假設。但您是否看到 2 月份的趨勢相對於 1 月份有所改善？二月感覺和一年前的二月相似嗎？還是更接近去年年底消費的強勁成長？

Well, I just guided for February, well, for the quarter, and that reflects the consumption of trends we've been seeing through yesterday. So I would say it's more in line with what we saw coming out last year throughout on average for the year.

嗯，我剛剛對二月份，嗯，這個季度進行了指導，這反映了我們昨天看到的趨勢的消耗。所以我想說，這更符合我們去年看到的全年平均。

Our next question comes from Mike Cikos of Needham & Co.

我們的下一個問題來自 Needham & Co. 的 Mike Cikos。

I did just want to follow up. I guess one of the dominant messages from management, if I just go back a quarter ago, really seem to be this theme about the growing use of unstructured data by your customers. And I don't know if it's just my read here, but it feels like it's a more muted message today. Maybe I'm missing something. But can you just put some finer points here as far as the trends that Snowflake is seeing specifically around unstructured data in Q4? And then I did have a follow-up.

我只是想跟進。我想，如果我回顧一個季度前，管理層傳達的主要訊息似乎確實是關於客戶越來越多地使用非結構化資料的主題。我不知道這是否只是我在這裡讀到的，但感覺今天這是一個更加低調的信息。也許我錯過了一些東西。但是，您能否就 Snowflake 在第四季度特別關注非結構化資料的趨勢提出一些更具體的觀點？然後我確實進行了後續行動。

Christian here. The momentum that we shared in the last quarter has carried forward on through this quarter. So there is no change in pace or interest. And if anything, the Document AI that Sridhar mentioned earlier in the call, where we have hundreds of customers lined up to be able to leverage this technology is all about extracting value and signal from unstructured documents. So the fact that it was not mentioned is not anything that should be read into. It's still a topic of high interest among many of our customers.

基督徒在這裡。我們在上個季度分享的勢頭在本季度得到了延續。所以節奏或興趣沒有改變。如果有什麼不同的話，那就是斯里達爾之前在電話會議中提到的文檔人工智慧，我們有數百名客戶排隊等待能夠利用這項技術，就是從非結構化文件中提取價值和訊號。因此，沒有提到這一點並不值得深究。這仍然是我們許多客戶高度感興趣的話題。

Got it. And I know that you guys have these assumptions as well in the guidance around the potential headwinds from some of your larger customers adopting or moving data to Iceberg. Maybe to parse through that a little bit, can you talk about what you've seen thus far as far as customers' behavior? Are they already doing this with iceberg? Or is it just an assumption that you guys have? I'm trying to get a sense as far as the behavior customers have exhibited as well as how much you guys are peppering into your guidance today when we think about that headwind as well as the rollout of tiered storage.

知道了。我知道你們在一些大客戶採用或將資料轉移到 Iceberg 的潛在阻力的指南中也有這些假設。也許要稍微分析一下，您能談談到目前為止您所看到的客戶行為嗎？他們已經對冰山這樣做了嗎？或者這只是你們的假設？我試圖了解客戶所表現出的行為，以及今天當我們考慮這一逆風以及分層存儲的推出時，你們在指導中投入了多少精力。

Yes. Iceberg is not GA yet. So a customer is not going to roll that into production until it's GA, and we do think it will be a gradual process of if they're going to move data out. But as soon as this GA new data can go into iceberg table. And we don't expect that to be GA until sometime around June time frame.

是的。 Iceberg 尚未正式發表。因此，在正式發布之前，客戶不會將其投入生產，而且我們確實認為，如果他們要將數據移出，這將是一個漸進的過程。但是一旦這個 GA 新資料就可以進入冰山表。我們預計要到六月左右的某個時間才會正式發布。

Okay. And so just to put a finer point on that, though, the headwind we're expecting from Iceberg, this is theoretical. You have not seen that behavior from customers just yet.

好的。不過，為了更詳細地說明這一點，我們預期冰山會帶來逆風，這是理論上的。您還沒有看到客戶的這種行為。

Correct. It is what we are expecting. I know the tiered -- the new tiered storage pricing that we're seeing that today. And I'll let Christian add some to that as well, too.

正確的。這正是我們所期待的。我知道分層——我們今天看到的新的分層儲存定價。我也會讓克里斯蒂安補充一些內容。

Yes. I would add that for many of our large customers, we have been in touch on their plans for adoption on Iceberg. So some of what you see in our guidance has factored in those intentions.

是的。我想補充一點，對於我們的許多大客戶來說，我們一直在聯繫他們在 Iceberg 上採用的計劃。因此，您在我們的指南中看到的一些內容已經考慮到了這些意圖。

Our next question comes from Michael Turrin of Wells Fargo.

我們的下一個問題來自富國銀行的麥可‧特林。

Maybe one on hiring. I think you mentioned 1,000 net new heads you're expecting to add in the coming year. How are you balancing, adding to R&D given all the new product efforts and AI interest with bringing on sales capacity if the market does start to turn? And on the FedRAMP high authorization, any commentary around that, what that could open up and if that scenario of potential investment as well?

也許是關於招募的。我想您提到了您預計在來年增加 1,000 名淨新負責人。鑑於所有新產品的努力和人工智慧的興趣，您如何平衡增加研發與市場確實開始轉變時帶來的銷售能力？關於 FedRAMP 的高度授權，對此有何評論，這可能會帶來什麼，以及潛在投資的情況嗎？

I would say a lot of our expensive hiring is in the R&D area and it will continue to be more in the AI ML space. These engineers are very expensive. With that said, we're still adding in the sales organization. And if we see an uptick in new customers and consumption patterns with our customers. We can easily dial that up just like we've dialed back our hiring in the past, but we are not going to sacrifice on R&D.

我想說的是，我們的許多昂貴的招募都在研發領域，並且將繼續更多地在人工智慧機器學習領域。這些工程師的費用非常昂貴。儘管如此，我們仍在增加銷售組織。如果我們看到新客戶和客戶的消費模式增加。我們可以輕鬆地提高這一點，就像我們過去減少招募一樣，但我們不會犧牲研發。

Our next question comes from Derrick Wood of TD Cowen.

我們的下一個問題來自 TD Cowen 的 Derrick Wood。

Mike, most other cloud consumption vendors are talking about stabilizing growth. You guys are still modeling pretty sharp deceleration over the next year. So this certainly sticks out and may bring about questions on maybe company-specific challenges. So first, are there any notable customer or workload losses that could be weighing on growth this year? And generally, how are you feeling about sales productivity and competitive win rates in the current environment?

麥克，大多數其他雲端消費供應商都在談論穩定成長。你們仍在對明年的急劇減速進行建模。因此，這肯定很突出，並且可能會帶來有關公司特定挑戰的問題。首先，是否有任何顯著的客戶或工作負載損失可能會影響今年的成長？一般來說，您對當前環境下的銷售效率和競爭獲勝率有何看法？

Absolutely no big competitive losses or workloads moving up that I'm aware of. This is all related to our model where a lot of the performance improvements that we have in our software go directly to the customer. And that's why I was pointing out, you saw there was a 62% year-over-year growth in jobs on a daily basis run on Snowflake versus only a 33% revenue growth. And we know there's a lot of performance improvements coming into play this year, coupled with Iceberg, coupled with tiered storage pricing that we rolled out. And I was able to roll out the tiered storage pricing because we were getting much better pricing out of the cloud vendors to us.

據我所知，絕對沒有重大的競爭損失或工作量增加。這都與我們的模型有關，我們軟體中的許多效能改進都直接惠及客戶。這就是為什麼我要指出，您會看到 Snowflake 上每天的工作同比增長 62%，而收入僅增長 33%。我們知道今年會有很多效能改進，再加上 Iceberg，以及我們推出的分層儲存定價。我之所以能夠推出分層儲存定價，是因為我們從雲端供應商那裡獲得了更好的定價。

Yes. Okay. It would be interesting to kind of see what the growth rates on just the workloads related to compute is relative to the storage drag. Maybe that's something you could give. But just -- how do you feel about ultimately kind of getting back to 30% longer term? And kind of what's the top 1 or 2 critical things that need to take place to get you there?

是的。好的。看看與計算相關的工作負載的成長率相對於儲存阻力的變化是很有趣的。也許這是你可以給予的東西。但是，從長期來看，您對於最終恢復到 30% 的成長率有何感想？為了實現這一目標，最重要的 1 或 2 件關鍵事情是什麼？

I would say the biggest thing is the uptick in consumption associated with all the new enhancements we have in our product, in particular, what we could see coming out of Cortex, what we could see coming out of Snowpark container services, and ultimately, what we could see in native app development on our platform.

我想說，最重要的是與我們產品中的所有新增強功能相關的消費上升，特別是我們可以看到 Cortex 帶來的東西、我們可以看到 Snowpark 容器服務帶來的東西，以及最終，什麼我們可以在我們的平台上看到本機應用程式開發。

Our next question comes from Gray Powell of BTIG.

我們的下一個問題來自 BTIG 的 Gray Powell。

I just had a quick follow-up on Snowpark. I think you called out Snowpark at a $70 million annual run rate one day back in December. I know that was sort of like a 1-day statistic. But I'm just trying to think through that. I mean the guide of 3% of revenue for fiscal '25, that's like $95 million, $100 million. I guess it just seems kind of conservative given that Snowpark was -- or that consumption of Snowpark is growing 45% quarter-over-quarter last quarter. Just how should we think about the level of conservatism on that assumption within your guidance?

我剛剛對 Snowpark 進行了快速跟進。我想你在 12 月的一天就以 7000 萬美元的年運營率宣布了 Snowpark。我知道這有點像一日統計數據。但我只是想思考一下。我指的是 25 財年營收的 3%，相當於 9,500 萬美元、1 億美元。我想這似乎有點保守，因為 Snowpark 的消費量在上個季度環比增長了 45%。在您的指導下，我們應該如何考慮該假設的保守程度？

Well, what I'm saying is we're guiding it to be 3% of product revenue this year, and you can infer what you want from the guidance.

嗯，我的意思是，今年我們將其指導為產品收入的 3%，您可以從指導中推斷您想要的內容。

We will take no further questions for today. So that concludes today's conference call. Thank you all for joining. You may now disconnect your lines.

今天我們不會再回答任何問題。今天的電話會議到此結束。感謝大家的加入。現在您可以斷開線路。