Google的AI搜索问题终于显示在收益报告中上升，一些投资者认为
关键点
- Google的股票在第三季度搜索收入仅增长12%后下降了多达10%，缺少了15%的分析家们所期望的。
- 管理层承认，快速增长的AI模式带来的收入低于传统的搜索。 相形之下，云收入逐年翻了一番多，达到307亿美元，上升了103%.
——Google在AI建设上翻了一番：仅第三季度资本支出就突破了514亿美元，公司预计2027年将"大幅增加".
Google失去了高达10%的市场价值，因为一个老投资者的恐惧似乎已经实现。
搜索收入在第三季度增长约12%，缺少了15%的分析人员所预期的，短缺可能表明Google的核心业务出现紧张。 AI动力搜索查询增长快，但仍能产生比传统搜索更少的钱。
Google在收入呼吁中部分证实了这一点。 执行人Philipp Schindler 称认为"AI模式查询的强劲增长由于货币化程度较低而产生了影响"等因素。 尽管如此，称公司仍对总体轨迹表示满意，并报告说，搜索查询的年增长率达到了五年来的最高水平。
部分真相是传统搜索已被宣布死亡多年，无论如何，并持续增长。 本季度的增长率略低于预期，但与近年相符合，Google才刚刚开始货币化AI模式。 在呼叫中，宣布公司计划扩大在美国和国际上的聊天员广告程序。
CEO Sundar Pitchai 称 Gemini的app已经达到了10亿用户，成为了公司历史上增长最快的产品。 这不是什么出人意料的事，因为它基本上是搜索的附加内容，而且搜索已经是巨大的。 Google自有型号的API每分钟处理近320亿令牌，是一年前的4倍多。 部分跳跃可能反映了由推理模型和代理工作流程驱动的象征性通货膨胀，而不仅仅是较高的使用率。
云收入超过翻倍，因为Google在接下来的事情上花费了很大
云分会逐年增长103%，达到307亿美元，Google也相应投入。 仅在第三季度，资本支出就达到514亿美元。 2026年全年预测仍为1,950亿至2,050亿美元，预计2027年将达到"显著高".
云长托马斯·库里安期望AI agents能推动云业的进一步增长，因为生产中更多的agents需要更多的计算来完成他们的任务，更多的存储agent内存，以及更多的数据库推理能力，他称。
Google通过自己的消费品和其他经营的基础设施，对双方有利。 Pichai 称公司正在建设规模较小、蒸馏版的新碳模型，以平衡质量和反应速度，并计划通过云运行时间向“十亿”用户提供服务。 完整的"碳"模型最初将只面向选定的网络安全专家和美国政府。
Google也想出售以其专有的TPU系统所建的AI基础设施。 管理层没有披露向客户以外地区出售这些系统的利润率，即使两次被问及。
AI News Without the Hype – 由人类治愈
订阅THE DECODER免费阅读，每周AI通讯，我们独家的"AI雷达"前沿报告每年6次，全面存档访问，访问我们的评论部分。
现在订阅
本文由 AI宝库自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Google's AI search problem finally shows up in the earnings report, some investors believe
Key Points
- Google's stock dropped as much as ten percent after search revenue grew just 12 percent in the third quarter, missing the 15 percent analysts had expected.
- Management acknowledged that the fast-growing AI mode brings in less revenue than traditional search. Cloud revenue, by contrast, more than doubled year over year to $30.7 billion, up 103 percent.
- Google is doubling down on AI buildout: capital spending hit $51.4 billion in the third quarter alone, and the company expects it to "increase significantly" in 2027.
Google lost up to ten percent of its market value because an old investor fear appears to be coming true.
Search revenue grew about 12 percent in the third quarter, missing the 15 percent analysts had expected, and the shortfall might point to a tension in Google's core business. AI-powered search queries are growing fast but still generate less money than traditional search.
Google partly confirmed this on the earnings call. Executive Philipp Schindler said that "the strong growth in AI mode queries had an impact because of the lower monetization," among other factors. The company nonetheless said it was satisfied with the overall trajectory and reported that the annual growth rate for search queries hit its highest level in five years.
And part of the truth is that traditional search has been declared dead for years and keeps growing anyway. This quarter's growth rate was slightly below expectations but in line with recent years, and Google is only beginning to monetize AI mode. On the call, the company announced plans to expand its ad program for chatbots in the U.S. and internationally.
CEO Sundar Pichai said the Gemini app has reached one billion users, making it the fastest-growing product in the company's history. That is hardly a surprise, given that it is essentially a search add-on, and search is already huge. Google's APIs for its own models process nearly 32 billion tokens per minute, more than four times as many as a year ago. Part of that jump likely reflects token inflation driven by reasoning models and agentic workflows rather than higher usage alone.
Cloud revenue more than doubles as Google spends heavily on what comes next
The cloud division grew 103 percent year over year to $30.7 billion, and Google is spending accordingly. Capital expenditures hit $51.4 billion in the third quarter alone. The full-year 2026 forecast remains at $195 billion to $205 billion, with a "significantly higher" figure expected for 2027.
Cloud chief Thomas Kurian expects AI agents to drive further growth in the cloud business, since more agents in production require more compute for their tasks, more storage for agent memory, and more database capacity for reasoning, he said.
Google stands to gain on both sides, through its own consumer products and through the infrastructure other companies run on. Pichai said the company is building smaller, distilled versions of its new Carbon model that balance quality and response speed, and plans to serve them to "billions" of users through an always-on cloud runtime. The full Carbon model will initially go only to selected cybersecurity experts and the U.S. government.
Google also wants to sell AI infrastructure built on its proprietary TPU systems. Management did not disclose the profit margin on selling those systems to outside customers, even when asked twice.
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
Subscribe now
Auto-fetched and translated by AI Gem Box. All rights belong to the original author.