Meta Platforms Inc (META) 2025 Q2 法說會逐字稿

Good afternoon. My name is Krista, 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) And this call will be recorded. Thank you very much. Kenneth Dorell, Meta's Director of Investor Relations, you may begin.

午安.我叫克里斯塔，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）本次通話將會被錄音。非常感謝。您可以請 Meta 投資者關係總監 Kenneth Dorell 開始。

Thank you. Good afternoon, and welcome to Meta's second quarter 2025 earnings conference call. Joining me today are Mark Zuckerberg, CEO; and Susan Li, CFO. Our remarks today will include forward-looking statements, which are based on assumptions as of today. Actual results may differ materially as a result of various factors, including those set forth in today's earnings press release and in our quarterly report on Form 10-Q filed with the SEC.

謝謝。下午好，歡迎參加 Meta 2025 年第二季財報電話會議。今天與我一起出席的還有執行長馬克·祖克柏和財務長蘇珊·李。我們今天的評論將包括前瞻性陳述，這些陳述基於今天的假設。由於各種因素，實際結果可能存在重大差異，包括今天的收益新聞稿和我們向美國證券交易委員會提交的 10-Q 表季度報告中所述的因素。

We undertake no obligation to update any forward-looking statement. 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 earnings press release and an accompanying investor presentation are available on our website at investor.atmeta.com. And now I'd like to turn the call over to Mark.

我們不承擔更新任何前瞻性聲明的義務。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。收益新聞稿及隨附的投資者簡報可在我們的網站 investor.atmeta.com 上查閱。現在，我想將電話會議交給馬克。

All right. Thanks, Ken. Thanks, everyone, for joining today. We had another strong quarter with more than 3.4 billion people using at least one of our apps each day and strong engagement across the board. Our business continues to perform very well, which enables us to invest heavily in our AI efforts.

好的。謝謝，肯。感謝大家今天的參與。我們又度過了一個表現強勁的季度，每天有超過 34 億人使用至少一款我們的應用程序，並且全員參與度很高。我們的業務繼續表現良好，這使我們能夠在人工智慧方面投入大量資金。

Over the last few months, we've begun to see glimpses of our AI systems improving themselves. And the improvement is slow for now, but undeniable and developing super intelligence, which we define as AI that surpasses human intelligence in every way, we think, is now in sight.

在過去的幾個月裡，我們開始看到人工智慧系統自我改進的跡象。雖然目前的進步很慢，但不可否認的是，發展超級智慧（我們將其定義為在各個方面超越人類智慧的人工智慧）已經指日可待。

Meta's vision is to bring personal super intelligence to everyone, so that people can direct it towards what they value in their own lives. And we believe that this has the potential to begin an exciting new era of individual empowerment. A lot has been written about all the economic and scientific advances that Superintelligence can bring, and I'm extremely optimistic about this.

Meta 的願景是將個人超級智慧帶給每個人，以便人們能夠將其引導到自己生活中所重視的事情上。我們相信，這有可能開啟一個令人興奮的個人賦權新時代。關於超級智慧能夠帶來的經濟和科學進步已經有很多文章進行了論述，我對此非常樂觀。

But I think that if history is a guide, then an even more important role will be how Superintelligence empowers people to be more creative, develop culture and communities, connect with each other and lead more fulfilling lives.

但我認為，如果以歷史為鑑，那麼超級智慧將發揮更重要的作用，即如何賦予人們更多創造力，發展文化和社區，相互聯繫，過著更充實的生活。

To build this future, we've established Meta Superintelligence Labs, which includes our foundations, product, and FAIR teams as well as a new lab that is focused on developing the next generation of our models. We're making good progress towards Llama 4.1 and 4.2, and in parallel, we are also working on our next generation of models that will push the frontier in the next year or so.

為了建立這個未來，我們建立了元超級智慧實驗室，其中包括我們的基礎、產品和 FAIR 團隊以及專注於開發下一代模型的新實驗室。我們在 Llama 4.1 和 4.2 的開發上取得了良好的進展，同時，我們也致力於開發下一代模型，這些模型將在未來一年左右突破技術前沿。

We are building an elite, talent-dense team Alexandr Wang is leading the overall team. Nat Friedman is leading our AI product in Applied Research. And Shengjia Zhao is Chief Scientist for the new effort.

我們正在打造一支精英、人才密集的團隊，Alexandr Wang 領導整個團隊。Nat Friedman 正在領導我們的應用研究人工智慧產品。趙勝嘉是這項新計畫的首席科學家。

They are all incredibly talented leaders, and I'm excited to work closely with them. and the world-class group of AI researchers and infrastructure and data engineers that we're assembling. I spent a lot of time building this team this quarter.

他們都是非常有才華的領導者，我很高興能與他們密切合作。以及我們正在組建的世界級人工智慧研究人員、基礎設施和資料工程師團隊。本季我花了很多時間來組建這個團隊。

And the reason that so many people are excited to join is because Meta has all of the ingredients that are required to build leading models and deliver them to billions of people. The people who are joining us are going to have access to unparalleled compute as we build out several multi-gigawatt clusters.

而這麼多人興奮加入的原因是，Meta 擁有建立領先模型並將其交付給數十億人所需的所有要素。隨著我們建立多個多千兆瓦集群，加入我們的人們將獲得無與倫比的運算能力。

Our Prometheus cluster is coming online next year, and we think it's going to be the world's first gigawatt-plus cluster. We're also building out Hyperion, which we'll be able to scale up to 5 gigawatts over several years, and we have multiple more Titan clusters in development as well.

我們的 Prometheus 集群將於明年上線，我們認為它將成為世界上第一個千兆瓦以上的集群。我們還在建造 Hyperion，幾年內我們將能夠將其擴展到 5 千兆瓦，而且我們還在開發多個 Titan 叢集。

We are making all these investments because we have conviction that Superintelligence is going to improve every aspect of what we do. From a business perspective, I mentioned last quarter that there are five basic opportunities that we are pursuing, improved advertising, more engaging experiences, business messaging, Meta AI and AI devices. So I can go into a bit of detail on each.

我們進行所有這些投資是因為我們堅信超級智慧將改善我們所做的各個方面。從業務角度來看，我在上個季度提到，我們正在追求五個基本機會，即改進的廣告、更具吸引力的體驗、商業訊息、元人工智慧和人工智慧設備。因此我可以對每一個進行詳細的介紹。

On advertising, the strong performance this quarter is largely thanks to AI unlocking greater efficiency and gains across our ad system. This quarter, we expanded our new AI-powered recommendation model for ads to new surfaces and improved its performance by using more signals and longer context. It's driven roughly 5% more ad conversions on Instagram and 3% on Facebook.

在廣告方面，本季的強勁表現很大程度上得益於人工智慧提高了我們廣告系統的效率並帶來了收益。本季度，我們將新的人工智慧廣告推薦模型擴展到新的領域，並透過使用更多訊號和更長的上下文來提高其效能。它推動 Instagram 上的廣告轉換率提高了約 5%，Facebook 上的廣告轉換率提高了 3%。

We're also seeing good progress with AI for ad creative with a meaningful percent of our ad revenue now coming from campaigns using one of our generative AI features. This is going to be especially valuable for smaller advertisers with limited budgets, while agencies will continue the important work to help larger brands apply these tools strategically.

我們也看到人工智慧在廣告創意方面取得了良好進展，目前我們很大一部分廣告收入來自使用我們的生成人工智慧功能之一的廣告活動。這對於預算有限的小型廣告商來說尤其有價值，而代理商將繼續進行重要工作，幫助大品牌策略性地應用這些工具。

The second opportunity is more engaging experiences. AI is significantly improving our ability to show people content that they're going to find interesting and useful.

第二個機會是更具吸引力的體驗。人工智慧極大地提高了我們向人們展示他們感興趣且有用的內容的能力。

Advancements in our recommendation systems have improved quality so much that has led to a 5% increase in time spent on Facebook and 6% on Instagram, just this quarter. There is a lot of potential for content itself to get better too, we're seeing early progress with the launch of our AI video editing tools across Meta AI and our new Edits app. And there's a lot more to do here.

我們的推薦系統的進步極大地提高了質量，僅在本季度，用戶在 Facebook 上花費的時間就增加了 5%，在 Instagram 上花費的時間增加了 6%。內容本身也有很大潛力可以變得更好，隨著我們在 Meta AI 和新 Edits 應用程式中推出 AI 影片編輯工具，我們看到了早期進展。這裡還有很多事情要做。

The third opportunity is business messaging I've talked before about how I believe every business will soon have a business AI, just like they have an e-mail address social media account and website. We are starting to see some product market fit in a number of countries where we're testing these agents, and we're integrating these business AIs into ads on Facebook and Instagram as well as directly into e-commerce websites.

第三個機會是商業訊息傳遞，我之前談過，我相信每個企業很快都會擁有一個商業人工智慧，就像他們擁有電子郵件地址、社群媒體帳號和網站一樣。我們開始在測試這些代理商的多個國家看到一些產品市場契合度，並且我們正在將這些商業人工智慧整合到 Facebook 和 Instagram 上的廣告以及直接整合到電子商務網站中。

The fourth opportunity is Meta AI. Its reach is already quite impressive with more than 1 billion monthly actives. Our focus is now deepening the experience in making Meta AI the leading personal AI. As we continue improving our models, we see engagement grow. So our next generation of models is going to continue to really help here.

第四個機會是Meta AI。它的覆蓋範圍已經相當令人印象深刻，每月活躍用戶超過 10 億。我們現在的重點是深化經驗，使 Meta AI 成為領先的個人人工智慧。隨著我們不斷改進模型，我們看到參與度不斷增長。因此，我們的下一代模型將繼續在這裡提供真正的幫助。

And the fifth opportunity is AI devices. We continue to see strong momentum with our Ray-Ban Meta glasses with sales accelerating.

第五個機會是人工智慧設備。我們繼續看到雷朋 Meta 眼鏡的強勁成長勢頭，銷量正在加速成長。

We are also launching new performance AI glasses with the Oakley Meta HSTNs, they have longer battery life, higher resolution camera, and are designed for sports. The percent of people using Meta AI is growing, and we are seeing new users AI retention increase too, which is a good sign for that continued use.

我們還推出了採用 Oakley Meta HSTN 的新型高性能 AI 眼鏡，它們具有更長的電池壽命、更高解析度的攝像頭，並且專為運動而設計。使用 Meta AI 的人數比例正在成長，我們看到新用戶 AI 的保留率也在增加，這對於持續使用來說是一個好兆頭。

I think that AI glasses are going to be the main way that we integrate Superintelligence into our day-to-day lives. So it's important to have all of these different styles and products that appeal to different people in different settings. Finally, we're seeing people continue to spend more time with our Quest ecosystem and the community continues to grow steadily.

我認為人工智慧眼鏡將成為我們將超級智慧融入日常生活的主要方式。因此，擁有這些不同的風格和產品來吸引不同環境下的不同人群是很重要的。最後，我們看到人們繼續在我們的 Quest 生態系統上花費更多時間，而社群繼續穩步發展。

We launched the Meta Quest 3S Xbox addition last month, and we're seeing record interest in cloud gaming. And beyond gaming, we continue to see a broader set of use cases with media and web browsing contributing a significant portion of engagement. We're going to have more to share on all of this, especially the Reality Labs work at Connect on September 17. So I encourage you all to tune into that.

我們上個月推出了 Meta Quest 3S Xbox 附加產品，我們看到人們對雲端遊戲的興趣空前高漲。除了遊戲之外，我們還看到了更廣泛的使用案例，其中媒體和網頁瀏覽貢獻了相當大的一部分參與度。我們將分享更多關於這些內容的信息，特別是 9 月 17 日在 Connect 上展示的 Reality Labs 工作。所以我鼓勵大家關注這一點。

Overall, this has been a busy quarter. Strong business performance and real momentum and assembling both the talent and the compute that we need to build personal super intelligence for everyone. I am very grateful to our teams who are working hard to deliver all of this, and thanks to all of you for being on this journey with us. And now here is Susan.

總體而言，這是一個繁忙的季度。強勁的業務表現和真正的發展勢頭，匯集了我們為每個人建立個人超級智慧所需的人才和運算能力。我非常感謝我們的團隊為實現這一切而努力工作，也感謝大家與我們一起踏上這段旅程。現在蘇珊就在這裡。

Thanks, Mark, and good afternoon, everyone. Let's begin with our consolidated results. All comparisons are on a year-over-year basis unless otherwise noted. Q2 total revenue was $47.5 billion, up 22% on both a reported and constant currency basis. Q2 total expenses were $27.1 billion, up 12% compared to last year.

謝謝，馬克，大家下午好。讓我們從綜合結果開始。除非另有說明，所有比較均以同比計算。第二季總營收為 475 億美元，按報告和固定匯率計算均成長 22%。第二季總支出為 271 億美元，比去年同期成長 12%。

In terms of the specific line items, cost of revenue increased 16% and driven mostly by higher infrastructure costs and payments to partners, partially offset by a benefit from the previously announced extension of several useful lives. R&D increased 23%, mostly due to higher employee compensation and infrastructure costs.

就具體項目而言，收入成本增加了 16%，主要原因是基礎設施成本和支付給合作夥伴的款項增加，但部分被先前宣布的延長幾項使用壽命帶來的好處所抵消。研發費用增加了 23%，主要原因是員工薪資和基礎設施成本增加。

Marketing and sales increased 9% primarily due to an increase in professional services related to our ongoing platform integrity efforts as well as marketing costs, partially offset by lower employee compensation. G&A decreased 27%, driven mostly by lower legal-related costs.

行銷和銷售額成長了 9%，主要原因是與我們持續的平台完整性努力相關的專業服務以及行銷成本的增加，但員工薪酬的降低部分抵消了這一增長。一般及行政費用下降了 27%，主要原因是法律相關成本下降。

We ended Q2 with over 75,900 employees, down 1% quarter over quarter, as the vast majority of the employees impacted by performance-related reductions earlier this year were no longer captured in our headcount.

我們第二季結束時員工人數超過 75,900 人，較上季下降 1%，因為今年稍早受到績效相關裁員影響的絕大多數員工已不再計入我們的員工總數。

This was partially offset by continued hiring in priority areas of monetization, infrastructure, Reality Labs, AI, as well as regulation and compliance. Second quarter operating income was $20.4 billion, representing a 43% operating margin.

這在一定程度上被貨幣化、基礎設施、現實實驗室、人工智慧以及監管和合規等優先領域的持續招聘所抵消。第二季營業收入為 204 億美元，營業利益率為 43%。

Our tax rate for the quarter was 11%, which reflects excess tax benefits from share-based compensation due to the increase in our share price versus prior periods. Net income was $18.3 billion or $7.14 per share. Capital expenditures, including principal payments on finance leases were $17 billion, driven by investments in servers, data centers, and network infrastructure.

我們本季的稅率為 11%，這反映了由於我們的股價與前期相比上漲而產生的基於股票的薪酬的超額稅收優惠。淨收入為 183 億美元，即每股 7.14 美元。資本支出（包括融資租賃的本金支付）為 170 億美元，主要由對伺服器、資料中心和網路基礎設施的投資推動。

Free cash flow was $8.5 billion. We repurchased $9.8 billion of our Class A common stock and paid $1.3 billion in dividends to shareholders. We also made $15.1 billion in non-marketable equity investments in the second quarter which includes our minority investment in Scale AI, along with other investment activities. We ended the quarter with $47.1 billion in cash and marketable securities and $28.8 billion in debt.

自由現金流為85億美元。我們回購了價值 98 億美元的 A 類普通股，並向股東支付了 13 億美元的股息。我們還在第二季進行了 151 億美元的非流通股權投資，其中包括對 Scale AI 的少數股權投資以及其他投資活動。本季結束時，我們的現金和有價證券為 471 億美元，債務為 288 億美元。

Moving now to our segment results. I'll begin with our Family of Apps segment. Our community across the Family of Apps continues to grow, and we estimate more than 3.4 billion people used at least one of our Family of Apps on a daily basis in June. Q2 total Family of Apps revenue was $47.1 billion, up 22% year over year. Q2 Family of Apps ad revenue was $46.6 billion up 21% or 22% on a constant currency basis. Within that revenue, the online commerce vertical was the largest contributor to year-over-year growth.

現在轉到我們的分部業績。我將從我們的應用程式系列部分開始。我們的應用程式系列社群正在不斷壯大，我們估計，6 月每天有超過 34 億人至少使用我們的一款應用程式系列。第二季應用程式系列總營收為 471 億美元，年增 22%。第二季應用程式系列廣告收入為 466 億美元，以固定匯率計算成長 21% 或 22%。在該收入中，線上商務垂直領域是年成長的最大貢獻者。

On a user geography basis, ad revenue growth was strongest in Europe and rest of world at 24% and 23%, respectively. North America and Asia Pacific grew 21% and 18%. In Q2, the total number of ad impressions served across our services increased 11%, with growth mainly driven by Asia Pacific.

從用戶地理來看，歐洲和世界其他地區的廣告收入成長最為強勁，分別為 24% 和 23%。北美和亞太地區分別成長21%和18%。第二季度，我們各項服務的廣告展示總數增加了 11%，成長主要來自亞太地區。

Impression growth accelerated across all regions due primarily to engagement tailwinds on both Facebook and Instagram and to a lesser extent, ad load optimizations on Facebook. The average price per ad increased 9%, benefiting from increased advertiser demand, largely driven by improved ad performance.

所有地區的展示次數成長均加速，主要得益於 Facebook 和 Instagram 的參與度提升，以及 Facebook 的廣告負載優化（在較小程度上）。受廣告主需求增加（主要受廣告效果改善推動）推動，每則廣告的平均價格上漲了 9%。

Pricing growth slowed modestly from the first quarter due to the accelerated impression growth in Q2. Family of Apps other revenue was $583 million, up 50% driven by WhatsApp paid messaging revenue growth as well as Meta Verified subscriptions. We continue to direct the majority of our investments toward the development and operation of our Family of Apps.

由於第二季印象成長加速，定價成長較第一季略有放緩。應用程式系列的其他收入為 5.83 億美元，成長 50%，主要得益於 WhatsApp 付費訊息收入成長以及 Meta Verified 訂閱。我們繼續將大部分投資用於應用程式系列的開發和營運。

In Q2, Family of Apps expenses were $22.2 billion, representing 82% of our overall expenses. Family of Apps expenses were up 14% and mainly due to growth in employee compensation and infrastructure costs, partially offset by lower legal-related costs.

第二季度，應用程式系列的支出為 222 億美元，占我們整體支出的 82%。應用程式系列費用上漲了 14%，主要原因是員工薪資和基礎設施成本的成長，但法律相關成本的降低部分抵消了這一成長。

Family of Apps operating income was $25 billion, representing a 53% operating margin. Within our Reality Labs segment, Q2 revenue was $370 million up 5% year over year due to increased sales of AI glasses, partially offset by lower Quest sales. Reality Labs expenses were $4.9 billion, up 1% year-over-year, driven by higher non-headcount-related technology development costs. Reality Labs operating loss was $4.5 billion.

應用程式系列營運收入為 250 億美元，營運利潤率為 53%。在我們的現實實驗室部門，第二季度營收為 3.7 億美元，年增 5%，這得益於 AI 眼鏡銷量的成長，但 Quest 銷量的下降部分抵消了這一增長。Reality Labs 的支出為 49 億美元，較去年同期成長 1%，主要原因是非員工相關技術開發成本的增加。Reality Labs 的營運虧損為 45 億美元。

Turning now to the business outlook. There are two 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。推動我們收入表現的主要因素有兩個：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, daily actives continue to grow across Facebook, Instagram, and WhatsApp as we make additional improvements to our recommendation systems and product experiences. We continue to see momentum with video engagement, in particular. In Q2, Instagram video time was up more than 20% year over year globally. We're seeing strong traction on Facebook as well, particularly in the US, where video time spent similarly expanded more than 20% year-over-year.

首先，隨著我們對推薦系統和產品體驗的進一步改進，Facebook、Instagram 和 WhatsApp 的每日活躍用戶持續成長。我們繼續看到視訊參與度的強勁成長勢頭。第二季度，Instagram 全球影片使用時間年增超過 20%。我們也看到了 Facebook 的強勁成長勢頭，尤其是在美國，人們觀看影片的時間年增了 20% 以上。

These gains have been enabled by ongoing optimizations to our ranking systems to better identify the most relevant content to show. We expect to deliver additional improvements throughout the year as we further scale up our models and make recommendations more adaptive to a person's interest within their session. Another emphasis of our recommendations work is promoting original content. On Instagram, over two-thirds of recommended content in the US now comes from original posts.

這些收益是透過我們對排名系統的持續優化來實現的，以便更好地識別要顯示的最相關內容。隨著我們進一步擴大模型規模，並使推薦更適應個人在會話中的興趣，我們期望全年能夠實現更多改進。我們的推薦工作的另一個重點是推廣原創內容。在 Instagram 上，美國目前超過三分之二的推薦內容來自原創貼文。

In the second half, we'll be focused on further increasing the freshness of original posts, so the right audiences can discover original content from creators soon after it is posted. We are also making good progress on our longer-term ranking innovations that we expect will provide the next leg of improvements over the coming years.

下半年，我們將致力於進一步提高原創貼文的新鮮度，以便合適的受眾能夠在創作者發布原創內容後儘快發現它。我們的長期排名創新也取得了良好進展，預計這將在未來幾年帶來下一階段的改善。

Our research efforts to develop cross-surface foundation recommendation models continue to progress. We are also seeing promising results from using LLMs in Threads recommendation systems. The incorporation of LLM are now driving a meaningful share of the ranking related time spent gains on Threads.

我們開發跨表面基礎推薦模型的研究工作正在不斷取得進展。我們也看到了在 Threads 推薦系統中使用 LLM 所帶來的有希望的結果。現在，法學碩士 (LLM) 的加入正在推動 Threads 上與排名相關的時間花費大幅增加。

We're now exploring how to extend the use of LLMs in recommendation systems to our other apps. We're leveraging Llama and several other back-end processes as well, including actioning bug reports, so we can identify and resolve recurring issues more quickly and efficiently.

我們現在正在探索如何將推薦系統中的 LLM 的使用擴展到我們的其他應用程式。我們也利用 Llama 和其他幾個後端流程，包括處理錯誤報告，以便我們能夠更快、更有效地識別和解決重複出現的問題。

This has resulted in top line bug reports in the US and Canada in Facebook feed and notifications dropping by roughly 30% over the past 10 months. The primary way we're using Llama in our apps today is to power Meta AI which is now available in over 200 countries and territories.

這導緻美國和加拿大 Facebook 動態消息和通知中的頂級錯誤報告在過去 10 個月內下降了約 30%。我們目前在應用程式中使用 Llama 的主要方式是為 Meta AI 提供支持，該技術目前已在 200 多個國家和地區推出。

WhatsApp continues to be the largest driver of queries as people message Meta AI directly for tasks, such as information gathering, homework assistance, and generating images. Outside of WhatsApp, we're seeing Meta AI become an increasingly valuable complement to our content discovery engines.

WhatsApp 繼續成為查詢的最大驅動力，因為人們直接向 Meta AI 發送訊息來完成資訊收集、家庭作業輔導和生成圖像等任務。在 WhatsApp 之外，我們看到 Meta AI 正在成為我們內容髮現引擎越來越有價值的補充。

Meta AI usage on Facebook is expanding as people use it to ask about post they see in feed, and find content across our platform in search. Another way we expect Meta AI will help with content discovery is through the automatic translation and dubbing of foreign language content into the audience's local language. We'll have more to share on our efforts there later this year.

隨著人們使用 Meta AI 來詢問他們在資訊流中看到的帖子，並在我們的平台上透過搜尋來尋找內容，Facebook 上的 Meta AI 的使用範圍正在不斷擴大。我們期望 Meta AI 能夠幫助發現內容的另一種方式是將外語內容自動翻譯並配音為觀眾的當地語言。今年晚些時候，我們將分享更多有關我們在那裡所做努力的資訊。

Moving to Reality Labs. The growth of Ray-Ban Meta sales accelerated in Q2, with demand still outstripping supply for the most popular SKUs despite increases to our production earlier this year. We're working to ramp supply to better meet consumer demand later this year. Now to the second driver of our revenue performance, increasing monetization efficiency. The first part of this work is optimizing the level of ads within organic engagement.

轉向現實實驗室。Ray-Ban Meta 的銷售成長在第二季加速，儘管我們今年稍早增加了產量，但最受歡迎的 SKU 的需求仍然超過供應。我們正在努力增加供應，以便在今年稍後更好地滿足消費者的需求。現在談談我們收入表現的第二個驅動力，即提高貨幣化效率。這項工作的第一部分是優化有機參與中的廣告水平。

We continue to optimize ad supply across each surface to better deliver ads at the time and place they are most relevant to people. In Q2, we also began introducing ads within feed on Threads and the updates tab of WhatsApp, which is a separate space away from people's chats.

我們將繼續優化各個平台上的廣告供應，以便在與人們最相關的時間和地點更好地投放廣告。在第二季度，我們也開始在 Threads 的動態流和 WhatsApp 的更新標籤中引入廣告，這是一個獨立於人們聊天空間的單獨空間。

As of May, advertisers globally can now run video and image ads to Threads users in most countries, including the United States. While ad supply remains low and Threads is not expected to be a meaningful contributor to overall impression growth in the near term, we are optimistic about the longer-term opportunity with Threads as the community and engagement grow and monetization scales.

自 5 月起，全球廣告商現在可以向包括美國在內的大多數國家的 Threads 用戶投放影片和圖片廣告。儘管廣告供應量仍然很低，而且預計 Threads 在短期內不會對整體印象增長做出有意義的貢獻，但隨著社區和參與度的增長以及貨幣化規模的擴大，我們對 Threads 的長期機會持樂觀態度。

On WhatsApp, we are rolling out ads and status and channels, along with channel subscriptions and the updates tab to help businesses reach the more than 1.5 billion daily actives who visit that part of the app.

在 WhatsApp 上，我們推出了廣告、狀態和頻道，以及頻道訂閱和更新標籤，以幫助企業接觸到每天訪問該應用程式該部分的超過 15 億活躍用戶。

We expect the introduction of ads and status will be gradual over the course of this year and next, with low levels of expected ad supply initially. We also expect WhatsApp ads and status to earn a lower average price than Facebook or Instagram ads for the foreseeable future, due in part towards WhatsApp skewed toward lower monetizing markets, and more limited information that can be used for targeting.

我們預計，今年和明年廣告和地位的推出將逐步進行，預計初期廣告供應量將較低。我們也預計，在可預見的未來，WhatsApp 廣告和狀態的平均價格將低於 Facebook 或 Instagram 廣告，部分原因是 WhatsApp 偏向於貨幣化程度較低的市場，並且可用於定位的資訊更有限。

Given this, we do not expect ads and status to be a meaningful contributor to total impressions or revenue growth for the next few years.

有鑑於此，我們預計未來幾年廣告和狀態不會對總印像或收入成長做出有意義的貢獻。

The second part of increasing monetization efficiency is improving marketing performance. There are three areas of this work that I'll focus on today, improving our ad systems advancing our ad products, including by building tools that assist in ads creation and evolving our ads platform to drive results that are optimized for each business' objectives.

提升變現效率的第二部分是提升行銷績效。今天我將重點討論這項工作的三個面向：改進我們的廣告系統，推進我們的廣告產品，包括建立有助於廣告創作的工具，以及改進我們的廣告平台，以推動針對每個業務目標進行最佳化的結果。

First is our ad systems where we're innovating in both the ads retrieval and ranking stages to serve more relevant ads to people. A lot of this work involves us continuing to advance the modeling innovations we've introduced previously, while expanding their adoption across our platform.

首先是我們的廣告系統，我們在廣告檢索和排名階段都進行了創新，以便為人們提供更相關的廣告。這項工作的大部分內容涉及我們繼續推動我們先前引入的建模創新，同時擴大它們在我們平台上的採用。

The Andromeda model architecture we began introducing in the second half of 2024 powers the ads retrieval stage of our ad system, where we select the few thousand most relevant ads from tens of millions of potential candidates.

我們從 2024 年下半年開始引入的 Andromeda 模型架構為我們的廣告系統的廣告檢索階段提供支持，我們從數千萬個潛在候選廣告中選出數千個最相關的廣告。

In Q2, we made enhancements to Andromeda that enabled it to select more relevant and more personalized ads candidates while also expanding coverage to Facebook Reels. These improvements have driven nearly 4% higher conversions on Facebook Mobile Feed and Reels.

在第二季度，我們對 Andromeda 進行了改進，使其能夠選擇更相關、更個人化的廣告候選對象，同時擴大對 Facebook Reels 的覆蓋範圍。這些改進使 Facebook 行動 Feed 和 Reels 的轉換率提高了近 4%。

Our new Generative Ads Recommendation System, or GEM, powers the ranking stage of our ad system, which is the part of the process after ads retrieval where we determine which ads to show someone from candidates suggested by our retrieval engine. In Q2, we improved the performance of GEM by further scaling our training capacity and adding organic and ads engagement data on Instagram.

我們的新生成廣告推薦系統 (GEM) 為我們的廣告系統的排名階段提供支持，該階段是廣告檢索之後的流程的一部分，我們會從檢索引擎建議的候選人中確定向某人展示哪些廣告。在第二季度，我們透過進一步擴大培訓能力並在 Instagram 上添加自然和廣告參與數據來提高 GEM 的表現。

We also incorporated new advanced sequence modeling techniques that helped us double the length of event sequences we use, enabling our systems to consider a longer history of the content or ads that a person has engaged with in order to provide better ad selections.

我們還採用了新的高級序列建模技術，幫助我們將使用的事件序列的長度增加了一倍，使我們的系統能夠考慮用戶參與的內容或廣告的更長歷史記錄，從而提供更好的廣告選擇。

The combination of these improvements increased ad conversions by approximately 5% on Instagram and 3% on Facebook Feed and Reels in Q2.

這些改進的結合使得第二季 Instagram 上的廣告轉換率提高了約 5%，Facebook Feed 和 Reels 上的廣告轉換率提高了 3%。

Finally, we expanded coverage of our Lattice model architecture in Q2. We first began deploying Lattice in 2023 with our later-stage ads ranking efforts, allowing us to run significantly larger models that generalize learnings across objectives and surfaces in place of numerous smaller ads models that have historically been optimized for individual objectives and surfaces. In April, we began deploying Lattice to earlier-stage ads ranking models as well.

最後，我們在第二季擴大了 Lattice 模型架構的覆蓋範圍。我們在 2023 年首次在後期廣告排名工作中部署 Lattice，這使我們能夠運行更大的模型，這些模型可以概括跨目標和介面的學習成果，取代歷史上針對單一目標和介面進行最佳化的眾多較小的廣告模型。從四月開始，我們也開始將 Lattice 部署到早期的廣告排名模型。

This is leading not only to greater capacity and engineering efficiency but also improved performance, with the recent Lattice deployments driving a nearly 4% increase in ad conversions across Facebook Feed and Reels in Q2. Next, ad products. Here, we're seeing strong momentum with our Advantage+ suite of AI-powered solutions.

這不僅帶來了更大的容量和工程效率，還提高了效能，最近的 Lattice 部署推動第二季度 Facebook Feed 和 Reels 上的廣告轉換率增長了近 4%。接下來是廣告產品。在這裡，我們看到了 Advantage+ 人工智慧解決方案套件的強勁發展勢頭。

In Q2, we completed the rollout of our streamlined campaign creation flow for Advantage+ sales and app campaigns, which makes it easier for advertisers to realize the performance benefits from Advantage+ by having it turned on at the beginning. We've seen lifts in advertiser adoption of sales and app campaigns since we've expanded availability, and are working to complete the rollout for leads campaigns in the coming months.

在第二季度，我們完成了 Advantage+ 銷售和應用廣告系列的簡化廣告系列創建流程的推出，這使得廣告商能夠更輕鬆地透過在一開始就啟用 Advantage+ 來實現其效能優勢。自從我們擴大了可用性以來，我們看到廣告主對銷售和應用程式廣告系列的採用率有所提升，並且我們正在努力在未來幾個月內完成潛在客戶廣告系列的推出。

Within our Advantage+ Creative Suite, adoption of GenAI and creative tools continues to broaden. Nearly 2 million advertisers are now using our video generation features, image animation and video expansion, and we're seeing strong results with our text generation tools as we continue to add new features. In Q2, we started testing AI-powered translation so that advertisers can automatically translate the caption of their ads to 10 different languages.

在我們的 Advantage+ Creative Suite 中，GenAI 和創意工具的採用範圍不斷擴大。目前，已有近 200 萬廣告商正在使用我們的影片生成功能、圖像動畫和影片擴展，並且隨著我們不斷添加新功能，我們的文字生成工具也取得了顯著的效果。在第二季度，我們開始測試人工智慧翻譯，以便廣告商可以自動將其廣告標題翻譯成 10 種不同的語言。

While it's early, we have seen promising performance lifts in our prelaunch tests. We're also continuing to see strong adoption of image expansion among small- and medium-sized advertisers, which speaks to how these tools help businesses who have fewer resources to develop creative.

雖然還處於早期階段，但我們在發布前測試中已經看到了令人鼓舞的效能提升。我們也繼續看到中小型廣告商大力採用圖像擴展，這說明這些工具如何幫助資源較少的企業開發創意。

With larger advertisers, we expect agencies will continue to be valuable partners in helping apply these new tools to drive performance. Outside of Advantage+, we're seeing good momentum in business messaging, particularly in the US, where click to message revenue grew more than 40% year over year in Q2.

對於規模較大的廣告商，我們預期代理商將繼續成為寶貴的合作夥伴，幫助他們應用這些新工具來提高績效。除了 Advantage+ 之外，我們還看到商業訊息傳遞領域發展勢頭良好，尤其是在美國，第二季點擊訊息收入較去年同期成長超過 40%。

The strong US growth is benefiting from a ramp in adoption of our website to message ads, which drive people to a business's website for more information before choosing to launch a chat with the business in one of our messaging apps.

美國市場的強勁成長得益於我們網站訊息廣告的普及，這種廣告可以吸引人們訪問企業網站獲取更多信息，然後再選擇透過我們的某個訊息應用程式與該企業進行聊天。

Finally, we continue to evolve our ads platform to drive results that are optimized for each business' objectives and the way they measure results. In Q2, we completed the global rollout of our incremental attribution feature, which is the only product on the market that optimizes for and reports on incremental conversions, which are conversions that would not have happened without a person seeing the ad.

最後，我們將繼續改進我們的廣告平台，以推動針對每個企業的目標和衡量結果的方式進行最佳化的結果。在第二季度，我們完成了增量歸因功能的全球推廣，這是市場上唯一針對增量轉換進行最佳化和報告的產品，增量轉換是如果沒有人看到廣告就不會發生的轉換。

We also launched omnichannel ads globally in Q2 and which enable advertisers to optimize for incremental sales, both in store and online with just one campaign. In tests, advertisers using omnichannel ads have seen a median 15% reduction in total cost per purchase compared to website-only optimization.

我們也在第二季在全球推出了全通路廣告，使廣告主只需一個廣告活動即可優化店內和線上的增量銷售。在測試中，使用全通路廣告的廣告主與僅進行網站優化相比，每次購買的總成本平均降低了 15%。

Next, I would like to discuss our approach to capital allocation. Our primary focus remains investing capital back into the business with infrastructure and talent being our top priorities.

接下來我想討論一下我們的資本配置方法。我們的首要重點仍然是將資本重新投入到業務中，而基礎設施和人才是我們的首要任務。

I'll start with hiring. Our approach to adding head count continues to be targeted at the company's highest priority areas. We expect talent additions across all of our priority areas will continue to drive overall head count growth through this year in 2026, while head count growth in our other functions remains constrained. Within AI, we've had a particular emphasis on recruiting leading talent within the industry, as we build out Meta Superintelligence Labs to accelerate our AI model development and product initiatives.

我先從招募開始。我們增加員工人數的方法仍然是針對公司最優先的領域。我們預計，到 2026 年，我們所有重點領域的人才增加將繼續推動整體員工數量的成長，而其他職能部門的員工數量成長仍然受到限制。在人工智慧領域，我們特別重視招募產業內的領先人才，同時我們建立了元超級智慧實驗室來加速我們的人工智慧模型開發和產品計畫。

Next, infrastructure. We expect having sufficient compute capacity will be central to realizing many of the largest opportunities in front of us over the coming years. We continue to see very compelling returns from our AI capacity investments in our core ads and organic engagement initiatives and expect to continue investing significantly there in 2026.

接下來是基礎建設。我們預計，擁有足夠的運算能力對於在未來幾年實現我們面臨的許多最大機會至關重要。我們繼續看到我們在核心廣告和有機參與計劃中的人工智慧能力投資獲得了非常可觀的回報，並預計在 2026 年將繼續在該領域進行大量投資。

We also expect that developing leading AI infrastructure will be a core advantage in developing the best AI models and product experiences. So we expect to ramp our investments significantly in 2026 to support that work.

我們也預計，開發領先的人工智慧基礎設施將成為開發最佳人工智慧模型和產品體驗的核心優勢。因此，我們預計在 2026 年將大幅增加投資以支持這項工作。

Moving to our financial outlook. We expect third quarter 2025 total revenue to be in the range of $47.5 billion to $50.5 billion. Our guidance assumes foreign currency is an approximately 1% tailwind to year-over-year total revenue growth, based on current exchange rates.

轉向我們的財務展望。我們預計 2025 年第三季總營收將在 475 億美元至 505 億美元之間。根據當前匯率，我們的指導假設是外幣將為年比總收入成長帶來約 1% 的順風。

While we are not providing an outlook for fourth quarter revenue, we would expect our year-over-year growth rate in the fourth quarter of 2025 to be slower than the third quarter as we lap a period of stronger growth in the fourth quarter of 2024.

雖然我們沒有提供第四季營收的前景，但我們預計 2025 年第四季的同比成長率將低於第三季度，因為我們將在 2024 年第四季迎來更強勁的成長期。

Turning now to the expense outlook. We expect full year 2025 total expenses to be in the range of $114 billion to $118 billion, narrowed from our prior outlook of $113 billion to $118 billion and reflecting a growth rate of 20% to 24% year over year. While we're still very early in planning for next year, there are a few factors we expect will provide meaningful upward pressure on our 2026 total expense growth rate.

現在來談談費用前景。我們預計 2025 年全年總支出將在 1,140 億美元至 1,180 億美元之間，低於先前預測的 1,130 億美元至 1,180 億美元，年成長率為 20% 至 24%。雖然我們對明年的規劃仍處於早期階段，但我們預計有幾個因素將對我們 2026 年的總支出成長率產生有意義的上行壓力。

The largest single driver of growth will be infrastructure costs, driven by a sharp acceleration in depreciation expense growth and higher operating costs as we continue to scale up our infrastructure fleet. Aside from infrastructure, we expect the second largest driver of growth to be employee compensation as we add technical talent in priority areas and recognize a full year of compensation expenses for employees hired throughout 2025.

成長的最大單一驅動力將是基礎設施成本，隨著我們繼續擴大基礎設施規模，折舊費用成長將急劇加速，營運成本也將上升。除基礎設施外，我們預計第二大成長動力將是員工薪酬，因為我們在重點領域增加了技術人才，並為 2025 年全年僱用的員工確認了全年的薪酬支出。

We expect these factors will result in a 2026 year-over-year expense growth rate that is above the 2025 expense growth rate.

我們預計這些因素將導致 2026 年的年比費用成長率高於 2025 年的費用成長率。

Turning now to the CapEx outlook. We currently expect 2025 capital expenditures, including principal payments on finance leases, to be in the range of $66 billion to $72 billion, narrowed from our prior outlook of $64 billion to $72 billion and up approximately $30 billion year over year at the midpoint.

現在來談談資本支出前景。我們目前預計 2025 年資本支出（包括融資租賃的本金支付）將在 660 億美元至 720 億美元之間，低於我們先前預測的 640 億美元至 720 億美元，中間值較上年同期增加約 300 億美元。

While the infrastructure planning process remains highly dynamic, we currently expect another year of similarly significant CapEx dollar growth in 2026 as we continue aggressively pursuing opportunities to bring additional capacity online to meet the needs of our AI efforts and business operations.

雖然基礎設施規劃過程仍然高度動態，但我們目前預計 2026 年資本支出將再次出現同樣顯著的成長，因為我們將繼續積極尋求機會，增加線上容量，以滿足我們的人工智慧工作和業務運營的需求。

On to tax. With the enactment of the new US tax law, we anticipate a reduction in our US federal cash tax for the remainder of the current year and future years.

談稅。隨著美國新稅法的頒布，我們預計今年剩餘時間和未來幾年美國聯邦現金稅將會減少。

There are several alternative ways of implementing the provisions of the act, which we are currently evaluating. While we estimate that the 2025 tax rate will be higher than our Q2 tax rate, we cannot quantify the magnitude at this time. In addition, we continue to monitor an active regulatory landscape, including the increasing legal and regulatory headwinds in the EU that could significantly impact our business and our financial results.

有幾種實施該法案規定的替代方法，我們目前正在評估。雖然我們估計 2025 年的稅率將高於我們的第二季稅率，但目前我們無法量化其幅度。此外，我們繼續關注活躍的監管環境，包括歐盟日益增加的法律和監管阻力，這些阻力可能會對我們的業務和財務表現產生重大影響。

For example, we continue to engage with the European Commission on our Less Personalized Ads offering or LPA, which we introduced in November 2024 and based on feedback from the European Commission in connection with the DMA. As the commission provides further feedback on LPA, we cannot rule out that it may seek to impose further modifications to it that would result in a materially worse user and advertiser experience.

例如，我們繼續與歐盟委員會就我們的低個人化廣告產品或 LPA 進行合作，該產品是我們於 2024 年 11 月推出的，並且基於歐盟委員會對 DMA 的反饋。隨著委員會對 LPA 提出進一步的回饋，我們不能排除其可能會對其進行進一步的修改，從而導致用戶和廣告商的體驗大幅惡化。

This could have a significant negative impact on our European revenue as early as later this quarter. We have appealed the European Commission's DMA decision, but any modifications to our model may be imposed during the appeal process.

這可能會對我們本季後期的歐洲收入產生重大負面影響。我們已經對歐盟委員會的 DMA 決定提出上訴，但在上訴過程中可能會對我們的模型進行任何修改。

In closing, this was another strong quarter for our business as our investments in infrastructure and technical talent continue to improve core ads performance and engagement on our platforms. We expect the significant investments we're making now will allow us to continue leveraging advances in AI to extend those gains and unlock a new set of opportunities in the years to come. With that, Krista, let's open up the call for questions.

總而言之，這對我們的業務來說又是一個強勁的季度，因為我們對基礎設施和技術人才的投資繼續改善了我們平台上的核心廣告效果和參與度。我們預計，我們現在所做的大量投資將使我們能夠繼續利用人工智慧的進步來擴大這些收益，並在未來幾年開啟一系列新的機會。克里斯塔，讓我們開始提問吧。

(Operator Instructions)

（操作員指示）

Eric Sheridan, Goldman Sachs.

高盛的艾瑞克·謝裡丹。

Mark, when you think about where the AI parts of your business have been evolving over the last three to six months, I wanted to know what your key learnings were as you went deep into that strategy that informed some of the shifts in both talent acquisition and compute coupled with some of the blogs you put out recently in terms of how that strategy might have evolved based on those key learnings.

馬克，當您思考過去三到六個月內您公司的人工智能部分是如何發展的時，我想知道您在深入研究該戰略時的主要收穫是什麼，這些戰略為人才招聘和計算方面的一些轉變提供了信息，同時您最近發布了一些博客，介紹了該戰略如何基於這些主要收穫而發展。

And Susan, building on Mark's comments on scaling talent and compute, I want to know if you go a little bit deeper and how we should be thinking about those two components driving some of the commentary you've given around OpEx and CapEx over the next 12 to 18 months.

蘇珊，基於馬克關於擴大人才和計算能力的評論，我想知道您是否可以更深入地探討一下，我們應該如何看待這兩個組成部分在未來 12 到 18 個月內推動您對運營支出和資本支出的一些評論。

Yeah, sure. I can start. At a high level, I think that there are all these questions that people have about what are going to be the time lines to get to really strong AI or Superintelligence or whatever you want to call it. And I guess that each step along the way so far, we've observed the more kind of aggressive assumptions or the fastest assumptions have been the ones that have most accurately predicted what would happen. And I think that, that just continued to happen over the course of this year, too.

是的，當然。我可以開始了。從高層次上講，我認為人們對實現真正強大的人工智慧或超級智慧（無論你想怎麼稱呼它）的時間線存在諸多疑問。我想，到目前為止，我們一路走來的每一步，都觀察到了更激進的假設或最快的假設，它們最準確地預測了將會發生什麼。我認為，這種情況在今年也將持續發生。

And so I've given a number of those anecdotes on these earnings calls in the past. And I think, certainly, some of the work that we're seeing with teams internally being able to adapt Llama 4 to build autonomous AI agents that can help improve the Facebook algorithm to increase quality and engagement, or like. I mean, that's like a fairly profound thing if you think about it. I mean it's happening in low volume right now. So I'm not sure that, that result by itself was a major contributor to this quarter's earnings or anything like that.

所以我過去在這些收益電話會議上講過一些這樣的軼事。我認為，我們看到內部團隊正在進行一些工作，能夠調整 Llama 4 來建立自主 AI 代理，從而幫助改進 Facebook 演算法，提高品質和參與度。我的意思是，如果你仔細想想，這是一件相當深刻的事。我的意思是，現在這種情況發生的頻率很低。所以我不確定，這個結果本身是否對本季的收益或類似因素做出了重大貢獻。

But I think the trajectory on this stuff is very optimistic. And I think it's one of the interesting challenges in running a business like this now is there's just a very high chance, it seems like the world is going to look pretty different in a few years from now. And on the one hand, there are all these things that we can do, there are improvements to our core products that exist.

但我認為這方面的發展軌跡非常樂觀。我認為現在經營這樣的企業面臨的一個有趣的挑戰是，很有可能幾年後世界就會變得非常不同。一方面，我們可以做所有這些事情，可以對現有的核心產品進行改進。

And then I think we have this principle that we believe in across the company, which we tell people take Superintelligence seriously. And the basic principle is this idea that we think that this is going to really shape all of our systems sooner rather than later, not necessarily on the trajectory of a quarter or two, but on the trajectory of a few years.

然後我認為我們整個公司都信奉這項原則，我們告訴人們要認真對待超級智慧。基本原則是，我們認為這很快就會真正塑造我們所有的系統，不一定是在一個或兩個季度的軌道上，而是在幾年的軌跡上。

And I think that, that's just going to change a lot of the assumptions around how different things work across the company. So anyway, I think it's basically just what we're continually observing how this works and what the trajectory or the pace of AI progress has been. I think it continues to be on the faster end.

我認為，這將改變許多關於公司內部不同事務如何運作的假設。所以無論如何，我認為這基本上就是我們不斷觀察它是如何運作的，以及人工智慧進步的軌跡或速度。我認為它將繼續保持較快的速度。

And that I think informs a lot of the decisions from everything from the importance and value of having the absolute best and most elite talent ends team at the company to making sure that we have a leading compute fleet so that the people here can do -- obviously, the researchers here have more compute per person to be able to leave their research and then roll it out to billions of people across our products, making sure that we build and drive these products through all the different things that we do, which I think is one of the things that our company is the best in the world at is basically when we take a technology, we're good at driving that through all of our apps and our ad systems and all that stuff, it's not just going to kind of sit on the line.

我認為這影響了很多決策，從擁有絕對最優秀、最精英的人才團隊的重要性和價值，到確保我們擁有領先的計算能力，以便這裡的人們可以做到——顯然，這裡的研究人員每人擁有更多的計算能力，能夠完成他們的研究，然後通過我們的產品將其推廣給數十億人，確保我們透過我們所做的所有不同的事情來建立和驅動這些產品，我認為這是我們公司在世界上最好的事情之一，基本上，當我們採用一項技術時，我們擅長透過我們所有的應用程式、廣告系統和所有這些東西來驅動它，它不僅僅是停留在線上。

I think that there's no other company, I think that is as good as us at kind of taking something and kind of getting it in front of billions of people. So yeah, I mean, we're just going to push very aggressively on all of that. But at some level, yeah, there's sort of the trajectory that we're seeing and those are the signals that we're seeing. But we're just trying to read it.

我認為沒有其他公司能夠像我們一樣擅長將某些東西呈現在數十億人面前。是的，我的意思是，我們將非常積極地推動這一切。但在某種程度上，是的，我們看到了某種軌跡，這些就是我們看到的訊號。但我們只是想讀。

Eric, for the second part of your question, we haven't, in fact, kicked off our budgeting process for 2026. So thinking about next year, there are clearly many, many moving pieces in a very dynamic operating environment.

艾瑞克，關於你問題的第二部分，實際上，我們還沒有啟動 2026 年的預算編制流程。因此，展望明年，在非常動態的營運環境中，顯然存在許多變化因素。

But there are certain aspects that we have some visibility into today, including the rough shape of our 2026 infrastructure plans. And that flows through into our expense expectations next year. And we also have some visibility into the compensation expense growth that we'll recognize from the AI talent that we're hiring this year.

但今天我們對某些方面已經有了一定的了解，包括 2026 年基礎設施計畫的大致情況。這會影響我們明年的支出預期。我們也對今年招募的人工智慧人才的薪酬費用成長有所了解。

And so those two things are part of why we gave a little bit of an early preview into the expectations for growth for 2026 total expenses as well as for 2026 CapEx. So on the total expenses side, as I mentioned, we expect infrastructure will be the single largest contributor to 2026 expense growth.

因此，這兩件事也是我們對 2026 年總支出以及 2026 年資本支出的成長預期進行提前預測的原因之一。因此，在總支出方面，正如我所提到的，我們預計基礎設施將成為 2026 年支出成長的最大單一貢獻者。

That's driven primarily by a sharp acceleration in depreciation expense growth in 2026, largely driven by recognizing incremental depreciation from assets that we purchased and placed in service in '26 as well as from infrastructure deployed through 2025 that will recognize a full year of depreciation next year. We also expect a greater mix of our CapEx to be in shorter-lived assets in 2025 and '26 than it has been in prior years.

這主要是由於 2026 年折舊費用增長急劇加速所致，這主要是由於我們在 26 年購買並投入使用的資產以及到 2025 年部署的基礎設施的增量折舊所致，這些基礎設施將在明年確認全年折舊。我們也預計，2025 年和 2026 年，我們的資本支出中用於短期資產的比例將比前幾年有所提高。

And then the other component of infra cost growth next year would come from higher operating expenses, including energy costs, leases, maintenance and operational expenses that are associated with maintaining that fleet.

明年基礎設施成本成長的另一個部分將來自更高的營運費用，包括與維護該車隊相關的能源成本、租賃、維護和營運費用。

And we also expect some increased spend on cloud services in '26 to meet our capacity needs as well as growth in network-related costs. So a lot going on, on the infrastructure side as it contributes to the 2026 total expense number.

我們也預計 26 年在雲端服務上的支出將有所增加，以滿足我們的容量需求以及網路相關成本的成長。因此，基礎設施方面有很多事情要做，因為它會影響 2026 年的總支出數字。

After that, employee compensation is the next largest driver of expense growth in '26. Again, driven primarily in the investments that we're making in technical talent, including recognizing a full year of compensation expense for the AI talent we hire this year. I realize this answer is getting a little long, so I'll try to wrap up quickly.

此後，員工薪酬是26年費用成長的第二大驅動力。再次強調，這主要得益於我們對技術人才的投資，包括對我們今年聘用的人工智慧人才確認全年的薪資費用。我意識到這個答案有點長，所以我會嘗試快速總結。

On the CapEx side, the big driver of our increased CapEx in '26 will be scaling GenAI capacity as we build out training capacity that's going to drive higher spend across servers, networking, data centers next year. We also expect that we're going to continue investing significantly in core AI in 2026. And again, this is a pretty very dynamic area of planning, but we wanted to share kind of our early thoughts as things are shaping up.

在資本支出方面，2026 年我們增加資本支出的主要驅動力將是擴大 GenAI 容量，因為我們正在建立培訓能力，這將推動明年在伺服器、網路、資料中心方面的支出增加。我們也預計，我們將在 2026 年繼續對核心人工智慧進行大量投資。再次強調，這是一個非常活躍的規劃領域，但我們希望在事情逐漸成形之際分享我們的早期想法。

Brian Nowak, Morgan Stanley.

摩根士丹利的布萊恩·諾瓦克。

I have two. The first one, Mark, just to kind of go back to the intelligence labs and sort of the vision for Superintelligence. As you sort of sit here now versus 12 months ago, can you just sort of walk us through any changes of technological constraints or technological gating factors that you are most focused on overcoming in the next 24 months that may have been different than they were in the past just to make sure you can really lead in the idea of Superintelligence over the next 10 years?

我有兩個。第一個，馬克，我想回顧一下智慧實驗室以及超級智慧的願景。您現在坐在這裡，與 12 個月前相比，您能否向我們介紹一下未來 24 個月您最關注的技術限製或技術限制因素的變化，這些變化可能與過去有所不同，只是為了確保您能夠在未來 10 年真正引領超級智能的理念？

And then the second one to Susan or Mark, one on the core, you've made so many improvements to the core to drive higher engagement, recommendations, et cetera. Can you just walk us through a couple of the factors you're still most excited about to come in the next 18 months that you think could drive a further lift to engagement on the core platform.

然後第二個是蘇珊或馬克，一個是關於核心的，你們對核心做出了很多改進，以推動更高的參與、推薦等等。您能否向我們介紹一下未來 18 個月內您最期待的幾個因素，您認為這些因素可以進一步提升核心平台的參與度？

Yeah, sure. I mean in terms of the research agenda and a bunch of the areas that we're very focused on, I do you think focusing on self-improvement is a very important area of research. And there's obviously different scaling paradigms, and I don't want to get too much into the detail of research that we're doing on this. But I think that for developing Superintelligence at some level, you're not just going to be learning from people because you're trying to build something that is fundamentally smarter than people.

是的，當然。我的意思是，就研究議程和我們非常關注的一系列領域而言，我認為關注自我提升是一個非常重要的研究領域。顯然有不同的擴展範例，我不想過多地談論我們在這方面所進行的研究的細節。但我認為，為了在某種程度上開發超級智能，你不能只是向人類學習，因為你試圖建構某種根本上比人類更聰明的東西。

So it's going to need to learn how to -- or you're going to need to develop a way for it to be able to improve itself. So that, I think, is a very fundamental thing. That is going to have very broad implications for how we build product, how we run the company, new things that we can invent new discoveries that can be made, society more broadly. I think that, that's just a very fundamental part of this.

所以它需要學習如何——或者你需要開發一種方法讓它能夠自我改進。所以我認為這是一件非常基本的事情。這將對我們如何製造產品、如何經營公司、如何發明新事物、如何進行新發現以及更廣泛的社會產生非常廣泛的影響。我認為，這只是其中非常基本的部分。

In terms of the shape of the effort overall, I guess I've just gotten a little bit more convinced around the ability for small talent-dense teams to be the optimal configuration for driving frontier research.

就整體努力的形式而言，我想我已經更加確信，小型人才密集團隊是推動前沿研究的最佳配置。

And it's a bit of a different setup than we have on our other world-class machine learning system. So if you look at like what we do in Instagram or Facebook or our ad system, we can very productively have many hundreds or thousands of people basically working on improving those systems, and we have very well-developed systems for kind of individuals to run tests and be able to test a bunch of different things.

它的設置與我們其他世界一流的機器學習系統略有不同。因此，如果你看看我們在 Instagram 或 Facebook 或我們的廣告系統中所做的事情，我們可以非常有效率地讓數百或數千人致力於改進這些系統，並且我們擁有非常完善的系統，供個人運行測試並能夠測試各種不同的東西。

You don't need every researcher there to have the whole system in their head. But I think for the leading research on Superintelligence, you really want the smallest group that can hold the whole thing in their head, which drives, I think, some of the physics around the team size and how -- and the dynamics around how that works. But I'll hand it over to Susan to talk about more of the practical stuff.

你不需要每個研究人員都了解整個系統。但我認為，對於超級智慧的領先研究，你真正需要的是最小的群體，能夠將整個事情記在腦子裡，我認為這推動了團隊規模的一些物理原理以及如何運作的動態。但我會把它交給蘇珊來談論更多實際的事情。

Brian, on the sort of forward-looking road map for the core recommendation engine. There are a handful of shorter-term things that we're focused on in the near term. One is we're focused on making recommendations even more adaptive to what a person is engaging with during their session, so that the recommendations we surface are the most relevant to what they're interested in at that moment.

布萊恩 (Brian) 就核心推薦引擎的前瞻性路線圖發表了看法。近期我們會專注在一些短期事務。一是，我們專注於使推薦更適應用戶在會話期間所參與的內容，以便我們提供的推薦與他們當時感興趣的內容最相關。

And we're making optimizations to help the best content from smaller creators break out by matching it to the right audiences sooner after it gets posted. And we're also working on improving the ability for our systems to discover more diversified and niche interest for each person through interest exploration and learning explicit user preferences.

我們正在進行優化，透過在小型創作者發布後儘快將最佳內容配對給合適的受眾，幫助其脫穎而出。我們也致力於提高我們的系統的能力，透過興趣探索和學習明確的使用者偏好，為每個人發現更多樣化和小眾的興趣。

We're also planning to scale up our models further and incorporate more advanced techniques that should improve the overall quality of recommendations. But we also have a lot of long-term bets in the hopper around areas like developing foundational models that will support recommendations across multiple services. Incorporating LLM more deeply into our recommendation systems.

我們還計劃進一步擴大我們的模型並融入更先進的技術，以提高建議的整體品質。但我們也在諸如開發支持跨多種服務推薦的基礎模型等領域進行了許多長期投資。將 LLM 更深入地融入我們的推薦系統。

And a big focus of this work is going to be on optimizing the systems to make them more efficient, so that we can continue to scale up the capacity that we use for our recommendation systems without eroding the ROI that we deliver.

這項工作的重點是優化系統，使其更有效率，這樣我們就可以繼續擴大推薦系統的容量，而不會影響我們提供的投資報酬率。

Doug Anmuth, JPMorgan.

摩根大通的道格安穆斯 (Doug Anmuth)。

One for Mark and one for Susan. Mark, Meta has been a huge proponent of open source AI how has your thinking changed here at all, just as you pursue superintelligence and push for even greater returns on your significant infrastructure investments. And then, Susan, your comments on '26 CapEx suggest more than $100 billion of spend next year potentially, do you continue to expect to finance all this yourself? Or could there be opportunities to partner here?

一份給馬克，一份給蘇珊。馬克，Meta 一直是開源人工智慧的巨大支持者，您的想法在這裡發生了怎樣的變化，就像您追求超級智慧並推動在重大基礎設施投資上獲得更大的回報一樣。然後，蘇珊，您對 26 年資本支出的評論表明明年的支出可能超過 1000 億美元，您是否仍然期望自己為所有這些支出提供資金？或是這裡有合作的機會嗎？

Yeah, I mean on open source, I don't think that our thinking has particularly changed on this. We've always open-sourced some of our models and not open source to everything that we've done. So I would expect that we will continue to produce and share leading open source models. I also think that there are a couple of trends that are playing out.

是的，我的意思是，關於開源，我認為我們的想法並沒有特別的改變。我們始終開源我們的一些模型，但並不開源我們所做的一切。因此我希望我們能夠繼續生產並分享領先的開源模型。我還認為，有幾種趨勢正在顯現。

One is that we're getting models that are so big that they're just not practical for a lot of other people to use. So we would kind of wrestle with whether it's productive or helpful to share that or if that's really just primarily helping competitors or something like that. So I think that there's that concern.

一是我們的模型太大，對很多人來說不實用。因此，我們會思考分享這些資訊是否有生產力或幫助，或者這是否真的只是主要幫助競爭對手或諸如此類的事情。所以我認為有這樣的擔憂。

And then obviously, as you approach real superintelligence, I think there is a whole different set of safety concerns that I think we need to take very seriously, that I wrote about in my note this morning. But I think the bottom line is, I would expect that we will continue open sourcing work.

顯然，當你接近真正的超級智慧時，我認為存在一系列完全不同的安全問題，我認為我們需要非常認真地對待這些問題，我在今天早上的筆記中寫到了這一點。但我認為最重要的是，我希望我們能夠繼續開源工作。

I expect us to continue to be a leader there. And I also expect us to continue to not open source everything that we do, which is a continuation of kind of what we've been kind of working on. And yeah, I mean, I think Susan will talk a little bit more about the infrastructure, but it really is a massive investment.

我希望我們能繼續在那裡保持領先。我也希望我們繼續不開源我們所做的一切，這是我們一直在努力的延續。是的，我的意思是，我認為蘇珊會更多地談論基礎設施，但這確實是一項巨大的投資。

We think it will be good over time. But we do take very seriously that this is a just massive amount of capital to convert into many gigawatts of compute which we think is going to help us produce leading research and quality products and running the business, I do look for opportunities to basically convert capital into quality of products that we can deliver for people.

我們認為隨著時間的推移它會變得更好。但我們確實非常重視，這只是一筆巨大的資本，可以轉化為數千兆瓦的計算能力，我們認為這將有助於我們進行領先的研究和生產高品質的產品，並且經營業務，我確實在尋找機會將資本轉化為我們可以為人們提供的高品質產品。

But this is certainly a massive. That we're kind of -- we're focused on and we want to make sure that what we build accrues to building the best products that we can deliver to the billions of people who use our services.

但這無疑是一個巨大的進步。我們專注於此，我們希望確保我們所建造的產品能夠為使用我們服務的數十億人提供最好的產品。

Doug, on your second question about how we expect to finance the growing CapEx next year, we certainly expect that we will finance some large share of that ourselves, but we're also exploring ways to work with financial partners to codevelop data centers.

道格，關於您的第二個問題，即我們預計如何為明年不斷增長的資本支出提供資金，我們當然希望自己能夠為其中很大一部分提供資金，但我們也在探索與金融合作夥伴合作開發資料中心的方式。

We don't have any finalized transactions to announce, but we generally believe that there will be models here that will attract significant external financing to support large-scale data center projects that are developed using our ability to build world-class infrastructure while providing us with flexibility should our infrastructure requirements change over time. So we are exploring many different paths.

我們沒有任何最終交易可以宣布，但我們普遍認為，這裡將會有一些模式可以吸引大量外部融資，以支持利用我們建設世界一流基礎設施的能力開發的大型數據中心項目，同時在我們的基礎設施需求隨時間變化時為我們提供靈活性。因此我們正在探索許多不同的道路。

Justin Post, Bank of America.

美國銀行的賈斯汀·波斯特 (Justin Post)。

I'll ask another one on the infrastructure. Mark, your spend is now approaching some of the biggest hyperscalers out there. Do you think of all this capacity mostly for internal uses? Or do you think there's a way to share or even come up with a business model, we're leveraging that capacity for external uses. And then Susan, when you think about the ROI on this CapEx, I'm sure you have internal models, I'm sure you can't share all that, but how are you thinking about the ROI? And are you optimistic about the long-term returns?

我會再問一個有關基礎設施的問題。馬克，你的支出現在已經接近一些最大的超大規模企業了。您是否認為所有這些容量主要用於內部用途？或者您認為有一種方法可以共享甚至提出一種商業模式，我們正在利用這種能力供外部使用。然後蘇珊，當您考慮這筆資本支出的投資回報率時，我相信您有內部模型，我相信您不能分享所有這些，但是您是如何考慮投資回報率的？您對長期回報是否樂觀？

Justin, I can go ahead and take a crack at both of those. And obviously, Mark, you should feel free to weigh in. Right now, we are focused on ensuring that we have enough capacity for our internal use cases, which includes both all of the core AI work that we do to support the recommendation engine work on the organic content side to support all the ads ranking and recommendation work.

賈斯汀，我可以繼續嘗試這兩個問題。顯然，馬克，你應該可以自由地發表意見。目前，我們專注於確保我們擁有足夠的容量來滿足內部用例的需求，其中包括我們所做的所有核心人工智慧工作，以支援有機內容方面的推薦引擎工作，以及支援所有廣告排名和推薦工作。

And then, of course, to make sure that we are building the training capacity that we think we need in order to build frontier AI models. And to make sure that we're preparing ourselves for the types of inference use cases that we think might -- that we might have ahead of us as we eventually focus not only on developing frontier models, but also how we can expand into the kinds of consumer use cases that we think will be hopefully live -- hopefully, widely useful and engaging for our users.

然後，當然，要確保我們正在建立我們認為需要的訓練能力，以建立前沿人工智慧模型。並確保我們為我們認為可能出現的推理用例類型做好準備——我們最終不僅關注開發前沿模型，還關注如何擴展到我們認為有望實現的消費者用例類型——希望這些用例對我們的用戶有廣泛的用途和吸引力。

So at present, we're not really thinking about external use cases on the infrastructure, but it's a good question.

因此，目前我們並沒有真正考慮基礎設施上的外部用例，但這是一個很好的問題。

On your second question, which is really around the sort of ROI on CapEx, there are a couple of things. So again, on the core AI side, we continue to see strong ROI. Our ability to measure that is quite good, and we feel sort of very good about the rigorous measurement and returns that we see there. On the GenAI side, we are clearly much, much earlier on the return curve and we don't expect that the GenAI work is going to be a meaningful driver of revenue this year or next year.

關於你的第二個問題，其實是關於資本支出的投資報酬率，有幾件事。因此，在核心人工智慧方面，我們持續看到強勁的投資報酬率。我們的衡量能力相當好，我們對所看到的嚴格測量和回報感到非常滿意。在 GenAI 方面，我們顯然處於回報曲線的早期階段，並且我們預計 GenAI 工作不會成為今年或明年收入的顯著推動力。

But we remain generally very optimistic about the monetization opportunities that will open up, and Mark spoke to them in his script, the sort of five pillars, so I won't repeat them here. And we think that over the medium- to long-term time frame, those are opportunities that are very adjacent and intuitive for where -- in terms of where our business is today, why they would be big opportunities for us and that there will be sort of big markets attached to each of them.

但我們總體上仍然對即將出現的貨幣化機會非常樂觀，馬克在他的腳本中談到了這些機會，即五大支柱，所以我就不再重複了。我們認為，從中長期來看，這些都是非常相鄰且直觀的機會——就我們目前的業務而言，為什麼它們對我們來說是巨大的機會，並且每個機會都會帶來巨大的市場。

So we, again, are also, I would say, the last thing I would add here is we are building the infrastructure with fungibility in mind. Obviously, there are a lot of things that you have to build up front in terms of the data center shells, the networking infrastructure, et cetera. But we will be ordering servers, which ultimately will be the biggest bulk of CapEx spend as we need them and when we need them and making sort of the best decisions at those times in terms of figuring out where the capacity will go to use.

因此，我想補充的最後一點是，我們在建造基礎設施時考慮到了可替代性。顯然，在資料中心外殼、網路基礎設施等方面，有很多東西需要預先建造。但我們將訂購伺服器，這最終將成為資本支出的最大一部分，因為我們需要它們，當我們需要它們時，我們會在那些時候做出最好的決定，以確定容量將用在何處。

Mark Shmulik, Bernstein.

馬克‧施穆里克，伯恩斯坦。

Mark, as you go after the Superintelligence vision, especially for those of us on the outside, what are kind of some of the markers or KPIs that you're tracking on whether you're on track in making progress? Is it really against kind of those five pillars you outlined above? Or should we be thinking more broadly?

馬克，當你追求超級智慧願景時，特別是對於我們這些局外人來說，你會追蹤哪些標記或關鍵績效指標 (KPI) 來判斷你是否朝著正確的方向前進？它真的違背了您上面概述的五大支柱嗎？或者我們應該思考得更廣泛？

And Susan, obviously AI delivering great ROI today, all those investments and also building towards kind of longer-term goals, just curious, has there just been any change or adjustment to how you think about the relationship between revenues or core business performance and the cadence of investments?

蘇珊，顯然人工智慧如今帶來了巨大的投資回報率，所有這些投資也都在朝著某種長期目標邁進，我只是好奇，您對收入或核心業務績效與投資節奏之間的關係的看法是否有任何變化或調整？

Yeah, in terms of what to look at, I mean, what I'm going to look at internally, the quality of the people on the teams, the quality of the models that we're producing, the rate of improvement of our other AI systems across the company and the extent to which the leading kind of foundation models that we're building contribute to improving all of the other AI systems and kind of everything that we're doing around the company.

是的，就要看什麼而言，我的意思是，我要從內部看什麼，團隊人員的素質，我們正在生產的模型的質量，整個公司其他人工智能係統的改進速度，以及我們正在構建的領先基礎模型對改進所有其他人工智能係統的貢獻程度，以及我們在公司周圍所做的一切。

Then I think you just get into our standard product and business playbook, which is translating that technology into new products, which will first scale to billions of people and then over time, we will monetize. But I think that there's going to be some lag in that, right?

然後我想你就會進入我們的標準產品和商業劇本，即將該技術轉化為新產品，這些產品將首先擴展到數十億人，然後隨著時間的推移，我們將實現貨幣化。但我認為這會有一些滯後，對嗎？

And that, I think, is kind of always the way that we work is, whether we're building some new social product or this something like Meta AI or a new product around this that we're going to work on getting to leading scale, building the highest quality product focused on that for a few years. And then once we're really confident in that position, then we'll focus on ramping up the business around it.

我認為，這始終是我們工作的方式，無論我們是在建立一些新的社交產品，還是像 Meta AI 這樣的產品，或者圍繞這個的新產品，我們都將努力達到領先規模，專注於此，打造最高品質的產品。一旦我們對該地位真正有信心，我們就會專注於擴大周圍的業務。

So I mean, going back to the last question a little bit, it's sort of when you compare this business to some of the cloud businesses, it's like we do have this delay where we focus on building research and then doing research and then ramping consumer products, and it often does take some period of time before we really are ramping up the business around it. I think that's kind of a known property of our business and the cycle around it.

所以我的意思是，稍微回到最後一個問題，當你將這項業務與一些雲端業務進行比較時，我們確實存在這種延遲，我們專注於建立研究，然後進行研究，然後推出消費產品，而且通常需要一段時間才能真正擴大圍繞它的業務。我認為這是我們業務及其周邊週期的已知特性。

But I guess, on the flip side, we believe that if you are building Superintelligence, you should use all of your GPUs to make it so that you're serving your customers really well with that. And we think that there's going to be a much higher return than we can do by generating that directly rather than just kind of renting or leasing out the infrastructure at other companies.

但我想，另一方面，我們相信，如果你正在建立超級智能，你應該使用所有的 GPU 來實現這一點，這樣你才能真正為你的客戶服務。我們認為，透過直接創造收益而不是僅僅租用或出租其他公司的基礎設施，我們可以獲得更高的回報。

On the second part of your question, we've said in the past that our primary focus from a profitability perspective is driving consolidated operating profit growth over time. And it won't be linear. In some years, we'll deliver above-average profit growth.

關於您問題的第二部分，我們過去曾說過，從獲利能力的角度來看，我們的主要關注點是推動綜合營業利潤的長期成長。而且它不是線性的。在某些年份，我們的利潤成長將高於平均水準。

And in years where we're making big investments, I think we will see that impact the amount of operating profit growth that we can deliver. And at the moment, we see a lot of attractive investment opportunities that we believe are going to set us up to deliver compelling profit growth in the coming years for all of our investors.

在我們進行大規模投資的年份，我認為我們將看到這對我們所能實現的營業利潤成長量的影響。目前，我們看到了許多有吸引力的投資機會，我們相信這些機會將幫助我們在未來幾年為所有投資者帶來可觀的利潤成長。

And so we're focused on constraining investments elsewhere as we pursue those investments. But we really believe that this is a time for us to really make investments in the future of AI as I think it will open up both new opportunities for us in addition to strengthen our core business.

因此，我們在進行投資時，重點是限制其他地方的投資。但我們確實相信，現在是我們真正投資人工智慧未來的時候了，因為我認為除了加強我們的核心業務之外，它還將為我們開闢新的機會。

Ron Josey, Citi.

花旗銀行的 Ron Josey。

Mark, I wanted to ask you on Meta AI and I think you talked about in the call just growing engagement overall, particularly on WhatsApp and now you have 1 billion users on the platform and the focus is now on driving personalization.

馬克，我想問您關於 Meta AI 的問題，我記得您在電話會議中談到了整體參與度的提高，特別是在 WhatsApp 上，現在該平台擁有 10 億用戶，現在的重點是推動個性化。

So I want to understand a little bit more how these next-gen models can help drive adoption here, particularly with Behemoth coming online at some point. And then as people are using that Meta AI with WhatsApp thoughts on search and queries and potentially monetizing that.

因此，我想進一步了解這些下一代模型如何幫助推動這裡的採用，特別是隨著 Behemoth 的某個時候上線。然後，人們將 Meta AI 與 WhatsApp 結合起來，思考搜尋和查詢，並有可能將其貨幣化。

Yeah, I'm not going to get super deep into the road map on this, but the basic -- we do see that as we continue improving the models behind Meta AI and post training and just engagement increases and as we swap in the updated models, when we go from Llama 4 to Llama 4.1 when we have that, we expect that just -- the models are inherently pretty general.

是的，我不會深入探討這個路線圖，但基本情況是——我們確實看到，隨著我們不斷改進 Meta AI 背後的模型和後期訓練，參與度不斷提高，並且隨著我們更換更新的模型，當我們從 Llama 4 升級到 Llama 4.1 時，我們預計——這些模型本質上是相當通用的。

So yeah, you focus on specific areas, but in general, just sort of gets better at a lot of different things that people want to ask it or want to do with it. And I think with each version, both like what we're doing on a week-to-week basis in terms of continuing to train it. And when we drop kind of new generations or big dot releases of each generation, that will improve engagement, too. So we're focused on that. I'm not going to go into the specific research areas or capabilities that we're planning on dropping in the future. But obviously, I'm pretty excited about it.

是的，你專注於特定的領域，但總的來說，它在人們想問它或想要用它做的許多不同的事情上都會變得更好。我認為對於每個版本，我們都在每週進行持續訓練。當我們推出新一代產品或每一代產品的重大發佈時，參與度也會提高。所以我們專注於此。我不會深入探討我們計劃在未來放棄的具體研究領域或能力。但顯然，我對此感到非常興奮。

Youssef Squali, Truist Securities.

Youssef Squali，Truist Securities。

So Mark, the rebound initiative has been a home market for you guys so far. Where are we on the development of glasses? Is that new computational platform that you've talked about in the past? Is it moving faster or slower than you thought? And as you leverage Meta AI, do you believe glasses will ultimately replace smartphones?

馬克，到目前為止，反彈計畫對你們來說還是本土市場。我們的眼鏡發展到了什麼程度？這就是您之前談到的新的運算平台嗎？它的移動速度比你想像的快還是慢？當您利用 Meta AI 時，您是否相信眼鏡最終會取代智慧型手機？

Or do you need the new form factor that's AI first? And then, Susan, just quickly, how do you guys see SBC progressing over the next couple of years? Is it fair to assume it will grow materially faster than the revenue and OpEx? And how do you minimize shareholder dilution?

或者您需要先採用 AI 的新形式？然後，蘇珊，請簡單問一下，你們認為 SBC 在未來幾年會取得怎樣的進展？是否可以合理地假設其成長速度將顯著快於收入和營運支出？那麼如何最大程度地減少股東權益稀釋呢？

Yeah, I can talk a bit about the glasses. Yeah, I mean, I'm very excited about the progress that we're making. I think both the Ray-Ban Metas and I'm very excited about the Oakley Meta, the HSTNs too, and other things that we have planned. Yeah, I mean, this product category is clearly doing quite well.

是的，我可以談談眼鏡。是的，我的意思是，我對我們所取得的進展感到非常興奮。我認為 Ray-Ban Metas 和 Oakley Meta、HSTN 以及我們計劃的其他產品都讓我非常興奮。是的，我的意思是，這個產品類別顯然表現相當不錯。

And I think it's good for a lot of things. It is stylish eyewear, so people like wearing them just as glasses. It has a bunch of interesting functionality. And then the use of Meta AI in them just continues to grow, and the percent of people who are using it for that on a daily basis is increasing, and that's all good to see.

我認為這對很多事情都有好處。它是一種時尚的眼鏡，所以人們喜歡像戴眼鏡一樣戴著它們。它具有許多有趣的功能。然後，其中元人工智慧的使用就不斷增長，每天使用它的人數比例也在增加，這一切都是令人高興的。

I mean I continue to think that glasses are basically going to be the ideal form factor for AI. At least you can let an AI see what you see throughout the day, hear what you hear, talk to you once you get a display in there, whether it's the kind of wide holographic field of view like we showed with Orion or just a smaller display that might be good for displaying some information.

我的意思是，我仍然認為眼鏡基本上是人工智慧的理想外形。至少你可以讓人工智慧看到你一整天所看到的東西，聽到你聽到的東西，一旦你在裡面安裝了顯示器，它就可以和你交談，無論是像我們在 Orion 上展示的那種寬全息視野，還是可能適合顯示某些信息的小型顯示器。

And that's also going to unlock a lot of value where you can just interact with an AI system throughout the day in this multimodal way. I can see the content around you. It can generate a UI for you to show you information and be helpful. I mean, I personally think that just like I work contact lenses, I feel like if I didn't have my vision corrected, I'd be sort of at a cognitive disadvantage going through the world.

這也將釋放很多價值，您可以全天以這種多模式方式與人工智慧系統進行互動。我可以看到你周圍的內容。它可以為您產生一個 UI 來向您顯示資訊並提供幫助。我的意思是，我個人認為，就像我戴隱形眼鏡一樣，如果我的視力沒有得到矯正，我在與外界交往時就會處於認知劣勢。

And I think in the future, if you don't have glasses that have AI or some way to interact with AI, I think you're kind of similar we'd probably be at a pretty significant cognitive disadvantage compared to other people who you're working with, or competing against.

我認為，在未來，如果你沒有配備人工智慧的眼鏡或某種與人工智慧互動的方式，那麼與你一起工作或競爭的其他人相比，你可能會處於相當明顯的認知劣勢。

So I think that this is a pretty fundamental form factor. There are a lot of different versions of it. Right now, we're building ones that I think are stylish, but aren't focused on the display. I think if there's a whole set of different things to explore with displays. This is kind of what we've been maxing out with Reality Labs over the last 5 to 10 years is basically doing the research on all of these different things.

所以我認為這是一個非常基本的形式因素。它有很多不同的版本。目前，我們正在打造一些我認為很時尚但不注重展示的產品。我認為如果有一整套不同的事物可以透過展示來探索。過去 5 到 10 年來，我們一直在 Reality Labs 上進行各種不同事物的研究。

And I don't know, 10 years ago, I would have -- like the other thing that's awesome about glasses is, they are going to be the ideal way to blend the physical and digital worlds together. So the whole metaverse vision, I think, is going to end up being extremely important, too, and AI is going to accelerate that, too.

我不知道，10 年前，我會——眼鏡的另一個奇妙之處在於，它們將成為融合物理世界和數位世界的理想方式。因此，我認為整個元宇宙願景最終也將變得極為重要，而人工智慧也將加速這一進程。

It just that if you'd asked me five years ago, whether we'd have kind of holograms that created immersive experiences or superintelligence first, I think most people would have thought that you'd get the holograms first. And it's this interesting kind of quirk of the tech industry that I think we're going to end up having really strong AI first. But because we've been investing in this, I think we're just several years ahead on building out glasses.

只是如果五年前你問我，我們是否會先擁有能夠創造沉浸式體驗或超級智慧的全像圖，我想大多數人都會認為你會先獲得全像圖。這是科技業的一個有趣的怪癖，我認為我們最終將首先擁有真正強大的人工智慧。但因為我們一直在這方面投資，我認為我們在眼鏡製造方面領先了幾年。

And I think that, that's something that we're excited to keep on investing in heavily because I think it's going to be a really important part of the future.

我認為，我們很高興繼續大力投資這一點，因為我認為這將成為未來非常重要的一部分。

Youssef, we didn't quite get your second question, do you mind just repeating it?

優素福，我們不太明白你的第二個問題，你介意重複一次嗎？

Sure. Just as you look at the spend on stock-based compensation over the next couple of years with all these hires, I'm assuming that we're going to see that materially or grow materially faster maybe than revenue and OpEx. And I just want to know what you guys are doing to plan to minimize shareholder dilution. Is it mostly buybacks or anything else?

當然。正如您在未來幾年內透過招聘看到股票薪酬支出一樣，我認為我們將看到這一支出出現實質增長，或者可能比收入和營運支出增長更快。我只是想知道你們正在採取什麼措施來盡量減少股東權益稀釋。主要是回購還是其他什麼？

Thanks, Youssef. So I mean, the impact of the sort of increased compensation costs, including SBC, of our AI hires this year is reflected in the revised 2025 expense outlook and in the comments I made about sort of 2026, expense outlook.

謝謝，優素福。所以我的意思是，今年我們招募的 AI 員工的薪資成本（包括 SBC）增加的影響反映在修訂後的 2025 年費用前景以及我對 2026 年費用前景的評論中。

Those are obviously a big driver of 2026 expense growth as we recognize the full year of compensation for the additional talent we're bringing on. Having said that, so we factored that into our sort of expense outlook. Having said that, we certainly -- we are very focused on making sure on keeping an eye on dilution.

顯然，這些是 2026 年費用成長的主要驅動力，因為我們要對所引進的額外人才進行全年補償。話雖如此，我們還是將其納入了我們的支出前景中。話雖如此，我們當然——我們非常注重確保關注稀釋。

And we generally believe that our strong financial position is going to allow us to support these investments, while continuing to repurchase shares as part of the sort of buyback program that offsets equity compensation and as well as provide quarterly cash dividend distributions to our investors.

我們普遍認為，我們強大的財務狀況將使我們能夠支持這些投資，同時繼續回購股票，作為抵消股權補償的回購計劃的一部分，並向我們的投資者提供季度現金股息分配。

Great. Thank you, everyone, for joining us today. We look forward to speaking with you again soon.

偉大的。感謝大家今天加入我們。我們期待很快再次與您交談。