Alphabet（GOOGL）— 2026-10-28 3Q26 法說會逐字稿（英文原文＋中譯）

日期：2026-10-28（美國時間盤後） 逐字稿為 Alphabet earnings call 音檔 AI 轉寫，英文原文與中譯並列；專有名詞與數字以公司 press release／10-Q 為準。

校正紀錄

本檔為 AI 語音轉寫，落地前已校正下列轉寫誤植；未列出者為原文。Memo 區塊與 Transcript 區塊同步修正。 - 00:36:26 Schindler「Search grew 12% year over year to 63 billion queries」→ to $63 billion（營收，非查詢數；與 CFO 口述 Search 及其他營收 633 億美元一致；中譯「630 億次查詢」→「630 億美元營收」） - 00:57:13 Pichai「later this year and into 2024」→ 2027（期間錯置，依時序推定；中譯同步） - 00:37:07「AI overuse」→ AI overviews（英文術語誤植） - 中譯產品名誤譯還原：「氬」→ Argon、「碳」→ Carbon、「雙子星 4」（指 Genie 4 世界模型一處）→ Genie 4；00:47:02 提問「千瓦」→ 吉瓦（GW） - 財務四項交叉驗算：①營收 1,276 億 − 營業成本 514 億 − 營業費用 356 億 ＝ 營業利益 406 億（31.8%）✓；②部門營業利益 Services 392 億 + Cloud 111 億 − Other Bets 20 億 − Alphabet 層級 77 億 ＝ 406 億 ✓；③廣告 633 + 112 + 75 ＝ 820 億，加訂閱／平台／裝置 143 億 ＝ Services 963 億 ✓；Cloud 111/307 ＝ 36.2% ✓、Services 392/963 ＝ 40.7% ✓；④營運現金流 423 億 − 資本支出 514 億 ＝ 自由現金流 −91 億 ✓。淨利與 EPS 於會上未報金額（僅稱因其他收支 −82 億美元而 YoY 下滑），不補推 - 與同場券商 call memo 交叉：營收、營業利益、Cloud backlog、資本支出指引、TPU 客戶數等數字一致，無需更動 - 保留原文未改：Kurian 備稿「Nearly 100 customers use TPUs today」與 Q&A「over 100 significant TPU customers」口徑略有差異，各自保留；Kurian「customers using it in Q3 2025 more than quadrupled their spend」年份存疑 [數字存疑]（Memo 區塊同）；提問中「Spark」、客戶名「Mirendil」「Crux AI」等依轉寫保留

警語：此 Memo 和逐字稿為 AI 生成，可能因為音檔品質導致轉錄逐字稿有誤，或是 AI 曲解原意／幻覺導致 Memo 內容有誤，請以原始音檔為準。

Memo 亮點 • Alphabet 在 2026 年第三季表現強勁，營收年增 25% 至 1,276 億美元，Google Cloud 營收則成長 103% 至 307 億美元。推出了 Gemini 4 Argon 與 Carbon 兩款模型，其中 Carbon 榮登人工分析智能指數第一名。Gemini 應用用戶數達 10 億，成為 Google 歷史上成長最快的產品。YouTube Music 與 Premium 全球訂閱用戶超過 1.5 億。Waymo 已行駛超過 3 億英里全自動駕駛，擴展至美國 15 個主要城市，並計劃國際擴張。 [CEO - 00:01:37] [CFO - 00:24:12] [CFO - 00:29:31] • Google Cloud 現有 18 條產品線，每條年營收均超過 10 億美元。近 90% 的財富 100 強企業使用 Gemini Enterprise，2025 年第三季使用該產品的客戶支出年增超過四倍（根據音訊資料推斷）。 [CEO - 00:16:34] • 營業利益增加 30% 至 406 億美元，營業利潤率為 31.8%。然而，淨利與每股盈餘因 Alphabet 對 SpaceX 投資產生 82 億美元未實現損失而下降。 [CFO - 00:25:31] • Alphabet 於本季末持有現金及可供出售證券共 2,651 億美元，長期負債為 1,251 億美元。 [CFO - 00:27:14] 總體經濟 • 外匯對年增率造成輕微逆風，對季增率影響較大。預期外匯逆風在第四季仍將帶來適度影響。 [CFO - 00:28:29] [CFO - 00:41:04] 地緣政治 • 在零售領域，部分亞太區廣告主受歐盟取消最低免稅額政策影響，導致該區廣告表現受挫。 [Chief Business Officer - 00:36:43] 產業循環週期 • 媒體與娛樂支出在第二季因 FIFA 世界盃異常強勁後回歸正常，該賽事六月比七月有更多比賽及廣告預算分配。 [Chief Business Officer - 00:36:53] 產業供需 • Google Cloud 的 AI 基礎設施與解決方案需求龐大，公司正努力增加產能以滿足長期需求。TPU 系統需求強勁且持續成長。 [CEO - 00:02:16] • Google 隨時可透過增加產能大幅滿足更多需求，但正謹慎管理各項參數與成長速度。 [CEO - 00:59:40] 產業狀況 • 網路安全威脅風險日增，推動對 Google AI 驅動安全平台的強勁需求，Gemini Flash 網路安全模型能自動化漏洞發現與修補，表現優於其他模型。 [CEO - 00:20:20] 獲利 • 營業利益增加 30% 至 406 億美元，營業利潤率為 31.8%。Cloud 營業利益為 111 億美元，年增超過三倍，營業利潤率從 23.7% 提升至 36.2%。 [CFO - 00:25:31] • 淨利與每股盈餘因 Alphabet 對 SpaceX 投資產生 82 億美元未實現損失而下降。 [CFO - 00:25:38] 公司展望、未來成長動能 • Alphabet 預期未來 12 個月 AI 模型開發將持續加速，受益於遞迴自我優化（RSI）。更高效模型即將推出，旨在協助開發者優化速度與成本，並部署至數十億用戶。 [CEO - 00:05:31] • TPU 系統銷售預計在 2026 年第四季持續擴大，2027 年將迎來顯著營收成長（根據音訊資料推斷）。資本支出預計在 2027 年大幅增加。 [CFO - 00:31:38] • Alphabet 預期將產生明顯高於資本成本的吸引人投資報酬率，並預計在整個投資週期中維持強勁營運現金流。 [CFO - 00:33:05] 市場競爭策略 • Google 正與包括 Apple 在內的競爭對手合作開發模型與雲端服務，視 TPU 系統為重要平台。公司目標是成為世界級開發者，並在最嚴苛工作負載中爭取客戶。 [CEO - 01:01:28] 公司對市場觀點的看法 • 用戶研究顯示，民眾對 AI 驅動服務中的廣告感到接受且高度相關。AI 模式下贊助與自然產品列表項目的點擊率已相近。 [Chief Business Officer - 00:11:34] 營收 • 合併營收為 1,276 億美元，年增 25%。Google 服務營收增 11% 至 963 億美元，整體廣告營收增 11%。Google Cloud 營收成長 103% 至 307 億美元。 [CFO - 00:24:19] 訂單 • Google Cloud 第三季訂單餘額為 5,610 億美元，大部分為典型 GCP 合約。超過 55% 訂單餘額預計於未來 24 個月內認列為營收。 [CFO - 00:30:32] 產品組合 • Google Cloud 現有 18 條產品線，每條年營收均超過 10 億美元。成長最強勁的驅動力為 Google Cloud Platform，在 AI 與核心 GCP 服務均展現強勁表現。 [CEO - 00:16:34] [CEO - 00:58:18] 營收地區 • 在零售領域，亞太區廣告主受歐盟取消最低免稅額政策影響，導致該區廣告表現受挫。 [Chief Business Officer - 00:36:43] 下游客戶 • BNP Paribas 將 Gemini Enterprise 部署給超過 65,000 名員工，以簡化工作流程並維持嚴格企業級安全與資料治理。Tata Steel 在九個月內部署超過 300 個專用 AI 代理，提升製造安全、優化供應鏈，並將客戶服務時間減半。 [CEO - 00:18:14] 同業 • Google 正與包括 Apple 在內的競爭對手合作開發模型與雲端服務，並與新興前沿模型公司、資本市場機構、大學及公共部門服務提供者合作。 [CEO - 01:01:28] [CEO - 00:22:58] 稼動率 • Google 努力增加產能以滿足 AI 基礎設施與解決方案的長期需求。Google Cloud 的第三方橋接產能僅於第三季末上線，對本季利潤率影響有限。 [CEO - 00:02:26] [CFO - 00:30:20] 存貨庫存 • 營運現金流因 TPU 系統銷售的庫存增加而受負面影響。預期庫存餘額將增加，因尚未銷售的 TPU 系統將列入庫存，現金流出於庫存建置時發生，隨系統交付與開票而回流。 [CFO - 00:26:08] [CFO - 00:49:26] 新產品／新技術 • 推出 Gemini 4 Argon 與 Carbon 模型，Carbon 具備先進網路安全與長期預測能力，榮登人工分析智能指數第一名。並宣佈 Genie 4，一款可即時模擬互動世界並具原生音訊的世界模型。 [CEO - 00:03:04] • 與 Meta 合作推出 TorchTPU，使用戶能利用熟悉工具在 TPU 上訓練、服務及部署 PyTorch 模型。 [CEO - 00:22:26] 產品價格 • TPU 8T 與 8i 分別提供較前代 TPU 高出 2.7 倍的訓練效能及 80% 更佳的推論價格效能。 [CEO - 00:22:10] 法規影響 • 於上訴程序結束後，支付 52 億美元現金以繳納 2018 年歐盟委員會罰款。 [CFO - 00:26:08] 成本毛利 • 營收總成本為 514 億美元，年增 24%。其他營收成本合計 351 億美元，年增 32%，主要因 TPU 系統銷售相關硬體成本、折舊增加、內容採購成本及技術基礎設施營運費用增加。Google 服務營業利潤率為 40.7%。 [CFO - 00:24:28] 折舊 • 第三季約 65% 技術基礎設施投資用於伺服器與網路設備，使用年限約六年。其餘 35% 用於資料中心，使用年限介於七至四十年。折舊增加，主因 AI 人才投資與基礎設施支出提升。 [CFO - 00:26:33] [CFO - 00:25:01] 匯率 • 外匯對年增率造成輕微逆風，對季增率影響較大。預期外匯逆風在第四季仍將帶來適度影響。 [CFO - 00:28:29] [CFO - 00:41:04] 銷售／管理／研發費用 • 研發費用增加 36%，主要因 AI 人才薪酬及折舊增加。銷售與行銷費用成長 22%，受 Gemini 投資及薪酬提升推動。一般及行政費用減少 17%，主要因去年第三季較高法律費用基期效應。 [CFO - 00:25:01] 擴廠規劃 • Waymo 目前在美國 15 個主要城市營運，並於聖安東尼奧、聖地牙哥及坦帕開放公眾服務，計劃擴展至加拿大、德國、新加坡及西班牙。 [CEO - 00:08:21] 營業利益 • 營業利益增加 30% 至 406 億美元，營業利潤率為 31.8%。Google 服務營業利益增 17% 至 392 億美元，營業利潤率為 40.7%。Cloud 營業利益為 111 億美元，年增超過三倍。 [CFO - 00:25:31] 業外損益 • 其他收益與費用為負 82 億美元，主要因 Alphabet 對 SpaceX 投資在第三季股價下跌後產生未實現損失。 [CFO - 00:25:38] 股利政策 • 董事會宣佈普通股季度現金股利每股 0.22 美元，將於十二月發放。 [CFO - 00:28:01] 現金、現金流 • Alphabet 於本季末持有現金及可供出售證券共 2,651 億美元，其中包含 904 億美元市值股權證券。第三季營運現金流為 423 億美元，過去十二個月累計為 1,796 億美元。自由現金流第三季為負 91 億美元，主因資本支出、TPU 系統銷售庫存增加及歐盟罰款支付。 [CFO - 00:25:56] 長短期借款 • 本季末長期負債為 1,251 億美元。 [CFO - 00:27:25] Transcript Presentation Operator - Operator

(00:00:00)Welcome, everyone. Thank you for standing by for the Alphabet Third Quarter 2026 Earnings Conference Call. At this time, all participants are in listen-only mode. After the speaker presentation, there will be a question-and-answer session. To ask a question during the session, please press star 1 on your telephone. I would now like to hand the conference over to your speaker today, Jim Friedland, Head of Investor Relations. Please go ahead.

(00:00:00)歡迎大家。感謝您耐心等待 Alphabet 2026 年第三季度財報電話會議。此時，所有參與者均為靜音模式。在演講者發表完後，將會有問答環節。在會議期間，如需提問，請在您的電話上按星號 1。現在，我想將會議交給今天的演講者，Jim Friedland，投資者關係負責人。請開始。

Jim Friedland - Alphabet, Investor Relation

(00:00:32)Thank you. Good afternoon, everyone, and welcome to Alphabet's third quarter 2026 earnings conference call. With us today are Sundar Pichai, Philipp Schindler, Thomas Kurian, and Anat Ashkenazi. Now, I'll quickly cover the safe harbor statement. Some of the statements we make today regarding our business, operations, and financial performance may be considered forward-looking. Such statements are based on current expectations and assumptions and are subject to various risks and uncertainties. Actual results could differ materially.

(00:00:32)謝謝。大家下午好，歡迎來到 Alphabet 2026 年第三季度財報電話會議。今天與我們在一起的有 Sundar Pichai、Philipp Schindler、Thomas Kurian 和 Anat Ashkenazi。現在，我將快速介紹安全港聲明。我們今天關於業務、運營和財務表現的某些聲明可能被視為前瞻性陳述。此類聲明基於當前的預期和假設，並受到各種風險和不確定性的影響。實際結果可能會有重大差異。

Jim Friedland - Alphabet, Investor Relation

(00:01:01)Please refer to our Forms 10-K and 10-Q, including the risk factors. We undertake no obligation to update any forward-looking statements. During this call, we will present both GAAP and non-GAAP financial measures. A reconciliation of non-GAAP to GAAP measures is included in today's earnings press release, which is available to the public through our investor relations website at abc.xyz/investor. Our comments will be based on year-over-year comparisons unless stated otherwise. And now, I'll turn the call over to Sundar.

(00:01:01)請參閱我們的 10-K 和 10-Q 表格，包括風險因素。我們不承擔更新任何前瞻性陳述的義務。在此次通話中，我們將呈現 GAAP 和非 GAAP 財務指標。非 GAAP 與 GAAP 指標的調整已包含在今天的財報新聞稿中，該新聞稿可通過我們的投資者關係網站 abc.xyz/investor 公開獲取。我們的評論將基於同比比較，除非另有說明。現在，我將把通話交給 Sundar。

Sundar Pichai - Alphabet, CEO

(00:01:35)Thanks, Jim. And hello, everyone. Q3 was a strong quarter for Alphabet. Our deep full-stack approach is driving momentum across the business, with many opportunities ahead. Let me give you a quick snapshot of the quarter. Alphabet's revenues grew 25% year over year. We are now at a half-trillion-dollar annual run rate. That's remarkable growth. Search has strong momentum. Our AI investments are driving increased usage and overall query growth, including commercial queries. In fact, our year-over-year query growth rate is the highest it has been in the last five years.

(00:01:35)謝謝，Jim。大家好。第三季度對 Alphabet 來說是一個強勁的季度。我們的深度全棧方法正在推動業務的動能，未來有許多機會。讓我給您快速介紹一下這個季度的情況。Alphabet 的收入同比增長了 25%。我們現在的年化運行率已達到五千億美元。這是驚人的增長。搜索業務勢頭強勁。我們的 AI 投資正在推動使用量增加和整體查詢增長，包括商業查詢。實際上，我們的同比查詢增長率是過去五年來的最高水平。

Sundar Pichai - Alphabet, CEO

(00:02:12)Google Cloud revenues grew 103%. Our AI infrastructure and solutions are the major drivers of cloud growth, and our customers see us as the leading provider, so demand is huge. We are working hard to bring more capacity online to meet this long-term demand. At the same time, we are seeing strong demand for our TPU systems across a wide range of customers, and we expect that to continue as we scale. I'm excited to welcome Thomas to the call; he will share more about Cloud's momentum. On our AI models, customers are using them more than ever.

(00:02:12)Google Cloud 的收入增長了 103%。我們的 AI 基礎設施和解決方案是雲增長的主要驅動力，我們的客戶將我們視為領先的供應商，因此需求巨大。我們正在努力增加更多的產能，以滿足這一長期需求。同時，我們看到各類客戶對我們的 TPU 系統有強勁需求，我們預計隨著規模擴大，這一需求將持續。我很高興歡迎 Thomas 參加通話；他將分享有關 Cloud 的動能的更多信息。在我們的 AI 模型上，客戶的使用量比以往任何時候都要多。

Sundar Pichai - Alphabet, CEO

(00:02:52)Our first-party model APIs are now processing nearly 32 billion tokens per minute, 40% more than last quarter, and more than four times what we processed a year ago. With Gemini 4 Argon and now Carbon, which we announced today, we are excited to be leading at the frontier. The Gemini app now has 1 billion users and is the fastest-growing product in Google's history. Lastly, we have surpassed 150 million YouTube Music and Premium subscribers globally, including trials. Today, I'll talk more about our AI progress, our global consumer platforms led by Search and YouTube, and then our long-term bets.

(00:02:52)我們的第一方模型 API 現在每分鐘處理近 320 億個標記，比上季度增長了 40%，是去年處理量的四倍多。隨著 Gemini 4 Argon 和今天宣佈的 Carbon，我們很高興能在前沿領域領先。Gemini 應用程序現在擁有 10 億用戶，是 Google 歷史上增長最快的產品。最後，我們的 YouTube Music 和 Premium 訂閱者數量已超過 1.5 億，包括試用用戶。今天，我將更多地談論我們的 AI 進展、由搜索和 YouTube 主導的全球消費者平台，以及我們的長期投資。

Sundar Pichai - Alphabet, CEO

(00:03:34)Starting with AI, highlighted by our Gemini 4 series of frontier models. We are seeing strong demand and usage for Argon, which we launched two weeks ago. In fact, on our developer platforms, usage in the first week ramped up three and a half times faster than Gemini 3, even with capacity constraints. Argon is performing very well in areas like finance, legal, cybersecurity, and science. Developers and cloud customers are already using it for tasks like untangling legacy codebases, accelerating scientific research, robotics, discovering product improvements, and domain-specific tasks requiring in-depth reasoning.

(00:03:34)從 AI 開始，重點是我們的 Gemini 4 系列前沿模型。我們看到 Argon 的需求和使用量強勁，這是我們兩週前推出的。實際上，在我們的開發者平台上，第一週的使用量增長速度是 Gemini 3 的三倍半，即使在產能受限的情況下。Argon 在金融、法律、網絡安全和科學等領域表現非常出色。開發者和雲客戶已經在使用它來處理如解開舊代碼庫、加速科學研究、機器人技術、發現產品改進以及需要深入推理的特定領域任務等任務。

Sundar Pichai - Alphabet, CEO

(00:04:15)And at Google, just in the last few weeks, teams across Chrome, Ads, and Search have used Argon to write and optimize code much faster, shipping 35% more code with 30% fewer errors. Argon is just the start of the Gemini 4 series. Today, we announced Carbon. It's our most advanced frontier model, featuring state-of-the-art cybersecurity and long-horizon capabilities. It ranks number one on the recently updated Artificial Analysis Intelligence Index with a score of 59. Carbon outperforms Argon across nearly every domain, including full-stack web development, multimodal understanding, complex knowledge work, and scientific reasoning.

(00:04:15)在 Google，僅在過去幾週內，Chrome、廣告和搜索團隊已經使用 Argon 更快地編寫和優化代碼，發佈的代碼量增加了 35%，錯誤減少了 30%。Argon 只是 Gemini 4 系列的開始。今天，我們宣佈了 Carbon。這是我們最先進的前沿模型，具備最先進的網絡安全和長期能力。它在最近更新的人工分析智能指數中排名第一，得分為 59。Carbon 在幾乎所有領域的表現都超過 Argon，包括全棧網頁開發、多模態理解、複雜知識工作和科學推理。

Sundar Pichai - Alphabet, CEO

(00:04:56)We are initially rolling out Carbon to a select group of trusted cyber defenders before expanding access; it is currently being tested by the U.S. government. Beyond our Gemini 4 series, yesterday we announced Genie 4, our most advanced world model yet, which simulates safe, realistic, interactive worlds with native audio in real time. This represents an important building block for training robots and agents. This rapid progress highlights our incredible momentum with these models. We have reached a point where AI is helping us increase both training and launch velocity.

(00:04:56)我們最初將 Carbon 推出給一小部分受信任的網絡防禦者，然後再擴大訪問範圍；目前它正在美國政府進行測試。除了我們的雙子星 4 系列，昨天我們還宣佈了 Genie 4，這是我們迄今為止最先進的世界模型，能夠實時模擬安全、真實和互動的世界，並具備原生音頻功能。這代表了訓練機器人和代理的重要基石。這一快速進展突顯了我們在這些模型上的驚人動能。我們已經達到了一個階段，人工智能正在幫助我們提高訓練和啟動的速度。

Sundar Pichai - Alphabet, CEO

(00:05:31)I expect continued acceleration over the next 12 months due to recursive self-improvement (RSI). We have more models coming soon, including more efficient versions designed to help developers optimize for speed and cost, and to deploy across our billions of users. Let's turn to Search, which brings the benefits of AI to more people than any other product in the world. Just like our shift to mobile over a decade ago, we are focused first on building the best search user experience, followed by monetization.

(00:05:31)我預計在接下來的 12 個月中，由於遞歸自我改進（RSI），將持續加速。我們即將推出更多模型，包括設計用於幫助開發者優化速度和成本的更高效版本，並在我們的數十億用戶中進行部署。讓我們來談談搜索，它將人工智能的好處帶給世界上比其他任何產品更多的人。就像十多年前我們轉向移動一樣，我們首先專注於打造最佳的搜索用戶體驗，然後再考慮變現。

Sundar Pichai - Alphabet, CEO

(00:06:01)We are creating one seamless experience across AI overviews, AI mode, and traditional search. People are increasingly using our AI features, searching in entirely new ways. Their queries are longer, personalized, and multimodal. Queries in AI mode are on average three times longer than traditional searches. One in five AI mode searches globally use voice, images, or documents. We have now surpassed 50 billion visual searches per month. One area I'm particularly excited about is helping people get things done—whether booking reservations, filling out forms, scheduling appointments, or making purchases.

(00:06:01)我們正在創造一個無縫的體驗，涵蓋人工智能概述、人工智能模式和傳統搜索。人們越來越多地使用我們的人工智能功能，以全新的方式進行搜索。他們的查詢更長、更個性化，並且是多模態的。在人工智能模式下的查詢平均比傳統搜索長三倍。全球每五個人工智能模式搜索中就有一個使用語音、圖像或文件。我們現在每月的視覺搜索已經超過 500 億次。我特別興奮的一個領域是幫助人們完成任務——無論是預訂、填寫表格、安排約會還是進行購買。

Sundar Pichai - Alphabet, CEO

(00:06:42)As AI becomes more personal and begins taking action for people and businesses, trust and security matter more than ever. People already trust Google to help them accomplish tasks in their daily lives, and our leadership here remains a key differentiator. We are focusing first on perfecting these agentic experiences with the Gemini agent for enterprise, because the standards for security, safety, and governance are extremely high. Next, we are bringing these features to Search and, of course, to our billion users on the Gemini app.

(00:06:42)隨著人工智能變得更加個性化並開始為人們和企業採取行動，信任和安全比以往任何時候都更重要。人們已經信任谷歌幫助他們完成日常生活中的任務，而我們在這方面的領導地位仍然是關鍵的區別因素。我們首先專注於完善這些企業代理體驗，因為安全性、穩定性和治理的標準極高。接下來，我們將這些功能帶入搜索，當然還有我們在雙子星應用上的十億用戶。

Sundar Pichai - Alphabet, CEO

(00:07:16)People are already using the Gemini app to do more than just ask questions. They use it to perform complex, multi-step agentic queries. Our Gemini 4 series has been a significant step forward and is already handling these complex tasks very well. Agents will serve as a unifying layer across our products, so you will see us extend these agentic experiences beyond Search and the Gemini app to Chrome, Android, and our Workspace products as well. Moving on to YouTube, which continues to lead in the living room as the number one streaming platform in the U.S. by watch time.

(00:07:16)人們已經在使用雙子星應用來做的不僅僅是提問。他們用它來執行複雜的多步代理查詢。我們的雙子星 4 系列已經是向前邁進的一大步，並且已經能很好地處理這些複雜任務。代理將作為我們產品之間的統一層，因此您將看到我們將這些代理體驗擴展到搜索和雙子星應用之外，包括 Chrome、Android 和我們的 Workspace 產品。接下來談談 YouTube，它在美國的客廳中繼續作為觀看時間最多的流媒體平台領先。

Sundar Pichai - Alphabet, CEO

(00:07:53)It is also experiencing strong subscriber growth, as I mentioned earlier. At Made on YouTube last month, we announced more than 30 new products and features for creators and viewers. One great example is custom feeds, where you can select a topic you love and receive a tailored video home feed based on that topic. Philipp will provide more details on YouTube. Now, onto our long-term bets. Waymo is on a strong trajectory. We have now driven more than 300 million fully autonomous miles. Waymo currently operates in 15 major U.S.

(00:07:53)正如我之前提到的，它也在經歷強勁的訂閱增長。在上個月的 YouTube 創作大會上，我們宣佈了超過 30 個新產品和功能，供創作者和觀眾使用。一個很好的例子是自定義資訊流，您可以選擇您喜愛的主題，並根據該主題接收量身定製的視頻首頁資訊流。Philipp 將提供更多有關 YouTube 的詳細信息。現在，讓我們談談我們的長期投資。Waymo 正處於強勁的發展軌跡上。我們現在已經駕駛了超過 3 億英里完全自動駕駛的里程。Waymo 目前在美國 15 個主要城市運營，並最近在聖安東尼奧、聖地亞哥和坦帕向公眾開放服務，計劃擴展到加拿大、德國、新加坡和西班牙。

Sundar Pichai - Alphabet, CEO

(00:08:21)cities and recently opened service to the public in San Antonio, San Diego, and Tampa, with planned expansions into Canada, Germany, Singapore, and Spain. We are also seeing strong progress across our other long-term initiatives. Isomorphic is using Argon to support drug design research. We demonstrated magic state cultivation for the first time in quantum computing and launched our TPUs into space with Project Suncatcher. It has been a busy quarter, and there is much more to come. I want to thank our employees for their work this quarter. Over to you, Philipp.

(00:08:21)我們在其他長期計劃上也看到了強勁的進展。Isomorphic 正在使用 Argon 來支持藥物設計研究。我們首次在量子計算中演示了魔法狀態培養，並將我們的 TPU 發射到太空，進行了 Suncatcher 計劃。這是一個繁忙的季度，還有更多的事情即將到來。我想感謝我們的員工在這個季度的努力。接下來請 Philipp。謝謝，Sundar，大家好。

Philipp Schindler - Alphabet, Chief Business Officer

(00:08:59)Thanks, Sundar, and hello, everyone. I'll highlight our momentum across Search, YouTube, and partnerships. We continue to accelerate the deployment of Gemini models across our entire ads infrastructure to boost performance in the three areas I mentioned earlier. These areas are ads quality, advertiser tools, and AI user experiences. First, ads quality. Integrating Gemini models into our ads infrastructure is improving how we decode user intent and enhance ads relevance. For example, in Q3, we made the biggest update to product listing ads in a decade, adding new AI capabilities to help users make more confident purchase decisions.

(00:08:59)我將強調我們在搜索、YouTube 和合作夥伴關係方面的動能。我們繼續加速在整個廣告基礎設施中部署雙子星模型，以提升我之前提到的三個領域的表現。這些領域是廣告質量、廣告主工具和人工智能用戶體驗。首先，廣告質量。將雙子星模型整合到我們的廣告基礎設施中正在改善我們解碼用戶意圖和增強廣告相關性的方式。例如，在第三季度，我們對產品列表廣告進行了十年來最大的更新，增加了新的人工智能功能，幫助用戶做出更有信心的購買決策。我們已經看到滿意度提高了 19%，而不滿意度下降了超過一半。

Philipp Schindler - Alphabet, Chief Business Officer

(00:09:37)Already, we've seen satisfaction increase by 19% and dissatisfaction drop by more than half. We're also seeing AI models enable better prediction of clicks and conversions, stronger overall performance, and deeper personalization of ads. For instance, through several innovations and personalization modeling, we delivered an 18% incremental advertiser value and an 8% increase in user sentiment on Discover ads in 2026. Overall, AI is accelerating our ads development and drives the majority of our product launches.

(00:09:37)我們也看到人工智慧模型能夠更好地預測點擊和轉換，整體表現更強勁，廣告的個性化程度更深。例如，通過幾項創新和個性化建模，我們在 2026 年實現了 18% 的增量廣告商價值，以及 8% 的用戶情感提升。總體而言，人工智慧正在加速我們的廣告開發，並驅動我們大多數產品的推出。

Philipp Schindler - Alphabet, Chief Business Officer

(00:10:10)Second, advertiser tools. We're ensuring every advertiser, from global brands to small businesses, can seamlessly participate in our AI services. AI Max and Performance Max continue to see strong adoption. To give marketers even more control, we're now introducing a way to test different budgets and ROI targets across multiple search campaigns in a single AI Max A-B test. To drive better performance, we integrated our AI creative tools directly into AI Max. In Q3 alone, Gemini Models helped advertisers generate over 1.5 billion campaign assets, up more than 7 times year over year.

(00:10:10)第二，廣告商工具。我們確保每位廣告商，從全球品牌到小型企業，都能無縫參與我們的人工智慧服務。AI Max 和 Performance Max 繼續受到強勁採用。為了給市場營銷人員更多控制權，我們現在推出了一種方法，可以在單一的 AI Max A-B 測試中測試不同的預算和投資回報目標，跨多個搜索活動。為了提高表現，我們將人工智慧創意工具直接整合到 AI Max 中。僅在第三季度，Gemini 模型幫助廣告商生成了超過 15 億個活動資產，同比增長超過 7 倍。

Philipp Schindler - Alphabet, Chief Business Officer

(00:10:50)For example, by using AI Max to capture long-form and multilingual queries, Telstra saw a 52% increase in conversions and a 36% reduction in cost per order, capturing key moments however customers search. Third, AI user experiences. Sundar mentioned the distinct behavioral changes we're seeing as users ask entirely new categories of questions. We think of search holistically, as a unified product spanning traditional results and AI-powered search. On AI mode, many of these queries are upper-funnel, allowing us to capture new high-value commercial intent.

(00:10:50)例如，通過使用 AI Max 捕捉長格式和多語言查詢，Telstra 看到轉換率提高了 52%，每訂單成本降低了 36%，無論客戶如何搜索，都能捕捉到關鍵時刻。第三，人工智慧用戶體驗。Sundar 提到我們看到的明顯行為變化，因為用戶提出了全新的問題類別。我們將搜索視為一個整體，作為一個統一的產品，涵蓋傳統結果和人工智慧驅動的搜索。在人工智慧模式下，許多這些查詢屬於上層漏斗，使我們能夠捕捉到新的高價值商業意圖。

Philipp Schindler - Alphabet, Chief Business Officer

(00:11:27)In addition, the majority of AI mode queries are incremental, creating more monetization opportunities over time. Additionally, our user research shows that people are comfortable with ads on these services and find them highly relevant. Click-through rates are already similar between sponsored and organic product list items within AI mode. Based on these collective results, we're now scaling high-quality ads within AI overviews in AI mode to more users and queries. This includes taking our monetization footprint global for the first time.

(00:11:27)此外，大多數人工智慧模式查詢都是增量的，隨著時間的推移創造了更多的貨幣化機會。此外，我們的用戶研究顯示，人們對這些服務上的廣告感到舒適，並認為它們高度相關。在人工智慧模式下，贊助和自然產品列表項的點擊率已經相似。根據這些綜合結果，我們現在正在將高質量廣告擴展到更多用戶和查詢中，並在人工智慧概覽中使用人工智慧模式。這包括首次將我們的貨幣化足跡擴展到全球。

Philipp Schindler - Alphabet, Chief Business Officer

(00:12:01)Today, we're announcing the expansion of ads in AI mode into international markets alongside multi-turn ads in AI mode in the US. As always, our approach is deliberate, prioritizing the user experience and engagement first, then scaling the formats that drive proven ROI for advertisers. That discipline is what gives us a strong, durable foundation for the future. As Sundar mentioned, we're in the early stages of agentic experiences across our services and excited about the commercial opportunities ahead.

(00:12:01)今天，我們宣佈將人工智慧模式下的廣告擴展到國際市場，同時在美國推出多輪廣告。一如既往，我們的做法是有意識的，首先優先考慮用戶體驗和參與度，然後擴展那些為廣告商帶來證明的投資回報的格式。這種紀律為我們未來提供了堅實而持久的基礎。正如 Sundar 提到的，我們在我們的服務中處於代理體驗的早期階段，對未來的商業機會感到興奮。

Philipp Schindler - Alphabet, Chief Business Officer

(00:12:32)The Universal Commerce Protocol, UCP, is being rapidly adopted by retailers. Today, UCP enables direct checkout for hundreds of thousands of brands across Google, including Chewy, Nike, Tapestry, and many more. We're combining this with the agent payments protocol as we build for the future of commerce. Our focus spans two fronts: unlocking agent capabilities that act on behalf of the consumer and designing practical tools that give shoppers confidence and control. We launched Universal Cart this quarter to support that second track, mirroring how people actually shop today across Google.

(00:12:32)通用商務協議（UCP）正在被零售商迅速採用。今天，UCP 使數十萬個品牌在 Google 上實現直接結帳，包括 Chewy、Nike、Tapestry 等等。我們將這一點與代理支付協議結合起來，為未來的商務建設。我們的重點涵蓋兩個方面：解鎖代表消費者行動的代理能力，以及設計實用工具，讓購物者感到信心和控制。我們在本季度推出了通用購物車，以支持第二條路徑，反映人們今天在 Google 上的實際購物方式。

Philipp Schindler - Alphabet, Chief Business Officer

(00:13:10)This allows retailers to maintain their relationships with customers while helping to capture sales and avoid abandoned carts. Moving to YouTube. We view YouTube as a flywheel of creators, viewers, and advertisers. First, creators. Every day, more than 20 million videos are uploaded. We are making creation easier with new generative AI tools in YouTube Studio, plus a conversational editing assistant in Shorts and YouTube Create. Looking at viewership, at Made on YouTube, we introduced short series where creators can structure their shorts into seasons and episodes.

(00:13:10)這使零售商能夠維持與客戶的關係，同時幫助捕捉銷售並避免購物車放棄。轉到 YouTube。我們將 YouTube 視為創作者、觀眾和廣告商的飛輪。首先，創作者。每天，有超過 2000 萬個視頻被上傳。我們通過 YouTube Studio 中的新生成式人工智慧工具，以及 Shorts 和 YouTube Create 中的對話編輯助手，使創作變得更容易。在觀看方面，在 Made on YouTube 中，我們推出了短系列，讓創作者可以將他們的短片結構化為季和集。

Philipp Schindler - Alphabet, Chief Business Officer

(00:13:47)Sequential playback makes it seamless for viewers to discover new shorts and keep watching. The third part is monetization, starting with ads. Our focus remains on delivering value across the entire funnel. In Q3, Living Room accounted for over one-third of total YouTube ads revenues. Small and medium advertisers continue to adopt Demand Gen, driving direct response. On YouTube, Wayfair's Demand Gen campaigns drove a 2.3 times higher return on ad spend compared with traditional video reach campaigns.

(00:13:47)連續播放使觀眾能夠無縫發現新短片並持續觀看。第三部分是貨幣化，從廣告開始。我們的重點仍然是提供整個漏斗的價值。在第三季度，客廳佔 YouTube 廣告收入的三分之一以上。中小型廣告商繼續採用需求生成，推動直接響應。在 YouTube 上，Wayfair 的需求生成活動相比傳統視頻觸及活動帶來了 2.3 倍的廣告支出回報。

Philipp Schindler - Alphabet, Chief Business Officer

(00:14:20)Our YouTube Creator Partnerships program connects brands with over 3 million creators to drive highly effective campaigns. Brands like Lexus continue to partner with creators like Colin and Samir, driving an 11% lift in brand consideration. More than 1.3 million channels are enrolled in YouTube Shopping globally. Building on this momentum, we recently announced that Amazon is joining YouTube Shopping. US creators can now tag Amazon products directly in their videos, Shorts, and live streams, driving direct conversions.

(00:14:20)我們的 YouTube 創作者夥伴計劃將品牌與超過 300 萬創作者聯繫起來，以推動高效的活動。像 Lexus 這樣的品牌繼續與 Colin 和 Samir 等創作者合作，推動品牌考慮度提升 11%。全球有超過 130 萬個頻道參加 YouTube 購物。在這一勢頭的基礎上，我們最近宣佈亞馬遜將加入 YouTube 購物。美國創作者現在可以在他們的視頻、Shorts 和直播中直接標記亞馬遜產品，推動直接轉換。

Philipp Schindler - Alphabet, Chief Business Officer

(00:14:53)Importantly, our subscription ecosystem continues to complement our advertising business. Subscriptions continue to grow faster than ads, led by YouTube Music Premium. YouTube Premium has more than 150 million subscribers. With Premium and Premium Lite, viewers have choice. As more people opt into Premium experiences, it benefits viewers, creators, and our business alike. Subscriptions help fund the creator ecosystem and drive the platform's progress. Finally, strategic partnerships. Businesses choose Google because we integrate cloud, AI, ads, and consumer platforms to accelerate their transformation.

(00:14:53)重要的是，我們的訂閱生態系統繼續補充我們的廣告業務。訂閱的增長速度繼續超過廣告，以 YouTube Music Premium 為首。YouTube Premium 擁有超過 1.5 億的訂閱者。通過 Premium 和 Premium Lite，觀眾有了選擇。隨著越來越多的人選擇高級體驗，這對觀眾、創作者和我們的業務都有好處。訂閱幫助資助創作者生態系統並推動平台的進步。最後，戰略夥伴關係。企業選擇谷歌，因為我們整合了雲端、人工智慧、廣告和消費者平台，以加速他們的轉型。

Philipp Schindler - Alphabet, Chief Business Officer

(00:15:32)We expanded our partnership with American Express to support the ongoing modernization of their contact centers and collaborate on agentic commerce across Search and Wallet. Unity signed an expanded multi-year cloud deal to utilize our full AI stack to power the next-generation game engine. Our commercial priorities are clear. We are focused on expanding advertiser ROI, systematically advancing AI monetization efficiency, and scaling ecosystem partnerships to support the next era of discovery. In closing, I want to thank Googlers everywhere for their incredible execution, and our customers and partners for their continued trust.

(00:15:32)我們擴大了與美國運通的合作，以支持他們聯絡中心的持續現代化，並在搜尋和錢包上合作進行代理商商務。Unity 簽署了一項擴展的多年雲端協議，以利用我們完整的人工智慧技術堆疊來驅動下一代遊戲引擎。我們的商業優先事項非常明確。我們專注於擴大廣告主的投資回報率，系統性地推進人工智慧的貨幣化效率，並擴大生態系統夥伴關係，以支持下一個發現時代。最後，我想感謝全球的谷歌員工們出色的執行，以及我們的客戶和合作夥伴對我們的持續信任。

Philipp Schindler - Alphabet, Chief Business Officer

(00:16:09)Thomas, over to you.

(00:16:09)Thomas，請你接著說。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:16:12)Our strong momentum is driven by our integrated AI portfolio of chips, models, data, security, and agent platforms, all designed to work seamlessly together. We have strong diversification across our product lines, industries, and customers. Google Cloud now has 18 product lines generating more than $1 billion each in annual revenue. Our growth this quarter was driven by strong performance in Google Cloud Platform, with strength in both AI and Core GCP. TPU system sales were also a significant contributor.

(00:16:12)我們強勁的動能來自於我們整合的人工智慧產品組合，包括晶片、模型、數據、安全性和代理平台，這些都旨在無縫協作。我們在產品線、行業和客戶方面有著強大的多樣化。谷歌雲端目前擁有 18 條產品線，每條年收入均超過 10 億美元。本季度的增長主要得益於谷歌雲端平台的強勁表現，人工智慧和核心 GCP 的表現均強勁。TPU 系統的銷售也是一個重要的貢獻者。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:16:56)Over the past 12 months, more than 2,600 customers have used at least 100 billion tokens, and nearly 75 have used at least 10 trillion tokens on Gemini Enterprise. Nearly 90% of the Fortune 100 use Gemini Enterprise, and customers using it in Q3 2025 more than quadrupled their spend year over year. One important use case for customers is knowledge work with Gemini. Across all our AI services, we have sold more than 100 million seats for knowledge work. Gemini Enterprise now includes the Gemini agent, deeply integrated across our entire cloud portfolio.

(00:16:56)在過去的 12 個月中，超過 2600 個客戶使用了至少 1000 億個代幣，近 75 個客戶在 Gemini Enterprise 上使用了至少 10 萬億個代幣。幾乎 90% 的《財富》100 強企業使用 Gemini Enterprise，而在 2025 年第三季度使用它的客戶年支出增長了四倍以上。對客戶來說，一個重要的使用案例是使用 Gemini 進行知識工作。在我們所有的人工智慧服務中，我們已經銷售了超過 1 億個知識工作席位。Gemini Enterprise 現在包括 Gemini 代理，深度整合在我們整個雲端產品組閤中。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:17:53)It can research and answer questions, write and run code, and automate process flows with one agent, one API, and a single prompt window. It provides companies with strong security and cost controls, along with the widest choice of frontier models. Customers include BNP Paribas, deploying Gemini Enterprise to more than 65,000 employees to streamline workflows while maintaining strict enterprise-grade security and data governance. Tata Steel deployed over 300 specialized AI agents in just nine months to enhance manufacturing safety, optimize supply chains, and reduce service times for customers by 50%.

(00:17:53)它可以研究和回答問題，編寫和運行代碼，並通過一個代理、一個 API 和一個單一的提示窗口自動化流程。它為公司提供強大的安全性和成本控制，以及最廣泛的前沿模型選擇。客戶包括法國巴黎銀行，將 Gemini Enterprise 部署給超過 65,000 名員工，以簡化工作流程，同時保持嚴格的企業級安全性和數據治理。塔塔鋼鐵在短短九個月內部署了超過 300 個專業人工智慧代理，以提高製造安全性、優化供應鏈，並將客戶的服務時間減少 50%。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:18:46)Ulta Beauty built the Ulta AI Shopping Assistant, resulting in a threefold higher conversion rate compared to traditional browsing on their site. Gemini is also accelerating our database, analytics, and cybersecurity products. Gemini now helps users find, analyze, and visualize information from both structured and unstructured data. Gemini leverages our Knowledge Catalog to locate the right sources and definitions, improving accuracy, lowering token costs, and making responses more reliable. Gemini uses our Lakehouse to access information from multiple systems without copying data.

(00:18:46)Ulta Beauty 建立了 Ulta AI 購物助手，與傳統瀏覽相比，轉換率提高了三倍。Gemini 也在加速我們的數據庫、分析和網絡安全產品。Gemini 現在幫助用戶從結構化和非結構化數據中查找、分析和可視化信息。Gemini 利用我們的知識目錄來定位正確的來源和定義，提高準確性，降低代幣成本，並使回應更可靠。Gemini 使用我們的 Lakehouse 從多個系統訪問信息，而無需複製數據。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:19:37)To support agentic analysis, BigQuery improved query speed by over 40% year-over-year while reducing query processing costs by more than 50% year-over-year. Customers include Deutsche Telekom, which uses BigQuery, Cloud Storage, and Knowledge Catalog to build a unified data lakehouse across 40-plus legacy on-premise systems. Etsy migrated three petabytes of data to the Unified Lakehouse. By using BigQuery and Iceberg, they now perform critical data joins up to 60% faster. The growing risk of cybersecurity threats is driving strong demand for our AI-powered security platform.

(00:19:37)為了支持代理分析，BigQuery 在查詢速度上同比提高了超過 40%，同時查詢處理成本同比降低了超過 50%。客戶包括德國電信，該公司使用 BigQuery、雲端存儲和知識目錄，在 40 多個舊有的本地系統之間建立統一的數據湖。Etsy 將三個 PB 的數據遷移到統一的 Lakehouse。通過使用 BigQuery 和 Iceberg，他們現在的關鍵數據聯接速度提高了 60%。網絡安全威脅的風險日益增加，推動了對我們的人工智慧驅動安全平台的強勁需求。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:20:30)Our Gemini Flash cyber models are frontier models that automate vulnerability discovery across 20 programming languages and patch 2.6 times more reliably than other models. Even better performance is expected as we roll out Argon and Carbon. Our security platform integrates threat intelligence, cyber response prioritization, automated scanning, code remediation, and monitoring. Gemini and Google Threat Intelligence autonomously analyze millions of external events daily with 98% accuracy. More than 95% of Wiz customers are adopting AI-powered features, and growth continues.

(00:20:30)我們的 Gemini Flash 網絡模型是前沿模型，能夠自動化發現 20 種編程語言中的漏洞，並且修補的可靠性比其他模型高出 2.6 倍。隨著我們推出 Argon 和 Carbon，預期會有更好的性能。我們的安全平台整合了威脅情報、網絡響應優先級、自動掃描、代碼修復和監控。Gemini 和谷歌威脅情報每天自主分析數百萬個外部事件，準確率達到 98%。超過 95% 的 Wiz 客戶正在採用人工智慧驅動的功能，增長持續。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:21:23)These security tools protect critical infrastructure, including financial institutions like Banorte and Jack Henry, and telecommunications providers such as AT&T. All of this is powered by our leading AI infrastructure, spanning TPU systems, Axion CPUs, the latest NVIDIA GPUs, and our agent compute engine, which is specifically designed for AI agents. TPU systems are specialized platforms built for AI, combining hardware, compute, storage, and networking with optimized software for training and inference.

(00:21:23)這些安全工具保護關鍵基礎設施，包括像 Banorte 和 Jack Henry 這樣的金融機構，以及 AT&T 等電信提供商。所有這些都由我們領先的人工智慧基礎設施提供支持，涵蓋 TPU 系統、Axion CPU、最新的 NVIDIA GPU 和我們的代理計算引擎，專為人工智慧代理設計。TPU 系統是專為人工智慧構建的專用平台，結合了硬體、計算、存儲和網絡，並配備優化的軟體以進行訓練和推理。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:22:10)TPU 8T and 8i offer up to 2.7 times better training performance, and 80% better inference price performance, respectively, compared to our previous generation of TPUs. To simplify workload porting to TPUs, we recently collaborated with Meta to deliver TorchTPU, enabling users to train, serve, and deploy PyTorch models on TPUs using familiar tools. Nearly 100 customers use TPUs today, and we are seeing strong and growing demand from many others. In addition to our work with Anthropic, both directly and via Broadcom, and public reference customers like Apple, who are scaling pre-training of their foundation models on the latest generation of TPUs,

(00:22:10)TPU 8T 和 8i 提供高達 2.7 倍的訓練性能，與我們前一代 TPU 相比，推理價格性能提高了 80%。為了簡化工作負載移植到 TPU，我們最近與 Meta 合作推出了 TorchTPU，使得用戶能夠使用熟悉的工具在 TPU 上訓練、服務和部署 PyTorch 模型。現在幾乎有 100 個客戶使用 TPU，我們看到來自許多其他客戶的強勁和增長需求。除了我們與 Anthropic 的合作，無論是直接還是通過 Broadcom，以及像 Apple 這樣的公共參考客戶，他們正在最新一代 TPU 上擴大基礎模型的預訓練，

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:23:14)we are also working with new frontier model companies, including Mirendil and Physical Intelligence; capital markets firms such as Citadel Securities and Jump Trading; universities like Caltech; and public sector service providers including Cassava and Cirrascale. Given continued strong and exhilarating demand, we have recently entered partnerships to deliver TPU systems with both Crux AI and Oracle, with more collaborations underway. Over to you, Anat.

(00:23:14)我們還在與新的前沿模型公司合作，包括 Mirendil 和 Physical Intelligence；資本市場公司如 Citadel Securities 和 Jump Trading；加州理工學院等大學；以及包括 Cassava 和 Cirrascale 在內的公共部門服務提供商。鑑於持續強勁和令人振奮的需求，我們最近與 Crux AI 和 Oracle 建立了合作夥伴關係，以提供 TPU 系統，並且還有更多合作正在進行中。請你，Anat。

Anat Ashkenazi - Alphabet, CFO

(00:23:54)Thank you, Thomas. My comments will focus on year-over-year comparisons for the third quarter unless I specify otherwise. I will begin with the results at the Alphabet level and then cover our segment results. I'll conclude with some commentary on our outlook for the fourth quarter and the full year 2026. We had a strong third quarter, delivering our 13th consecutive quarter of double-digit revenue growth. Consolidated revenues were $127.6 billion, up 25% in both reported and constant currency. Total cost of revenue was $51.4 billion, an increase of 24%.

(00:23:54)謝謝你，Thomas。我的評論將專注於第三季度的年比年比較，除非我另有說明。我將從 Alphabet 的整體結果開始，然後涵蓋我們的各個部門結果。我將以對第四季度和 2026 年全年展望的一些評論作結。我們在第三季度表現強勁，實現了第 13 個連續季度的雙位數收入增長。綜合收入為 1276 億美元，按報告和固定貨幣計算均增長 25%。總收入成本為 514 億美元，增長 24%。

Anat Ashkenazi - Alphabet, CFO

(00:24:33)Traffic acquisition costs (TAC) were $16.4 billion, up 10%. Other cost of revenues totaled $35.1 billion, up 32%, driven primarily by hardware costs related to TPU system sales, increased depreciation, content acquisition costs, and technical infrastructure operations expenses. Total operating expenses rose 20% to $35.6 billion. Research and development expenses increased by 36%, driven by compensation primarily for AI talent investments, as well as higher depreciation. Sales and marketing expenses grew 22% due to investments in Gemini and increased compensation.

(00:24:33)流量獲取成本（TAC）為 164 億美元，增長 10%。其他收入成本總計 351 億美元，增長 32%，主要受到與 TPU 系統銷售相關的硬體成本、折舊增加、內容獲取成本和技術基礎設施運營費用的推動。總營運費用增長 20%，達到 356 億美元。研究和開發費用增長 36%，主要受到對 AI 人才投資的補償以及更高的折舊推動。銷售和市場營銷費用增長 22%，因為對 Gemini 的投資和補償增加。

Anat Ashkenazi - Alphabet, CFO

(00:25:19)General and administrative expenses decreased 17%, mainly due to lapping higher legal expenses from Q3 last year, partially offset by current year legal charges. Operating income increased 30% to $40.6 billion, with an operating margin of 31.8%. Other income and expenses were negative $8.2 billion, primarily due to an unrealized loss from our investment in SpaceX following a decline in its stock price during the third quarter. Net income and earnings per share decreased due to the decline in other income and expenses.

(00:25:19)一般和行政費用減少 17%，主要是因為與去年第三季度較高的法律費用相比，部分被當年法律費用抵消。營運收入增長 30%，達到 406 億美元，營運利潤率為 31.8%。其他收入和費用為負 82 億美元，主要是因為我們對 SpaceX 的投資在第三季度股價下跌後產生的未實現損失。淨收入和每股收益因其他收入和費用的下降而減少。

Anat Ashkenazi - Alphabet, CFO

(00:25:56)We generated strong operating cash flow of $42.3 billion in the third quarter and $179.6 billion over the trailing 12 months. Operating cash flow was negatively impacted by inventory buildup for TPU system sales and a $5.2 billion cash payment for a 2018 European Commission fine following the conclusion of the appeal process. Capital expenditures were $51.4 billion in the third quarter, with the vast majority spent on technical infrastructure to support our AI investments. Approximately 65% of our technical infrastructure investments in Q3 were related to servers and networking equipment, which generally have useful lives of six years.

(00:25:56)我們在第三季度產生了 423 億美元的強勁營運現金流，過去 12 個月的總額為 1796 億美元。營運現金流受到 TPU 系統銷售的庫存積壓和 52 億美元的現金支付（用於 2018 年歐洲委員會的罰款，隨著上訴程序的結束）負面影響。第三季度的資本支出為 514 億美元，絕大多數用於支持我們的 AI 投資的技術基礎設施。我們在第三季度的技術基礎設施投資中，大約 65% 與伺服器和網絡設備相關，這些設備的使用壽命通常為六年。

Anat Ashkenazi - Alphabet, CFO

(00:26:45)The remaining 35% was for data centers, which have useful lives ranging from seven to 40 years. We had negative free cash flow of $9.1 billion in the third quarter, driven by our capital investments. Free cash flow in the third quarter was also affected by the inventory buildup to support TPU system sales and the payment of the European Commission fine. Free cash flow was $19.7 billion over the trailing 12 months. We ended the quarter with $265.1 billion in cash and marketable securities, including $90.4 billion in marketable equity securities.

(00:26:45)剩餘的 35% 用於數據中心，其使用壽命範圍從七年到 40 年不等。我們在第三季度的自由現金流為負 91 億美元，主要是由於我們的資本投資。第三季度的自由現金流也受到支持 TPU 系統銷售的庫存積壓和支付歐洲委員會罰款的影響。過去 12 個月的自由現金流為 197 億美元。我們在本季度結束時擁有 2651 億美元的現金和可交易證券，其中包括 904 億美元的可交易股權證券。

Anat Ashkenazi - Alphabet, CFO

(00:27:25)Long-term debt stood at $125.1 billion. We have also entered into multi-year agreements to secure capacity for technical infrastructure and inventory, as well as leases and energy. At the end of the third quarter, we had future purchase commitments of $924 billion and pre-commencement leases and other obligations totaling $152 billion. These amounts represent the maximum contractual obligations. However, many of these agreements include variable pricing and volume flexibility, which could result in lower actual spending.

(00:27:25)長期債務為 1251 億美元。我們還簽訂了多年協議，以確保技術基礎設施和庫存的容量，以及租賃和能源。在第三季度結束時，我們的未來購買承諾為 9240 億美元，預租賃和其他義務總計 1520 億美元。這些金額代表最大合同義務。然而，許多這些協議包括可變定價和數量靈活性，這可能導致實際支出降低。

Anat Ashkenazi - Alphabet, CFO

(00:28:01)As mentioned in our press release, our Board of Directors declared a quarterly cash dividend on our common stock of $0.22 per share, payable in December. Turning to segment results, Google Services revenues increased 11% to $96.3 billion, reflecting strong growth in search and subscriptions. Total advertising revenues rose 11%, with growth across all major verticals. As we noted in our outlook last quarter, foreign exchange was a slight headwind on year-over-year growth rates and had a more significant impact sequentially.

(00:28:01)正如我們在新聞稿中提到的，我們的董事會宣佈每股普通股現金股息為 0.22 美元，將於 12 月支付。轉向各部門結果，Google Services 的收入增長 11%，達到 963 億美元，反映出搜索和訂閱的強勁增長。總廣告收入增長 11%，所有主要垂直領域均有增長。正如我們在上季度的展望中提到的，外匯對年比年增長率造成了輕微的阻力，並在環比上有更顯著的影響。

Anat Ashkenazi - Alphabet, CFO

(00:28:39)Google Search and other advertising revenues increased 12% to $63.3 billion, with retail contributing the most, followed by finance. YouTube advertising revenues grew 9% to $11.2 billion, driven primarily by brand advertising and notable strength in living room engagement. Network advertising revenues were $7.5 billion, up 1%. Subscription platforms and devices revenues increased 11% this quarter to $14.3 billion, driven by strong growth in YouTube subscriptions—particularly YouTube Music and Premium—and strong demand for AI plans in Google One.

(00:28:39)Google Search 和其他廣告收入增長 12%，達到 633 億美元，零售貢獻最大，其次是金融。YouTube 廣告收入增長 9%，達到 112 億美元，主要受到品牌廣告和客廳參與度顯著增強的推動。網絡廣告收入為 75 億美元，增長 1%。訂閱平台和設備的收入在本季度增長 11%，達到 143 億美元，主要受到 YouTube 訂閱（特別是 YouTube Music 和 Premium）和 Google One 的 AI 計劃需求強勁的推動。

Anat Ashkenazi - Alphabet, CFO

(00:29:22)Google Services operating income increased 17% to $39.2 billion, with an operating margin of 40.7%. The Google Cloud segment delivered record growth in the third quarter. Cloud revenues were up 103% to $30.7 billion, driven by strong growth in Google Cloud Platform (GCP), including AI infrastructure, AI solutions, and core GCP services. TPU system sales were also a significant contributor. GCP accounted for the vast majority of revenue generated in Q3. Cloud revenue growth, excluding the impact of TPU systems, remained exceptionally strong.

(00:29:22)Google Services 的營運收入增長 17%，達到 392 億美元，營運利潤率為 40.7%。Google Cloud 部門在第三季度實現了創紀錄的增長。雲端收入增長 103%，達到 307 億美元，主要受到 Google Cloud Platform（GCP）強勁增長的推動，包括 AI 基礎設施、AI 解決方案和核心 GCP 服務。TPU 系統銷售也是一個重要貢獻者。GCP 佔第三季度產生的收入的絕大多數。除去 TPU 系統的影響，雲端收入增長依然非常強勁。

Anat Ashkenazi - Alphabet, CFO

(00:30:06)Cloud operating income was $11.1 billion, more than tripling year over year. Operating margin increased from 23.7% in the third quarter last year to 36.2%. The third-party bridging capacity, which we discussed on our prior call, came online only at the very end of Q3 and therefore had limited negative impact on our margin this quarter. Google Cloud's backlog was $561 billion in the third quarter, with the majority related to typical GCP contracts. We expect to recognize just over 55% of the total backlog as revenue over the next 24 months.

(00:30:06)雲端營運收入為 111 億美元，同比增長超過三倍。營運利潤率從去年第三季度的 23.7% 增至 36.2%。我們在之前的電話中討論的第三方橋接能力，僅在第三季度末才上線，因此對本季度的利潤率影響有限。Google Cloud 的未來訂單在第三季度為 5610 億美元，其中大部分與典型的 GCP 合同相關。我們預計在接下來的 24 個月內將超過 55% 的總未來訂單確認為收入。

Anat Ashkenazi - Alphabet, CFO

(00:30:49)In Other Bets, revenues were $372 million, and the operating loss was $2 billion as we continue to expand Waymo's business and invest in key Other Bets. At the Alphabet level, the operating loss was $7.7 billion, driven by shared AI R&D expenses. I will now turn to our outlook for the remainder of 2026. First, regarding revenues, at current spot rates, we expect slightly larger FX headwinds to our consolidated revenues in Q4 compared to the prior quarter. This impact will be primarily seen in Search and YouTube ads.

(00:30:49)在其他業務方面，收入為 3.72 億美元，營運虧損為 20 億美元，因為我們持續擴展 Waymo 的業務並投資於關鍵的其他業務。在 Alphabet 層面，營運虧損為 77 億美元，主要是由於共享的人工智慧研發費用。現在我將轉向我們對 2026 年剩餘時間的展望。首先，關於收入，根據當前的即期匯率，我們預計第四季度的合併收入將面臨略大的外匯逆風，與前一季度相比。這一影響主要體現在搜尋和 YouTube 廣告上。

Anat Ashkenazi - Alphabet, CFO

(00:31:28)In Google Services, we will continue to lap the search performance acceleration that began in Q3 last year, which will pressure revenue growth in Q4. In Google Cloud, we expect TPU system sales to continue scaling in Q4, followed by a meaningful revenue ramp in 2027. Regarding cloud margins, given the supply-constrained environment, we began utilizing third-party compute capacity at the end of Q3 and will ramp up usage in Q4. We expect margin headwinds in the near term as we utilize this capacity. Moving to investments, our full-year 2026 CAPEX guidance range of $19.5 to $20.5 billion remains unchanged.

(00:31:28)在 Google 服務方面，我們將繼續面對去年第三季度開始的搜尋表現加速，這將對第四季度的收入增長造成壓力。在 Google Cloud 方面，我們預計 TPU 系統的銷售將在第四季度持續增長，並在 2027 年實現顯著的收入增長。關於雲端利潤率，鑑於供應受限的環境，我們在第三季度末開始利用第三方計算能力，並將在第四季度增加使用。我們預計在短期內使用這些能力會面臨利潤率的逆風。轉向投資，我們對 2026 年的全年資本支出指導範圍為 195 億至 205 億美元，保持不變。

Anat Ashkenazi - Alphabet, CFO

(00:32:12)We continue to expect CAPEX to increase significantly in 2027 and will provide further details on our Q4 call. In terms of expenses, the significant increase in our investment in technical infrastructure will continue to weigh on the P&L through higher depreciation and related data center operating costs, such as energy. Before we move to Q&A, let me share a few thoughts on how we view return on invested capital and free cash flow. We are uniquely positioned to monetize our AI investments across many products that reach billions of users and enterprises.

(00:32:12)我們仍然預計 2027 年的資本支出將顯著增加，並將在第四季度的電話會議中提供進一步的細節。在支出方面，我們在技術基礎設施上的投資大幅增加，將繼續通過更高的折舊和相關的數據中心運營成本（如能源）對損益表造成壓力。在我們進入問答環節之前，讓我分享一些關於我們如何看待投資回報率和自由現金流的想法。我們在許多產品中獨特地定位於將我們的人工智慧投資變現，這些產品觸及數十億用戶和企業。

Anat Ashkenazi - Alphabet, CFO

(00:32:49)Our AI investments in Search are driving our highest query growth in half a decade. Google Cloud has generated tens of billions of dollars in AI revenues year-to-date, and Google One is our fastest-growing subscription product, driven by demand for AI plans. Throughout this investment cycle, we expect to generate an attractive ROIC that remains meaningfully above our cost of capital and anticipate strong cash flow from operations. Our sizable investments in technical infrastructure and inventory build of TPU systems as we scale that business will put pressure on free cash flow.

(00:32:49)我們在搜尋方面的人工智慧投資正在推動我們五年來最高的查詢增長。截至目前，Google Cloud 已經產生了數百億美元的人工智慧收入，而 Google One 是我們增長最快的訂閱產品，這是由於對人工智慧計劃的需求。在這一投資週期中，我們預計將產生具有吸引力的投資回報率，並且將顯著高於我們的資本成本，預期運營現金流強勁。我們在技術基礎設施和 TPU 系統庫存上的大量投資，隨著業務的擴展，將對自由現金流造成壓力。

Anat Ashkenazi - Alphabet, CFO

(00:33:27)Q3 was a strong quarter, highlighted by our innovation and disciplined execution across all areas of the business, including the launch of Gemini 4. I want to take this opportunity to thank our employees for their contributions to our performance. We will now open the call for Q&A.

(00:33:27)第三季度是一個強勁的季度，突顯了我們在所有業務領域的創新和嚴謹執行，包括 Gemini 4 的推出。我想藉此機會感謝我們的員工對我們表現的貢獻。現在我們將開放問答環節。

Operator - Operator

(00:33:47)Thank you. As a reminder, to ask a question, please press star 1 on your telephone. To prevent background noise, please mute your line once your question has been stated. Our first question comes from Doug Anmuth with JPMorgan. Your line is now open.

(00:33:47)謝謝。提醒一下，若要提問，請在您的電話上按星號 1。為了防止背景噪音，請在您提出問題後靜音您的電話。我們的第一個問題來自 JPMorgan 的 Doug Anmuth。您的電話線現在已開放。

Q&A Doug Anmuth - JPMorgan, Analyst

(00:34:10)Thanks for taking the questions, one for Sundar and one for Philipp. Sundar, on a recent podcast, you said that 2027 could be the golden age of productivity for Google because of the new class of models and how they'll integrate with your other products. I was hoping you could elaborate on that and share more about how you view the opportunity ahead. And then, Philipp, just on query growth—the fastest in five years—but with some deceleration in search growth. Can you help us understand the disconnect there?

(00:34:10)感謝您接受提問，這裡有一個問題給 Sundar，還有一個給 Philipp。Sundar，在最近的一個播客中，您提到 2027 年可能是 Google 的生產力黃金時代，因為新一類模型及其如何與您的其他產品整合。我希望您能詳細說明一下，並分享更多關於您如何看待未來的機會。然後，Philipp，關於查詢增長——五年來最快——但搜尋增長卻有所減緩。您能幫助我們理解這裡的脫節嗎？

Doug Anmuth - JPMorgan, Analyst

(00:34:42)Is it simply that fewer ads are being shown in AI overviews and AI mode, which is now becoming a larger percentage of overall queries?

(00:34:42)是否只是因為在人工智慧概述和人工智慧模式中顯示的廣告較少，而這些現在成為整體查詢的更大比例？

(00:34:51)Thanks. Thanks, Doug. We definitely started seeing a step change with 3.8 Flash internally, as our models improved at long-horizon tasks, agent decoding, and so forth. But I see a huge inflection point once we started deploying Argon internally. It is one of the most meaningful internal productivity levers we've seen in software engineering, directly tracking high-quality code check-ins, the rate of check-ins, and reducing rollbacks. More importantly, people are tackling... ...massive problems, and we are beginning to see innovation on the product and consumer experience side from deploying those models.

(00:34:51)謝謝。謝謝，Doug。我們確實開始看到內部 3.8 Flash 的步驟變化，因為我們的模型在長期任務、代理解碼等方面有所改善。但我看到一個巨大的轉折點，當我們開始在內部部署 Argon 時。這是我們在軟體工程中看到的最有意義的內部生產力槓桿之一，直接追蹤高質量的代碼提交、提交的速度，並減少回滾。更重要的是，人們正在解決......巨大的問題，我們開始看到從部署這些模型中產生的產品和消費者體驗的創新。

Sundar Pichai - Alphabet, CEO

(00:35:57)All this happened only about four weeks ago. We are now deploying Carbon internally, which is a meaningful step up from Argon. As we roll it out and make it widely available to Googlers with more compute, I expect—this is what I referred to—and I'm even more excited since I said that in the podcast. I view a golden year ahead in terms of increased productivity and innovation. Philipp?

(00:35:57)所有這一切發生在大約四週前。我們現在正在內部部署 Carbon，這是比 Argon 更有意義的一步。隨著我們推廣並使其廣泛提供給 Google 員工，擁有更多計算能力，我預期——這就是我所提到的——自從我在播客中提到這一點以來，我更感到興奮。我認為未來一年將是生產力和創新的黃金年。Philipp？

Philipp Schindler - Alphabet, Chief Business Officer

(00:36:26)Yeah, Doug, on the second part of your question, in Q3, Search grew 12% year over year to $63 billion. When looking at the change in growth rate from Q2 to Q3, there were a few drivers. First, FX was the single largest driver of search deceleration relative to Q2. Second, there were vertical-specific dynamics. In retail, some of our APAC advertisers were impacted by the EU eliminating its de minimis exemption. This change affected advertising performance in that region. In media and entertainment, spending normalized after an unseasonably strong Q2, driven by the FIFA World Cup, which featured more matches and consequently more advertising budgets allocated in June than in July.

(00:36:26)是的，Doug，關於您問題的第二部分，在第三季度，搜尋年增長 12%，達到 630 億美元營收。在查看第二季度到第三季度的增長率變化時，有幾個驅動因素。首先，外匯是相對於第二季度搜尋減速的最大驅動因素。其次，存在特定垂直領域的動態。在零售方面，我們的一些亞太廣告客戶受到歐盟取消其最低免稅額的影響。這一變化影響了該地區的廣告表現。在媒體和娛樂領域，支出在一個異常強勁的第二季度後恢復正常，這是由於國際足總世界盃的推動，該賽事在六月的比賽數量更多，因此在六月分配的廣告預算也比七月多。

Philipp Schindler - Alphabet, Chief Business Officer

(00:37:07)That also relates to the next part of your question. It's worth noting that the strong growth in AI mode queries impacted results due to the lower monetization we observed in AI mode. To be clear, we feel very confident about our trajectory here. Our ad experiments have shown promising results, and we are now scaling ads within both AI overviews and AI mode to more users and queries, including, as I mentioned, international markets and multi-turn ads in the U.S.

(00:37:07)這也與您問題的下一部分有關。值得注意的是，AI 模式查詢的強勁增長影響了結果，因為我們觀察到 AI 模式的變現能力較低。明確來說，我們對我們的發展路徑非常有信心。我們的廣告實驗顯示出良好的結果，現在我們正在將廣告擴展到更多用戶和查詢中，包括我提到的國際市場和美國的多輪廣告。

Operator - Operator

(00:37:38)Our next question comes from Brian Nowak with Morgan Stanley. Your line is now open.

(00:37:38)我們的下一個問題來自摩根士丹利的布萊恩·諾瓦克。您的線路現在開放。

Brian Nowak - Morgan Stanley, Analyst

(00:37:46)Thank you for taking my questions. I have two questions. Let me start with agentic. You rolled out Spark to some users throughout the year. I know, Sundar, you've spoken extensively about shipping more agentic products to a broader user base. Can you help us understand some of the key technological hurdles involved that you need to overcome to ship agentic products to the full user base? Is it model efficiency? Is it capacity? Is it the readiness of the technologies? What technological steps do we need to take to achieve that?

(00:37:46)感謝您回答我的問題。我有兩個問題。讓我先從代理產品開始。您在整個年度中向一些用戶推出了 Spark。我知道，蘇達爾，您已經廣泛談論將更多代理產品推向更廣泛的用戶基礎。您能幫助我們理解一些關鍵的技術障礙嗎？您需要克服這些障礙才能將代理產品推向全體用戶基礎？是模型效率嗎？是容量嗎？是技術的準備程度嗎？我們需要採取哪些技術步驟來實現這一目標？

Brian Nowak - Morgan Stanley, Analyst

(00:38:16)And then, secondly, regarding Doug's question on search, Anat, can you help us better understand the search performance comparison factor you mentioned as impacting search? How should we think about that as we approach Q4 and Q1, especially when considering the comparisons? Thank you.

(00:38:16)然後，第二，關於道格的搜索問題，阿納特，您能幫助我們更好地理解您提到的影響搜索的搜索性能比較因素嗎？當我們接近第四季度和第一季度時，尤其是在考慮比較時，我們應該如何看待這一點？謝謝。

Sundar Pichai - Alphabet, CEO

(00:38:34)Thanks, Brian. First, on agents, this is an exciting and dynamic area. When I look at our new models and see their ability to handle complex tasks automatically and execute them with very high quality, it's clear that the agentic capabilities we've envisioned will become real, and I couldn't be more excited to bring next-generation agentic experiences to our users. To address your question, I would highlight a few points. The Gemini 4 model series has been particularly important. We are developing distilled versions of our carbon models to deploy efficiently at a billion-user scale.

(00:38:34)謝謝，布萊恩。首先，關於代理，這是一個令人興奮和充滿活力的領域。當我查看我們的新模型，看到它們自動處理複雜任務並以非常高的質量執行時，很明顯，我們所設想的代理能力將變為現實，我對將下一代代理體驗帶給我們的用戶感到無比興奮。為了回答您的問題，我想強調幾個要點。Gemini 4 模型系列特別重要。我們正在開發我們的 Carbon 模型的精簡版本，以便在十億用戶的規模上高效部署。

Sundar Pichai - Alphabet, CEO

(00:39:25)This applies to compute, latency, and quality considerations. That has been a key area of investment. It's clear that having a capable model at scale is necessary but not sufficient. You also need a robust harness. You need a highly secure runtime, and this is where Google Cloud Platform (GCP) plays a critical role. GCP provides each user with a secure runtime, 24/7 cloud availability, and the right ecosystem of skills and connectors. All of this forms the foundation. We have been working hard and decided to roll out the technology to enterprise first, which we announced with our Gemini agent for enterprise.

(00:39:25)這適用於計算、延遲和質量考量。這一直是投資的關鍵領域。很明顯，擁有一個可擴展的強大模型是必要的，但並不足夠。您還需要一個穩健的框架。您需要一個高度安全的運行環境，而這正是 Google Cloud Platform （GCP）發揮關鍵作用的地方。GCP 為每位用戶提供安全的運行環境、24/7 的雲可用性，以及合適的技能和連接器生態系統。所有這些構成了基礎。我們一直在努力工作，並決定首先將技術推向企業，這是我們與企業的 Gemini 代理一起宣佈的。

Sundar Pichai - Alphabet, CEO

(00:40:06)This is because enterprises have the highest standards for security, governance, and controls. The rollout has been well-received, and we are in the process of expanding it. Later this year, we will bring the same experience to consumers, starting with the 1 billion users of our Gemini app. We already see users in the app moving beyond chat to what we call task and agent-like experiences. So, the trust users have in our products and the fact that all their context is readily available, I believe, positions us well.

(00:40:06)這是因為企業對安全性、治理和控制的標準最高。推出得到了良好的反響，我們正在擴大這一範圍。在今年晚些時候，我們將把相同的體驗帶給消費者，首先是我們的 Gemini 應用的十億用戶。我們已經看到應用中的用戶超越了聊天，進入我們所稱的任務和代理類似的體驗。因此，用戶對我們產品的信任以及他們所有的上下文隨時可用的事實，我相信，讓我們的定位非常良好。作為一家公司，我們專注於在我們的產品中提供統一的體驗。

Sundar Pichai - Alphabet, CEO

(00:40:42)As a company, we are focused on delivering a unified experience across our products. You should expect to see significant progress later this year.

(00:40:42)您應該期待在今年晚些時候看到顯著的進展。謝謝。

Anat Ashkenazi - Alphabet, CFO

(00:40:56)Thank you. Regarding your second question about my comments, I believe you are referring to the outlook for search. There are a couple of dynamics I mentioned. One is the sequential impact of FX headwinds, which we also observed this quarter compared to the previous quarter. Given the current spot rates, we expect this to present a modest headwind to our Q4 results. When we examine the year-over-year comparisons, This quarter, we began to lap the acceleration in search growth that occurred in late 2025.

(00:40:56)關於您第二個問題的評論，我相信您是指搜索的前景。我提到過幾個動態。一個是外匯逆風的連續影響，我們在本季度也觀察到了這一點，與上一季度相比。根據目前的即期匯率，我們預期這將對我們的第四季度業績造成輕微的阻力。當我們檢視年對年的比較時，本季度，我們開始比較 2025 年底出現的搜尋增長加速。

Anat Ashkenazi - Alphabet, CFO

(00:41:31)We saw about three points of growth compared to Q3, and we anticipate continuing to lap that growth acceleration in Q4 by a couple of points. Great.

(00:41:31)與第三季度相比，我們看到約三個百分點的增長，我們預期在第四季度將繼續以幾個百分點的速度增長。很好。

Brian Nowak - Morgan Stanley, Analyst

(00:41:46)Thank you both.

(00:41:46)謝謝你們兩位。

Operator - Operator

(00:41:53)Our next question comes from Eric Sheridan of Goldman Sachs. Your line is now open.

(00:41:53)下一個問題來自高盛的埃裡克·謝裡丹。您的線路現在開放。

Eric Sheridan - Goldman Sachs, Analyst

(00:42:01)Thank you for taking my question. Sundar, based on your recent comments about compute constraints, could you please refresh us on your key priorities for allocating compute resources as they become available across the organization? And secondly, perhaps for you or Thomas, how is your perspective evolving regarding the application of compute to meet Google Cloud customer needs relative to external demands? Thanks so

(00:42:01)謝謝您回答我的問題。Sundar，根據您最近關於計算資源限制的評論，您能否告訴我們在整個組織中分配計算資源的主要優先事項？其次，或許是您或託馬斯，您對於計算資源的應用以滿足 Google Cloud 客戶需求相對於外部需求的看法是如何演變的？謝謝你，

Sundar Pichai - Alphabet, CEO

(00:42:32)much. Thank you, Eric. I'll begin, and then Thomas will address the second part. Certainly, compute allocation is an area where we are dedicating significant attention due to the forward planning required. I would start by saying we feel fortunate to have an extraordinary set of opportunities ahead. We operate a large-scale consumer business reaching billions, a hyperscale cloud business with strong momentum, and several other critical areas across Google and Alphabet. We always engage in long-range planning, so each of our— business units submits long-range plans guided by a disciplined ROIC framework.

(00:42:32)謝謝你，埃裡克。我先開始，然後託馬斯會回答第二部分。當然，計算資源的分配是我們非常重視的領域，因為這需要前瞻性的規劃。我想首先說，我們感到幸運，面前有一系列非凡的機會。我們經營著一個大型消費者業務，覆蓋數十億人，一個具有強勁動能的超大規模雲業務，以及谷歌和 Alphabet 的其他幾個關鍵領域。我們始終進行長期規劃，因此我們的每個—業務單位都提交由嚴謹的 ROIC 框架指導的長期計劃。

Sundar Pichai - Alphabet, CEO

(00:43:20)Broadly, this is a long-term planning process rather than a short-term allocation exercise. Our AI models underpin everything from Search and YouTube to Cloud and Waymo. They form the foundation, and we view this as a highly leveraged platform. Our full-stack approach drives significant efficiency and provides substantial optionality across diverse opportunities. This foundation is central, and we plan for it with a long-term perspective. As I mentioned, we take a disciplined, long-range planning approach across our businesses.

(00:43:20)總體而言，這是一個長期規劃過程，而不是短期的資源分配練習。我們的 AI 模型支撐著從搜尋和 YouTube 到雲端和 Waymo 的所有業務。它們形成了基礎，我們將其視為一個高度槓桿化的平台。我們的全棧方法驅動了顯著的效率，並在多樣的機會中提供了可觀的選擇性。這一基礎是核心，我們以長期的視角進行規劃。正如我提到的，我們在各業務中採取嚴謹的長期規劃方法。

Sundar Pichai - Alphabet, CEO

(00:44:04)We are seeing very robust demand signals across many areas. We view this as an extraordinary opportunity set and are investing accordingly. Thomas.

(00:44:04)我們在許多領域看到非常強勁的需求信號。我們將此視為一個非凡的機會集，並相應地進行投資。託馬斯。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:44:20)Eric, regarding cloud demand, we continuously manage our portfolio to ensure a broad customer base. Our customers range from digital natives and AI labs to traditional enterprises and financial and capital markets clients. It is a diverse set of customers. Many have begun using TPU systems within our cloud, with over 100 significant TPU customers starting on our cloud platform. They learn to utilize TPUs effectively. Remember, a TPU system combines hardware, system software, compilers, and includes optical networking and other components, assembled as a complete system.

(00:44:20)埃裡克，關於雲端需求，我們不斷管理我們的投資組合，以確保廣泛的客戶基礎。我們的客戶範圍從數位原生者和 AI 實驗室到傳統企業以及金融和資本市場客戶。這是一個多樣化的客戶群。許多人已經開始在我們的雲端中使用 TPU 系統，超過 100 個重要的 TPU 客戶已經在我們的雲端平台上啟動。他們學會有效地利用 TPU。請記住，TPU 系統結合了硬體、系統軟體、編譯器，並包括光纖網路和其他組件，組裝成一個完整的系統。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:45:12)We have taken a deliberate approach to delivering TPUs to customers. First, we focus on building the right silicon. We have TPU versions for training (TPU-T) and inference (TPU-I). Both offer substantial price-performance and efficiency improvements over prior generations. Second, we optimize the software. Software optimization not only improves system performance, but also ensures seamless portability. Recently, we introduced PyTorch support, significantly expanding our market by optimizing PyTorch on TPU.

(00:45:12)我們採取了一種有意識的方法來向客戶提供 TPU。首先，我們專注於構建合適的矽晶片。我們有用於訓練（TPU-T）和推理（TPU-I）的 TPU 版本。兩者在價格性能和效率上都比前幾代有顯著改善。其次，我們優化軟體。軟體優化不僅提高了系統性能，還確保了無縫的可攜性。最近，我們引入了對 PyTorch 的支持，通過在 TPU 上優化 PyTorch 顯著擴大了我們的市場。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:45:54)Third, we are broadening distribution. Previously, TPUs were available only through our cloud. Now, we offer TPU systems to customers for deployment in their own data centers. We are making TPU systems available for on-premises use. Additionally, we are expanding availability to other providers. For example, we announced a partnership with Oracle and, together with Blackstone, launched an agreement for Crux AI, a joint venture to offer customers TPU systems. We are seeing extraordinary demand and strong growth from AI labs, capital markets, financial service institutions, and enterprises.

(00:45:54)第三，我們正在擴大分銷。之前，TPU 僅通過我們的雲端提供。現在，我們向客戶提供 TPU 系統，以便在他們自己的數據中心進行部署。我們正在提供 TPU 系統以供本地使用。此外，我們正在擴大對其他供應商的可用性。例如，我們宣佈與甲骨文的合作，並與黑石集團共同啟動了一項關於 Crux AI 的協議，這是一個合資企業，旨在為客戶提供 TPU 系統。我們看到來自 AI 實驗室、資本市場、金融服務機構和企業的需求異常強勁，增長迅速。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:46:36)If you consider these 100-plus customers expanding to many more and scaling their use of TPU systems, we anticipate strong demand in the future.

(00:46:36)如果考慮到這 100 多位客戶擴展到更多並擴大他們對 TPU 系統的使用，我們預計未來會有強勁的需求。

Operator - Operator

(00:46:55)Our next question comes from Ross Sandler with Barclays. Your line is now open.

(00:46:55)我們的下一個問題來自巴克萊的 Ross Sandler。您的電話線現在已開通。

Ross Sandler - Barclays, Analyst

(00:47:02)Thanks. I wanted to continue with the TPU resale business topic you just mentioned. So, Anat, last quarter you mentioned that revenue would be recognized on system sales, I believe. If a gigawatt of TPU full systems costs around $35 billion for the compute and all related components, what percentage of that are we actually recognizing as revenue? And how does that flow through in terms of margin profile relative to where the Google Cloud segment currently sits? And then

(00:47:02)謝謝。我想繼續討論您剛提到的 TPU 轉售業務主題。所以，Anat，上個季度您提到收入將在系統銷售時確認，我相信。如果一個吉瓦（GW）的 TPU 完整系統的成本約為 350 億美元，包括計算和所有相關組件，我們實際上認可的收入百分比是多少？而這在利潤率方面如何流入，與目前 Google Cloud 部門的情況相比？然後

Ross Sandler - Barclays, Analyst

(00:47:45)I guess the second part of the question is, as you help companies like Anthropic bring on more of their own first-party infrastructure, what does that mean for Alphabet's CapEx intensity as you start to hand off some future CapEx to the AI labs, potentially reducing your own data center CapEx intensity? Your thoughts on that would be appreciated. Thank you.

(00:47:45)我想問題的第二部分是，當您幫助像 Anthropic 這樣的公司引入更多他們自己的第一方基礎設施時，這對 Alphabet 的資本支出強度意味著什麼，因為您開始將一些未來的資本支出交給 AI 實驗室，可能會減少您自己的數據中心資本支出強度？您對此的看法將不勝感激。謝謝。

Anat Ashkenazi - Alphabet, CFO

(00:48:13)Yeah, let me combine these two questions. First, we don't provide specific financials at the product level, but let me explain how to think about the financials associated with TPU system sales. As a reminder, and Thomas can elaborate, a TPU system is not just a chip; it's a specialized system combining multiple components, including hardware, compute, storage, and networking. So, we are selling much more than just a chip. Software, training, and inference capabilities add significant value to the customer purchasing a TPU system.

(00:48:13)是的，讓我將這兩個問題結合起來。首先，我們不提供產品層級的具體財務數據，但讓我解釋一下如何看待與 TPU 系統銷售相關的財務數據。提醒一下，Thomas 可以詳細說明，TPU 系統不僅僅是一個芯片；它是一個專門的系統，結合了多個組件，包括硬體、計算、存儲和網絡。因此，我們銷售的遠不止是一個芯片。軟體、訓練和推理能力為購買 TPU 系統的客戶增添了顯著的價值。

Anat Ashkenazi - Alphabet, CFO

(00:48:56)Regarding your second question, when we sign an agreement to sell TPU systems, that agreement first appears in our backlog. Then, we begin building inventory. Think of this from a CapEx versus cash perspective. We build inventory ahead of recognizing the revenue stream. I believe I mentioned this on the previous call. You will see inventory balances increase; previously, inventory primarily consisted of some hardware products. Now, those TPU systems not yet sold will appear in inventory. Cash goes out as we build up the inventory base.

(00:48:56)關於您的第二個問題，當我們簽署協議以銷售 TPU 系統時，該協議首先會出現在我們的待處理訂單中。然後，我們開始建立庫存。從資本支出與現金的角度來看這一點。我們在確認收入流之前建立庫存。我相信我在之前的電話中提到過這一點。您將看到庫存餘額增加；之前，庫存主要由一些硬體產品組成。現在，尚未售出的 TPU 系統將出現在庫存中。隨著我們建立庫存基礎，現金會流出。

Anat Ashkenazi - Alphabet, CFO

(00:49:41)Then, as we deliver and bill customers for these TPU systems, cash comes back in. So, this is not traditional CapEx as you might think, but rather a different cash flow cycle. Since we are ramping up this business, you see increased cash needs... reflected in our cash flow statements and inventory build. To date, some system sales have been recorded as revenue, with more expected through 2027.

(00:49:41)然後，當我們交付並向客戶開具這些 TPU 系統的帳單時，現金會迴流。所以，這不是您可能想的傳統資本支出，而是一種不同的現金流循環。由於我們正在擴大這項業務，您會看到現金需求增加……反映在我們的現金流報表和庫存增長中。到目前為止，一些系統銷售已被記錄為收入，預計到 2027 年會有更多。

Operator - Operator

(00:50:34)Our next question comes from Justin Post with Bank of America. Your line is now open.

(00:50:34)我們的下一個問題來自美國銀行的 Justin Post。您的電話線現在已開通。

Justin Post - Bank of America, Analyst

(00:50:42)Great. Thank you. I have some questions regarding the agentic transition. You did launch the Universal Shopping Cart. Clearly, people are liking and using agentic use cases. As that usage grows, how do you plan to manage the search transition to agentic use? Do you think this transition could be a headwind? And second, do you think agentic could actually drive more monetization of your traffic over the long term? Then, I'll return to TPU margins. We're wondering... Do you think this is a really good business for you, and why?

(00:50:42)很好。謝謝。我有一些關於代理轉型的問題。您確實推出了通用購物車。顯然，人們喜歡並使用代理用例。隨著使用量的增長，您計劃如何管理搜索向代理使用的轉型？您認為這一轉型可能會成為阻力嗎？第二，您認為代理實際上能在長期內推動更多的流量變現嗎？然後，我將回到 TPU 利潤率。我們想知道……您認為這對您來說是一個非常好的業務，為什麼？

Justin Post - Bank of America, Analyst

(00:51:15)If you can't provide margins, why do you believe this is a good way to utilize all this capacity? Thank you.

(00:51:15)如果您無法提供利潤率，為什麼您認為這是利用所有這些產能的好方法？謝謝。

Philipp Schindler - Alphabet, Chief Business Officer

(00:51:22)Justin. We're in the early days of agentic technology, but to address the second part of your question, we view agentic experiences as clearly additive. Our focus is on building trusted, secure, and large-scale experiences that users love, anchored on agentic commerce foundational blocks like universal commerce, the cart, agents, payments protocols, and so on. When users shop, there's an important part of the journey they don't enjoy—the grunt work—which is fantastic to automate with AI. However, we also want to enhance, not just automate away, the discovery experience that people genuinely enjoy in many cases.

(00:51:22)Justin。我們正處於代理技術的早期階段，但針對您問題的第二部分，我們認為代理體驗顯然是附加的。我們的重點是建立用戶喜愛的可信、安全和大規模的體驗，這些體驗以代理商務的基礎組件為基礎，如通用商務、購物車、代理、支付協議等。當用戶購物時，有一個重要的旅程部分是他們不喜歡的——繁瑣的工作——這是用人工智慧自動化的絕佳機會。然而，我們也希望增強，而不僅僅是自動化人們在許多情況下真正喜愛的發現體驗。

Philipp Schindler - Alphabet, Chief Business Officer

(00:52:01)Users appreciate choice and want to consider factors beyond price, such as brand loyalty and customer service. We're architecting these agentic workflows intentionally to provide users the time and space to explore those choices. For example, an agent could respond with: here’s the insurance product we found and another strong option—one with a higher premium but a lower deductible, currently offering a special discount. Since there is almost never only one perfect option for a decision, we are confident we can provide compelling advertising models for agentic use cases.

(00:52:01)用戶欣賞選擇，並希望考慮超越價格的因素，如品牌忠誠度和客戶服務。我們正在有意地設計這些代理工作流程，以提供用戶探索這些選擇的時間和空間。例如，代理可以回應說：這是我們找到的保險產品，還有另一個強勁的選擇——一個保費較高但免賠額較低的選擇，目前提供特別折扣。由於幾乎不會只有一個完美的選擇來做決策，我們有信心能為代理使用案例提供引人注目的廣告模型。

Philipp Schindler - Alphabet, Chief Business Officer

(00:52:42)And obviously, you can imagine extending beyond us with subscriptions and other payment methods. I'm genuinely excited about the advertising opportunities in an agentic world.

(00:52:42)顯然，您可以想像超越我們的訂閱和其他支付方式。我對代理世界中的廣告機會感到非常興奮。

Sundar Pichai - Alphabet, CEO

(00:52:51)Overall, I would add and say, I view it as an expansion opportunity. Just like a few years ago, incorporating AI chat use cases in search allowed people to ask more and do more, which is reflected in our usage. I think agents will similarly expand the types of problems we solve for our users. People often underestimate all the components needed to deliver these complex user experiences. Search may look simple on the surface, but it involves deep investments and technologies built over time, like the Knowledge Graph.

(00:52:51)總的來說，我想補充說，我將其視為擴展機會。就像幾年前，在搜索中整合人工智慧聊天用例使人們能夠詢問更多並做更多，這在我們的使用中得到了體現。我認為代理將類似地擴展我們為用戶解決的問題類型。人們常常低估了提供這些複雜用戶體驗所需的所有組件。搜索在表面上看起來簡單，但它涉及深度的投資和隨時間建立的技術，如知識圖譜。

Sundar Pichai - Alphabet, CEO

(00:53:41)This underpins our understanding of the world, among other things. So, I'm excited about agents and view them as a new frontier that brings new use cases to search. We are eager to explore this further.

(00:53:41)這是我們理解世界的基礎，還有其他方面。因此，我對代理感到興奮，並將其視為帶來新用例的搜索新前沿。我們渴望進一步探索這一點。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:53:56)Regarding TPU systems and margins, we do not break out margins for individual products or infrastructure components. However, we see several benefits from designing and manufacturing TPU systems. To clarify, the TPU system is not just a chip but a complete system. Given the price-performance benefits of TPU systems relative to competitors, growing demand for accelerated compute, and scarcity of such compute, we are focused on improving margins as we scale. Two additional points. First, when Anat mentioned procuring components to build inventory, we only do so after signing contracts.

(00:53:56)關於 TPU 系統和利潤率，我們不會為單個產品或基礎設施組件分解利潤率。然而，我們看到設計和製造 TPU 系統的幾個好處。為了澄清，TPU 系統不僅僅是一個晶片，而是一個完整的系統。鑑於 TPU 系統相對於競爭對手的價格性能優勢、對加速計算的需求增長以及此類計算的稀缺性，我們專注於隨著規模擴大而改善利潤率。兩個額外的要點。首先，當 Anat 提到採購組件以建立庫存時，我們僅在簽署合同後進行。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:54:48)Therefore, all TPU system inventory is associated with committed contracts. Additionally, placing TPU systems in third-party facilities and customer data centers broadens our addressable markets. Some industries have data on-premises that cannot be moved to the cloud, so this allows us to bring accelerated compute to the data. It also reduces the need for us to pre-build facilities to host these TPU systems.

(00:54:48)因此，所有 TPU 系統庫存均與已承諾的合同相關聯。此外，將 TPU 系統放置在第三方設施和客戶數據中心擴大了我們的可服務市場。一些行業擁有無法移動到雲端的本地數據，因此這使我們能夠將加速計算帶到數據所在的地方。這也減少了我們預建設施以容納這些 TPU 系統的需求。

Operator - Operator

(00:55:28)Our next question comes from Mark Shmulik with AllianceBernstein. Your line is now open.

(00:55:28)我們的下一個問題來自 AllianceBernstein 的 Mark Shmulik。您的線路現在開放。

Mark Shmulik - AllianceBernstein, Analyst

(00:55:36)Yes, thanks for taking my questions. I'll focus on a couple of related topics. Philipp, regarding agentic technology, have you observed anything surprising or changed your view based on early agentic-style use cases on Gemini AI mode and some third-party integrations with Google properties? We've discussed commerce extensively, but perhaps you could also comment on consumer productivity. Then, Thomas or Anat, could you share insights on Google Cloud's growth rate excluding TPU system sales, and how much capacity constraints have affected it?

(00:55:36)是的，謝謝您回答我的問題。我將專注於幾個相關主題。Philipp，關於代理技術，您是否觀察到任何驚訝的事情，或根據早期代理風格的用例在 Gemini AI 模式和一些與 Google 屬性的第三方集成中改變了您的看法？我們已經廣泛討論了商務，但也許您可以對消費者生產力發表評論。然後，Thomas 或 Anat，您能否分享有關 Google Cloud 的增長率（不包括 TPU 系統銷售）以及容量限制對其影響的見解？

Mark Shmulik - AllianceBernstein, Analyst

(00:56:10)As some third-party capacity comes online in Q4, is it fair to expect acceleration in Google Cloud's business? Thank you.

(00:56:10)隨著一些第三方容量在第四季度上線，是否可以合理預期 Google Cloud 的業務會加速？謝謝您。

Sundar Pichai - Alphabet, CEO

(00:56:20)I'll address the first part regarding early agent-style use cases. We are definitely seeing activity. Recently, in the Gemini app, we've added a task tab where users type agent queries. This allows us to observe usage patterns. Looking at all this, users have high expectations and want to be delighted. What encourages repeat use is the agent's ability to handle longer tasks and deliver satisfactory outcomes. We've seen similar patterns in other areas, including coding. When performance is around 70%, it seems ineffective until a threshold is crossed, after which usage really takes off.

(00:56:20)我將針對早期代理風格的用例回答第一部分。我們確實看到了活動。最近，在 Gemini 應用中，我們添加了一個任務標籤，讓用戶輸入代理查詢。這使我們能夠觀察使用模式。看著這一切，用戶有很高的期望，並希望感到驚喜。促使重複使用的因素是代理能夠處理更長的任務並提供令人滿意的結果。我們在其他領域也看到了類似的模式，包括編碼。當性能約為 70% 時，似乎效果不佳，直到超過某個閾值，之後使用量才真正上升。

Sundar Pichai - Alphabet, CEO

(00:57:13)We are closely learning from this, identifying gaps, and laying foundational work as I mentioned earlier. I'm confident this area will see significant progress across our platforms later this year and into 2027. Transitioning to TPUs, one comment I'd make is that in the semiconductor industry, economies of scale are critical. I would highlight our capability to deploy TPUs both within our data centers and in other data centers as a key factor in scaling the product, which directly improves our economics.

(00:57:13)我們正在密切學習這一點，識別差距，並如我之前提到的那樣奠定基礎工作。我相信這一領域在我們的平台上將在今年晚些時候和 2027 年看到顯著進展。轉向 TPU，我想說的是，在半導體行業，規模經濟至關重要。我想強調我們在數據中心內部以及其他數據中心部署 TPU 的能力，這是擴大產品規模的關鍵因素，直接改善了我們的經濟效益。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:58:05)To address the question about cloud growth, the Google Cloud segment delivered record growth in the third quarter, with cloud revenues increasing 103% to $30.7 billion. The strongest growth driver was Google Cloud Platform, showing strength in both AI and core GCP services. GCP accounted for the vast majority of revenue generated in Q3, and cloud revenue growth, excluding TPU system sales, remained very strong. Demand was also robust across our product lines. We observe three key trends. First, as customers use Gemini, we see direct revenue growth from Gemini itself.

(00:58:05)為了回答有關雲端增長的問題，Google Cloud 部門在第三季度實現了創紀錄的增長，雲端收入增長了 103%，達到 307 億美元。最強的增長驅動因素是 Google Cloud Platform，在人工智慧和核心 GCP 服務方面均顯示出強勁的表現。GCP 佔據了第三季度產生的絕大多數收入，雲端收入增長（不包括 TPU 系統銷售）仍然非常強勁。我們的產品線需求也非常強勁。我們觀察到三個關鍵趨勢。首先，隨著客戶使用 Gemini，我們看到 Gemini 本身的直接收入增長。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:58:52)Additionally, Gemini usage drives growth in other parts of our portfolio. For example, when building agents, customers use compute resources to perform tasks, leading to increased compute purchases. They also use storage to maintain memories for the agents. Databases serve as context repositories for agents to reason upon. Given the rise in cybersecurity threats, we are seeing strong growth in our cybersecurity portfolio. We are confident in both the growth and diversity of our business. We are also seeing new customers adopting GCP, accelerating growth and adoption rates.

(00:58:52)此外，Gemini 的使用推動了我們產品組合其他部分的增長。例如，在構建代理時，客戶使用計算資源來執行任務，這導致計算購買的增加。他們還使用存儲來維護代理的記憶。數據庫作為代理推理的上下文存儲庫。鑑於網絡安全威脅的上升，我們在網絡安全產品組閤中看到了強勁的增長。我們對業務的增長和多樣性充滿信心。我們還看到新的客戶採用 GCP，加速了增長和採用率。

Thomas Kurian - Alphabet, CEO, Google Cloud

(00:59:34)This represents an additional growth vector for us. At any given time, we could serve significantly more demand with additional capacity, but we carefully manage all parameters and growth.

(00:59:34)這為我們提供了額外的增長向量。在任何時候，我們可以用額外的產能來滿足顯著更多的需求，但我們會仔細管理所有參數和增長。

Operator - Operator

(00:59:55)Our last question comes from Michael Nathanson with MoffettNathanson. Your line is now open.

(00:59:55)我們的最後一個問題來自 MoffettNathanson 的 Michael Nathanson。您的線路現在開放。

Michael Nathanson - MoffettNathanson, Analyst

(01:00:03)Thank you. I have two questions for Sundar. In an earlier response to Eric's question about TPU and cloud, there was a willingness to collaborate with some competitors or companies we might view as competitors in models and cloud. How do you think about the philosophy of working with competitors, and why is this strategy an opportunity rather than a long-term risk? Secondly, regarding agentic consumer behavior, one of the debates centers on safety and trust. Could you share what you've learned so far about safety and trust?

(01:00:03)謝謝。我有兩個問題要問 Sundar。在之前對 Eric 有關 TPU 和雲的問題的回應中，表達了與一些競爭對手或我們可能視為競爭對手的公司在模型和雲方面合作的意願。您如何看待與競爭對手合作的理念，為什麼這一策略是一個機會而不是長期風險？其次，關於代理消費者行為，一個辯論的焦點是安全性和信任。您能分享一下到目前為止您對安全性和信任的理解嗎？

Michael Nathanson - MoffettNathanson, Analyst

(01:00:40)Are these variables important to consumers? If so, how are you protecting consumer trust and privacy as you develop agentic consumer tools? Thank you.

(01:00:40)這些變數對消費者來說重要嗎？如果是這樣，您在開發代理消費者工具時如何保護消費者的信任和隱私？謝謝。

Sundar Pichai - Alphabet, CEO

(01:00:52)Thanks, Michael. Those are great questions. Regarding the first question, we've long viewed what we build as either product categories or platform categories. It's important to clearly distinguish between the two. When building platforms, we focus on making them excellent and ensuring they work for everyone. We have a long history of operating platforms, including browsers, operating systems, and a highly successful cloud platform. In all cases, we strive to serve all players and be a great partner. You have seen us collaborate on models with other companies, including Apple, with whom we compete in many areas.

(01:00:52)謝謝你，Michael。這些問題非常好。關於第一個問題，我們已經長期以來將我們所構建的視為產品類別或平台類別。清楚區分這兩者非常重要。在構建平台時，我們專注於使其卓越並確保它們對所有人都有效。我們在運營平台方面有著悠久的歷史，包括瀏覽器、操作系統和一個非常成功的雲平台。在所有情況下，我們努力服務所有參與者，並成為一個出色的合作夥伴。您已經看到我們與其他公司（包括 Apple）在模型上合作，儘管我們在許多領域與其競爭。

Sundar Pichai - Alphabet, CEO

(01:01:37)Similarly, we consider our TPU systems an important platform we are building. We aim to be world-class in how we develop it. This means subjecting it to market forces, testing, and competing for customers with the most demanding workloads. This approach ensures our teams remain at the technological frontier. That's how we think about it. Over time, this strategy expands opportunities, and we've succeeded with it over a long period. Regarding agentic behavior, to address the earlier question, I should emphasize that people are clearly concerned about safety and trust when approaching agentic experiences.

(01:01:37)同樣，我們認為 TPU 系統是我們正在構建的重要平台。我們的目標是在開發方面達到世界一流。這意味著要讓它接受市場力量、測試，並與最苛刻的工作負載的客戶競爭。這種方法確保我們的團隊保持在技術前沿。這就是我們的思考方式。隨著時間的推移，這一策略擴大了機會，我們在長期內取得了成功。關於代理行為，為了回答之前的問題，我應該強調，人們在接觸代理體驗時顯然對安全性和信任感到擔憂。

Sundar Pichai - Alphabet, CEO

(01:02:31)Safety and trust are very significant factors in how users want to engage with these technologies. Over time, we have built trust by delivering safe experiences across many consumer products at scale, serving billions of users, while integrating end-to-end safety in our models and agents. We deploy these with extensive testing and real-time monitoring as we scale. I believe all of this will become a significant differentiator. We already hear positive feedback from our users, and I think they will show their preference as this frontier evolves.

(01:02:31)安全性和信任是用戶希望與這些技術互動的非常重要的因素。隨著時間的推移，我們通過在大規模上提供安全的消費者產品來建立信任，服務數十億用戶，同時在我們的模型和代理中整合端到端的安全性。我們在擴展時進行廣泛的測試和實時監控。我相信這一切將成為一個重要的區別因素。我們已經聽到了用戶的正面反饋，我認為隨著這一領域的發展，他們會表現出自己的偏好。

Michael Nathanson - MoffettNathanson, Analyst

(01:03:33)Thank you.

(01:03:33)謝謝。

Operator - Operator

(01:03:36)Thank you. That concludes our question and answer session for today. I will now turn the conference back over to Jim Friedland for any final remarks.

(01:03:36)謝謝。今天的問答環節到此結束。現在我將會議交回給 Jim Friedland 進行最後的發言。

Jim Friedland - Alphabet, Investor Relation

(01:03:46)Thank you all for joining us today. We look forward to speaking with you again during our fourth quarter 2026 call. Thank you, and have a good evening.

(01:03:46)感謝大家今天的參與。我們期待在 2026 年第四季度的電話會議中再次與您交談。謝謝，祝您晚上愉快。

Operator - Operator

(01:03:56)Thank you, everyone. This concludes today's conference call. Thank you for your participation. You may now disconnect.

(01:03:56)謝謝大家。今天的電話會議到此結束。感謝您的參與。您現在可以斷開連接。