Baidu’s AI restructuring meets a search business in transition
TMTPost reports that Baidu is reorganizing its AI business by replacing letter-based job grades with levels 5–12, merging professional and management tracks below middle management, and moving foundational-model decisions closer to technical teams. The report says AI revenue surpassed traditional advertising in Q1 2026 while profit declined, as AI cloud growth offset falling search advertising and fragmented AI product entry points. It argues that Baidu’s central challenge is connecting organizational reform with a more coherent product strategy as search becomes an embedded capability inside AI applications.
Baidu’s shift from an advertising-led search company toward AI infrastructure and application products makes its organizational structure and product execution relevant to the broader Chinese AI market.
(本文作者为 Tech商业,钛媒体经授权发布)
文 | Tech商业
原基础模型研发部负责人吴甜的离开,为文心过去一年的承压添了一个人事注脚。。但人事从来只是表层。真正值得追问的是:当百度在2026年连续拆掉职级标签、合并管理通道、把基础模型研发交给一位29岁的外部博士时,它拆掉的究竟是什么?是“楚门世界”的墙,还是墙倒之后,露出了更难以修补的地基?
百度AI的困境,不在某一个模型版本的跑分,也不在某一位高管的去留。它藏在一组同时发生、方向相反的事实里:AI收入首次超过传统广告,但利润在收缩;组织层级被大幅压平,但产品入口到年中才勉强统一;搜索的垄断地位尚未被正面攻破,但“搜索”这件事本身正在被拆解成按次计价的商品。旧引擎在失速,新引擎在轰鸣,而连接两者的传动轴,始终没有装上。