转自：经济日报新闻客户端 AI原生营销服务方面，2025全年收入同比增长301%。2025年12月，百度App月活用户数达6.79亿，文心 助手月活用户数达2.02亿。春节红包活动启动以来，文心助手月活跃用户同比增长4倍。2025年，慧播 星数字人开播规模同比增长202%，收入同比增长228%，已赋能30多个行业。 AI应用2025全年收入突破100亿元。据IDC《中国无代码生成式应用开发平台实测报告》，无代码生成 平台秒哒在平台功能和应用质量方面，表现出行业领先的性能。季度内，全球首个自我演化超级智能体 百度伐谋已获得超2000家企业申请试用，并发布"同舟生态伙伴计划"。近期，百度文库、百度网盘整合 成为个人超级智能事业群（PSIG），进一步加快AI应用创新。 同时，萝卜快跑四季度全球无人驾驶出行服务次数达340万，同比增长超200%，季度内每周出行次数峰 值超30万。财报显示，萝卜快跑加速其向全球多个市场的扩张。截至2026年2月，萝卜快跑累计提供全 球出行服务次数超2000万，足迹已覆盖全球26个城市。目前，萝卜快跑自动驾驶总里程累计超3亿公 里，全无人驾驶里程超1.9亿公里。 近期，文心5.0正式版上 ...

编辑 | 虞尔湖 出品 | 潮起网「于见专栏」 刚刚过去的2025年，人工智能行业的发展如日中天。在各大玩家的共同推动下，整个AI行业的商业化，也正在加速。而百度作为最早躬身入局的行业头部 玩家，更像是行业风向标，其一举一动，也牵动着业界的神经。 2月26日，百度发布了2025年第四季度及全年财报，显示2025年总营收达1291亿元，AI业务营收达400亿元；四季度，百度总营收327亿元，同比增长5%，AI 业务收入占百度一般性业务收入的43%，超出市场预期。 | | | 截至以下日期止三個月 | | | | 截至以下日期止十二個月 | | | --- | --- | --- | --- | --- | --- | --- | --- | | | 2024年 | 2025年 | 2025年 | 2025年 | 2024年 | 2025年 | 2025年 | | | 12月31日 | 9月30日 | 12月31日 | 12月31日 | 12月31日 | 12月31日 | 12月31日 | | | 人民幣元 | 人民幣元 | 人民幣元 | 美元 | 人民幣元 | 人民幣元 | 美元(4 | | 收入 | 34,1 ...

2月26日，百度集团股份有限公司（以下简称"百度"）发布2025年第四季度及全年财报。2025年，百度 总营收达1291亿元，AI业务营收达400亿元；第四季度，百度总营收327亿元，AI业务收入占百度一般 性业务收入的43%。 百度创始人李彦宏表示："2025年是AI成为百度新核心的关键一年。AI云势头强劲，百度凭借差异化的 全栈端到端AI能力赢得越来越多的企业认可。百度的AI应用组合持续扩展，满足了企业和个人的多样 化需求。萝卜快跑进一步巩固了其全球领先地位，以行业领先的规模运营，同时加速向新市场的国际扩 张。与此同时，AI原生营销服务持续增长，为长期发展开启了新的可能性。" 在AI应用方面，2025年百度AI应用全年收入突破100亿元。四季度内，自我演化超级智能体百度伐谋已 获得超2000家企业申请试用，并发布"同舟生态伙伴计划"。近期，百度文库、百度网盘整合成为个人超 级智能事业群（PSIG），进一步加快AI应用创新。此外，文心5.0在近期正式上线。作为原生全模态大 模型，该模型参数达2.4万亿，采用原生全模态统一建模技术，支持文本、图像、音频、视频等多种信 息的输入与输出。 在Robotaxi方面， ...

2月26日，百度发布2025年第四季度及全年财报，2025年总营收达1291亿元，AI业务营收达400亿元； 2025年第四季度总营收为327亿元，AI业务收入占一般性业务收入的43%，超出市场预期。 "2025年是AI成为百度新核心的关键一年。AI云势头强劲，我们凭借差异化的全栈端到端AI能力赢得越 来越多的企业认可。我们的AI应用组合持续扩展，满足了企业和个人的多样化需求。萝卜快跑进一步 巩固了其全球领先地位，以行业领先的规模运营，加速向新市场的国际扩张。与此同时，AI原生营销 服务持续增长，为长期发展开启了新的可能性。"百度创始人李彦宏表示，"随着'以AI为先'的战略清晰 成型，我们对在AI时代创造持久价值的能力充满信心。" 【导读】百度最新财报：2025年营收为1291亿元，第四季度AI业务收入占比为43% 萝卜快跑加速扩张 去年四季度，萝卜快跑全球无人驾驶出行服务次数达340万次，同比增长超200%，季度内每周出行次数 峰值超30万次。 财报显示，萝卜快跑加速向全球多个市场扩张。截至2026年2月，萝卜快跑累计提供全球出行服务次数 超2000万次，足迹已覆盖全球26个城市。 目前，萝卜快跑自动驾驶总 ...

Core Insights - Baidu's total revenue for 2025 reached 129.1 billion yuan, with AI business revenue accounting for 40 billion yuan, exceeding market expectations [1] - The company emphasizes that 2025 is a pivotal year for AI to become its new core, with strong momentum in AI cloud services and a growing portfolio of AI applications [1] - AI cloud revenue grew by 34% year-on-year, with high-performance computing subscription revenue increasing by 143% in Q4 [1] Group 1: Financial Performance - In Q4, Baidu's total revenue was 32.7 billion yuan, with AI business revenue constituting 43% of its general business revenue [1] - AI-native marketing services saw a revenue increase of 301% for the entire year of 2025 [2] - The total revenue from AI applications surpassed 10 billion yuan in 2025 [2] Group 2: User Engagement and Market Expansion - As of December 2025, Baidu App had 679 million monthly active users, while Wenxin Assistant reached 202 million monthly active users, with a fourfold increase since the launch of the Spring Festival red envelope activity [2] - The autonomous driving service, "Luobo Kuaipao," recorded 3.4 million service instances in Q4, marking a year-on-year growth of over 200% [2] - Luobo Kuaipao has expanded its global reach, providing over 20 million service instances across 26 cities by February 2026 [2] Group 3: Technological Advancements - The official version of Wenxin 5.0 was launched, featuring 2.4 trillion parameters and supporting multiple forms of input and output [3] - Wenxin 5.0 has achieved top rankings in various domestic benchmarks, surpassing several mainstream models [3] - Baidu's ESG performance has improved, as noted in the 2026 Global Edition of the S&P Global Sustainability Yearbook [3]

Core Insights - Baidu's total revenue for 2025 reached 129.1 billion yuan, with AI business revenue at 40 billion yuan, indicating a strong performance in AI sectors [1][4] - In Q4 2025, Baidu's total revenue was 32.7 billion yuan, a 5% year-over-year increase, with AI business revenue accounting for 43% of general business income, exceeding market expectations [1][4] AI Business Growth - The AI cloud segment saw a revenue increase of 34% year-over-year for 2025, with Q4 subscription revenue from high-performance computing facilities growing by 143%, accelerating from 128% in Q3 [1][4] - Baidu's AI-native marketing services experienced a remarkable 301% year-over-year revenue growth in 2025 [5] - The AI application revenue surpassed 10 billion yuan in 2025, showcasing the company's expanding capabilities in AI solutions [5] User Engagement and Expansion - By December 2025, Baidu App had 679 million monthly active users, while Wenxin Assistant reached 202 million monthly active users, with a fourfold increase in active users since the launch of the Spring Festival red envelope activity [5] - The "Luobo Kuaipao" autonomous driving service recorded 3.4 million global rides in Q4, more than doubling year-over-year, with a peak of over 300,000 rides per week [6] - As of February 2026, "Luobo Kuaipao" had provided over 20 million global rides, covering 26 cities worldwide [6] Technological Advancements - The total mileage for "Luobo Kuaipao" autonomous driving exceeded 300 million kilometers, with fully autonomous driving mileage surpassing 190 million kilometers [3][6] - The self-evolving super-intelligent agent "Baidu Famou" received applications for trials from over 2,000 enterprises, indicating strong interest in innovative AI solutions [5]

Core Insights - The 2025 "Smart Life" annual case event revealed 20 companies and products, highlighting the deep integration of technology and daily life, with a focus on AI products that enhance empathy and user experience [1] - The event emphasizes the importance of practical applications of smart technology in addressing real-life issues and improving user convenience, showcasing a shift from tool-based to ecosystem-based technology [1] Group 1: Key Trends in Smart Technology - The integration of AI products is becoming more pronounced, with a focus on empathy and user-centric solutions [1][3] - Future products are expected to deeply incorporate large model capabilities from the design phase, leading to a new generation of AI-native applications that transform user habits [2] Group 2: Notable Award Winners - Taobao Flash Purchase was recognized as the annual instant retail innovation platform [5] - Ant Financial's "Afu" was awarded as the annual family health guardian AI butler [5] - Kuaishou's "wow" intelligent agent was named the annual commercial AI assistant [6] - Hisense's AI Life Butler received recognition as the annual smart appliance innovation case [12] - DJI's Osmo Action 6 was awarded for its contribution to smart imaging accessibility [13]

Core Viewpoint - Beijing is positioning itself as the "Artificial Intelligence Capital" by leveraging its unique advantages in talent density, full-stack ecosystem, and industrial clusters, aiming for a core AI industry scale exceeding 1 trillion yuan within two years [2][11]. Group 1: AI Industry Development - The AI core industry in Beijing is projected to grow from approximately 450 billion yuan in 2025 to over 1 trillion yuan by 2027, indicating a doubling of the industry size [12]. - The city has established itself as a global AI innovation hub, with 1.5 million AI scholars, accounting for 30% of the national total, and 148 individuals listed among the world's most influential AI scholars [4][10]. - Beijing's AI ecosystem includes 209 registered large models, representing nearly 30% of the national total, with significant contributions from various local companies and institutions [10]. Group 2: Technological Innovations - The launch of the AI video generation model Vidu by Shengshu Technology marks a significant advancement, enabling the generation of high-quality videos in just 8 seconds, a drastic reduction from the previous 900 seconds [6][7]. - The collaboration between domestic GPU companies and research institutions has led to the successful training of advanced AI models, showcasing the capabilities of local technology [8][10]. - The FlagOS system software stack developed by Zhiyuan Research Institute serves as a universal language connecting domestic AI chips and large models, enhancing efficiency and reducing costs in the AI industry [5][10]. Group 3: Market Opportunities - The consumer market for AI applications is expanding, with companies like Kuaishou's Keling AI achieving significant revenue growth, indicating a robust demand for AI-driven solutions [13]. - The integration of AI in sectors such as automotive design is streamlining processes, reducing time for tasks like wind resistance validation from 10 hours to just 1 minute [12]. - The development of innovative AI models and applications is creating new market opportunities, positioning Beijing as a leader in AI technology and application [13].

Core Insights - The article discusses the challenges and advancements in the development of Agentic models, emphasizing that the main bottleneck is not the models themselves but the replication of real-world environments and stable access to external interfaces and databases [2][4][5] - It highlights the current limitations of general-purpose models in achieving industrial-level performance across various vertical agent scenarios, suggesting that tailored models for specific applications are more effective [5][12] - The article also explores the evolution of multi-modal models, indicating that while there have been significant advancements, a unified modeling approach for understanding and generating across modalities remains a key goal for the future [17][20] Group 1: Agentic Models - The primary focus is on enhancing models to perform effectively in various vertical agent scenarios, particularly in coding applications [4] - Current general-purpose models lack the capability to achieve stable generalization across diverse environments, necessitating the customization of models for specific applications [5] - The complexity of real-world environments, including external dependencies and interfaces, poses significant challenges for training agentic models [5][6] Group 2: Multi-Modal Models - The transition from single-modal to multi-modal models has introduced visual capabilities into language models, with a focus on aligning text and visual tokens [17][18] - Despite advancements, the industry faces challenges in scaling multi-modal models due to the difficulty in obtaining high-quality, aligned data [18] - Future directions include the pursuit of unified modeling that integrates generation and understanding capabilities, although current results indicate that separate optimization yields better performance [20][21][22] Group 3: Reinforcement Learning and Training Efficiency - The article emphasizes the importance of reinforcement learning systems for continuous model iteration in specific scenarios, with a focus on high efficiency and throughput [6][9] - The scaling of reinforcement learning has not yet reached a consensus in the industry, but there is recognition of its potential to enhance model capabilities significantly [10][11] - Efficient training processes, particularly in generating diverse paths for evaluation, are critical for the success of reinforcement learning in agentic models [9] Group 4: Future Trends and Directions - The article predicts that the development of agentic models with stable and accurate tool-calling capabilities will expand beyond coding applications to a broader range of real-world APIs [28] - The concept of "world models" is discussed, highlighting the evolution from language models to dynamic models that understand physical world operations [26] - The integration of tools into agent development is seen as a crucial pathway for enhancing model capabilities, reflecting the importance of tool usage in human intelligence evolution [25]

Core Insights - The article highlights the significant advancements in AI agents, particularly the launch of Qianwen App, which integrates various services like food delivery and flight booking, marking a comprehensive AI shopping experience [1][3] - The rise of AI agents is seen as a transformative force in the AI industry, shifting from mere tool applications to value creation, with strong policy support and market dynamics driving this evolution [1][2] Group 1: AI Agent Development - Qianwen App has opened functionalities for food delivery, shopping, and travel services, allowing users to place orders through simple commands, showcasing the seamless integration of AI in daily tasks [3] - Major companies, including Alibaba and Tencent, are rapidly developing their AI agent frameworks, with a focus on addressing challenges in building, running, and managing these agents [4] - The AI agent market in China is projected to reach 4.75 billion yuan in 2024, with a growth rate of 64.4%, and is expected to approach 15 billion yuan by 2026 [5][6] Group 2: Multi-Scenario Applications - AI agents are evolving from content generation to goal-oriented functionalities, enhancing their decision-making and real-time interaction capabilities across various industries [7] - In manufacturing, AI agents can predict equipment failures, reducing downtime by 50% through real-time monitoring and predictive maintenance [7] - The financial sector is increasingly adopting AI agents for customer service, risk management, and loan processing, with over 60% of banks implementing AI customer service solutions [8] Group 3: Challenges and Future Directions - The development of AI agents is recognized at the national policy level, with goals set for deep integration into key sectors by 2027 and a fully empowered smart economy by 2030 [10] - There are concerns about the emergence of "pseudo AI agents" that do not offer true intelligence but rather basic automation, highlighting the need for genuine innovation [11] - The industry faces challenges in data quality and ecosystem collaboration, which are crucial for the effective deployment of AI agents in complex scenarios [12]