惠普企業 (HPE) 2024 Q1 法說會逐字稿

Good afternoon, and welcome to the First Quarter Fiscal 2024 Hewlett Packard Enterprise Earnings Conference Call. My name is Gary, and I'll be your conference moderator for today's call. (Operator Instructions) As a reminder, this conference is being recorded for replay purposes. I would now like to turn the presentation over to your host for today's call, Ms. Shannon Cross, Senior Vice President and Chief Strategy Officer and Investor Relations. Please proceed.

下午好，歡迎參加 2024 財年第一季惠普企業財報電話會議。我叫加里，我將擔任今天電話會議的主持人。 （操作員指示）提醒一下，本次會議將會被錄音以供重播。現在，我想將演講交給今天電話會議的主持人、高級副總裁兼首席策略長和投資者關係部主任 Shannon Cross 女士。請繼續。

Good afternoon. I'd like to welcome you to our fiscal 2024 first quarter earnings conference call with Antonio Neri, HPE's President and Chief Executive Officer; and Marie Myers, HPE's Chief Financial Officer. After 25 years on Wall Street and over 20 years covering HP, I'm very excited to join HPE as Chief Strategy Officer. I look forward to working with Jeff Kvaal and the rest of the IR team, and I look forward to seeing many of you in the months ahead.

午安.歡迎您參加我們 2024 財年第一季財報電話會議，與 HPE 總裁兼執行長 Antonio Neri 一起；以及 HPE 財務長 Marie Myers。在華爾街工作了 25 年並在惠普工作了 20 多年之後，我非常高興加入 HPE 擔任首席策略長。我期待與 Jeff Kvaal 和 IR 團隊的其他成員合作，並期待在未來幾個月內與你們見面。

Before handing the call to Antonio, let me remind you that this call is being webcast. A replay of the webcast will be available shortly after the call concludes. We have posted the press release and the slide presentation accompanying the release on our HPE Investor Relations web page. Elements of the financial information referenced on this call are forward-looking and are based on our best view of the world and our businesses as we see them today. HPE assumes no obligation and does not intend to update any such forward-looking statements.

在將電話交給安東尼奧之前，請允許我提醒您，該通話正在進行網路直播。通話結束後不久將提供網路直播的重播。我們已在 HPE 投資者關係網頁上發布了新聞稿和隨附的幻燈片簡報。本次電話會議中引用的財務資訊部分是前瞻性的，基於我們對當今世界和我們業務的最佳看法。 HPE 不承擔任何義務，也不打算更新任何此類前瞻性聲明。

We also note that the financial information discussed on this call reflects estimates based on information available at this time and could differ materially from the amounts ultimately reported in HPE's quarterly report on Form 10-Q for the fiscal quarter ended January 31, 2024. For a more detailed information, please see the disclaimers on the earnings materials relating to forward-looking statements that involve risks, uncertainties and assumptions. Please refer to HPE's filings with the SEC for a discussion of these risks.

我們還注意到，本次電話會議上討論的財務信息反映了基於​​目前可用信息的估計，可能與 HPE 2024 年 1 月 31 日結束的財政季度的 10-Q 表季度報告中最終報告的金額存在重大差異。請參閱 HPE 向美國證券交易委員會提交的文件，以了解有關這些風險的討論。

For financial information we have expressed on a non-GAAP basis, we have provided reconciliations to the comparable GAAP information on our website. Please refer to the tables and slide presentation accompanying today's earnings release on our website for details. Throughout this conference call, all revenue growth rates, unless otherwise noted, are presented on a year-over-year basis and adjusted to exclude the impact of currency. Finally, Antonio and Marie will reference our earnings presentation and their prepared comments.

對於我們以非 GAAP 基礎表達的財務訊息，我們已在我們的網站上提供了與可比較 GAAP 資訊的對帳。有關詳細信息，請參閱我們網站上今天的收益報告中附帶的表格和幻燈片演示。在整個電話會議中，除非另有說明，所有收入增長率均按同比計算，並經過調整以排除貨幣的影響。最後，安東尼奧和瑪麗將參考我們的收益報告和他們準備好的評論。

Before handing the call to Antonio, let me remind you that this call is being webcast. A replay of the webcast will be available shortly after the call concludes. We have posted the press release and the slide presentation accompanying the release on our HPE Investor Relations web page.

在將電話交給安東尼奧之前，請允許我提醒您，該通話正在進行網路直播。通話結束後不久將提供網路直播的重播。我們已在 HPE 投資者關係網頁上發布了新聞稿和隨附的幻燈片簡報。

With that, let me turn it over to Antonio.

現在，讓我把話題交給安東尼奧。

Thank you, Shannon. Good afternoon, and thank you for joining us today. In the first quarter, we are proud to have outpaced our profitability expectations while advancing our long-term strategy. We also continue to scale our recurring revenue, achieving the second highest year-over-year growth rate since we started tracking ARR in late 2019. This is a promising indicator for our ongoing portfolio shift to higher-margin revenues. But overall, Q1 revenue performance did not meet our expectations.

謝謝你，香農。下午好，感謝您今天加入我們。第一季度，我們很自豪地超越了我們的獲利預期，同時推進了我們的長期策略。我們也持續擴大經常性收入，實現了自 2019 年底開始追蹤 ARR 以來第二高的年成長率。但整體來看，Q1營收表現並未達到我們的預期。

During this call, I will address 3 key points. First, I will touch on our revenue, which was lower than expected in large part because network and demand softened industry-wide and because the timing of several large GPU acceptances shifted. Additionally, we did not have the GPU supply we wanted, curtailing our revenue upside. Second, I will address how we are streamlining our reporting segments, accelerating our new specialized sales model, managing our spending and reinforcing execution discipline. Third and most importantly, I will discuss the progress we're making in executing our long-term strategy, which we remain confident in.

在這次通話中，我將討論3個重點。首先，我要談談我們的收入，我們的收入低於預期，很大程度上是因為整個行業的網路和需求疲軟，以及幾個大型 GPU 的驗收時間發生了變化。此外，我們沒有得到我們想要的 GPU 供應，這限制了我們的收入成長。其次，我將談談我們如何精簡報告部門、加速推動新的專業銷售模式、管理支出並加強執行紀律。第三，也是最重要的一點，我將討論我們在執行長期策略方面所取得的進展，我們對此仍然充滿信心。

Let me first address revenue in the quarter. Similar to peers in the market, we saw campus networking product demand weaken and the decline later in the quarter was greater than expected. This was a large headwind relative to our expectations. Customers are taking longer to digest prior orders than we had anticipated, which partially offset the benefit of our backlog entering the quarter. And Europe and Asia were areas of relative softness. We expect weakness in the networking market to persist, which is likely to impact revenue through fiscal year 2024. That said, we anticipate some improvement late in fiscal year 2024 as inventory clears and we ramp into the purchasing season for state and local education customers in the United States.

首先讓我談談本季的收入。與市場同業類似，我們看到校園網路產品需求減弱，且本季後期的降幅大於預期。與我們的預期相比，這是一個巨大的阻力。客戶消化之前的訂單所花的時間比我們預期的要長，這部分抵消了我們進入本季的積壓訂單帶來的好處。歐洲和亞洲則是相對疲軟的地區。我們預計網路市場的疲軟態勢將持續，這可能會影響到 2024 財年的收入。

AI server demand remains very strong, evidenced by our growing cumulative order book. However, GPU availability remains tight, and our delivery timing has also been affected by the increasing length of time customers required to set up the data center space, power and cooling requirements needed to run these systems. As a result, overall AI server orders conversion was below our expectations. However, the AI contribution to ARR is increasing.

人工智慧伺服器的需求仍然非常強勁，我們不斷增長的累積訂單證明了這一點。然而，GPU 的可用性仍然很緊張，而且我們的交貨時間也受到客戶設定運行這些系統所需的資料中心空間、電力和冷卻要求的時間越來越長的影響。因此，整體 AI 伺服器訂單轉換率低於我們的預期。然而，AI對ARR的貢獻正在增加。

We continue to prioritize profitability. I am pleased that we have delivered non-GAAP gross margin of 36.2%, which is up 200 basis points year-over-year and above 600 basis points from fiscal year 2018. This performance helped our Q1 non-GAAP diluted net earnings per share grow to $0.48, which was above the midpoint of our guidance range despite lower-than-expected revenue, illustrating the positive impact of our pivot to higher growth, higher margin revenue.

我們繼續優先考慮盈利能力。我很高興我們實現了 36.2% 的非 GAAP 毛利率，比去年同期增長 200 個基點，比 2018 財年增長 600 個基點以上。

We know that the current environment will require continued discipline in how we execute. We are accelerating new specialized sales motions to maximize opportunities and improve order linearity across our portfolio. Improved cost management will remain an important competency for us in fiscal year 2024.

我們知道，目前的環境要求我們在執行上繼續嚴格規範。我們正在加速新的專業銷售舉措，以最大限度地利用機會並提高我們整個產品組合的訂單線性。改善成本管理仍將是我們在 2024 財政年度的重要能力。

We also found opportunity to streamline our reporting segments. We have now combined the Compute and HPC and AI segments into a single server segment that integrates general-purpose computing, high-performance computing, supercomputing and AI systems. This will enable us to maximize the opportunities across the entire AI life cycle, from training to tuning to inferencing, and execute with agility. And as previously discussed, we have simplified our hybrid cloud strategy by putting all related products, software and services into one business unit. Our new Hybrid Cloud segment will further accelerate customer adoption of the HPE GreenLake hybrid cloud platform.

我們也發現了簡化報告部門的機會。現在，我們將運算、HPC 和 AI 部分合併為一個伺服器部分，整合通用運算、高效能運算、超級運算和 AI 系統。這將使我們能夠最大限度地利用整個人工智慧生命週期中的機會，從訓練到調整到推理，並靈活地執行。如同前面所討論的，我們透過將所有相關產品、軟體和服務納入一個業務部門，簡化了我們的混合雲策略。我們新的混合雲部門將進一步加速客戶對 HPE GreenLake 混合雲平台的採用。

Turning to our strategy. While we are experiencing cyclicality in submarkets, I am more confident than ever in our long-term strategy that is aligned to key market mega trends. In edge, over the last several quarters, we have gained share in the campus networking markets and our strategic investments have paid off. Most recently, we have seen strong growth in SASE, an offering bolstered by our acquisition of Silver Peak in 2020 and Axis Security in 2023. Our sales pipeline for our private 5G offering is also growing rapidly, following our acquisition of Athonet in 2023.

談談我們的策略。雖然我們正在經歷子市場的周期性變化，但我對我們與主要市場大趨勢一致的長期策略比以往任何時候都更有信心。在邊緣運算方面，在過去的幾個季度中，我們在校園網路市場中獲得了份額，我們的策略投資也獲得了回報。最近，我們看到 SASE 的強勁成長，這得益於我們在 2020 年收購 Silver Peak 和 2023 年收購 Axis Security。

In Hybrid Cloud, HPE GreenLake continues to resonate in the marketplace and was the primary driver of the highest Q1 year-over-year rise in ARR over the 4-plus years we have been reporting it. Our ARR grew 41% year-over-year to more than $1.4 billion in Q1, and we continue to expect ARR growth of 35% to 45% as we look ahead. We're also capitalizing on cross-selling opportunities when customers come to us for our AI solutions and realize we can meet their storage needs as well.

在混合雲領域，HPE GreenLake 繼續在市場上引起共鳴，並且是我們報告的 4 年多來第一季 ARR 年成長最高的主要推動力。我們的 ARR 在第一季年增 41%，達到 14 億美元以上，我們預計未來 ARR 成長率仍將達到 35% 至 45%。當客戶向我們尋求人工智慧解決方案並意識到我們也可以滿足他們的儲存需求時，我們也利用交叉銷售機會。

In AI, we are capturing the explosion in demand for AI systems. Our cumulative accelerator processing unit orders rose to $4 billion in the quarter, driven by demand across HPE Cray EX and XD solutions as well as HPE ProLiant Gen11 AI-optimized servers. APU orders represent nearly 25% of our total server orders since the first quarter of fiscal 2023. Our pipeline is large and growing across the entire AI life cycle from training to tuning to inferencing. We are starting to see AI demand pull-through for other solutions, including storage.

在人工智慧領域，我們正在捕捉人工智慧系統需求的爆炸性成長。受 HPE Cray EX 和 XD 解決方案以及 HPE ProLiant Gen11 AI 優化伺服器的需求推動，本季我們的累積加速器處理單元訂單增至 40 億美元。自 2023 財年第一季以來，APU 訂單占我們伺服器訂單總量的近 25%。我們開始看到人工智慧對其他解決方案（包括儲存）的需求拉動。

We expect our Server and Hybrid Cloud segments to grow sequentially through the fiscal year. Server revenue stands to benefit from AI system demand, improving GPU supply, and our continued mix shift to HPE ProLiant Gen11. Hybrid Cloud will benefit from continued HPE GreenLake storage demand and the rising productivity of our specialized sales force.

我們預計我們的伺服器和混合雲部門將在整個財年內持續成長。伺服器收入將受益於 AI 系統需求、GPU 供應改善以及我們持續向 HPE ProLiant Gen11 轉變。混合雲將受益於持續的 HPE GreenLake 儲存需求以及我們專業銷售團隊不斷提高的生產力。

Our customers continue to validate our value proposition. As one example, we are building for Eni, one of the world's largest energy providers, a new HPE Cray EX supercomputer that will reach more than half an exaflop performance. The system will be one of the most powerful in the world for enterprise use and will accelerate AI-driven scientific discovery to advance efforts in energy.

我們的客戶不斷驗證我們的價值主張。舉例來說，我們正在為全球最大的能源供應商之一埃尼公司 (Eni) 打造一台新型 HPE Cray EX 超級計算機，其性能將達到半百億億次浮點運算以上。該系統將成為全球最強大的企業用系統之一，並將加速人工智慧驅動的科學發現，推動能源事業的發展。

We also have been awarded the deal for Poland's most powerful supercomputer system located at the Academic Computer Centre CYFRONET of the AGH University of Science and Technology. Based on the HPE Cray EX supercomputer and HPE Slingshot interconnect fabric with NVIDIA Grace Hopper GPUs, the system will be used to support modeling simulation and AI-driven scientific research needs, including training and tuning of large language models.

我們還贏得了波蘭最強大的超級電腦系統合同，該系統位於 AGH 科技大學學術電腦中心 CYFRONET。該系統基於 HPE Cray EX 超級電腦和配備 NVIDIA Grace Hopper GPU 的 HPE Slingshot 互連結構，將用於支援建模模擬和 AI 驅動的科學研究需求，包括大型語言模型的訓練和調整。

We're also seeing growth in AI inferencing. For example, Coles Supermarkets, a leading Australian retailer, implemented an HPE ProLiant Gen11 AI inferencing solution in Q1, which helps with video imaging to reduce store stock lost to theft and infrared scanning at the checkout.

我們也看到人工智慧推理的成長。例如，澳洲領先的零售商 Coles Supermarkets 在第一季實施了 HPE ProLiant Gen11 AI 推理解決方案，該解決方案透過視訊成像幫助減少因盜竊和結帳時紅外線掃描而損失的商店庫存。

In the quarter, we expanded our strategic collaboration with NVIDIA targeting the enterprise segment of the market. We introduced a preconfigured solution for enterprise customers to fine-tune AI large language models with their private data to accelerate inferencing. Our HPE machine learning development environment software and unique HPE supercomputer and IP are critical parts of the solution, alongside NVIDIA AI Enterprise software. We have built a strong and growing sales pipeline for this new offering.

本季度，我們擴大了與 NVIDIA 的策略合作，瞄準企業市場領域。我們為企業客戶推出了預先配置的解決方案，以便他們使用私有資料對 AI 大型語言模型進行微調，從而加速推理。我們的 HPE 機器學習開發環境軟體和獨特的 HPE 超級電腦和 IP 與 NVIDIA AI Enterprise 軟體一起是此解決方案的關鍵部分。我們為這項新產品建立了強大且不斷成長的銷售管道。

We also continue to build out our HPE GreenLake hybrid cloud platform services. Earlier this quarter, we announced an expanded HPE GreenLake for file storage that is designed for generative AI. This solution is highly differentiated through a high-performant file system solution specifically designed for AI applications. We believe it better positions us to take market share in storage by addressing the previously underserved segment of the file market. Innovation like this continue to attract customers to HPE GreenLake platform, which connects 3.8 million network devices and supports more than 31,000 customer organizations, up approximately 8% from last quarter.

我們也繼續建構我們的 HPE GreenLake 混合雲平台服務。本季早些時候，我們宣布推出擴展的 HPE GreenLake 文件儲存產品，專為生成式 AI 而設計。該解決方案透過專為AI應用設計的高效能檔案系統解決方案實現了高度差異化。我們相信，透過解決文件市場中先前服務不足的部分，我們可以更好地佔據儲存市場份額。這樣的創新繼續吸引客戶使用 HPE GreenLake 平台，該平台連接了 380 萬台網路設備並支援超過 31,000 個客戶組織，比上一季成長約 8%。

One new HPE GreenLake customer is the U.S. Navy Fleet Numerical Meteorology and Oceanography Center, which produces critical models of weather and ocean conditions for the U.S. and coalition forces worldwide. They turned to HPE GreenLake to improve predictability, accuracy and speed of their modeling while reducing costs. This week, I attended the Mobile World Congress where AI and private 5G were the key topics among telcos and service providers alike. With our pending acquisition of Juniper Networks, HPE's portfolio will expand to better serve these unique customers from the edge of the network to core 5G to cloud.

HPE GreenLake 的一個新的客戶是美國海軍艦隊數值氣象學和海洋學中心，該中心為美國和世界各地的聯軍製作關鍵的天氣和海洋狀況模型。他們採用 HPE GreenLake 來提高建模的可預測性、準確性和速度，同時降低成本。本週，我參加了世界行動通訊大會，其中人工智慧和私有 5G 是電信營運商和服務供應商討論的關鍵主題。隨著我們即將收購瞻博網路 (Juniper Networks)，HPE 的產品組合將得到擴展，從而更好地服務於從網路邊緣到核心 5G 再到雲端的這些獨特客戶。

Customers were very interested in our integration of Athonet private 5G capabilities into our Intelligent Edge portfolio as well as in our cloud Open RAN and vRAN solutions. They were also very eager to explore the massive new market opportunity AI inferencing presents at the edge of the network. That is one of the reasons why we're so excited about our pending Juniper Networks acquisition. Combining our complementary portfolios will supercharge HPE's edge-to-cloud strategy, accelerating our entire portfolio with AI-enabled innovation. When our proposed acquisition closes, we will create a new networking innovator with our comprehensive portfolio for customers and partners. The transaction is expected to double the size of our networking business, which will be the core foundation of HPE, covering the anticipated $180 billion market opportunity with our combined IP.

客戶對我們將 Athonet 私有 5G 功能整合到我們的智慧邊緣產品組合以及我們的雲端 Open RAN 和 vRAN 解決方案中非常感興趣。他們也非常渴望探索人工智慧推理在網路邊緣帶來的巨大新市場機會。這就是我們對即將進行的瞻博網路收購感到如此興奮的原因之一。結合我們互補的產品組合將增強 HPE 的邊緣到雲端策略，並透過 AI 支援的創新加速我們整個產品組合的發展。當我們提議的收購完成後，我們將為客戶和合作夥伴打造一個擁有全面產品組合的全新網路創新企業。此次交易預計將使我們的網路業務規模擴大一倍，這將成為HPE的核心基礎，透過我們的合併IP覆蓋預計1800億美元的市場機會。

From a financial perspective, this transaction is also compelling for our shareholders. In the first year post close, we expect accretion to non-GAAP earnings per share and in the long term, higher non-GAAP gross and operating margins. We are working to secure regulatory approvals in several jurisdictions. We are hopeful that regulators will recognize that this acquisition is centered around driving further innovation for our customers. We continue to expect that the transaction will close later this calendar year or in early calendar 2025.

從財務角度來看，這筆交易對我們的股東來說也具有吸引力。在交易完成後的第一年，我們預計非公認會計準則每股收益將會增加，並且從長期來看，非公認會計準則毛利率和營業利潤率將會更高。我們正在努力獲得多個司法管轄區的監管批准。我們希望監管機構能夠認識到此次收購的核心在於為我們的客戶推動進一步的創新。我們仍然預計該交易將在今年稍後或 2025 年初完成。

In summary, Q1 2024 was a mixed quarter for HPE. We achieved strong profitability and drove new record ARR growth with overall revenue short-term expectations given the softening networking market, GPU deal timing and to some extent, GPU availability. We are focused on execution as we navigate the fluctuations in demand we see in certain areas of the market. Marie will take you through our adjusted guidance, which reflects our latest thinking about the year ahead.

總而言之，2024 年第一季對 HPE 來說是一個好壞參半的季度。考慮到網路市場的疲軟、GPU 交易時機以及某種程度上的 GPU 可用性，我們實現了強勁的盈利能力並推動了新創紀錄的 ARR 成長，整體收入短期預期達到。在應對市場某些領域的需求波動時，我們專注於執行。瑪麗將帶您了解我們調整後的指導方針，這反映了我們對未來一年的最新想法。

This quarter is a moment in time and does not at all dampen our confidence in the future ahead of us. We have taken the right actions to maximize value for our shareholders. The work we are doing now, combined with our technological edge and our strategy that has never been more relevant, will position us to convert on the long-term opportunities in front of us across edge, hybrid cloud, and AI.

本季是一個重要時刻，但這絲毫不會削弱我們對未來的信心。我們採取了正確的行動來為股東實現價值最大化。我們現在所做的工作，加上我們的技術優勢和前所未有的重要策略，將使我們能夠抓住邊緣、混合雲和人工智慧領域面臨的長期機會。

Before we transition, I'm delighted to welcome Marie Myers as our new CFO. Having worked with her at HP before the separation, it is a pleasure to partner with her again. I admire her passion for and skill at fueling innovation and performance. I am confident that Marie is a great fit for this role and expect she will help drive the next phase of growth and shareholder return for HPE.

在我們過渡之前，我很高興歡迎瑪麗·邁爾斯擔任我們的新任財務長。在分拆之前我曾與她在惠普共事過，很高興能再次與她合作。我欽佩她推動創新和績效的熱情和技巧。我相信瑪麗非常適合這個職位，並期望她將幫助推動 HPE 下一階段的成長和股東回報。

I will now turn the call over to her for details about our segments and our outlook. Marie?

現在我將把電話轉給她，以了解有關我們的部門和前景的詳細資訊。瑪麗？

Thank you, Antonio. I'm pleased to be with you today on my first earnings call as HPE's CFO. I've long admired HPE's impressive transformation, and there has never been a more exciting time to be part of this company. We have a growing addressable market, a proven strategy and a differentiated portfolio that is levered to long-term market trends around networking, hybrid cloud and AI. I believe we have a significant opportunity ahead of us. I'm excited to partner with Antonio and the rest of the outstanding HPE team to capitalize on this opportunity and drive value for our shareholders.

謝謝你，安東尼奧。我很高興今天能夠以 HPE 財務長的身份與大家一起參加我的第一次財報電話會議。我一直很欽佩 HPE 令人矚目的轉型，成為這家公司的一員是再令人興奮不過的事情了。我們擁有不斷成長的潛在市場、行之有效的策略和差異化的產品組合，並順應網路、混合雲和人工智慧等長期市場趨勢。我相信我們面前有一個重大機會。我很高興能與安東尼奧和其他優秀的 HPE 團隊合作，利用這個機會為我們的股東創造價值。

And as Antonio mentioned, we have much to be proud of. Financial highlights in the quarter included record gross margins and expense discipline, which helped lift non-GAAP EPS to the high end of our guidance range. Demand for our traditional server and storage products has stabilized. Demand for our HPE GreenLake offerings was evident in a healthy ARR growth and demand for our AI systems remains robust.

正如安東尼奧所說，我們有很多值得驕傲的事。本季度的財務亮點包括創紀錄的毛利率和費用控制，這有助於將非公認會計準則每股收益提升至我們指導範圍的高點。我們傳統伺服器和儲存產品的需求已經穩定下來。我們對 HPE GreenLake 產品的需求體現在健康的 ARR 成長中，並且對我們的 AI 系統的需求仍然強勁。

However, demand in Intelligent Edge did soften due to customer digestion of strong product shipments in fiscal year '23, which is lasting longer than we initially anticipated and is the primary reason Q1 revenue came in below our expectations. GPU availability and deal timing also contributed. We are taking swift action to address these headwinds by curtailing costs and driving efficiencies across the business.

然而，由於客戶消化了 23 財年強勁的產品出貨量，智慧邊緣的需求確實有所減弱，這種持續時間比我們最初預期的要長，也是第一季營收低於我們預期的主要原因。 GPU 的可用性和交易時機也起到了一定作用。我們正在迅速採取行動，透過削減成本和提高整個業務的效率來應對這些阻力。

With that, let's take a closer look at the details of the quarter. Revenue fell 14% year-over-year in constant currency to $6.8 billion. Please recall, we had significant backlog consumption in Q1 '23, particularly in traditional servers and storage. Backlog has now largely normalized across our business with the exception of our APU products.

因此，讓我們仔細看看本季的細節。以固定匯率計算，營收年減 14% 至 68 億美元。請回想一下，我們在 23 年第一季有大量積壓消費，特別是在傳統伺服器和儲存方面。除 APU 產品外，我們整個業務的積壓訂單現在已基本恢復正常。

We have strong momentum in HPE GreenLake. ARR exceeded $1.4 billion in Q1. Storage and networking software and services are the fastest growth elements of ARR. Our software and services mix rose 400 basis points year-over-year to 69%. ARR is the best indicator of our model transformation to our as-a-service offerings. This growth validates what our customers are telling us, that HPE GreenLake is a differentiated value proposition in the market.

HPE GreenLake 發展勢頭強勁。第一季的 ARR 超過 14 億美元。儲存和網路軟體及服務是ARR成長最快的元素。我們的軟體和服務組合年增400個基點，達到69%。 ARR 是我們轉向即服務產品模式的最佳指標。這一成長驗證了客戶告訴我們的事實，即 HPE GreenLake 是市場上的差異化價值主張。

Our Q1 non-GAAP gross margin was 36.2%. It rose 200 basis points year-over-year driven by a mix shift to Intelligent Edge and favorable input cost management. We are pleased that our non-GAAP gross margin is up 600 basis points from fiscal year '18 to a company record. This illustrates the ongoing pivot to higher growth, higher margin revenue across our portfolio.

我們第一季的非 GAAP 毛利率為 36.2%。受向智慧邊緣組合轉變和有利的投入成本管理的推動，該業務同比增長了 200 個基點。我們很高興，我們的非公認會計準則毛利率較 2018 財年上升了 600 個基點，創下了公司最高紀錄。這顯示我們的投資組合正在不斷轉向更高的成長、更高的利潤收入。

Given the current networking market dynamics, we are taking this moment to streamline and simplify our operations. We are focused on controlling the things we can control. Prudent cost management and disciplined execution are important regardless of the macro environment and are even more critical at times like these.

鑑於目前的網路市場動態，我們正在利用這一時機簡化和精簡我們的營運。我們專注於控制我們能夠控制的事情。無論宏觀環境如何，審慎的成本管理和嚴格的執行都很重要，而在這樣的時刻就顯得更加關鍵。

The benefits of our focus are already evident in our Q1 non-GAAP operating expenses, which declined 4.7% year-over-year at 9.5% sequentially to $1.7 billion. The strong Q1 non-GAAP gross margins and OpEx discipline led to Q1 non-GAAP operating margins of 11.5%, which is down only 30 basis points year-over-year despite less revenue scale. Favorable timing on some corporate expenses was a tailwind to Q1 operating expense and will appear in Q2. GAAP EPS of $0.29 and non-GAAP EPS of $0.48 exceeded the midpoint of our guidance range. Our diluted share count was approximately 1.3 billion, and non-GAAP diluted net earnings per share exclude $251 million in costs primarily from stock-based compensation expense, amortization of intangibles and noncash loss on investments.

我們專注的優勢已經在我們的第一季非 GAAP 營業費用中得到明顯體現，該費用年減 4.7%，環比下降 9.5%，至 17 億美元。強勁的 Q1 非 GAAP 毛利率和營運支出約束導致 Q1 非 GAAP 營業利潤率為 11.5%，儘管收入規模較小，但同比僅下降 30 個基點。一些公司費用的有利時機對第一季的營運費用產生了順風作用，並將在第二季顯現。 GAAP 每股收益為 0.29 美元，非 GAAP 每股收益為 0.48 美元，均超過了我們預期範圍的中點。我們的稀釋股數約為 13 億股，非 GAAP 稀釋每股淨收益不包括 2.51 億美元的成本，這些成本主要來自股票薪酬費用、無形資產攤銷和投資非現金損失。

As we manage the business with focus and discipline, we will also invest to capitalize on the sizable opportunities associated with moving through the interrelated inflection points in networking, hybrid cloud and AI all at the same time. HPE is evolving into a more simple, more agile company that is even better positioned to pursue our growth opportunities and to evolve our mix of products, services and software, and to drive structural higher profitability.

在我們專注且嚴謹地管理業務的同時，我們還將進行投資，以利用同時在網路、混合雲和人工智慧相互關聯的轉折點中帶來的巨大機會。 HPE 正在發展成為一家更簡單、更敏捷的公司，它更有能力追求成長機會，改進我們的產品、服務和軟體組合，並提高結構性的獲利能力。

Turning to our segment results. In addition to the new segments we discussed at SAM, we also created a new service segment that combines our prior compute and HPC and AI segment under one streamlined segment that offers solutions for our customers' training, tuning, and inferencing AI needs across their AI life cycle. Server revenues were $3.4 billion in the quarter, which was down 23% year-over-year. This new segment had a difficult year-over-year compare as, in Q1 '23, the business made significant progress against its backlog and benefited from HPE Cray EX revenue for Frontier.

談到我們的分部表現。除了我們在 SAM 上討論的新細分市場之外，我們還創建了一個新的服務細分市場，將我們先前的運算、HPC 和 AI 細分市場整合到一個精簡的細分市場中，為我們的客戶在整個 AI 生命週期中的訓練、調整和推理 AI 需求提供解決方案。本季伺服器營收為 34 億美元，年減 23%。這個新部門的年比表現很艱難，因為在 23 年第一季度，該業務在積壓訂單方面取得了重大進展，並受益於 Frontier 的 HPE Cray EX 收入。

We are capturing the robust growth in AI demand. Our cumulative APU orders since Q1 '23, which include APU attached products and services in our HPE Cray EX, XD and ProLiant systems, rose approximately $500 million sequentially to now $4 billion. Our pipeline is robust and GPU supply, while still constrained, is improving. Our APU product revenue increased sequentially to well over $400 million. We are now converting our APU orders into revenues, and yet very strong demand means our APU backlog is over $3 billion.

我們正在捕捉人工智慧需求的強勁成長。自 23 年第一季以來，我們的累計 APU 訂單（包括 HPE Cray EX、XD 和 ProLiant 系統中的 APU 連線產品和服務）較上季成長約 5 億美元，目前達到 40 億美元。我們的管道非常強大，GPU 供應雖然仍然受到限制，但正在改善。我們的APU產品營收連續成長至4億美元以上。我們現在正在將 APU 訂單轉化為收入，但強勁的需求意味著我們的 APU 積壓訂單超過 30 億美元。

Our pipeline is well above that. AI revenue in the quarter included our first revenues from AI cloud offering and from our HPE GreenLake win with a large hyperscale customer. Our traditional high-performance computing and supercomputing revenue fell seasonally, sequentially following a strong Q4. Revenue from our traditional server business increased sequentially. We expect this trend to continue given a structural mix shift to higher AUP Gen11 and rising input costs. Gen11 servers nearly doubled sequentially to 30% of the mix in Q1.

我們的管道遠高於這個水準。本季的 AI 收入包括我們從 AI 雲端產品和 HPE GreenLake 與大型超大規模客戶合作中獲得的第一筆收入。我們傳統的高效能運算和超級運算收入在第四季度強勁成長之後出現季節性下降。我們傳統伺服器業務的收入連續增加。鑑於結構性組合轉向更高的 AUP Gen11 和投入成本的上升，我們預計這一趨勢將會持續下去。 Gen11 伺服器在第一季較上季成長近一倍，達到 30%。

Including HPE, Cray XD takes the mix of next-generation servers to 44%. We are encouraged that our Gen11 pipeline is starting to include AI inferencing activity and enterprise applications. We expect AI inferencing to gather momentum in fiscal year '24.

包括 HPE 在內，Cray XD 的下一代伺服器佔比達到 44%。我們很高興看到我們的 Gen11 管道開始包含 AI 推理活動和企業應用程式。我們預計人工智慧推理將在 24 財年獲得發展動力。

As a reminder, our Gen11 services come with an attached subscription to our compute ops management software, which lifts our margin structure. Our Q1 operating margin was 11.4%. While up sequentially, the margin was down 430 basis points year-over-year given declining revenues. Intelligent Edge revenues were $1.2 billion or up 2% year-over-year. Demand for our campus switching and WiFi products eased materially, particularly in Europe and Asia, and was the largest contributor to our Q1 revenue gap. We continue to see mid-single-digit growth benefits from our existing backlog but expect these to normalize going forward as we are now approaching our typical range.

提醒一下，我們的 Gen11 服務附帶我們的計算操作管理軟體的訂閱，這提升了我們的利潤結構。我們第一季的營業利益率為 11.4%。利潤率雖然環比上升，但由於收入下降，比去年同期下降了 430 個基點。 Intelligent Edge 營收為 12 億美元，年增 2%。我們校園交換和 WiFi 產品的需求大幅下降，尤其是在歐洲和亞洲，這是我們第一季收入差距的最大原因。我們繼續從現有的積壓訂單中看到中等個位數的成長收益，但預計這些收益將在未來正常化，因為我們現在正接近我們的典型範圍。

Our total channel inventory is within the normal range overall, inclusive of a pocket in the SMB business. We also continue to make progress in our new TAMs of data center networking, private 5G and SASE. The Intelligent Edge portfolio of subscription revenue grew well above 50% as we are benefiting from growing attach from our strong fiscal year '23 revenue growth. We are pleased with our 29.4% operating margin, which was up 1,000 basis points year-over-year. The new Hybrid Cloud segment includes our storage businesses and now the HPE GreenLake portion of our Server business. Revenues of $1.2 billion were down 10% year-over-year.

我們的總通路庫存總體上處於正常範圍內，其中包括中小企業業務的一部分。我們在資料中心網路、私有 5G 和 SASE 等新的 TAM 方面也持續取得進展。由於我們受益於 23 財年強勁的營收成長，智慧邊緣訂閱營收組合成長遠超 50%。我們對 29.4% 的營業利潤率感到滿意，比去年同期成長了 1,000 個基點。新的混合雲部門包括我們的儲存業務以及現在的伺服器業務的 HPE GreenLake 部分。營收為 12 億美元，年減 10%。

Our traditional storage business was down year-over-year on difficult compares, given backlog consumption in Q1 '23. Total electro subscription revenue grew over 100% year-over-year and is an illustration of our long-term transition to an as-a-service model across our businesses. We are starting to see AI server demand pull through interest in our file storage portfolio. We are also already seeing some cross-selling benefits of integrating the majority of our HPE GreenLake offering into a single business unit.

由於 23 年第一季的積壓消費，我們的傳統儲存業務較去年同期下滑，這很困難。總電子訂閱收入年增超過 100％，這表明我們的業務正在長期向「即服務」模式轉型。我們開始看到人工智慧伺服器需求拉動我們對文件儲存產品組合的興趣。我們也看到將大部分 HPE GreenLake 產品整合到單一業務部門所帶來的一些交叉銷售優勢。

Our operating margin was 3.8%, which was down 200 basis points year-over-year. Lower revenues and a high mix of third-party product impacted margins in the traditional storage business. Our HPE Financial Services revenue was down 2% year-over-year and financing volume was $1.4 billion. Our operating margin of 8.5% was up 130 basis points year-over-year. We are successfully passing through interest rate increases, and our asset management margins are returning to normal. Our Q1 loss ratio remained steady at 0.5%.

我們的營業利益率為 3.8%，年減 200 個基點。收入下降和第三方產品混合度高影響了傳統儲存業務的利潤率。我們的HPE金融服務收入較去年同期下降2%，融資額為14億美元。我們的營業利益率為 8.5%，比去年同期上升了 130 個基點。我們正成功度過升息階段，資產管理利潤率正在恢復正常。我們第一季的損失率保持穩定在 0.5%。

Turning now to cash flow and capital allocation. We generated $64 million in cash flow from operations and consumed $482 million in free cash flow this quarter. HPE typically consumes cash in the first half of the year and generates cash in the second half. Our first quarter free cash flow benefited from some prepayments associated with pending HPE GreenLake deals and HPE Cray XD shipments. Our cash conversion cycle was 7 days, which is a reduction of 8 days from Q1 '23. Our days payable and days of inventory were both higher, given the lower revenue and strong demand for APUs. We returned $172 million in capital to shareholders in Q1, primarily through our dividend.

現在來談談現金流和資本配置。本季度，我們產生了 6,400 萬美元的營運現金流，並消耗了 4.82 億美元的自由現金流。 HPE 通常在上半年消耗現金，在下半年產生現金。我們第一季的自由現金流受益於與待定的 HPE GreenLake 交易和 HPE Cray XD 出貨相關的一些預付款。我們的現金轉換週期為 7 天，比 23 年第一季減少了 8 天。由於營收較低且 APU 需求強勁，我們的應付帳款天數和庫存天數均較高。我們在第一季向股東返還了 1.72 億美元的資本，主要透過股息的方式。

Before I discuss our outlook, let me first recap the key drivers that factor into our expectations for Q2 and the full year. For server, we expect improving GPU supply to drive sequential revenue increases through fiscal year '24. Given improving supply and the timing of installations, we expect segment revenue and therefore, corporate revenue to be heavily weighted to the second half of the year. We expect the blended margin of the new segment to be flattish for the remainder of the fiscal year.

在討論我們的展望之前，讓我先回顧一下影響我們對第二季和全年預期的關鍵驅動因素。對於伺服器，我們預計 GPU 供應的改善將推動 24 財年營收的連續成長。鑑於供應的改善和安裝時間的安排，我們預計分部收入以及公司收入將主要集中在下半年。我們預計新部門的混合利潤率將在本財年剩餘時間內保持穩定。

For Intelligent Edge, we expect the market to remain soft throughout the year. Our cost reduction efforts will take time to show their benefits, which will result in margin pressure in Q2. Later in our fiscal year, we expect our normal channel inventory position and seasonal strength in the education market to be a modest revenue tailwind. We expect the full year margin to be in the mid-20% range. For Hybrid Cloud, we expect sequential increases through the year as our traditional storage business improves and HPE GreenLake momentum continues. We expect meaningful progress through the year.

對於智慧邊緣，我們預計市場將全年保持疲軟。我們的成本削減努力需要時間才能顯現成效，這將導致第二季的利潤壓力。在我們的財政年度後期，我們預計我們正常的通路庫存狀況和教育市場的季節性實力將為收入帶來溫和的順風。我們預計全年利潤率將在 20% 左右。對於混合雲，隨著我們傳統儲存業務的改善和 HPE GreenLake 的持續發展勢頭，我們預計全年將實現連續成長。我們期待今年能夠取得有意義的進展

With that context, let me turn to our outlook. For Q2, we expect revenues in the range of $6.6 billion to $7 billion or up slightly sequentially at the midpoint. We expect GAAP diluted net EPS to be between $0.20 and $0.25, and our non-GAAP diluted net EPS between $0.36 and $0.41. In part, given the higher prepayments in Q1, we expect negative free cash flow in Q2.

在此背景下，讓我來談談我們的展望。對於第二季度，我們預計營收將在 66 億美元至 70 億美元之間，或比中位數略有上升。我們預計 GAAP 稀釋淨每股收益在 0.20 美元至 0.25 美元之間，非 GAAP 稀釋淨每股收益在 0.36 美元至 0.41 美元之間。部分原因是，考慮到第一季的預付款較高，我們預計第二季的自由現金流將為負值。

Let me remind you that we have 2 drivers of potential lumpiness in our business. One is the timing and customer acceptances of large Cray wins, including Cray XD wins, which should ramp beginning in Q2. The second is the start of certain large HPE GreenLake deals. For some of these large deals, our segments recognize the related hardware on installation. In our consolidated results, we eliminate the segment revenue and recognize it over time. Both large deals and higher eliminations are indicators of our confidence in the second half of the fiscal year.

讓我提醒你一下，我們的業務中存在兩個潛在的不穩定性因素。一是 Cray 大單的時機和顧客接受度，包括 Cray XD 大單，應該會從第二季開始增加。第二個是某些大型 HPE GreenLake 交易的開始。對於其中一些大型交易，我們的部門會識別安裝時相關的硬體。在我們的合併結果中，我們消除了分部收入並隨著時間的推移確認它。大型交易和更高的抵銷金額都顯示了我們對財年下半年的信心。

For fiscal year '24, we are revising our outlook, primarily given the current networking market headwinds. Let me remind you that we are excluding the H3C earnings and gain on sale from our non-GAAP results beginning in fiscal year '24. We now expect revised ranges for constant currency revenue and non-GAAP operating profit growth of 0% to 2%, GAAP diluted net EPS of $1.81 to $1.91 and non-GAAP diluted net EPS of $1.82 to $1.92.

對於24財年，我們正在修改我們的展望，主要考慮到當前網路市場的逆風。讓我提醒您，從 24 財年開始，我們將從非 GAAP 業績中排除 H3C 的收益和銷售收益。我們現在預計，以固定匯率計算的收入和非 GAAP 營業利潤的修訂範圍為 0% 至 2%，以 GAAP 計算的稀釋淨每股收益為 1.81 美元至 1.91 美元，以非 GAAP 計算的稀釋淨每股收益為 1.82 美元至 1.92 美元。

The mix shift from Intelligent Edge to Server should also weigh on our gross margins. We now expect the full year non-GAAP gross margin to be slightly down from our prior full year expectation of 35%. We expect the impact of the cost actions we have initiated to materialize in the second half and leave fiscal year '24 OpEx to be flat to down from fiscal '23 OpEx. We expect our operating margin to be flattish year-over-year. We now expect OI&E to be $200 million to $250 million headwind versus our prior $300 million expectation, given better Q1 free cash flow and more favorable interest rate assumptions.

從智慧邊緣到伺服器的混合轉變也會對我們的毛利率產生影響。我們現在預計全年非 GAAP 毛利率將略低於我們先前全年預期的 35%。我們預計，我們已啟動的成本行動的影響將在下半年實現，並使 24 財年的營運支出與 23 財年的營運支出持平或下降。我們預計我們的營業利潤率將與去年同期持平。鑑於第一季自由現金流改善和更有利的利率假設，我們現在預計 OI&E 將出現 2 億至 2.5 億美元的逆風，而我們之前的預期是 3 億美元。

We expect the effect on currency to be immaterial. We expect free cash flow to be at least $1.9 billion in fiscal year '24. We expect significantly stronger free cash flow in the second half of the year led by improvements in inventories as AI service shipments ramp. We reiterate our commitment to our dividend, which was raised 8% from fiscal year '23 to fiscal year '24; debt repayment to maintain an investment-grade credit rating; and in long term, returning capital to shareholders through share repurchases.

我們預期這對貨幣的影響不大。我們預計24財年的自由現金流至少達到19億美元。我們預計，隨著人工智慧服務出貨量的增加，庫存的改善將帶動下半年自由現金流顯著增強。我們重申對股利的承諾，從 23 財年到 24 財年，股利增加了 8%；償還債務以維持投資等級信用評等；長遠來看，透過股票回購向股東返還資本。

To conclude, we executed well in Q1 amidst a challenging market backdrop. We are pleased with our margin and EPS results while understanding our slower networking product demand and GPU availability and timing impacted our revenue performance. We have taken prompt action to further reduce our costs and continue to manage our expenses prudently while we advance our long-term strategy. AI server demand is strong. Demand has stabilized for our traditional server and storage products, and our HPE GreenLake momentum is robust.

總而言之，在充滿挑戰的市場背景下，我們在第一季表現良好。我們對我們的利潤率和每股盈餘結果感到滿意，同時也了解到我們的網路產品需求放緩以及 GPU 可用性和時間安排影響了我們的收入表現。我們迅速採取行動，進一步降低成本，並在推動長期策略的同時繼續審慎管理開支。 AI伺服器需求強勁。我們傳統伺服器和儲存產品的需求已經穩定，我們的 HPE GreenLake 勢頭強勁。

We will continue to invest in IT inflections, networking, hybrid cloud and AI, to drive our pivot to higher growth, higher margin revenue. I look forward to engaging with you in the months ahead, as does our new Chief Strategy Officer, Shannon Cross, who has joined us following a distinguished career as a Wall Street analyst and now has oversight of our IR function.

我們將繼續在 IT 轉型、網路、混合雲和人工智慧方面進行投資，以推動我們實現更高的成長、更高的利潤收入。我期待在未來幾個月與您合作，我們的新任首席策略長 Shannon Cross 也同樣期待，他在加入我們之前曾是華爾街的傑出分析師，現在負責監督我們的 IR 職能。

We will appreciate your input and questions along the way, and we can get started with that right now. I'll open it up now for your questions about the quarter.

我們將非常感謝您在過程中的意見和問題，我們現在就可以開始處理。我現在開始回答大家關於本季的問題。

(Operator Instructions) The first question today is from Meta Marshall with Morgan Stanley.

（操作員指示）今天的第一個問題來自摩根士丹利的 Meta Marshall。

Maybe on the GPU available -- or GPU delays that you're seeing in far as acceptance, just what are you seeing in terms of -- you identified that power and some of these things were conditions for what the delays were. But just what are you seeing in terms of how long those delays are going to take and how long the delays and the acceptances are going to be?

也許在 GPU 可用方面 - 或者就接受度而言您所看到的 GPU 延遲，只是您所看到的 - 您確定了功率以及其中一些因素是造成延遲的條件。但是，您認為這些延遲會持續多久以及延遲和接受會持續多久？

Well, thank you, Meta. So as I said in my prepared remarks, we had a couple of deals that slipped from Q1 into future quarters because customers are taking a little bit longer to prepare the data center space, getting the power and the cooling ready. And obviously, those deals will come as we complete those installations.

好吧，謝謝你，Meta。正如我在準備好的演講中所說的那樣，我們有幾筆交易從第一季推遲到了未來幾個季度，因為客戶需要更長的時間來準備資料中心空間、電力和冷卻系統。顯然，這些交易將會在我們完成這些安裝後達成。

And then on the GPU side, obviously, we continue to experience a tight environment, although we are seeing some improvements. We have already a lot of GPUs that we already built, but the customers will take time to accept those systems. But the reality is that we need more supply against the backlog that we announced today, which was $3 billion at the end of the Q1.

然後在 GPU 方面，顯然，我們繼續經歷緊張的環境，儘管我們看到了一些改進。我們已經建立了很多 GPU，但客戶需要時間來接受這些系統。但現實情況是，我們需要更多的供應來彌補我們今天宣布的積壓訂單，截至第一季末積壓訂單金額為 30 億美元。

So that's what we see today. And as we go forward, we expect that improvement to happen, and that's why we are confident on the conversion of the GPU orders into revenue as we go along, not just because of the GPU availability but also the acceptances.

這就是我們今天所看到的情況。隨著我們繼續前進，我們預計這種改善將會發生，這就是為什麼我們對 GPU 訂單轉化為收入充滿信心，這不僅是因為 GPU 的可用性，還因為接受度。

And the next question is from Amit Daryanani with Evercore.

下一個問題來自 Evercore 的 Amit Daryanani。

I guess, Antonio, maybe just talk about, if I look at the revenue shortfall in Jan quarter, how much of that do you think is because your customers are pushing out their delivery schedules because there isn't enough power versus you just didn't have enough GPUs? If you were to think about those 2 buckets. And then as you think about the full year guide, perhaps I didn't appreciate this. But can you just talk about what are you expecting for the networking segment, Intelligent Edge to do in the 0% to 2% guide right now?

我想，安東尼奧，也許可以談談，如果我看一下 1 月份季度的收入缺口，您認為其中有多少是因為您的客戶因為電力不足而推遲了交貨計劃，還是因為您沒有足夠的 GPU？如果您考慮一下這兩個桶。然後，當您考慮全年指南時，也許我並不欣賞這一點。但是，您能否談談您對網路領域智慧邊緣目前在 0% 到 2% 範圍內的表現有何期望？

So I'll take -- I'll answer the first part of your question around the revenue. So in the first quarter with respect to what drove the $300 million on the revenue, it was mostly actually networking. We did have one deal that moved out. I think Antonio mentioned that in his prepared remarks. And then in terms of just how we're thinking about the second half of the guide, we are expecting a very strong second half, and that's predominantly driven actually by AI systems revenue. We are expecting networking to be slightly more favorable in the back half, but we really see the trough of networking in Q2, Amit.

所以我會——我將回答你關於收入的問題的第一部分。因此，就第一季 3 億美元收入的推動因素而言，主要實際上來自於網路。我們確實有一筆遷出交易。我認為安東尼奧在他的準備好的演講中提到了這一點。然後就我們如何看待指南的下半部分而言，我們預計下半部分將表現非常強勁，而這主要得益於人工智慧系統收入。我們預計下半年網路狀況會稍微好一些，但我們確實看到第二季度網路狀況的低谷，阿米特。

And the next question is from Simon Leopold with Raymond James.

下一個問題來自雷蒙德詹姆斯 (Raymond James) 的西蒙利奧波德 (Simon Leopold)。

I wanted to see if we could drill down a little bit in terms of understanding what's changed in the Intelligent Edge versus 90 days ago. And 2 elements are crossing my mind. One is really around the pending Juniper deal, whether that's influencing customers to maybe hold off purchases because of the uncertainties that might be affecting their decision-making as to what happens after you're combined. And the other part is just wondering, if there's inventory that's been sitting in the channel, why didn't you know about it or why didn't you see it. Just trying to get an understanding of sort of what you've learned over these last 90 days.

我想看看我們是否可以深入了解與 90 天前相比智慧邊緣發生了哪些變化。目前有兩個元素浮現在我的腦海中。一個是圍繞即將達成的 Juniper 交易，這是否會影響客戶，使他們可能推遲購買，因為這些不確定性可能會影響他們對合併後會發生什麼的決策。而另一部分人只是想知道，如果有庫存一直在渠道中，為什麼你不知道或為什麼沒有看到它。只是想了解過去 90 天您學到了什麼。

Thank you, Simon. So first of all, we saw an acceleration of the demand softness in the back -- at the end of the Q1, really in January, whether it's now people coming back, but obviously, the reality is we saw that as a headwind to our revenue in Q1. I have to say, we do not have a channel inventory problem. Actually, we are in great position. Particularly from the enterprise customers and the enterprise product, we do not have that issue at this point in time.

謝謝你，西蒙。首先，我們看到需求疲軟加速——在第一季末，實際上是在 1 月份，不管現在人們是否回來了，但顯然，現實情況是，我們認為這對我們第一季的收入造成了阻力。我必須說，我們沒有渠道庫存問題。事實上，我們處於非常有利的地位。特別是從企業客戶和企業產品來看，我們目前還沒有這個問題。

What we do see is customers are taking longer with the product we already shipped to them to install it and eventually go through the next cycle. And that's why we said with Marie, we're going to start seeing a slight improvement on the back half with Q2 being the trough. And part of the back half also is the traditional buying season in the United States with state, local education. The pipeline is very good. We have not lost one single deal that I can point to, neither because of the slowdown or customers deferring, not because of the announcement of the acquisition of Juniper.

我們確實發現，客戶需要花更長的時間來安裝我們已經發送給他們的產品，並最終經歷下一個週期。這就是為什麼我們說，對於瑪麗來說，我們將開始看到後半部分略有改善，而第二季度是低谷。下半年部分地區也是美國傳統的購屋旺季，包括州和地方教育。管道非常好。就我所知，我們沒有失去任何一筆交易，無論是因為經濟放緩或客戶推遲，還是因為宣布收購瞻博網路。

And maybe, Simon, just to add to Antonio's comments, the only place where we saw a slightly elevated pocket of inventory was in SMB, which is a pretty small part of our business.

也許，西蒙，只是為了補充安東尼奧的評論，我們看到庫存略有增加的唯一地方是中小企業，這只是我們業務的一小部分。

Yes.

是的。

And the next question is from Toni Sacconaghi with Bernstein.

下一個問題來自伯恩斯坦的托尼·薩科納吉。

Sorry, I just have one clarification and a question. Marie, on the (inaudible) that the backlog drop down has contributed, I think, mid-single digits to the Intelligent Edge or something more broad than that. So can you just comment or clarify exactly what the backlog contribution was? And I suppose there's none going forward.

抱歉，我只想澄清一件事並提出一個問題。瑪麗，關於（聽不清楚）積壓訂單的減少，我認為，對智慧邊緣或比這更廣泛的東西的貢獻是中等個位數。那麼能否評論一下或澄清一下積壓貢獻到底是什麼？我想，未來也不會有什麼進展了。

And then just on my question, you sound pretty excited about sequential growth over the course of the year, both in servers and storage. Maybe you can just elaborate on why you see that. Is servers -- do you see sequential growth in traditional servers, non-accelerated (inaudible) and what (inaudible)

然後就我的問題而言，您似乎對今年伺服器和儲存的連續成長感到非常興奮。也許您可以詳細說明您看到這種情況的原因。伺服器－您是否看到傳統伺服器、非加速伺服器（聽不清楚）以及哪些（聽不清楚）的連續成長？

So maybe I'll take the first part of the question, and I'll turn to Antonio for the second part. So look, in terms of the backlog, we really don't disclose the backlog on edge. But I think what we've said is that we've seen our backlog sort of revert back to normalized levels with the exception, obviously, of our APU or AI system. So that's how we're thinking about the backlog.

因此，也許我會回答問題的第一部分，然後我會請安東尼奧回答第二部分。所以，就積壓訂單而言，我們確實沒有透露邊緣積壓訂單的情況。但我認為，我們已經看到我們的積壓訂單已經恢復到正常水平，當然，APU 或 AI 系統除外。這就是我們對積壓問題的看法。

I think in terms of just some context and commentary, you've seen in the industry that the market has definitely softened. And I think as Antonio said earlier, we saw that late in the quarter. So that's how we characterize the networking demand. And I would say that we do expect the trough in Q2 and to be slightly more favorable in the back half. So that's how we're thinking about the networking market playing out for the year. I think I'll turn to Antonio to comment on servers.

我認為，就一些背景和評論而言，你已經看到行業市場確實已經疲軟。我認為正如安東尼奧之前所說，我們在本季末看到了這一點。這就是我們對網路需求的定義。我想說的是，我們確實預計第二季會出現低谷，而下半年的情況會略有好轉。這就是我們對今年網路市場前景的看法。我想我會讓安東尼奧對服務器進行評論。

Yes, Toni. Thank you for the question. So on the server and storage side, obviously, we see signs of stabilization. It has been now a couple of quarters with sequential order improvement. But the reality is that, as we said in our opening remarks, we continue to see the mix shift in the traditional servers, as you call it, to Gen11. By the end of the year, we should be approximately 60% of the way there. Obviously, those servers come with a different set of structural configurations and pricing, which obviously is higher.

是的，托尼。感謝您的提問。因此，在伺服器和儲存方面，我們顯然看到了穩定的跡象。現在已經有幾季的訂單連續改善了。但現實情況是，正如我們在開場白中所說的那樣，我們繼續看到傳統伺服器（正如您所說的）的混合轉變到 Gen11。到今年年底，我們應該已經完成約 60% 的進度。顯然，這些伺服器具有不同的結構配置，定價也顯然更高。

At the same time, we're going to see cost inflation. We believe that's going to be the case, and therefore, we have to eventually pass those as well. But the number of units have been very stable or slightly improving. And that's an important indicator because, ultimately, that also drives our attach rate of our Operational Services, which in the quarter was very, very good. So that's why we are confident there's sequential improvement from here on.

同時，我們將看到成本上漲。我們相信情況會如此，因此，我們最終也必須通過這些。但單位數量一直非常穩定或略有改善。這是一個重要的指標，因為最終它也會推動我們營運服務的附加率，而本季的附加率非常非常好。所以我們相信從現在開始情況會逐漸改善。

And then on the storage side, obviously, AI is going to be a pull-through demand for us. We introduced a new offer now specifically for file. And remember that HPE Alletra will continue to grow from here on. And a portion of that HPE Alletra revenue is also in the ARR because remember that software now is completely disaggregated from the solution itself, which means you have the CapEx portion of the revenue recognized in quarter and the subscription part of the software amortized over the period of the contract. So that's why ARR is growing because of the subscription in networking, which was up significantly, the storage and obviously, AI now also contributing to the ARR as well.

然後在儲存方面，顯然，人工智慧將成為我們的拉動需求。我們現在專門針對文件推出了一項新優惠。請記住，HPE Alletra 將從此繼續發展。 HPE Alletra 收入的一部分也包含在 ARR 中，因為請記住軟體現在已經與解決方案本身完全分離，這意味著您在季度中確認了收入的資本支出部分，而軟體的訂閱部分則在合約期間攤銷。這就是為什麼 ARR 會成長，因為網路訂閱量大幅增加，存儲，以及顯然 AI 也為 ARR 做出了貢獻。

The next question is from Aaron Rakers with Wells Fargo.

下一個問題來自富國銀行的 Aaron Rakers。

I wanted to ask about the server market, maybe 2 parts. I guess when we look at some of your peers, I mean, it seems to be that the lead times have improved on some of the GPUs, particularly the H100. But I'd be curious of kind of like can you talk a little bit about what you've seen on lead times there in terms of your ability to deliver on some of this backlog, how that's changed over the course of this last quarter? And then any thoughts on traditional server recovery? How do we think about the pace of that embedded in your expectations looking through this year?

我想問一下伺服器市場，大概有兩個部分。我想，當我們看看你們的一些同行時，我的意思是，似乎一些 GPU 的交貨時間有所改善，特別是 H100。但我很好奇，您能否談談您在交付週期方面看到的情況，以及您完成部分積壓訂單的能力，以及上個季度的情況有何變化？您對傳統伺服器復原有什麼看法？展望今年，您如何看待這項進程？

Yes, sure. I think I covered the larger part of that question with Toni's question about sequential improvement in the traditional server, which obviously is still very CPU centric. On a combined server, right, now 25% of the total volume is APUs, which obviously GPUs is the biggest portion of it. So we expect that sequential improvement driven by recovery in demand in units and then obviously the shift to Gen11, which is important in this transition.

是的，當然。我認為 Toni 關於傳統伺服器的連續改進的問題已經涵蓋了該問題的大部分內容，顯然傳統伺服器仍然以 CPU 為中心。在組合伺服器上，現在總量的 25% 是 APU，顯然 GPU 是其中最大的一部分。因此，我們預計，單位需求的復甦將推動連續的改善，然後顯然轉向 Gen11，這在這一轉變中非常重要。

On the GPU lead times, they have come down but it's still elevated. We're talking about 20-plus weeks at least lead time. And so -- and then it's going to be a combination of multiple type of GPUs, right, because there is still demand from the prior generation to H100. Obviously, the majority of the demand today is on H100. And going forward, we're going to have the Grace Hopper H200 and others, right, including MI300X and the like.

就 GPU 交付週期而言，它們已經縮短，但仍然較高。我們談論的是至少20多周的準備時間。所以 — — 然後它將是多種類型 GPU 的組合，對吧，因為從上一代到 H100 仍然有需求。顯然，今天的大部分需求是在 H100 上。展望未來，我們將推出 Grace Hopper H200 以及其他產品，包括 MI300X 等。

And the difference for us is that because we have a unique networking interconnect fabric, we can support all of them. So that's an important differentiation that I think everyone needs to remind because, while a lot of the volume today is NVIDIA and then on the supercomputing, which is also an AI business, by the way, we support all 3 of them. And so that, for us, gives us the optionality to convert the orders that we have and future orders we see in the pipeline with a little bit more flexibility, I will say.

而對我們來說，不同之處在於，由於我們擁有獨特的網路互連結構，因此我們可以支援所有這些。所以，我認為每個人都需要提醒這一點，這是一個重要的區別，因為雖然今天很大一部分業務來自 NVIDIA，然後在超級運算（也是 AI 業務）上，我們支持這三家公司。因此，對我們來說，這為我們提供了選擇權，讓我們能夠以更靈活的方式轉換我們現有的訂單和未來正在醞釀的訂單。

The next question is from Wamsi Mohan with Bank of America.

下一個問題來自美國銀行的 Wamsi Mohan。

You said some of your demand in AI systems is coming in via GreenLake. Can you help us understand the linkage between your view of AI revenue and ARR growth?

您說過，您對人工智慧系統的部分需求是透過 GreenLake 實現的。您能幫助我們了解您對 AI 收入和 ARR 成長的看法之間的關聯嗎？

Yes, Wamsi. I can start and then, Marie, feel free to add. I mean, the fact of the matter is that when you look at that $4 billion cumulative orders, a significant portion is going to go through the HPE GreenLake. If you recall last year, I announced that a hyperscaler placed an order with us, and that order is going through the GreenLake platform. And so that's why you see a breakdown over time of the AI GPU orders going through the ARR, which is fine.

是的，Wamsi。我可以開始，然後，瑪麗，請隨意添加。我的意思是，事實是，當你看到那 40 億美元的累積訂單時，很大一部分將透過 HPE GreenLake 進行。如果你還記得的話，去年我宣布一個超大規模提供者向我們下了訂單，而且該訂單正在透過 GreenLake 平台進行。這就是為什麼您會看到 AI GPU 訂單隨著時間的推移而透過 ARR 細分的原因，這很好。

Ultimately, they give us the ability to attach other services, which is important to remember here because, remember, when it goes to the HPE GreenLake, in many cases, we are actually running those systems for the customer. It's not just shipping the system to the customer. We actually put it in a location, whereas our data center footprint with our cooling and power, and then we attach our services, which are the run time plus other things we do.

最終，它們使我們能夠連接其他服務，這一點很重要，因為請記住，當它進入 HPE GreenLake 時，在許多情況下，我們實際上是在為客戶運行這些系統。這不僅僅是將系統運送給客戶。我們實際上把它放在一個位置，而我們的資料中心佔用我們的冷卻和電源，然後我們連接我們的服務，即運行時間加上我們所做的其他事情。

And why it's important also, the growth in ARR, because that drives a margin expansion and accretion over time. So that's what's going on in addition to the fact that obviously now we crossed 31,000 customers on HPE GreenLake platform. To put it in context, Wamsi, that's almost 3,000 customers in 1 quarter. I mean, 8% up quarter-over-quarter, that's 3,000 customers.

這也是為什麼 ARR 的成長很重要，因為它會隨著時間的推移推動利潤率的擴大和成長。這就是目前的情況，除此之外，顯然我們現在在 HPE GreenLake 平台上的客戶已經超過 31,000 名。具體來說，Wamsi，這意味著一個季度內就有近 3,000 名客戶。我的意思是，比上一季成長了 8%，也就是 3,000 名客戶。

And everything we do from the software perspective and now it's just subscription, whether you sell HPE ProLiant Gen11, let's say, an AI-optimized server, the software to connect the server actually runs to GreenLake. Obviously, the Alletra software runs through the GreenLake. A lot of the Aruba software, including Aruba Central, a subscription and now you have AI as well.

我們所做的一切都是從軟體角度進行的，現在只是訂閱，無論您銷售的是 HPE ProLiant Gen11，還是 AI 優化的伺服器，連接伺服器的軟體實際上都運行在 GreenLake 上。顯然，Alletra軟體是透過GreenLake運行的。許多 Aruba 軟體（包括 Aruba Central）都需要訂閱，現在您還可以擁有 AI。

And maybe, Wamsi, I'll just put a couple of numbers around the APU or the AI system orders that we saw, too. So we ended the quarter actually with $3 billion in backlog, so we really nearly tripled actually year-on-year. And in terms of just the link back to ARR, we shipped around $400 million in revenue, but we had incremental revenue actually that went into ARR. And that sort of underscores the growth that we saw in ARR and expect to see going forward as well, Wamsi.

也許，Wamsi，我也會對我們看到的 APU 或 AI 系統訂單給出幾個數字。因此，我們在本季結束時的積壓訂單量實際上為 30 億美元，與去年同期相比實際上增長了近三倍。就 ARR 的連結而言，我們的收入約為 4 億美元，但實際上我們有增量收入進入了 ARR。這在某種程度上凸顯了我們在 ARR 中看到的成長，並且預計未來也會出現這樣的成長，Wamsi。

The next question is from Samik Chatterjee with JPMorgan.

下一個問題來自摩根大通的 Samik Chatterjee。

I guess, Antonio, you referenced the increase in demand on AI that you see related to inferencing workloads on the enterprise side. Can you maybe talk a bit in terms of how these deployments are looking different to what you've been doing on the AI training side and maybe with some hyperscalers? And given the lead times, is this demand going to more materialize in relation to revenue more in fiscal '25? Or is that just a fair estimate about -- given the lead times?

我想，安東尼奧，您提到了與企業方面推理工作負載相關的人工智慧需求的增加。您能否談談這些部署與您在人工智慧訓練方面以及與一些超大規模企業所做的事情有何不同？考慮到交貨時間，這種需求是否會在 25 財年的收入中進一步實現？或者這只是一個合理的估計——考慮到交貨時間？

Yes. No, thank you. That's an excellent question. Obviously, I spoke about the AI life cycle, training, tuning, and inferencing. Obviously, the training side has been more focused on the hyperscalers, so Tier 2, Tier 3 type of providers or companies that are funded well to build these large language models.

是的。不，謝謝。這是一個非常好的問題。顯然，我談到了人工智慧的生命週期、訓練、調整和推理。顯然，訓練方面更加專注於超大規模提供商，因此 Tier 2、Tier 3 類型的提供者或資金充足的公司可以建立這些大型語言模型。

But when you look at enterprise, most of the enterprises are going to take a model and fine-tune it to give a context to the model with their data. And that, it can happen in multiple locations, right? It can happen in their data centers or potentially in a colo or, in some cases, in the public cloud, but we see more focus on where they can control the data in a secure environment.

但是當你看企業時，大多數企業都會採用一個模型並對其進行微調，以利用他們的數據為模型提供背景。而且，它可以在多個地方發生，對嗎？它可能發生在他們的資料中心，也可能發生在託管伺服器中，或者在某些情況下發生在公有雲中，但我們看到他們更加關注在安全的環境中控制資料的位置。

And then the inferencing side, it can happen in a data center or in a public cloud once they are all trained but also at the edge. In fact, we showcase a lot of the inferencing cases at the edge in Mobile Congress at the edge of the network. Think about use cases like the Coles Supermarket, right? So there is a lot of data to the video footage captured in the stores. That video footage needs to be inferenced right there at that given moment with 0 latency in order to deliver the outcome. And there are [others one] in manufacturing and the like.

然後是推理方面，一旦它們都經過訓練，它就可以在資料中心或公有雲中進行，但也可以在邊緣進行。事實上，我們在行動大會上展示了許多在網路邊緣進行邊緣推理的案例。想想像科爾斯超市這樣的用例，對吧？商店中捕獲的視訊片段包含大量數據。為了得出結果，需要在特定時刻立即推斷該影片片段，並且沒有任何延遲。還有其他一些行業，像是製造業等等。

And in fact, one of the use cases we saw also for inferencing is a large bank that now are doing some fine-tuning and inferencing to do risk management and other things. So we -- I will say we are kind of getting into it. I will say the growth will happen in the second half in '25. Definitely, the lead times will play a role. But I'm very encouraged about the momentum we see and the opportunity we have also with the combination of Juniper because most of these inferencing requires the network connectivity to deliver it. And that to me is one of the thesis why we went ahead with that acquisition.

事實上，我們看到的推理用例之一是一家大型銀行現在正在進行一些微調和推理以進行風險管理和其他事情。所以我們—我想說我們正在進入這個領域。我想說成長將在2025年下半年發生。毫無疑問，交貨時間會起到一定的作用。但我對我們所看到的發展動能以及與瞻博網路的結合所帶來的機會感到非常鼓舞，因為大多數推理都需要網路連結來實現。對我來說，這就是我們進行這項收購的理由之一。

The next question is from Tim Long with Barclays.

下一個問題來自巴克萊銀行的提姆朗。

Just on the -- another one on the AI server side. Can you talk a little bit about how you guys are thinking about profitability for these businesses as we get more accelerated compute in your servers? And if you can also kind of break that down between maybe the Cray businesses and the standard compute, is there going to be more of a margin gap in those 2 businesses when looking at more traditional going to accelerated?

就在——AI 伺服器端的另一個。您能否談談，當我們在伺服器上獲得更快速的計算時，你們如何考慮這些業務的盈利能力？如果您可以將其細分為 Cray 業務和標準運算業務，那麼從更傳統的加速發展來看，這兩大業務的利潤差距是否會更大？

Sure, I can start. I will say, listen, I think if you look at our Server segment that we just published, we delivered very strong performance. I mean we are in the target range we committed a while back of 11% to 13%. And that -- the fact that we bring them together give us the flexibility, opportunity to maximize the blended margin here as we go forward. To give a reference, when you sell an EX system, generally, it's a liquid cool system that tends to gravitate to the supercomputing side or large AI clusters of thousands of GPUs.

當然，我可以開始。我想說，聽著，如果你看一下我們剛剛發布的伺服器部分，你會發現我們的表現非常強勁。我的意思是，我們正處於我們先前承諾的 11% 到 13% 的目標範圍內。而且，將它們整合在一起這一事實為我們提供了靈活性和機會，可以在未來的發展中最大化混合利潤。舉個參考，當你銷售 EX 系統時，通常它是一種液體冷卻系統，傾向於超級運算方面或數千個 GPU 的大型 AI 叢集。

But an EX system supports today up to 80,000 GPUs in 1 system, and that's because of our interconnect fabric, HPE Slingshot. In fact, some of those systems have 80,000 GPUs and maybe 40,000 CPUs in 1 coherent system. But then you have other customers [are] many 2,000 or 4,000 GPUs. And depending on which location they pick, they need liquid cooling, we deploy those. Now generally speaking, the XD platform, the Cray XD platform is the one that has the density and is more air cool oriented and ability to mix-match different configurations, and that's where the vast majority of the action is today in AI.

但是，如今 EX 系統可以在 1 個系統中支援多達 80,000 個 GPU，這是因為我們的互連結構 HPE Slingshot。事實上，一些系統在 1 個連貫系統中擁有 80,000 個 GPU 和大約 40,000 個 CPU。但其他客戶也有很多 2,000 或 4,000 個 GPU。根據他們選擇的位置，他們需要液體冷卻，我們會部署這些液體冷卻。現在一般來說，XD 平台，Cray XD 平台具有密度高、更注重空氣冷卻、能夠混合搭配不同配置的特點，這就是當今 AI 領域絕大多數行動所在。

And ProLiant Gen11, actually, is more used for inferencing or in some areas of fine-tuning as well. So we have the flexibility to meet all those demands with our unique IP. And on top of that, we actually lay our Machine Learning Development Environment. In fact, there are customers that come to us just for the MLDE environment. Later on, we pull the server. Now on AUPs, I will tell you that when you sell an XD, depending on the [capacity], it can be 20x the value of a traditional server with CPUs. And in EX, it can be up to 35x. And so as we go forward, the ability to optimize margin through the configs and attach the services, whether it's our data center services plus the software and the Operational Services allows us to really drive the best outcome for our shareholders.

而ProLiant Gen11實際上更多地用於推理或某些微調領域。因此，我們可以靈活地利用我們獨特的 IP 來滿足所有這些需求。在此基礎上，我們實際上建立了機器學習開發環境。事實上，有些客戶就是為了 MLDE 環境來找我們的。稍後我們拉取伺服器。現在關於 AUP，我會告訴你，當你銷售 XD 時，根據 [容量]，它的價值可能是傳統帶有 CPU 的伺服器的 20 倍。在 EX 中，最高可達 35x。因此，隨著我們不斷前進，透過配置和附加服務來優化利潤的能力，無論是我們的資料中心服務還是軟體和營運服務，都使我們能夠真正為股東帶來最好的結果。

And that final question will come from Lou Miscioscia from Daiwa Capital Markets.

最後一個問題來自大和資本市場的 Lou Miscioscia。

Antonio, I guess the question I have is since you're talking a lot about data centers, I'm wondering what's going to happen. Is -- are the vast majority of x86 applications going to start to shift over to really be accelerated with GPUs due to the concern of more (inaudible) coming to end? And what I'm asking is not really inference and it's not training. These are just normal applications, sort of like the same way architectures shifted from IBM mainframes or PA years and years ago to x86, eventually to that and cloud. Do you think that that's going to shift over to running on GPUs?

安東尼奧，我想我的問題是，既然你談論了很多關於資料中心的事情，我想知道接下來會發生什麼。是否會因為擔心更多（聽不清楚）即將終結而開始真正轉向使用 GPU 加速？我所要求的其實並不是推理，也不是訓練。這些只是普通的應用程序，有點像多年前架構從 IBM 大型主機或 PA 轉向 x86，最終轉向雲端運算。您認為這會轉移到在 GPU 上運行嗎？

Well, thank you for the question. I think we need to understand there are 2 worlds that will coexist. There is the cloud-native world. Think about the cloud-native world where you have thousands and thousands of applications running on thousands and thousands of servers and they share everything. That architecture will exists for a long, long time because it's cost efficient, and the realization is that those applications rollout were designed for that type of environment, for the traditional monolithic approach to more a cloud-centric approach.

嗯，謝謝你的提問。我認為我們需要了解兩個世界將會共存。這就是雲端原生世界。想像一下雲端原生世界，成千上萬個應用程式在成千上萬台伺服器上運行，它們共享一切。這種架構將會長期存在，因為它具有成本效益，人們認識到這些應用程式的推出是為這種類型的環境而設計的，即從傳統的單一方法轉向以雲端為中心的方法。

And then you have these AI applications where you may have one application, only one, running on thousands and thousands of servers, which have accelerated compute. And it's a little bit farfetched to say everything is going to move there. I argue that you will have inferencing solutions that a CPU will be just fine.

然後你有這些人工智慧應用程序，其中可能有一個應用程序，只有一個，運行在成千上萬台伺服器上，這些伺服器已經加速了計算。說一切都會轉移到那裡有點牽強。我認為你會推斷 CPU 就可以解決問題。

Think about your phone, right? The phone will have, at some point, the ability to manage a large language model, let's say, 20 billion or 30 billion parameters; or the PC, maybe in the 80 billion to 100 billion parameters. But when you go up higher than that, obviously, you need potentially a server at the edge. And an 8 -- what I'll call 8-way GPU will be the right way to go.

想想你的手機，對吧？在某種程度上，手機將有能力管理大型語言模型，比方說 200 億或 300 億個參數；或者個人計算機，可能有 800 億到 1000 億個參數。但是當你將其提升到更高水平時，顯然你可能需要一個邊緣伺服器。而 8——我稱之為 8 路 GPU——才是正確的選擇。

So I argue there will be a mix in the transition here for a long period of time. Not everything will go to a GPU. It also depends how these large language models and all the AI applications get constructed.

因此我認為，在很長一段時間內，這裡的過渡將會是混合的。並非所有東西都會流向 GPU。這也取決於這些大型語言模型和所有人工智慧應用程式如何建構。

Now you asked us -- you made another interesting point, which I want to make sure the -- all of you remember. We, as a company, have 2 now public instances of AI, power with renewal energies where we are supporting some of these customers, including a hyperscaler and going forward, enterprise customers because they don't have the space and the cooling and they're understanding how to run the system of scale. That's a unique differentiation Hewlett Packard Enterprise have in addition to build systems and ship them. And I think that's an opportunity for us because that will drive stickiness to our HPE GreenLake platform, which obviously will drive recurring revenues but better attach of software and services down the road. And Juniper will play a huge role in that environment.

現在您問我們——您提出了另一個有趣的觀點，我想確保——大家都記住這一點。作為一家公司，我們現在有 2 個公共的 AI 實例，它們利用可再生能源提供動力，我們正在為其中一些客戶提供支持，包括超大規模客戶以及未來的企業客戶，因為他們沒有空間和冷卻系統，並且他們正在了解如何運行規模系統。這是惠普企業除了建構系統並運送系統之外的獨特差異化。我認為這對我們來說是一個機會，因為這將提高我們 HPE GreenLake 平台的黏性，這顯然將帶來經常性收入，但將來還能更好地整合軟體和服務。瞻博網路將在該環境中發揮巨大作用。

Thank you, Lou. Let me now turn it back to Antonio for concluding remarks.

謝謝你，盧。現在我請安東尼奧作最後發言。

Well, thank you, Shannon. And thank you, everyone. I know you have been covering multiple calls today. I know it's late on the East Coast, but I will leave you with few comments. Number one, we have the right strategy and the right team at the right time. This quarter, obviously, it was a little bit mixed because of the revenue. But remember, lower revenue also went through the ARR, so we need to understand that implication going forward.

好吧，謝謝你，香農。謝謝大家。我知道您今天已經接聽了多通電話。我知道東海岸已經很晚了，但我還是想給你們幾個評論。首先，我們在正確的時間制定了正確的策略並組建了正確的團隊。顯然，本季的營收情況有些好壞參半。但請記住，較低的收入也經過了 ARR 的處理，因此我們需要了解這一影響。

I'm very confident about the future. And the moves we have made and continue to make, including the Juniper acquisition, will allow us to participate in this inflection point with a unique IP. Everybody, obviously, is focused about this AI momentum and the server side, but you need more than servers. AI will drive the need for more ports. That means you need more networking bandwidth. That's for sure.

我對未來充滿信心。我們已經採取和繼續採取的舉措，包括收購瞻博網絡，將使我們能夠以獨特的 IP 參與這一轉折點。顯然，每個人都專注於人工智慧的發展勢頭和伺服器端，但你需要的不僅僅是伺服器。人工智慧將推動更多港口的需求。這意味著您需要更多的網路頻寬。這是肯定的。

Also, let's not forget, we need to do this responsibly. One of the things I'm really proud about our company is the commitment to social responsibility. Doing all of this, addressing the sustainability and the ethical challenges and the responsibility around AI. But listen, just we came out 2 weeks ago where HP was ranked #1 in just capital, ranking something that when you are proud of it and I know shareholder value, all of that. We have to take some actions here.

此外，請不要忘記，我們需要負責任地去做這件事。我對我們公司真正感到自豪的事情之一就是我們對社會責任的承諾。做到這一切，解決人工智慧的可持續性、道德挑戰和責任問題。但聽著，就在兩週前，我們公佈了惠普在資本方面的排名第一的消息，這個排名讓你感到自豪，我知道股東價值，所有這些。我們必須在此採取一些行動。

We are really focused on the strong execution and discipline, something we have shown now for 6 years-plus. And that's why I'm confident in the adjusted guidance we provided with Marie. And as we get into '25, obviously with the pending acquisition, I feel HPE will be even in a stronger position as we get through 2024.

我們真正注重的是強大的執行力和紀律性，這是我們六年來已經展現出來的。這就是我對我們為瑪麗提供的調整指導充滿信心的原因。隨著我們進入2025年，顯然隨著收購的臨近，我認為到2024年，HPE的地位將會更加穩固。

So thank you for your time and hope to connect with you soon.

非常感謝您抽出時間，希望能夠盡快與您聯繫。

Ladies and gentlemen, this concludes our call for today. Thank you. You may now disconnect.

女士們、先生們，今天的通話到此結束。謝謝。您現在可以斷開連線。