Meta Platforms Inc (META) 2023 Q2 法說會逐字稿

Good afternoon. My name is Dave, and I will be your conference operator today. At this time, I would like to welcome everyone to the Meta Second Quarter Earnings Conference Call. (Operator Instructions) This call will be recorded. Thank you very much.

午安.我叫戴夫，今天我將擔任您的會議主持人。現在，我歡迎大家參加 Meta 第二季財報電話會議。 （接線員指示）此通話將會被錄音。非常感謝。

Ken Dorell, Meta's Director of Investor Relations, you may begin.

Meta 投資者關係總監 Ken Dorell，您可以開始。

Thank you. Good afternoon, and welcome to Meta Platforms Second Quarter 2023 Earnings Conference Call. Joining me today to discuss our results are Mark Zuckerberg, CEO; and Susan Li, CFO.

謝謝。下午好，歡迎參加 Meta Platforms 2023 年第二季財報電話會議。今天與我一起討論我們業績的有執行長馬克·祖克柏；以及財務長 Susan Li。

Before we get started, I would like to take this opportunity to remind you that our remarks today will include forward-looking statements. Actual results may differ materially from those contemplated by these forward-looking statements.

在我們開始之前，我想藉此機會提醒大家，我們今天的發言將包括前瞻性陳述。實際結果可能與這些前瞻性陳述所預期的結果有重大差異。

Factors that could cause these results to differ materially are set forth in today's press release and in our quarterly report on Form 10-Q filed with the SEC.

可能導致這些結果出現重大差異的因素已在今天的新聞稿和我們向美國證券交易委員會提交的 10-Q 表季度報告中列出。

Any forward-looking statements that we make on this call are based on assumptions as of today, and we undertake no obligation to update these statements as a result of new information or future events.

我們在本次電話會議上所做的任何前瞻性陳述均基於今天的假設，我們不承擔因新資訊或未來事件而更新這些陳述的義務。

During this call, we will present both GAAP and certain non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures is included in today's earnings press release. The press release and an accompanying investor presentation are available on our website at investor.fb.com.

在本次電話會議中，我們將介紹 GAAP 和某些非 GAAP 財務指標。今天的收益新聞稿中包含了 GAAP 與非 GAAP 指標的對帳。新聞稿和隨附的投資者介紹可在我們的網站 investor.fb.com 上查閱。

And now I'd like to turn the call over to Mark.

現在我想把電話轉給馬克。

Thanks, Ken. Thanks, everyone, for joining.

謝謝，肯。謝謝大家的參與。

This was a good quarter for our business. We're seeing strong engagement trends across our apps. There are now more than 3.8 billion people who use at least one of our apps every month. Facebook now has more than 3 billion monthly actives, with daily actives continuing to grow around the world, including in the U.S. and Canada.

對於我們的業務來說，這是一個好的季度。我們發現我們的應用程式的參與趨勢非常強勁。目前，每月有超過 38 億人至少使用我們的一款應用程式。 Facebook 目前每月活躍用戶超過 30 億，全球每日活躍用戶持續成長，包括美國和加拿大。

In addition to our core products performing well, I think we have the most exciting road map ahead that I've seen in a while. We've got continued progress on Threads, Reels, Llama 2 and some groundbreaking AI products in the pipeline, as well as the Quest 3 launch coming up this fall. We're heads down executing on all of this right now, and it's really good to see the decisions and investments we've made start to play out.

除了我們的核心產品表現良好之外，我認為我們擁有我近期見過的最令人興奮的發展路線圖。我們在 Threads、Reels、Llama 2 和一些突破性的 AI 產品研發方面不斷取得進展，並且將於今年秋季推出 Quest 3。我們現在全力以赴地執行所有這些計劃，很高興看到我們所做的決定和投資開始發揮作用。

On Threads, briefly, I'm quite optimistic about our trajectory here. We saw unprecedented growth out of the gate. And more importantly, we're seeing more people coming back daily than I'd expected. And now we're focused on retention and improving the basics. And then after that, we'll focus on growing the community to the scale that we think is going to be possible. Only after that are we going to focus on monetization. We've run this playbook many times before with Facebook, Instagram, WhatsApp, Stories, Reels and more, and this is as good of a start as we could have hoped for. So I'm really happy with the path that we're on here.

關於 Threads，簡而言之，我對我們的發展軌跡非常樂觀。我們看到了前所未有的成長。更重要的是，我們看到每天回來的人數比我預期的要多。現在我們專注於保留和改善基礎知識。此後，我們將致力於將社區發展到我們認為可能的規模。只有在那之後我們才會專注於貨幣化。我們之前已經在 Facebook、Instagram、WhatsApp、Stories、Reels 等平台上多次運行過這個劇本，這是一個正如我們所希望的那樣好的開始。所以我對我們現在所走的路感到非常高興。

One note I want to mention about the Threads launch related to our Year of Efficiency is that the product was built by a relatively small team on a tight time line. We've already seen a number of examples of how our leaner organization and some of the cultural changes that we've made can build higher-quality products faster. And this is probably the biggest example so far.

關於我們效率年期間推出的 Threads 產品，我想指出的是，該產品是由一個相對較小的團隊在緊迫的時間內打造的。我們已經看到了許多例子，表明我們精簡的組織和所做的一些文化變革能夠更快地生產出更高品質的產品。這可能是迄今為止最大的例子。

The Year of Efficiency was always about 2 different goals: becoming an even stronger technology company; and improving our financial results so we can invest aggressively in our ambitious long-term road map. Now that we've gotten through the major layoffs, the rest of 2023 will be about creating stability for employees, removing barriers that slow us down, introducing new AI-powered tools to speed us up and so on.

效率年始終圍繞著兩個不同的目標：成為一家更強大的科技公司；並改善我們的財務業績，以便我們能夠積極投資於我們雄心勃勃的長期路線圖。現在我們已經度過了大規模裁員，2023 年剩餘的時間將致力於為員工創造穩定性，消除阻礙我們前進的障礙，引入新的人工智慧工具來加快我們的速度等等。

Over the next few months, we're going to start planning for 2024, and I'm going to be focused on continuing to run the company as lean as possible for these cultural reasons, even though our financial results have improved. I expect that we're still going to hire in key areas, but newly budgeted head count growth is going to be relatively low.

在接下來的幾個月裡，我們將開始規劃 2024 年，儘管我們的財務表現有所改善，但基於這些文化原因，我將專注於繼續盡可能精簡地經營公司。我預計我們仍將在關鍵領域招募員工，但新預算的員工人數成長將相對較低。

Now that said, as part of this year's layoffs, many teams chose to let people go in order to hire different people with different skills that they need, so much of that hiring is going to spill into 2024.

話雖如此，作為今年裁員的一部分，許多球隊選擇解僱一些員工，以便聘請他們需要的具有不同技能的不同人才，因此大部分招募將延續到 2024 年。

The other major budget point that we're working through is what the right level of AI CapEx is to support our road map. Since we don't know how quickly our new AI products will grow, we may not have a clear handle on this until later in the year.

我們正在研究的另一個主要預算點是，什麼樣的 AI 資本支出水準才適合支持我們的路線圖。由於我們不知道我們的新人工智慧產品將成長多快，我們可能要到今年晚些時候才能清楚掌握這一點。

Now moving on to our product road map, I've said on a number of these calls that the 2 technological waves that we're riding are AI in the near term and the metaverse over the longer term. Investments that we've made over the years in AI, including the billions of dollars we've spent on AI infrastructure, are clearly paying off across our ranking and recommendation systems and improving engagement and monetization.

現在談談我們的產品路線圖，我在許多電話會議上都說過，我們正在乘上的兩波科技浪潮是近期的人工智慧和長期的元宇宙。我們多年來在人工智慧領域的投資，包括在人工智慧基礎設施上投入的數十億美元，顯然在我們的排名和推薦系統中獲得了回報，並提高了參與度和貨幣化程度。

AI-recommended content from accounts you don't follow is now the fastest-growing category of content on Facebook's feed. Now since introducing these recommendations, they've driven a 7% increase in overall time spent on the platform. This improves the experience because you can now discover things that you might not have otherwise followed or come across.

來自您未追蹤的帳號的 AI 推薦內容現已成為 Facebook 動態中成長最快的內容類別。自從推出這些建議以來，它們已經推動該平台的整體使用時間增加了 7%。這改善了體驗，因為您現在可以發現一些您可能不會關注或遇到的事情。

Reels is a key part of this discovery engine. And Reels plays exceed 200 billion per day across Facebook and Instagram. We're seeing good progress on Reels monetization as well, with the annual revenue run rate across our apps now exceeding $10 billion, up from $3 billion last fall.

Reels 是這個發現引擎的關鍵部分。 Reels 在 Facebook 和 Instagram 上的每日播放量超過 2000 億次。我們也看到 Reels 貨幣化取得了良好進展，我們應用程式的年收入運行率現已超過 100 億美元，高於去年秋季的 30 億美元。

Beyond Reels, AI is driving results across our monetization tools through our automated ads products, which we call Meta Advantage. Almost all our advertisers are using at least one of our AI-driven products.

除了 Reels 之外，人工智慧還透過我們的自動廣告產品（我們稱之為 Meta Advantage）推動我們的貨幣化工具取得成果。幾乎我們所有的廣告商都在使用我們的至少一款人工智慧產品。

We've also deployed Meta Lattice, a new model architecture that learns to predict an ad's performance across a variety of data sets and optimization goals. And we introduced AI Sandbox, a testing playground for generative AI-powered tools like automatic text variation, background generation and image outcropping.

我們還部署了 Meta Lattice，這是一種新的模型架構，可以學習預測廣告在各種資料集和最佳化目標中的效果。我們推出了 AI Sandbox，這是一個用於測試生成 AI 工具（如自動文字變更、背景生成和圖像露出）的測試場地。

Business messaging is another key piece of our monetization strategy. We recently announced that the 200 million users of our WhatsApp Business app will now be able to create Click-to-WhatsApp ads for Facebook and Instagram without needing a Facebook account. This is a pretty big unlock, particularly in countries where WhatsApp is often the first step to bringing a business online.

商業訊息傳遞是我們貨幣化策略的另一個關鍵部分。我們最近宣布，WhatsApp Business 應用程式的 2 億用戶現在無需 Facebook 帳號即可為 Facebook 和 Instagram 創建 Click-to-WhatsApp 廣告。這是一個相當大的解鎖，特別是在那些通常使用 WhatsApp 作為企業上線第一步的國家。

Paid messaging is a bit earlier, but it's also showing good adoption. The number of businesses using our paid messaging products has doubled year-over-year.

付費訊息傳遞出現得稍早一些，但也顯示出良好的普及度。使用我們的付費訊息產品的企業數量比去年同期增加了一倍。

While we're on messaging, I'll mention that we started rolling out Channels on WhatsApp last month. It's a simple, reliable and private way to receive important updates from people and 1organizations. And I'm quite excited for more people to try it as we bring the product to more countries through the rest of this year.

當我們討論訊息傳遞時，我要提一下，我們上個月開始在 WhatsApp 上推出頻道功能。這是一種簡單、可靠且私密的方式，可以接收來自個人和組織的重要更新。隨著我們在今年剩餘時間將該產品推廣到更多國家，我非常高興有更多的人嘗試它。

Beyond the recommendations and ranking systems across our products, we're also building leading foundation models to support a new generation of AI products.

除了我們產品的推薦和排名系統之外，我們還在建立領先的基礎模型來支援新一代人工智慧產品。

We've partnered with Microsoft to open source Llama 2, the latest version of our large language model, and to make it available for both research and commercial use. We have a long history of open sourcing our infrastructure and AI work from PyTorch, which is the leading machine learning framework, to models like Segment Anything, ImageBind and DINO to basic infrastructure as part of the Open Compute Project.

我們與微軟合作，開源 Llama 2（我們的大型語言模型的最新版本），並將其用於研究和商業用途。我們長期致力於開源我們的基礎設施和人工智慧工作，從領先的機器學習框架 PyTorch，到 Segment Anything、ImageBind 和 DINO 等模型，再到作為開放運算專案一部分的基本基礎設施。

And we found that open-sourcing our work allows the industry, including us, to benefit from innovations that come from everywhere. And these are often improvements in safety and security, since open-source software is more scrutinized and more people can find and identify fixes for issues.

我們發現，開源我們的工作可以讓包括我們在內的整個產業受益於來自世界各地的創新。這些通常是安全性方面的改進，因為開源軟體受到更嚴格的審查，並且更多的人可以找到並確定問題的解決方案。

The improvements also often come in the form of efficiency gains, which should hopefully allow us and others to run these models with less infrastructure investment going forward. So I'm really looking forward to seeing the improvements that the community makes to Llama 2.

這些改進通常也以效率提升的形式出現，希望這能讓我們和其他人在未來以更少的基礎設施投資來運作這些模型。所以我真的很期待看到社區對 Llama 2 所做的改進。

We are also building a number of new products ourselves using Llama that will work across our services. I'm going to share more details on that later this year. But you can imagine lots of ways that AI can help people connect and express themselves in our apps: creative tools that make it easier and more fun to share content; agents that act as assistants, coaches, or that can help you interact with businesses and creators; and more. And these new products will improve everything that we do across both mobile apps and the metaverse, helping people create worlds and the avatars and objects that inhabit them as well.

我們也利用 Llama 自行建置了許多可在我們的服務中運作的新產品。我將在今年稍後分享更多相關細節。但你可以想像人工智慧可以透過多種方式幫助人們在我們的應用程式中建立聯繫並表達自己：讓內容共享變得更容易、更有趣的創意工具；充當助手、教練或可以幫助您與企業和創作者互動的經紀人；等等。這些新產品將改善我們在行動應用程式和元宇宙中所做的一切，幫助人們創造世界以及居住在其中的化身和物體。

Our investments in AI continue. We remain fully committed to the metaverse vision as well. We've been working on both of these 2 major priorities for many years in parallel now. And in many ways, the 2 areas are overlapping and complementary.

我們對人工智慧的投資仍在繼續。我們依然全力致力於元宇宙願景。多年來，我們一直在同時致力於這兩個主要優先事項。在許多方面，這兩個領域是重疊和互補的。

The next big thing on the Reality Labs side is the launch of our Quest 3 mixed-reality headset at Connect. It's our most powerful headset yet, with better displays and resolution and next-gen Qualcomm chipset with twice 0the graphics performance. It will also have the best immersive content library out there, and it's 40% thinner than Quest 2. Its mixed reality seamlessly blends your physical world and the virtual one by intelligently understanding the physical space around you. We pioneered mixed reality with our Quest Pro headset, and Quest 3 takes that to the next level.

Reality Labs 方面的下一個重大事件是在 Connect 上推出我們的 Quest 3 混合實境耳機。這是我們迄今為止功能最強大的耳機，具有更好的顯示效果和分辨率，以及圖形性能兩倍的下一代高通晶片組。它還將擁有最好的沉浸式內容庫，並且比 Quest 2 薄 40%。它的混合實境透過智慧地理解您周圍的實體空間，將您的物理世界和虛擬世界無縫融合。我們利用 Quest Pro 耳機率先實現了混合現實，而 Quest 3 將其提升到了一個新的水平。

Now others in the industry are, of course, working on bringing mixed reality to the market, too. But Quest 3 is going to be the first mainstream accessible device that we expect many millions of people will get to experience this technology with.

現在，業內其他公司當然也致力於將混合實境推向市場。但我們預計 Quest 3 將成為第一款主流無障礙設備，數百萬人將能夠體驗這項技術。

The metaverse content and software vision continues coming together as well. We recently announced that Roblox is coming to Quest with an open beta on App Lab. For Horizon, the team is focused on retention right now, and we're making good progress on that.

元宇宙內容和軟體願景也持續融合。我們最近宣布 Roblox 將登陸 Quest，並在 App Lab 上進行公開測試。對於 Horizo​​n，團隊目前專注於保留，並且我們在這方面取得了良好的進展。

We made big improvements on Avatars as well, and that's going to be a bridge between our mobile apps and our VR and mixed-reality experiences. We'll have a lot more to share on both our metaverse and AI work coming up at our Connect Conference, which we're going to be hosting in Hacker Square at our headquarters on September 27. It's going to be a good one, so I hope you tune in.

我們對 Avatars 也做出了很大的改進，它將成為我們的行動應用程式和 VR 及混合實境體驗之間的橋樑。我們將在 9 月 27 日於總部駭客廣場舉辦的 Connect 大會上分享更多關於元宇宙和人工智慧工作的內容。這將是一場精彩的會議，希望您能關注。

All right. To wrap up, I just want to say that I'm really proud of our teams for everything we've accomplished so far this year. It's been a tough year in a lot of ways, but it's also been an impactful one. I'm quite optimistic about the road ahead, and I'm grateful to all of you for being on this journey with us.

好的。總而言之，我只想說，我為我們團隊今年迄今為止所取得的一切成就感到非常自豪。從很多方面來說，這是艱難的一年，但也是影響深遠的一年。我對未來的道路充滿樂觀，並感謝大家與我們一起踏上這段旅程。

And now I'm going to hand it over to Susan.

現在我要把它交給蘇珊。

Thanks, Mark, and good afternoon, everyone.

謝謝，馬克，大家下午好。

Let's begin with our consolidated results. All comparisons are on a year-over-year basis, unless otherwise noted.

讓我們從綜合結果開始。除非另有說明，所有比較均以同比為基礎。

Q2 total revenue was $32 billion, up 11% or 12% on a constant currency basis. Q2 total expenses were $22.6 billion, up 10% compared to last year.

第二季總營收為 320 億美元，以固定匯率計算成長 11% 或 12%。第二季總支出為 226 億美元，比去年同期成長 10%。

In terms of the specific line items, cost of revenue increased 15%, driven primarily by infrastructure-related costs. R&D increased 8%, driven mainly by head count-related costs from our Reality Labs and Family of Apps segments as well as restructuring costs. Marketing and sales decreased 12%, due mostly to lower marketing spend and payroll-related costs. And G&A increased 39%, due primarily to an increase in legal accruals, which was partially offset by lower payroll-related costs.

從具體項目來看，收入成本增加了 15%，主要原因是基礎設施相關成本。研發費用增加了 8%，主要原因是現實實驗室和應用程式系列部門的員工人數相關成本以及重組成本。行銷和銷售下降了 12%，主要原因是行銷支出和工資相關成本降低。一般及行政費用增加了 39%，主要原因是法律應計費用增加，但工資相關成本的降低部分抵消了這一增長。

We ended the second quarter with over 71,400 employees, down 7% from the first quarter. Our second quarter head count still included roughly 1/2 of the approximately 10,000 employees impacted by the 2023 layoffs. We expect that our third quarter head count will no longer include the vast majority of impacted employees.

截至第二季度，我們的員工總數超過 71,400 人，較第一季下降 7%。我們第二季的員工總數仍然包括受 2023 年裁員影響的約 10,000 名員工中的約一半。我們預計第三季的員工總數將不再包括絕大多數受影響的員工。

Second quarter operating income was $9.4 billion, representing a 29% operating margin.

第二季營業收入為 94 億美元，營業利益率為 29%。

Our tax rate for the quarter was 16%. This is lower than our previous full year outlook as our higher share price provided a higher tax deduction and lowered our taxes.

我們本季的稅率為 16%。這低於我們先前的全年預期，因為我們更高的股價提供了更高的稅收減免並降低了我們的稅收。

Net income was $7.8 billion or $2.98 per share.

淨收入為 78 億美元，即每股 2.98 美元。

Capital expenditures, including principal payments on finance leases, were $6.4 billion, driven by investments in data centers, servers and network infrastructure.

資本支出（包括融資租賃的本金支付）為 64 億美元，主要由資料中心、伺服器和網路基礎設施的投資推動。

Free cash flow was $11 billion, significantly benefiting from a deferral of income taxes that we expect will be paid in the fourth quarter.

自由現金流為 110 億美元，這主要得益於我們預計在第四季度支付的所得稅延期。

We repurchased $793 million of our Class A common stock in the second quarter and ended the quarter with $53.4 billion in cash and marketable securities.

我們在第二季回購了價值 7.93 億美元的 A 類普通股，本季末我們持有價值 534 億美元的現金和有價證券。

Moving now to our segment results.

現在轉到我們的分部業績。

I'll begin with our Family of Apps segment. Our community across the Family of Apps continues to grow. We estimate that approximately 3.07 billion people used at least 1 of our Family of Apps on a daily basis in June and that approximately 3.88 billion people used at least 1 on a monthly basis.

我將從我們的應用程式系列部分開始。我們的應用程式家族社群正在不斷壯大。我們估計，6 月大約有 30.7 億人每天至少使用 1 款我們的應用程式系列，大約有 38.8 億人每月至少使用 1 款我們的應用程式系列。

Facebook continues to grow globally, and engagement remains strong. For the first time, we crossed 3 billion monthly active users, with Facebook MAU ending at 3.03 billion in June, up 3% or 96 million compared to last year. Facebook daily active users were 2.06 billion, up 5% or 96 million. DAUs represented approximately 68% of MAUs.

Facebook 在全球持續成長，參與度依然強勁。截至 6 月份，Facebook 的每月活躍用戶首次突破 30 億，達到 30.3 億，較去年同期成長 3%（增加 9,600 萬）。 Facebook每日活躍用戶數為20.6億，成長5%或9,600萬。 DAU 約佔 MAU 的 68%。

Q2 total Family of Apps revenue was $31.7 billion, up 12% year-over-year.

第二季應用程式系列總營收為 317 億美元，年增 12%。

Q2 Family of Apps ad revenue was $31.5 billion, up 12% or 13% on a constant currency basis.

第二季應用程式系列廣告收入為 315 億美元，以固定匯率計算成長 12% 或 13%。

Within ad revenue, the online commerce vertical was the largest contributor to year-over-year growth, followed by entertainment and media and CPG. Online commerce benefited from strong spend among advertisers in China, reaching customers in other markets.

在廣告收入中，線上商務垂直領域是年成長的最大貢獻者，其次是娛樂和媒體以及快速消費品。網路商務受益於中國廣告商的強勁支出，從而觸及了其他市場的客戶。

On a user geography basis, ad revenue growth was strongest in Rest of World at 16%, followed by Europe, North America and Asia Pacific at 14%, 11% and 10%, respectively. Foreign currency was a headwind to advertising revenue growth in all international regions.

從用戶地理來看，全球其他地區的廣告收入成長最為強勁，為 16%，其次是歐洲、北美和亞太地區，分別為 14%、11% 和 10%。外幣是所有國際地區廣告收入成長的阻力。

In Q2, the total number of ad impressions served across our services increased 34%, and the average price per ad decreased 16%. Impression growth was primarily driven by Asia Pacific and Rest of World. The year-over-year decline in pricing was driven by strong impression growth, especially from lower monetizing surfaces and regions. While overall pricing remains under pressure from these factors, we believe our ongoing improvements to ad targeting and measurement are continuing to drive improved results for advertisers.

第二季度，我們各項服務的廣告展示總數增加了 34%，而每個廣告的平均價格下降了 16%。印象增長主要受到亞太地區和世界其他地區的推動。價格同比下降的原因是印象數量的強勁增長，尤其是來自貨幣化程度較低的表面和地區的印像數增長。儘管整體定價仍然受到這些因素的壓力，但我們相信，我們對廣告定位和衡量標準的持續改進將繼續為廣告商帶來更好的業績。

Family of Apps other revenue was $225 million in Q2, up 3%, as strong business messaging revenue growth from our WhatsApp Business platform was partially offset by a decline in other line items.

應用程式系列其他收入在第二季為 2.25 億美元，成長 3%，因為我們的 WhatsApp Business 平台的強勁商業訊息收入成長被其他項目的下降部分抵消。

We continue to direct the majority of our investments toward the development and operation of our Family of Apps. In Q2, Family of Apps expenses were $18.6 billion, representing approximately 82% of our overall expenses. FoA expenses were up 8% due primarily to legal-related expenses and restructuring charges, partially offset by a decrease in non-headcount-related operating expenses, including marketing.

我們繼續將大部分投資用於應用程式系列的開發和營運。第二季度，應用程式系列的支出為 186 億美元，約占我們整體支出的 82%。 FoA 費用上漲 8%，主要原因是法律相關費用和重組費用，但部分被包括行銷在內的非員工相關營運費用的減少所抵銷。

Family of Apps operating income was $13.1 billion, representing a 41% operating margin.

應用程式系列營運收入為 131 億美元，營運利潤率為 41%。

Within our Reality Labs segment, Q2 revenue was $276 million, down 39% due to lower Quest 2 sales. Reality Labs' expenses were $4 billion, up 23% due to lapping a reduction in Reality Labs' loss reserves in Q2 of last year as well as growth in employee-related costs.

在我們的現實實驗室部門，第二季度營收為 2.76 億美元，由於 Quest 2 銷量下降而下降 39%。 Reality Labs 的支出為 40 億美元，成長 23%，原因是去年第二季 Reality Labs 的虧損準備金減少以及員工相關成本的成長。

Reality Labs operating loss was $3.7 billion.

Reality Labs 的營運虧損為 37 億美元。

Turning now to the business outlook. There are 2 primary factors that drive our revenue performance: our ability to deliver engaging experiences for our community; and our effectiveness at monetizing that engagement over time.

現在談談商業前景。推動我們的收入表現有兩個主要因素：我們為社區提供引人入勝的體驗的能力；以及我們隨著時間的推移將這種參與貨幣化的有效性。

On the first, overall engagement within Facebook and Instagram remains strong. Reels continues to grow and drive incremental engagement. On Facebook Feed, in particular, recommended content from accounts you don't follow has increased significantly over the past year while also becoming more incremental to engagement, demonstrating that people are getting added value from discovering content from unconnected accounts. Looking forward, we're optimistic about our ability to increase that value even further by leveraging advanced AI techniques to improve recommendations.

首先，Facebook 和 Instagram 的整體參與度依然強勁。 Reels 持續成長並推動參與度的提升。特別是在 Facebook Feed 上，來自您未關注的帳戶的推薦內容在過去一年中顯著增加，同時參與度也越來越高，這表明人們從發現未關聯帳戶的內容中獲得了附加價值。展望未來，我們堅信我們能夠利用先進的人工智慧技術來改進推薦，從而進一步提高這一價值。

In addition to improving the value people get within our Family of Apps today, we're also investing in entirely new experiences for the future. We're standing up infrastructure to support new AI-powered products across our services, which will give people more tools to express themselves and connect. And we've been pleased with the initial reception of our new stand-alone app, Threads, since its launch earlier this month. Our focus now is on further developing this into a product that will be valuable for a large set of people over time.

除了提高人們目前在我們的應用程式系列中獲得的價值之外，我們還在投資未來的全新體驗。我們正在建立基礎設施來支援我們服務中新的人工智慧產品，這將為人們提供更多表達自我和建立聯繫的工具。自本月初推出以來，我們的新獨立應用程式 Threads 的初步反應令我們非常滿意。我們現在的重點是進一步將其開發成一種能夠隨著時間的推移為廣大民眾帶來價值的產品。

Moving to the other driver of revenue, improving monetization. Here, we're focused on improving monetization efficiency of products that monetize at lower rates today, like Reels and our messaging services, and more broadly, driving measurable performance and returns for our advertisers.

轉向另一個收入驅動力，提高貨幣化。在這裡，我們專注於提高目前以較低利率盈利的產品（如 Reels 和我們的消息服務）的盈利效率，更廣泛地說，為我們的廣告商帶來可衡量的業績和回報。

On Reels, we are making good progress on monetization with more than 3/4 of our advertisers now using Reels ads. We remain focused on further reducing the Reels revenue headwind and narrowing the monetization efficiency gap with our more mature surfaces. However, we continue to expect time on Reels will monetize at a lower rate than Stories and Feed for the foreseeable future since people scroll more slowly through video content.

在 Reels 上，我們在貨幣化方面取得了良好的進展，目前超過 3/4 的廣告商都在使用 Reels 廣告。我們仍然專注於進一步減少 Reels 的收入阻力，並縮小與我們更成熟的平台之間的貨幣化效率差距。然而，我們仍然預計，在可預見的未來，Reels 上的時間貨幣化率將低於 Stories 和 Feed，因為人們滾動瀏覽影片內容的速度較慢。

Within messaging, billions of people and millions of businesses use our messaging services every day to connect. We see a significant opportunity to build tools and functionality for businesses to help facilitate those interactions and are seeing early but promising progress with WhatsApp's paid messaging solution today.

在訊息傳遞領域，數十億人和數百萬企業每天都使用我們的訊息傳遞服務進行聯繫。我們看到了為企業建立工具和功能以幫助促進這些互動的重要機遇，並且今天我們看到了 WhatsApp 付費訊息解決方案的早期但有希望的進展。

In terms of our work to drive measurable performance for advertisers, it's concentrated in 2 primary areas: AI; and on-site conversions. We're leveraging AI to move our systems towards using fewer, larger models that enable us to leverage learnings across product surfaces and deploy improvements more quickly, broadly and efficiently.

就我們為廣告主推動可衡量績效的工作而言，主要集中在兩個領域：人工智慧；以及現場轉換。我們正在利用人工智慧使我們的系統使用更少、更大的模型，這使我們能夠利用跨產品領域的學習成果並更快、更廣泛、更有效地部署改進。

We're also leveraging AI to power advanced ads products, like Advantage+ shopping, which continues to gain adoption. We're seeing this work translate into results for advertisers as conversion growth remained strong in Q2.

我們也利用人工智慧來支援先進的廣告產品，例如 Advantage+ 購物，該產品的採用率正在不斷提高。我們看到這項工作為廣告商帶來了成果，因為第二季的轉換率成長依然強勁。

In terms of driving on-site conversions, we continue to see strong results with click-to-messaging ads and are well positioned, given our suite of messaging applications.

在推動現場轉換方面，我們繼續看到點擊訊息廣告的強勁效果，並且憑藉我們的訊息應用程式套件，我們處於有利地位。

Daily Click-to-WhatsApp ads revenue continues to grow very quickly at over 80% year-over-year. We also recently started testing the ability to buy Click-to-WhatsApp ads directly from the WhatsApp Business app, which now has more than 200 million monthly users. Looking ahead, we're focused on enabling businesses to optimize for conversions further down the funnel in our messaging applications. We're also investing in scaling other on-site objectives like lead generation and shops ads.

每日點擊 WhatsApp 廣告收入持續快速成長，較去年同期成長超過 80%。我們最近也開始測試直接從 WhatsApp Business 應用程式購買 Click-to-WhatsApp 廣告的功能，目前每月擁有超過 2 億用戶。展望未來，我們致力於幫助企業在我們的訊息傳遞應用程式中進一步優化轉換。我們也投資擴大其他現場目標，如潛在客戶開發和商店廣告。

Before turning to our revenue outlook, I'd also like to talk about our investment philosophy. We expect to bring the discipline and habits that we built during this Year of Efficiency with us as we plan for the future. At the same time, we remain focused on investing in the significant opportunities ahead. Part of supporting these initiatives will come from prioritizing them against other areas of work and shifting resources. However, in some cases, they will require incremental investment. This is particularly true in the areas we see the most significant opportunity, which include AI and the metaverse.

在談到我們的收入展望之前，我還想談談我們的投資理念。我們希望在規劃未來時，能夠運用我們在這效率年中所建立的紀律和習慣。同時，我們仍將重點投資未來的重大機會。支持這些措施的部分措施將來自於優先考慮這些措施而不是其他工作領域並轉移資源。然而，在某些情況下，它們將需要增量投資。在我們認為最具機會的領域尤其如此，其中包括人工智慧和元宇宙。

As I mentioned last quarter, we also remain focused on modestly evolving our capital structure over time. We were pleased to execute our second bond offering in May and expect a measured pace of future debt raises as we work toward improving our overall cost of capital while maintaining a positive or neutral net cash balance.

正如我上個季度提到的，我們也將繼續致力於隨著時間的推移適度改善我們的資本結構。我們很高興在五月進行了第二次債券發行，並預計未來債務成長將保持平穩步伐，因為我們致力於改善整體資本成本，同時保持正或中性的淨現金餘額。

In addition, we continue to monitor the active regulatory landscape. With respect to EU-U.S. data transfers, we saw a positive development with the European Commission's adoption of a final adequacy decision, which allows us to continue to provide our services in Europe. This is good news, though broadly speaking, we continue to see increasing legal and regulatory headwinds in the EU and the U.S. that could significantly impact our business and our financial results.

此外，我們將繼續監測活躍的監管環境。關於歐盟-美國數據傳輸方面，隨著歐盟委員會通過最終充分性決定，我們看到了積極的進展，這使我們能夠繼續在歐洲提供服務。這是個好消息，儘管從廣義上講，我們繼續看到歐盟和美國的法律和監管阻力不斷增加，這可能會對我們的業務和財務表現產生重大影響。

Turning now to the revenue outlook.

現在來談談收入前景。

We expect third quarter 2023 total revenue to be in the range of $32 billion to $34.5 billion. Our guidance assumes a foreign currency tailwind of approximately 3% to year-over-year total revenue growth in the third quarter based on current exchange rates.

我們預計 2023 年第三季總營收將在 320 億美元至 345 億美元之間。根據當前匯率，我們的預測是，外幣將推動第三季總收入年增約 3%。

Turning now to the expense outlook.

現在來談談費用前景。

We anticipate that our full year 2023 total expenses will be in the range of $88 billion to $91 billion, increased from our prior range of $86 billion to $90 billion due to legal-related expenses recorded in Q2. This outlook includes approximately $4 billion of restructuring costs related to facilities consolidation charges and severance and other personnel costs. We expect Reality Labs; operating losses to increase year-over-year in 2023.

我們預計，2023 年全年總支出將在 880 億美元至 910 億美元之間，高於先前的 860 億美元至 900 億美元，原因是第二季記錄了法律相關支出。這項預期包括與設施整合費用、遣散費和其他人員費用有關的約 40 億美元的重組成本。我們期待現實實驗室；2023年營業虧損將年增。

While we are not providing a quantitative outlook beyond 2023 at this point, we expect a few factors to be drivers of total expense growth in 2024 as we continue to invest in our most compelling opportunities, including AI and the metaverse.

雖然我們目前還沒有提供 2023 年以後的量化展望，但我們預計，隨著我們繼續投資於最具吸引力的機會，包括人工智慧和元宇宙，一些因素將成為 2024 年總費用成長的驅動因素。

First, we expect higher infrastructure-related costs next year. Given our increased capital investments in recent years, we expect depreciation expenses in 2024 to increase by a larger amount than in 2023. We also expect to incur higher operating costs from running a larger infrastructure footprint.

首先，我們預計明年基礎設施相關成本將會更高。鑑於我們近年來增加的資本投資，我們預計 2024 年的折舊費用將比 2023 年增加更多。我們也預計，營運更大的基礎設施足跡將產生更高的營運成本。

Second, we anticipate growth in payroll expenses as we evolve our workforce composition toward higher-cost technical roles.

其次，隨著我們的勞動力結構轉向成本更高的技術職位轉變，我們預期薪資支出將會成長。

Finally, for Reality Labs, we expect operating losses to increase meaningfully year-over-year due to our ongoing product development efforts in AR, VR and our investments to further scale our ecosystem.

最後，對於 Reality Labs，由於我們在 AR、VR 領域持續進行的產品開發工作以及進一步擴大生態系統的投資，我們預計營運虧損將比去年同期大幅增加。

Turning now to the CapEx outlook.

現在來談談資本支出前景。

We expect capital expenditures to be in the range of $27 billion to $30 billion, lowered from our prior estimate of $30 billion to $33 billion. The reduced forecast is due to both cost savings, particularly on non-AI servers, as well as shifts in CapEx into 2024 from delays in projects and equipment deliveries rather than a reduction in overall investment plans.

我們預計資本支出將在 270 億美元至 300 億美元之間，低於先前估計的 300 億美元至 330 億美元。預測下調的原因是成本節省，尤其是在非人工智慧伺服器上，以及由於專案和設備交付延遲導致資本支出轉移到 2024 年，而不是整體投資計劃的減少。

Looking ahead, while we continue to refine our plans as we progress throughout the year, we currently expect total capital expenditures to grow in 2024, driven by our investments across both data centers and servers, particularly in support of our AI work.

展望未來，雖然我們在全年不斷完善計劃，但我們目前預計 2024 年總資本支出將會成長，這得益於我們對資料中心和伺服器的投資，特別是對人工智慧工作的支援。

On to tax. Absent any changes to U.S. tax law, we expect the tax rate for the rest of the year to be similar to Q2 2023.

談稅。如果美國稅法沒有任何變化，我們預計今年剩餘時間的稅率將與 2023 年第二季相似。

In closing, Q2 was a good quarter for our business. We're executing well across our core priorities and are continuing to make progress on delivering exciting new experiences for our community.

總而言之，第二季度對於我們的業務來說是一個良好的季度。我們在核心優先事項上執行得非常好，並繼續在為社區提供令人興奮的新體驗方面取得進展。

With that, Dave, let's open up the call for questions.

戴夫，讓我們開始提問吧。

(Operator Instructions) Your first question comes from the line of Brian Nowak with Morgan Stanley.

（操作員指示）您的第一個問題來自摩根士丹利的 Brian Nowak。

Maybe a 2-parter for Mark. Mark, you've spoken a few times over the last few months on AI agents and sort of building different types of assistance through Llama. I guess the 2-part question is, can you sort of talk to us a little bit about some of the use cases or consumer advertiser offerings you're most excited about enabling for some of these AI agents?

對馬克來說，也許是兩個部分。馬克，過去幾個月你曾多次談到人工智慧代理以及如何透過 Llama 建構不同類型的援助。我想這個問題分為兩個部分，您能否向我們簡單介紹一下您最感興趣的為這些 AI 代理提供的一些用例或消費者廣告商產品？

And then the second one, just to sort of level set on timing a bit. Can you just help us understand some of the larger technological barriers in your mind the teams will need to overcome to really scale these types of agent products across the ecosystem?

然後是第二個，只是為了稍微調整一下時間水準。您能否幫助我們了解您認為團隊需要克服的一些較大的技術障礙，以便在整個生態系統中真正擴展這些類型的代理產品？

Sure. So I think the main thing that I'll say on this for now is tune into Connect coming up in September. We're going to have a lot more to talk about on the road map and products that we're launching. We wanted to get the Llama 2 model out now. That's going to be -- that's going to underpin a lot of the new things that we're building. And now we're nailing down a bunch of these additional products, and this is going to be stuff that we're working on for years. But I think a lot of the journey will kind of start later this year when we start rolling out some of these things.

當然。因此我認為現在我要說的主要內容是關注即將於 9 月舉行的 Connect。我們將就路線圖和即將推出的產品進行更多討論。我們現在想推出 Llama 2 型號。這將是——這將是我們正在建立的許多新事物的基礎。現在我們正在確定一系列附加產品，這些產品將是我們多年來致力於研發的。但我認為，當我們開始推出其中的一些產品時，許多旅程將在今年稍後開始。

But overall, I think that there are 3 basic categories of products or technologies that we're planning on building with generative AI. One are around different kinds of agents, which I'll talk about in a second.

但總的來說，我認為我們計劃利用生成式人工智慧來建構 3 個基本類別的產品或技術。其中一個是關於不同類型的代理，我稍後會討論。

Two are just kind of generative AI-powered features. So some of the canonical examples of that are things like in advertising, helping advertisers basically run ads without needing to supply as much creative or, say, if they have an image but it doesn't fit the format, be able to fill in the image for them. So I talked about that a little bit upfront in my comments. But there's stuff like that across every app.

兩種都只是由人工智慧驅動的生成功能。因此，一些典型的例子就是在廣告領域，幫助廣告商基本上投放廣告而不需要提供太多的創意，或者說，如果他們有圖像但不符合格式，就可以為他們填充圖像。所以我在評論中提前談論了這一點。但每個應用程式都有類似的東西。

And then the third category of things, I'd say are broadly focused on productivity and efficiency internally. So everything from helping engineers write code faster to helping people internally understand the overall knowledge base at the company and things like that. So there's a lot to do on each of those zones.

然後，我認為第三類事情主要集中在內部的生產力和效率。因此，從幫助工程師更快地編寫程式碼到幫助公司內部人員了解公司的整體知識庫等等。因此，每個區域都有很多事情要做。

For AI agents, specifically, I guess what I'd say is, one of the things that's different about how we think about this compared to some others in the industry is we don't think that there's going to be one single AI that people interact with, just because there are all these different entities on a day-to-day basis that people come across, whether they're different creators or different businesses or different apps or things that you use.

具體來說，對於人工智慧代理，我想說的是，我們與業內其他人對此的看法不同之處在於，我們不認為人們會與單一的人工智慧進行交互，因為人們每天都會遇到各種各樣的實體，無論它們是不同的創造者、不同的企業、不同的應用程式還是你使用的東西。

So I think that there are going to be a handful of things that are just sort of focused on helping people connect around expression and creativity and facilitating connections. I think there are going to be a handful of experiences around helping people connect with the creators who they care about and helping creators foster their communities.

所以我認為將會有一些事情專注於幫助人們在表達和創造力方面建立聯繫並促進聯繫。我認為將會有一些經驗來幫助人們與他們關心的創作者建立聯繫並幫助創作者培養他們的社群。

And then the one that I think is going to have the fastest direct business loop is going to be around helping people interact with businesses. And you can imagine a world on this, where, over time, every business has as an AI agent that basically people can message and interact with. And it's going to take some time to get there, right? I mean, this is going to be a long road to build that out.

我認為最快的直接商業循環就是幫助人們與企業互動。你可以想像這樣一個世界，隨著時間的推移，每個企業都會擁有一個人工智慧代理，人們可以與之發送訊息並進行互動。到達那裡需要一些時間，對嗎？我的意思是，要實現這一目標還有很長的路要走。

But I think that that's going to improve a lot of the interactions that people have with businesses, as well as if that does work, it should alleviate one of the biggest issues that we're currently having around messaging monetization is that in order to -- for a person to interact with a business, it's quite human labor-intensive for a person to be on the other side of that interaction, which is one of the reasons why we've seen this take off in some countries where the cost of labor is relatively low.

但我認為這將改善人們與企業之間的許多互動，而且如果這確實有效，它應該可以緩解我們目前在消息貨幣化方面遇到的最大問題之一，即為了使一個人與企業進行互動，互動另一端的人需要耗費大量的人力，這也是我們看到這種技術在一些勞動力成本相對較低的國家蓬勃發展的原因之一。

But you can imagine in a world where every business has an AI agent, that we can see the kind of success that we're seeing in Thailand or Vietnam with business messaging could kind of spread everywhere. And I think that's quite exciting.

但你可以想像，在一個每個企業都有一個人工智慧代理的世界裡，我們可以看到我們在泰國或越南看到的那種商業訊息傳遞的成功可以傳播到世界各地。我認為這非常令人興奮。

But overall, I think we're going to follow a playbook that's similar to what we normally do on products. In terms of how quickly some of these new products scale, that's one of the big unknowns for their business and one of the things that we're debating heavily when thinking through the amount of AI CapEx to bring online, because the reality is we just don't know how quickly these will scale. And we want to have the capacity in place in case they scale very quickly. But because they're kind of brand-new things, there aren't that many precedents for things like this, it's actually quite hard to forecast.

但總的來說，我認為我們將遵循與我們通常在產品上所做的類似的策略。至於這些新產品的擴展速度有多快，這是其業務面臨的一大未知數，也是我們在考慮上線人工智慧資本支出數額時激烈爭論的問題之一，因為事實上我們不知道這些產品的擴展速度有多快。我們希望擁有足夠的能力，以應對它們快速擴張的情況。但因為它們是一種全新的事物，所以這樣的先例並不多，實際上很難預測。

So I don't know, Susan, if there's anything on that last point that you want to jump in on. But I think that's probably all I'll say on this for now. But I'm really excited about the segment. This is just going to be -- it fits into all the different products that we're building really well. I think that this is going to both complement and touch and transform every single thing that we're doing, and I'm really excited for it. So tune in to Connect.

所以，蘇珊，我不知道您是否想就最後一點發表任何意見。但我想目前我大概就說這麼多了。但我對這個片段確實很興奮。這將是——它非常適合我們正在製造的所有不同產品。我認為這將會補充、觸及和改變我們正在做的每一件事，我對此感到非常興奮。因此請關注 Connect。

Your next question comes from the line of Eric Sheridan with Goldman Sachs.

您的下一個問題來自高盛的 Eric Sheridan。

Maybe one for Mark, one for Susan. Mark, just following up on Brian's question. I did want to ask just to draw that out a little bit more, how you think about the extensions of the developer community sort of growing up around a platform like Llama 2.

也許一個給馬克，一個給蘇珊。馬克，我只是想回答布萊恩的問題。我確實想問一下，只是為了進一步闡述這一點，您如何看待圍繞 Llama 2 這樣的平台發展起來的開發者社群的擴展。

And we were intrigued by the Microsoft announcement. You traditionally have been more of a consumer-facing company with product. Could this provide you an avenue to be more of an enterprise-facing company over the long term? And is there a strategy there that maybe we haven't seen from you in the past? I'd love to flesh that out a little bit.

微軟的聲明讓我們很感興趣。傳統上，你們是一家面向消費者、提供產品的公司。從長遠來看，這能否為您提供一條成為面向企業公司的途徑？是否存在我們過去從未見過的策略？我很想進一步充實這一點。

And Susan, just one for you. The VRL losses just continue to build. And I think we continue to struggle a little bit of what the drivers of those losses are and how should we think about some of the components driving the losses versus elements of earning a return on those losses over the medium to long term, as Mark sort of framed the time frame around metaverse. So any color there about sort of rate of change or components of the losses in VRL, I think would be super helpful.

蘇珊，我只想告訴你一件事。 VRL 損失還在持續增加。我認為我們仍在努力弄清楚造成這些損失的因素是什麼，以及我們應該如何看待造成這些損失的一些因素，以及在中長期內從這些損失中獲得回報的因素，正如馬克圍繞元宇宙所設定的時間框架一樣。因此，我認為，任何有關 VRL 中變化率或損失成分的顏色都會非常有幫助。

Yes, sure. I'll take a cut at both of those, and Susan can jump into the second one if there's anything that she wants to add there as well. All right. So Llama is an open-source project, which is a little bit different from building out a developer platform, although there will be an ecosystem around this.

是的，當然。我會對這兩個進行刪減，如果蘇珊還想加什麼的話，她也可以跳到第二個。好的。因此，Llama 是一個開源項目，與建構開發者平台略有不同，儘管圍繞它會有一個生態系統。

What we've seen around open-source work that we've done, which we've done a lot of in our core infrastructure work, design of servers and data centers and basic infrastructure as well as in AI. And I pointed out some of these in my remarks upfront, like PyTorch, and just a bunch of other models that we've released recently.

我們看到，我們在開源工作方面已經做了很多工作，包括核心基礎設施工作、伺服器和資料中心的設計以及基礎設施以及人工智慧。我在前面的評論中指出了其中的一些，例如 PyTorch，以及我們最近發布的許多其他模型。

One of the things that we've seen is that when you release these projects publicly and in open source, there tend to be a few categories of innovations that the community makes. So on the one hand, I think it's just good to get the community standardized on the work that we're doing. That helps with recruiting because a lot of the best people want to come and work at the place that is building the things that everyone else uses. It makes sense that people are used to these tools from wherever else they're working. They can come here and build here.

我們看到的情況之一是，當你公開和開源發布這些專案時，社群往往會做出幾類創新。因此，一方面，我認為讓社區對我們所做的工作進行標準化是件好事。這有助於招聘，因為許多優秀人才都想來到生產其他人都在使用的產品的地方工作。無論人們在何處工作，他們都會習慣使用這些工具，這是有道理的。他們可以來這裡建造。

But in terms of improvements that we expect to see from the community, there are really a few different types. One is specifically around safety and security. That's a very important issue in AI. In open-source software overall, we've just seen through the history here that open-source software gets scrutinized more, and therefore, ends up being more secure and safer. And we think that there's a very good chance that's the likely outcome here. And whatever improvements that people make to harden it in the community, we'll be able to roll that into our work, both for the first-party products that we'll launch as well as making it easy to propagate that across the industry and make the AI that everyone uses safer.

但就我們期望從社區看到的改進而言，確實存在幾種不同的類型。其中一個就是專門關於安全和保障。這是人工智慧中一個非常重要的問題。總體而言，就開源軟體而言，我們剛剛透過歷史看到，開源軟體受到了更嚴格的審查，因此最終變得更加安全。我們認為這很有可能是最終的結果。無論人們為在社區中強化它而做出什麼改進，我們都能夠將其融入到我們的工作中，既用於我們將要推出的第一方產品，也使其易於在整個行業中傳播，並使每個人使用的人工智能更安全。

I would also hope to see efficiency improvements. I mean even from the initial Llama that was just released as a research project, some of the improvements are around things like quantization and being able to run the model way more efficiently.

我也希望看到效率的提升。我的意思是，即使是從剛剛作為研究項目發布的初始 Llama 開始，一些改進也圍繞著量化和能夠更有效地運行模型等方面。

Now we're spending so much on AI CapEx that all the help that we can get from the community to make this stuff more efficient to run will be helpful not just for individuals to be able to run powerful models on their laptop or locally or without a huge amount of compute, so individuals can afford it, but that should hopefully translate over time into just a more efficient infrastructure for us, which hopefully could be quite a big savings. So that's more what we're seeing there.

現在我們在人工智慧資本支出上投入瞭如此之多，以至於我們可以從社區獲得的所有幫助都將使這些東西運行得更高效，這不僅有助於個人能夠在筆記型電腦上或本地運行強大的模型，或者不需要大量的計算，因此個人可以負擔得起，而且希望隨著時間的推移，這能轉化為我們更高效的基礎設施，希望這可以節省一大筆錢。這就是我們在那裡看到的更多內容。

We partnered with Microsoft specifically because we don't have a public cloud offering. So this isn't about us getting into that. It's actually the opposite. We want to work with them because they have that and others have that, and that was the thing that we aren't planning on building out.

我們與微軟合作是因為我們沒有公有雲產品。所以這與我們無關。事實恰恰相反。我們想與他們合作，因為他們有這個，其他人也有這個，而這是我們不打算拓展的東西。

But one of the things that you might have noticed is, in addition to making this open through the open-source license, we did include a term that for the largest companies, specifically ones that are going to have public cloud offerings, that they don't just get a free license to use this. They'll need to come and make a business arrangement with us.

但您可能已經注意到的一件事是，除了透過開源許可證開放之外，我們還加入了一個條款，即對於最大的公司，特別是那些將提供公有雲服務的公司，他們不能只是獲得免費許可來使用它。他們需要來與我們達成商業協議。

And our intent there is we want everyone to be using this. We want this to be open. But if you're someone like Microsoft or Amazon or Google, and you're going to basically be reselling these services, that's something that we think we should get some portion of the revenue for.

我們的目的是希望每個人都能使用它。我們希望這是公開的。但如果你是微軟、亞馬遜或谷歌這樣的公司，你基本上是轉售這些服務，我們認為我們應該從中獲得一定比例的收入。

So those are the deals that we intend to be making, and we've started doing that a little bit. I don't think that that's going to be a large amount of revenue in the near term. But over the long term, hopefully, that can be something. Not sure if there's anything else to add on that, but Susan, you can jump in if you want.

這些都是我們打算達成的交易，而且我們已經開始做一些這方面的工作了。我不認為短期內這會帶來大量收入。但長遠來看，希望這能有所作為。不確定是否還有其他要補充的，但是蘇珊，如果你願意的話，你可以加入。

On RL, let me just jump over to that for a second. I don't think we're going to break out numbers right now but I'll defer to Susan on that. I mean one big thing for the next year is the launch of Quest 3, right? And basically, this is going to be the biggest headset that we've released since 2020 when we came out with Quest 2. And there are just a lot of expenses related to bringing that to market. And I think that's, at least in the near term, going to be one of the major drivers of this. So I think that that's probably the main thing that I would point to there.

關於 RL，讓我先簡單談談。我認為我們現在不會公佈具體數字，但我會聽從蘇珊的意見。我的意思是明年的一件大事就是推出 Quest 3，對吧？基本上，這將是我們自 2020 年推出 Quest 2 以來發布的最大的耳機。而將其推向市場需要大量的費用。我認為，至少在短期內，這將是這個問題的主要驅動因素之一。所以我認為這可能是我要指出的主要內容。

We're continuing to -- I continue to think, looking at the VR work that we're doing and the AR work, some of the neural interfaces work, I think we're leading in these areas. It's been good to see what others in the industry have done because, to some degree, it gives us more confidence that we're on the right track.

我們正在繼續——我一直在思考，看看我們正在做的 VR 工作和 AR 工作，以及一些神經介面工作，我認為我們在這些領域處於領先地位。看到業內其他人所做的事情是件好事，因為在某種程度上，它讓我們更相信我們走在正確的道路上。

And even though I know from an investor standpoint, most people aren't investing on quite as long of a time horizon as we are here, so I kind of get that a lot of investors might want to see us spending less here in the near term. My view is that we are leading in these areas. I believe that they're going to be big over time. I think we've shown that we can deliver good business results in the near term while investing ambitiously in the long term. So I'm planning on continuing to do that, and I do continue to believe that, over time, we will be happy that we did that.

儘管我從投資者的角度知道，大多數人的投資期限並不像我們這麼長，所以我明白很多投資者可能會希望看到我們在短期內減少支出。我的觀點是我們在這些領域處於領先地位。我相信隨著時間的推移它們會變得越來越大。我認為我們已經證明，我們可以在短期內取得良好的業務成果，同時在長期內進行雄心勃勃的投資。所以我計劃繼續這樣做，我確實相信，隨著時間的推移，我們會很高興我們這樣做。

So I don't know if, Susan, if there's anything you wanted to add there.

所以我不知道蘇珊，您是否還有什麼要補充的。

Sure. Let me just add a couple of things. Very quickly on the first question, I just want to clarify that we don't expect that Llama is going to result in a separate top line enterprise revenue line, so just making sure we're totally clear.

當然。我再補充幾點。關於第一個問題，我只想快速澄清一下，我們不認為 Llama 會帶來單獨的企業收入，因此請確保我們完全清楚。

On the second question about the growth in Reality Labs' operating losses in 2024, Mark alluded to the fact that this is an ambitious long-term horizon, multifaceted road map. There are lots of components to the Reality Labs portfolio across VR, AR, metaverse social platforms, neural interfaces. And we really have a long-term time horizon for evaluating the return on our investments here.

關於第二個問題，即 Reality Labs 2024 年的營運虧損成長情況，馬克暗示這是一個雄心勃勃的長期願景，是一個多方面的路線圖。 Reality Labs 的產品組合包含許多組件，包括 VR、AR、元宇宙社交平台和神經介面。我們確實有一個長期的時間範圍來評估我們在此的投資回報。

So in the near term, we're focused on growing adoption of the existing products, and we're constantly learning more about demand and use cases that inform our future plans. But a lot of the investment that's driving the growth here is around conducting the fundamental R&D to solve hard technology problems that are going to enable our vision here. A lot of it is around clearing technical hurdles that will make subsequent devices smaller, cost less, weigh less, et cetera. So that's really on the Reality Labs side.

因此，在短期內，我們將專注於擴大現有產品的採用，並不斷了解需求和用例，為我們的未來計劃提供資訊。但推動這裡成長的大量投資都圍繞著進行基礎研發，以解決實現我們願景的難題。其中很大一部分是為了清除技術障礙，使後續設備更小、更便宜、更輕等等。這確實是現實實驗室方面的事情。

Again, as we said, we expect operating losses to increase meaningfully year-over-year in 2024. In 2024, specifically, I think that will be driven by a combination of both head count-related and operating costs. But again, our ambitions in Reality Labs haven't changed, and it continues to be a significant long-term opportunity for us.

再次，正如我們所說，我們預計 2024 年的營運虧損將比去年同期大幅增加。具體到 2024 年，我認為這將受到員工人數相關成本和營運成本的共同影響。但同樣，我們對現實實驗室的雄心壯志沒有改變，它對我們來說仍然是一個重要的長期機會。

Your next question comes from the line of Mark Shmulik with Bernstein.

您的下一個問題來自伯恩斯坦的馬克·施穆里克 (Mark Shmulik)。

Mark, things move quickly. Like a month ago, we weren't talking about Reels. A quarter ago, we weren't talking about Llama. And start of the year, we were barely talking about AI. If you think back to kind of just how you were prioritizing your time at the start of the year to kind of where you are today and how that's kind of changed, would love to just get some color on kind of the changing priorities and really where you're spending your time.

馬克，事情進展很快。大約一個月前，我們還沒有談論 Reels。一個季度前，我們還沒有談論過 Llama。今年年初，我們幾乎沒有談到人工智慧。如果您回想一下從年初到現在您的時間安排情況以及時間安排發生了怎樣的變化，我很想了解一下時間安排的變化以及您實際上把時間花在了哪裡。

And then the second question, really impressive, obviously, acceleration of growth in kind of the core ad business. Would love to just get some color kind of beyond the vertical contribution. What are you seeing in terms of kind of adoption of Advantage+ products really driving some of that improved growth, especially what we've seen relative to the rest of the sector?

第二個問題確實令人印象深刻，顯然是核心廣告業務的成長加速。希望能夠獲得超越垂直貢獻的某種顏色。就 Advantage+ 產品的採用而言，您認為哪些因素真正推動了成長，尤其是相對於該行業的其他產品？

Yes. I can start with the first and then Susan can take the second. I actually think our priorities have been pretty consistent for a few years now. I think the way that people hear them might be different. But for example, last year, I think we were getting quite a bit of critique for the volume of AI CapEx spending that we were doing. And now, obviously, people want to understand where that's going and where we think that trajectory is going and want to make it as efficient as possible. But I think, at this point, it's more well understood that I think that was a good investment, and it's driving results in the near term and enabling us to build some of the new experiences that I think we all think are pretty fundamental.

是的。我可以從第一個開始，然後蘇珊可以進行第二個。我實際上認為幾年來我們的優先事項一直非常一致。我認為人們聽到它們的方式可能會有所不同。但例如，去年，我認為我們因人工智慧資本支出的數額而受到不少批評。現在，顯然人們想了解它的發展方向以及我們認為的軌跡，並希望使其盡可能高效。但我認為，在這一點上，大家更容易理解，我認為這是一項很好的投資，它將在短期內推動業績，並使我們能夠建立一些我認為我們都認為非常重要的新體驗。

I don't think Reels is new, although maybe you meant Threads when you said that. Threads, I would say, it's not that -- Threads is like -- it's not a massive project. It was -- I've thought for a while that the opportunity around public conversations and kind of a text-based product was bigger than what had been executed yet in the market. So we had a relatively small team work on that.

我不認為 Reels 是新產品，儘管您這麼說也許是指 Threads。我想說的是，Threads 並不是一個大型專案。是的——我一直認為，圍繞公共對話和基於文字的產品的機會比市場上已經實現的機會更大。因此，我們有一個相對較小的團隊來做這件事。

And this year, I think there have been 2 things that have been -- that have just vastly exceeded my expectations. One was Llama. The initial model, we thought it was interesting. But the scale of adoption, even just for the research version, really spurred us to go do Llama 2, and that was vastly more than we expected.

今年，我認為有兩件事大大超出了我的預期。其中一個是駱駝。我們認為最初的模型很有趣。但即使只是研究版本，其採用規模也確實促使我們開發 Llama 2，這遠遠超出了我們的預期。

And the second is Threads has been dramatically more than we expected in terms of the adoption and the rate of that. We thought this was going to be kind of a project that we just -- we had a small team working on for a while, but it really kind of blew up and created a big opportunity immediately.

第二，Threads 的採用率和普及率遠遠超出了我們的預期。我們原以為這只是一個小團隊在做的一個項目，但它真的取得了成功，並立即創造了一個巨大的機會。

But no, I mean, I think over time, you should expect that we're going to focus on AI and the metaverse. AI includes both all of the ranking and recommendation systems that power the core apps, so all the content that you're seeing in Facebook and Instagram, all the ads content that you're seeing. It also underpins all the safety systems that we build.

但不，我的意思是，我認為隨著時間的推移，你應該預料到我們會專注於人工智慧和元宇宙。 AI 包括支援核心應用程式的所有排名和推薦系統，因此您在 Facebook 和 Instagram 上看到的所有內容以及您看到的所有廣告內容。它也支撐著我們所建造的所有安全系統。

And increasingly, it's all the generative AI stuff. So all the agents, all the generative features that we're going to be rolling out and a lot of the other work that's going to underlie some of the efficiency stuff that we're doing internally. And then the metaverse stuff, I think we've talked about for a while. So I don't think that there's much change there, except that I'd say the signals that we're getting from the market are -- it's certainly not getting adopted a lot faster than we expected. So that's sort of the somewhat sobering signal.

而且越來越多地，它都是生成性人工智慧的東西。因此，我們將要推出的所有代理、所有生成功能以及許多其他工作都將成為我們內部正在進行的一些效率工作的基礎。然後關於元宇宙的事情，我想我們已經討論了一段時間了。所以我不認為那裡會發生太大的變化，只是我想說我們從市場得到的訊號是——它的採用速度肯定不會比我們預期的快很多。所以這多少是令人警醒的訊號。

But on the other side, I think a lot of companies that otherwise are doing, we respect and do great work, we don't necessarily view as building things that are ahead of where we are, which gives us confidence that we think that the long-term thesis still will hold there. We think that we're going to be the leaders in it. And nothing that we're seeing from the market makes us rethink that in a fundamental way.

但另一方面，我認為很多公司在其他方面做得很好，我們尊重它們，它們也做得很好，我們不一定認為它們在建立領先於我們的東西，這讓我們有信心，我們認為長期論點仍然成立。我們認為我們將成為該領域的領導者。而我們從市場上看到的一切都沒有讓我們從根本上重新思考這一點。

So I think we're going to continue focusing on AI and the metaverse as the 2 big thrusts of what we're doing, and all the other things kind of fall out of that.

因此，我認為我們將繼續專注於人工智慧和元宇宙作為我們正在做的兩大重點，而其他所有事情都將由此產生。

And Mark, I'll take the second question that you asked around the Q2 revenue acceleration. So I think there were a couple of parts. One was around the acceleration in the core ads business. The second was on the Advantage+ products.

馬克，我想回答你關於第二季營收加速成長的第二個問題。所以我認為有幾個部分。一是核心廣告業務的加速發展。第二個是關於 Advantage+ 產品。

In terms of the Q2 revenue acceleration, I'd highlight there are a few factors driving that. The first is, frankly, we're lapping a weaker demand period, including the first full quarter of the war in Ukraine and the suspension of our services in Russia.

就第二季營收加速而言，我想強調的是有幾個因素推動了這一成長。首先，坦白說，我們正在經歷一段需求疲軟的時期，包括烏克蘭戰爭的第一個完整季度以及我們在俄羅斯的服務暫停。

Second, we saw increased supply and improvements to ad performance, including improved Reels monetization as we continue to work down the Reels revenue headwind.

其次，我們看到供應量增加和廣告效果改善，包括隨著我們繼續努力降低 Reels 收入逆風，Reels 貨幣化得到改善。

And third, there were lower FX headwinds for us this quarter. So those were all 3 things that helped drive the revenue acceleration in Q2.

第三，本季我們面臨的外匯逆風較小。所以這三件事都有助於推動第二季營收的成長。

In terms of the question about Advantage+ specifically, Advantage+ is one of multiple AI-powered ad products that we have right now in the market. With Advantage+ specifically, we're seeing strong adoption and particular success with the e-commerce and retail verticals, and we've seen good traction with other verticals like CPG, especially DTC brands. And we're continuing to launch features to unlock use cases for advertisers and make it easier for them to adopt Advantage+ campaigns and measure their performance gains.

關於 Advantage+ 的具體問題，Advantage+ 是我們目前在市場上推出的多種人工智慧廣告產品之一。具體來說，憑藉 Advantage+，我們看到其在電子商務和零售垂直領域的廣泛應用和成功案例，並且我們也看到了其在 CPG 等其他垂直領域，尤其是 DTC 品牌的良好發展勢頭。我們將繼續推出新功能，為廣告主解鎖用例，讓他們更容易採用 Advantage+ 活動並衡量其績效提升。

So we've got a lot in the pipeline there that we're excited about for both the Advantage+ shopping campaigns specifically, which were our first sort of foray into this area, but then the Advantage+ portfolio more broadly.

因此，我們有許多計劃正在籌備中，我們對 Advantage+ 購物活動感到興奮，這是我們首次涉足該領域，但隨後是更廣泛的 Advantage+ 產品組合。

That basically enables us to take that same playbook of helping advertisers iterate and test very quickly and apply it to many different steps of the end-to-end ad buying experience.

這基本上使我們能夠採用相同的策略來幫助廣告商快速迭代和測試，並將其應用於端到端廣告購買體驗的許多不同步驟。

So the feedback and results that we've seen from advertisers is good, and we think it's a really promising area that we're continuing to invest in. But it's one of many ways that we're using AI to continue to sort of help make our ads systems and recommendations and ranking engines more performant to deliver better measurement and results to advertisers.

因此，我們從廣告商看到的回饋和結果是好的，我們認為這是一個非常有前景的領域，我們將繼續投資。但這只是我們使用人工智慧的眾多方法之一，以繼續幫助我們的廣告系統、推薦和排名引擎更加高效，從而為廣告商提供更好的衡量標準和結果。

Your next question comes from the line of Justin Post with Bank of America Merrill Lynch.

您的下一個問題來自美銀美林的賈斯汀波斯特 (Justin Post)。

I guess I want to follow up on Reality Labs passing $40 billion in losses and increasing annually next year. Just think about -- maybe help us understand the ROI on the business, how you're thinking about that investment, either on a stand-alone basis or as a complement to the Family of Apps, if you're thinking about it from an investor perspective.

我想跟進一下 Reality Labs 明年的虧損情況，預計超過 400 億美元，並且每年都會增加。想想看—如果您從投資者的角度考慮，也許可以幫助我們了解業務的投資回報率，您如何看待這項投資，無論是作為獨立投資還是作為應用程式系列的補充。

Well, I think it's both over time. The primary way that I think about this is that we've been able to show that we can -- we've been quite successful at building large-scale social experiences within the constraints of platforms that, often, our competitors are defining. I think we're going to be able to do even better work. And there's a lot of things that I would like to see us build that we just can't because we're -- of the ways that we're constrained by the competitors who build these platforms.

嗯，我認為隨著時間的推移，兩者都會改變。我對此的主要看法是，我們已經能夠證明我們可以——我們在平台限制內建立大規模社交體驗方面相當成功，而這些限制通常是由我們的競爭對手定義的。我認為我們能夠做得更好。我希望我們能夠建立很多東西，但我們卻無法實現，因為我們受到了建構這些平台的競爭對手的限制。

And over time, the main way that I think about this is from a product and mission perspective is we're here to build awesome experiences that help people connect. I think helping to shape the next platform is going to unlock that in a profound way for decades to come. And that's what I'm here to do and I think what a lot of people are here to do. So that's kind of the first principles part of this.

隨著時間的推移，我主要從產品和使命的角度來思考這個問題，我們來這裡是為了創造幫助人們建立聯繫的絕妙體驗。我認為，幫助塑造下一個平台將在未來幾十年內以深刻的方式解鎖這個問題。這就是我來這裡要做的，我想這也是很多人來這裡要做的事。這就是其中的第一原則部分。

From a business perspective, I think that there's going to end up being a large business component of this that is Reality Labs-specific. Directly, the products that we're building there and having that be a good business, but I also think that a lot of this is going to be that it unlocks a lot of value for the other experiences, the current apps that we have, what you think of it as the Family of Apps, where we currently, just a lot of the potential value, whether it's just engagement that could be created, features that we would love to build that we're not allowed to because Apple or others just don't allow us to build those things. And I think that, that's really unfortunate for the industry.

從商業角度來看，我認為這最終會成為 Reality Labs 特有的一個大型商業組成部分。直接而言，我們在那裡開發的產品以及擁有的好生意，但我還認為，這在很大程度上將為其他體驗釋放出很多價值，我們目前擁有的應用程序，你認為的應用程序系列，我們目前有很多潛在的價值，無論是可以創造的參與度，還是我們想要構建的功能，我們不允許這樣做，因為蘋果或其他人不允許我們構建這些東西。我認為這對這個行業來說真的很不幸。

I think that there's a lot of lost engagement that would have happened, a lot of the monetization and value that gets created. I mean just look at what happened with App Tracking Transparency, the massive value destruction for small businesses because of the rules that were set by another platform.

我認為這會導致大量的參與度喪失，以及大量創造的貨幣化和價值損失。我的意思是，看看應用程式追蹤透明度發生了什麼，由於另一個平台製定的規則，小型企業的價值遭受了巨大的破壞。

And that's not -- in a future version of the world, it's not -- that would be recognized not by device sales or some new, like, experience that we're building in Reality Labs. You would basically see that accrue through more efficient monetization of the apps and experiences that we're building in other places. So I think it's going to be both.

而在未來的世界裡，這並不是透過設備銷售或我們在現實實驗室中建構的一些新體驗來識別的。您基本上會看到，透過我們在其他地方建立的應用程式和體驗的更有效的貨幣化，收益會不斷增加。所以我認為兩者皆有。

But I think this is just a very fundamental thing. This is a very long-term bet. At a deep level, I understand the discomfort that a lot of investors have with it because it's just outside of the model of, I think, even most long-term investors, how you would think about this.

但我認為這只是一件非常基本的事情。這是一項非常長期的賭注。從深層上講，我理解許多投資者對此感到不適，因為我認為，這超出了大多數長期投資者的思維模式。

And look, I mean, I can't guarantee you that I'm going to be right about this bet. I do think that this is the direction that the world is going in. There are 1 billion or 2 billion people who have glasses today. I think, in the future, they're all going to be smart glasses.

聽著，我的意思是，我不能向你保證我對這個賭注的判斷是正確的。我確實認為這是世界的發展方向。今天有10億或20億人戴眼鏡。我認為，未來它們都會是智慧眼鏡。

And all the time that we spend on TVs and computers, I think that's going to get more immersive and look something more like VR in the future. And I think that what we're seeing is richer ways for people to communicate across even the mobile apps that we have, going from text to photos to videos, just a continual trend towards being more immersive.

而且我們花在電視和電腦上的所有時間，我認為未來將會變得更加身臨其境，看起來更像虛擬實境。我認為，我們看到人們透過現有的行動應用程式進行交流的方式越來越豐富，從文字到照片再到視頻，這是一種持續的、更具沉浸感的趨勢。

So like all of these trends line up to make me think that this is the right thing. I think we're going to be happy that we did this. So that's kind of how I think about it, both from kind of a mission and product perspective, a business perspective, an investment time frame perspective. But I also understand the discomfort that some folks have with something that I can't put exact numbers on a near-term time horizon around.

所以所有這些趨勢都讓我認為這是正確的事情。我想我們會很高興我們這樣做。這就是我對此的看法，從使命和產品角度、商業角度、投資時間框架角度。但我也理解有些人對某些事情感到不舒服，因為我無法在短期內給出確切的數字。

Your next question comes from the line of Doug Anmuth with JPMorgan.

您的下一個問題來自摩根大通的 Doug Anmuth。

One for Mark, one for Susan. Mark, you touched on it a bit, but can you just talk a little bit more about what you've learned from Threads early on, not just for that product specifically but also as you look to leverage the Family of Apps platform to launch additional products over time.

一個給馬克，一個給蘇珊。馬克，你稍微提到了這一點，但你能否再多談一下你早期從 Threads 中學到的東西，不僅僅是針對該產品，還包括你希望利用 Family of Apps 平台隨著時間的推移推出更多產品。

And then, Susan, you lowered the CapEx outlook this year by $3 billion. Can you just provide more color on the delays in projects and equipment in '23? And is there any way to frame how meaningful the CapEx increase could be in '24?

然後，蘇珊，你將今年的資本支出預期下調了 30 億美元。您能否提供更多有關 23 年專案和設備延遲的詳細資訊？有什麼方法可以說明 24 年資本支出增加的意義嗎？

Yes. On Threads, it's maybe too early to do this kind of analysis. I mean, I'm -- on the one hand, we've tried a bunch of stand-alone experiences over time. And in general, we haven't had a lot of success with building kind of stand-alone apps. The biggest exception to that, of course, is Messenger, but that started off as functionality inside Facebook and was spun out.

是的。對於 Threads 來說，現在進行此類分析可能還為時過早。我的意思是，一方面，我們隨著時間的推移嘗試了許多獨立的體驗。總的來說，我們在建立獨立應用程式方面並沒有取得太大的成功。當然，最大的例外是 Messenger，但它最初是 Facebook 內部的功能，後來被分拆出來。

So part of me wonders if this is just a kind of classic venture capital portfolio question, where you try a bunch of things and a bunch of them don't work. And then every once in a while, 1 hits and is a much bigger success. It could be that.

所以我有點懷疑這是否只是一個經典的創投組合問題，你嘗試了很多方法，但其中許多都沒有成功。然後每隔一段時間，就會有 1 次成功，並且取得更大的成功。有可能。

Or it could just be that this is such an idiosyncratic case because of all of the factors that are happening around Twitter or X, I guess, it's called now. So it's hard to say. I mean, but I think when something works or doesn't, you can often point to the reason why it did or didn't. And I think it is an interesting intellectual question of whether you could have known that a priori, that, that was actually going to be the case.

或者這可能只是一個特殊的案例，因為圍繞 Twitter 或 X 發生的所有因素，我猜，它現在被稱為。所以很難說。我的意思是，但我認為當某件事成功或失敗時，你通常可以指出它成功或失敗的原因。我認為這是一個有趣的智力問題，即你是否可以先驗地知道事實確實如此。

But -- so I'm not sure. But also, rather than trying to kind of analyze that, I'd say we have a lot of work to do to really make Threads reach its full potential. That's not a foregone conclusion yet, even though I think we're off to a great start, and I'm optimistic that, over time, this could be a fifth great app in the Family of Apps. But we have a lot of basic work to do.

但是——所以我不確定。但是，與其嘗試分析這一點，不如說我們還有很多工作要做才能真正發揮 Threads 的全部潛力。儘管我認為我們已經有了一個良好的開端，但這還不是一個定論，而且我樂觀地認為，隨著時間的推移，這可能會成為應用程式家族中的第五個偉大的應用程式。但我們還有很多基礎工作要做。

And we have a basic playbook here, which is build an experience. It's got to be something that people like so it has to have product market fit. Once you get that, it's not always retentive. So a lot of people might like an experience but you need to kind of tune it so that way, the numbers work so that people who use it are continuing. We feel like we're getting to a good place on that with Threads. There's still a lot of basic functionality to build.

我們這裡有一個基本的劇本，就是建立一種體驗。它必須是人們喜歡的東西，所以它必須適合產品市場。一旦你明白了這一點，它並不總是能被記住的。因此，很多人可能喜歡某種體驗，但你需要對其進行調整，以便數字發揮作用，以便使用它的人可以繼續使用。我們覺得透過 Threads 我們在這方面取得了很好的進展。還有很多基本功能要建置。

Once we feel like we're in a very good place on that, then I'm highly confident that we're going to be able to pour enough gasoline on this to help it grow once we get to the point where we feel good that everyone who's using it is going to continue using it at a high rate. And then a few years, once we get to the point where it's at hundreds of millions of people, assuming we can get there, then we'll worry about monetization. But I mean, that's basically the playbook that we're focused on.

一旦我們感覺到我們處於一個非常好的位置，那麼我非常有信心，我們將能夠為它注入足夠的動力，幫助它發展，一旦我們達到了讓我們感覺良好的程度，使用它的每個人都會繼續以高頻率使用它。幾年後，一旦我們達到了擁有數億用戶的水平（假設我們能夠實現這一目標），我們就會擔心貨幣化問題。但我的意思是，這基本上就是我們關注的劇本。

And so rather than thinking about right now, like what does this mean for other things like it that we can build, I'd say we're really just focused on taking this opportunity, which is an awesome one that we didn't expect to this scale and making sure we make the most of this and execute it. But I do think it has been sort of this weird anomalous thing in the tech industry that there hasn't been an app for public discussions like this that has reached 1 billion people.

因此，與其現在思考這對我們可以建立的其他類似事物意味著什麼，不如說我們真的只是專注於抓住這個機會，這是一個我們沒想到會達到如此規模的絕佳機會，並確保我們充分利用它並執行它。但我確實認為，科技業中有一種奇怪的異常現象，即還沒有一款可以覆蓋 10 億人的用於公開討論的應用程式。

When I look at all the different social experiences, it just seems like there should be 1 like this. I think there are a lot of reasons that you can point to why that might have not been the case historically. But it's awesome that we get a chance to work on this, and I'm really optimistic about where we are. But it's going to be a long road ahead.

當我回顧所有不同的社交經驗時，似乎應該有一個這樣的經驗。我認為有很多理由可以解釋為什麼歷史上的情況可能並非如此。但我們很高興能有機會從事這項工作，我對我們目前的狀況非常樂觀。但前面的路還很長。

And Doug, on your second question about our 2023 CapEx forecast and the impact on '24, so I'll start with 2023. The reduced forecast that we gave for '23, I mentioned on the call, is driven both by some cost savings, particularly on non-AI servers, where we're -- we previously had some underutilized capacity, and we've been identifying ways to be more efficient in the way that we allocate that capacity towards all of our various needs, as well as shifts in CapEx into 2024 that's coming from delays in data center projects and server deliveries. And that will just push that associated CapEx, which we were planning for in '23 into '24.

道格，關於你的第二個問題，即我們對 2023 年資本支出的預測及其對 2024 年的影響，我先從 2023 年說起。我在電話會議上提到，我們對 2023 年的預測有所下調，這既是由於一些成本節約，特別是在非人工智慧伺服器上，我們之前有一些未充分利用的產能，我們一直在尋找更有效地分配這些產能以滿足我們各種需求的方式，也是由於資料中心專案和伺服器交付的延遲導致的 2024 年資本支出的變化。這只會將我們在 23 年計劃的相關資本支出推至 24 年。

We're still working on our '24 CapEx plans. We haven't yet finalized that, and we'll be working on that through the course of this year. But I mentioned that we expect that CapEx in '24 will be higher than in '23. We expect both data center spend to grow in '24 as we ramp up construction on sites with the new data center architecture that we announced late last year.

我們仍在製定我們的「24 資本支出計畫」。我們尚未最終確定，但今年我們將致力於此。但我提到，我們預期 24 年的資本支出將高於 23 年。隨著我們加大去年年底宣布的新資料中心架構的建設力度，我們預計 2024 年這兩個資料中心的支出都將成長。

And then we certainly also expect to invest more in servers in 2024 for both AI workloads to support all of the AI work that we've talked about across the core AI ranking, recommendation work, along with the next-gen AI efforts. And then, of course, also our non-AI workloads, as we refresh some of our servers and add capacity just to support continued growth across the site.

然後，我們當然也希望在 2024 年為 AI 工作負載在伺服器上投入更多資金，以支援我們在核心 AI 排名、推薦工作以及下一代 AI 工作中討論的所有 AI 工作。當然，還有我們的非人工智慧工作負載，因為我們刷新了一些伺服器並增加了容量，以支援整個網站的持續成長。

Your next question comes from the line of Youssef Squali with Truist.

您的下一個問題來自 Truist 的 Youssef Squali。

One for Mark and one for Susan maybe. Mark, you touched on this a little earlier, but there is this open-source versus closed-source debate going on with Meta as one of the companies on one side of it with Llama and Llama 2. How do you see this market evolving over time? Do you see one approach as being potentially superior to the other, and therefore, would be likely to garner greater adoption? Or is this really a win-win situation, just considering how early we are in this process and how large the potential TAM is?

也許一個給馬克，一個給蘇珊。馬克，你之前提到過這個問題，但開源與閉源的爭論一直存在，Meta 是其中一家支持 Llama 和 Llama 2 的公司。你認為這個市場會如何發展？您是否認為一種方法可能優於另一種方法，因此可能會獲得更廣泛的採用？或者這真的是一個雙贏的局面，只是考慮到我們處於這個過程的早期階段以及潛在的 TAM 有多大？

And then Susan, your guidance for Q3 implies at the midpoint about a 20% growth, which is a pretty steep acceleration from Q2 at 11% or 12%. Could you maybe just help us understand the drivers of that acceleration? And as we look into Q4, the comps remain pretty easy. Is that sustainable until year-end?

然後蘇珊，您對第三季度的指導意味著中點增長約 20%，這比第二季度的 11% 或 12% 有了相當大的加速。您能否幫助我們了解這種加速的驅動因素？當我們展望第四季時，我們發現比較結果仍然相當簡單。這種狀況能持續到年底嗎？

Yes, I can speak briefly to the first one. I do think that there will continue to be both open and closed AI models. I think there are a bunch of reasons for this. There are obviously a lot of companies that their business model is to build a model and then sell access to it. So for them, making it open would undermine their business model.

是的，我可以簡單談談第一個問題。我確實認為開放和封閉的人工智慧模型將繼續存在。我認為這有很多原因。顯然，許多公司的商業模式是先建立一個模型，然後出售該模型的使用權。因此對他們來說，開放將會破壞他們的商業模式。

That is not our business model. We want to have the -- like we view the model that we're building as sort of the foundation for building products. So if by sharing it, we can improve the quality of the model and improve the quality of the team that we have that is working on that, that's a win for our business of basically building better products. So I think you'll see both of those models.

這不是我們的商業模式。我們希望將我們正在建立的模型視為建立產品的基礎。因此，如果透過共享，我們可以提高模型的品質並提高致力於此的團隊的質量，那麼這對於我們建立更好產品的業務來說是一個勝利。所以我認為你會看到這兩種模型。

There are also some important safety questions that I think we'll need to continue thinking about over time. There are a number of people who are out there saying that once the AI models get past a certain level of capability, it can become dangerous for them to become just in the hands of everyone openly. I think -- what I think is pretty clear is that we're not at that point today. I think that there's consensus generally among people who are working on this in the industry and policy folks that we're not at that point today. And it's not exactly clear at what point you reach that.

我認為還有一些重要的安全問題我們需要持續思考。許多人表示，一旦人工智慧模型的能力達到一定水平，那麼將其公開掌握在每個人手中就會變得非常危險。我認為——我認為很清楚的是，我們今天還沒有到達那個地步。我認為，從事這項工作的業內人士和政策制定者普遍認為，我們目前還沒有達到那個程度。並且並不完全清楚什麼時候才能達到這個目標。

So I think there are people who are kind of making that argument in good faith, who are actually concerned about the safety risk. So I think that there are probably some businesses that are out there making that argument because they want it to be more closed, because that's their business, so I think we need to be wary of that.

所以我認為有些人是出於善意提出這種論點，他們實際上擔心安全風險。所以我認為可能有些企業提出這樣的論點是因為他們希望更加封閉，因為這是他們的業務，所以我認為我們需要對此保持警惕。

But I think for both of those reasons, the business reasons and the safety reasons, I think there will continue to be a mix of open and closed models. And it's not just going to be that like 1 thing is what everyone uses. I think different businesses will use different things for different reasons.

但我認為，出於商業原因和安全原因，我認為開放和封閉模式將繼續混合存在。這並不是說每個人都會使用某一種東西。我認為不同的企業會因為不同的原因使用不同的東西。

And by open sourcing Llama now, we're not -- on the second question around safety, we intentionally did not go out there and say we're going to open source every single thing in the future because we do want to have the space to be able to look at how the safety landscape evolves. And if we think that we do cross some kind of critical threshold in the future, it may not be the right thing to open source it in the future.

現在透過開源 Llama，我們並不是——關於安全性的第二個問題，我們故意沒有公開表示我們將來會開源所有東西，因為我們確實希望有空間來觀察安全形勢如何發展。如果我們認為未來我們確實跨越了某種關鍵的門檻，那麼將來開源它可能不是正確的做法。

But for our business model, at least, since we're not selling access to this stuff, it's a lot easier for us to share this with the community because it just makes our products better and other people's. And that, I think, is a really healthy dynamic for the industry.

但至少對於我們的商業模式而言，由於我們不會出售這些東西的使用權，因此我們可以更輕鬆地與社群分享這些內容，因為這只會讓我們的產品和其他人的產品變得更好。我認為這對產業來說是一個非常健康的動態。

And thanks, Youssef, for your second question, which I'll take. You asked about our Q3 revenue outlook and maybe also what that implies looking forward into Q4. So on the further acceleration that we've guided to in Q3, first of all, I'd just point out that Q3 '22 revenue declined 4.5% year-over-year. So we're really lapping a much weaker demand period a year ago, and we've certainly seen demand this year stabilize, and so it's really a much easier compare.

尤素夫，謝謝你的第二個問題，我會回答。您詢問了我們對第三季營收的展望，以及這對第四季的展望意味著什麼。因此，關於我們預計第三季的進一步加速，首先，我要指出的是，22 年第三季的營收年減了 4.5%。因此，我們實際上正在經歷一年前需求較弱的時期，我們確實看到今年的需求趨於穩定，因此這確實是一個更容易比較的結果。

We also are expecting that currency is going to flip to a 3-point tailwind from a 1-point headwind last quarter. And finally, as I've mentioned earlier on this call, we've been investing for a long time and I think executing really well across our core monetization work and continuing to grow engagement. And I think we're going to benefit from all of the investments that we've made in those historical and key priorities for us.

我們也預計，貨幣匯率將從上個季度的 1 個百分點的逆風轉為 3 個百分點的順風。最後，正如我之前在電話會議上提到的，我們已經投資了很長時間，我認為我們在核心貨幣化工作中表現非常出色，並且參與度不斷提高。我認為我們將從我們在這些歷史性和關鍵性優先事項上所做的所有投資中受益。

In terms of what this means for Q4, we're not sharing a Q4 revenue outlook yet. There are obviously some tailwinds to year-over-year growth in Q4. Again, the same point about a weaker compare applies. And at current rates, FX would be a larger tailwind in Q4 than we're expecting it to be in Q3.

至於這對第四季意味著什麼，我們尚未分享第四季的營收前景。第四季年增率顯然存在一些順風因素。再次，關於較弱的比較的相同觀點也適用。以目前的利率，外匯對第四季的推動作用將比我們預期的第三季更大。

But we've had sizable fluctuations in advertiser demand over the last year. It's been a pretty volatile period. So while we're seeing strong advertiser demand now, and that's certainly informing our outlook, it's harder to predict as we look further forward, and there are a wide range of possible outcomes in Q4.

但去年我們的廣告需求出現了相當大的波動。這是一個相當動盪的時期。因此，雖然我們現在看到強勁的廣告商需求，並且這無疑為我們的前景提供了信息，但展望未來，預測起來會更加困難，第四季度可能會出現各種各樣的結果。

Dave, we have time for 1 last question.

戴夫，我們還有時間回答最後一個問題。

That will come from the line of Brad Erickson with RBC Capital Markets.

這項提議將由加拿大皇家銀行資本市場 (RBC Capital Markets) 的布拉德·埃里克森 (Brad Erickson) 提出。

Just had a follow-up, I guess, on the CapEx. Just in that planning process you keep alluding to for the '24 CapEx outlook, I think a lot of your commentary here implies that new products, it sounds like, are going to be really central to instructing that number. Just curious how dependent it also is based on just views of revenue growth in '24 and the relationship there.

我想，只是對資本支出進行了跟進。就在您不斷提到的「24 年資本支出展望」的規劃過程中，我認為您在這裡的許多評論都暗示新產品聽起來將對指導這一數字至關重要。只是好奇它在多大程度上依賴對 24 年收入成長的看法以及那裡的關係。

Yes. Thank you, Brad. I can go ahead and take that one. So our growth in AI investments is really the thing that is driving the growth in our 2024 CapEx outlook. And I think there are a couple of components to that. There is both the core AI work, which powers our ranking and recommendation systems, which underpins both a lot of our content ranking and engagement growth as well as the monetization work.

是的。謝謝你，布拉德。我可以繼續拿那個。因此，我們在人工智慧方面的投資成長才是真正推動我們 2024 年資本支出前景成長的因素。我認為這有幾個因素。核心人工智慧工作為我們的排名和推薦系統提供支持，這為我們大量的內容排名和參與度成長以及貨幣化工作奠定了基礎。

And that's an area where we're able to measure the ROI of our investments there, and we feel good about the ROI of those investments. And we want to continue investing appropriately to drive revenue growth, at the same time being mindful of our desire to, over the long term, decrease capital intensity.

這是我們能夠衡量投資報酬率的領域，我們對這些投資的投資報酬率感到滿意。我們希望繼續進行適當的投資以推動收入成長，同時牢記我們希望長期降低資本密集度的願望。

There's also another component, which is the next-generation AI efforts that we've talked about around advanced research and gen AI, and that's a place where we're already standing up training clusters and inference capacity. But we don't know exactly what we'll need in 2024 since we don't have any at-scale deployments yet of consumer business-facing features. And the scale of the adoption of those products is ultimately going to inform how much capacity we need.

還有另一個組成部分，即我們圍繞高級研究和新一代人工智慧所討論的下一代人工智慧工作，而這正是我們已經建立訓練集群和推理能力的地方。但我們並不確切知道 2024 年我們需要什麼，因為我們還沒有大規模部署面向消費者業務的功能。這些產品的採用規模最終將決定我們需要多少容量。

We think these are both going to be compelling investment opportunities. And some of the AI capacity is fungible, so if we don't need -- end up needing some of the some of the capacity for our gen AI work, we'll be able to allocate it to our core AI work, supporting ads and engagement. But we're really still working on our '24 plans. We will have a clearer and more quantitative outlook as those plans shape up. But we are mindful of our intention to reduce the capital intensity of these investments over time.

我們認為這都將是極具吸引力的投資機會。而且部分人工智慧能力是可替代的，所以如果我們不需要——最終不需要部分人工智慧能力用於我們的新一代人工智慧工作，我們將能夠將其分配給我們的核心人工智慧工作，支援廣告和參與。但我們實際上仍在努力實現我們的“24”計劃。隨著這些計劃的逐漸成形，我們將擁有更清晰、更量化的前景。但我們牢記，我們的目標是隨著時間的推移降低這些投資的資本密集度。

Great. Thank you for joining us today. We appreciate your time, and we look forward to speaking again soon.

偉大的。感謝您今天加入我們。感謝您抽出時間，我們期待很快再次與您交談。

And this concludes today's conference call. Thank you for joining us. You may now disconnect your lines.

今天的電話會議到此結束。感謝您加入我們。現在您可以斷開線路了。