百度Q1财报超预期：「云+无人驾驶」双轮驱动，连获多家长线基金大幅增持

Core Viewpoint - Baidu's AI-first strategy positions the company to seize long-term growth opportunities in the AI era, as evidenced by strong financial performance and advancements in AI technology [2][5]. Financial Performance - In Q1 2025, Baidu reported total revenue of 32.5 billion yuan, with core revenue growing 7% year-on-year to 25.5 billion yuan, exceeding market expectations [4]. - Core net profit for the first quarter increased by 48% year-on-year to 7.63 billion yuan [4]. Business Segments - Baidu's intelligent cloud business saw a remarkable revenue growth of 42% year-on-year in Q1 [4][11]. - The "LuoBo Kuaipao" service provided over 1.4 million rides globally in Q1, marking a 75% year-on-year increase [4]. AI Model Development - Baidu launched four major AI models within 40 days, including the multi-modal Wenxin 4.5 and the deep thinking model X1, enhancing capabilities in language understanding and reasoning [7][8]. - The Wenxin 4.5 Turbo model offers improved performance and reduced costs, with input prices dropping by 80% to 0.8 yuan per million tokens [9]. Investment and Market Position - Long-term funds have significantly increased their holdings in Baidu, reflecting confidence in the company's dual-engine strategy of "cloud + autonomous driving" [5]. - Baidu's intelligent cloud revenue growth outpaced that of international giants like Google and Microsoft, with a 42% increase compared to Google's 28% and Microsoft's 20% [11]. Autonomous Driving Expansion - The "LuoBo Kuaipao" service is expanding internationally, with recent partnerships in Dubai and Abu Dhabi for autonomous driving services [13][16]. - The sixth-generation vehicles of "LuoBo Kuaipao" have a cost advantage, being 30% cheaper than Tesla and 1/7th the operational cost of Waymo [15]. Technological Infrastructure - Baidu emphasizes the importance of continuous CAPEX investment in technology infrastructure to maintain a competitive edge in AI [10][11]. - The company has successfully launched self-developed computing clusters to support the training of large models, enhancing its capabilities in AI applications [11].