Alibaba Qianwen Core Team Turmoil: Key Figure Leaves, AI Strategy Faces Critical Test

On March 4, 2026, Alibaba Qwen (formerly Tongyi Qwen, Qwen) experienced a major personnel shakeup. Technical leader Lin Junyang publicly announced his resignation on his personal social media. Later that day, training lead Yu Bowen and core member Li Kaixin also left simultaneously.

Adding to this, Hu Bin, who previously joined Meta in January as the head of Qwen Code, and other core technical staff from Qwen have all left in bulk. This core team, responsible for developing the open-source large model with over 600 million downloads worldwide, now faces a restructuring crisis.

Lin Junyang, one of Alibaba’s youngest P10 technical experts, was a key creator of Qwen from the ground up. He led model development, iteration, and open-source deployment throughout the process. His departure is seen industry-wide as the “soul leaving” of the Qwen team. It is reported that the main reason for the mass departure is the upcoming organizational restructuring at Tongyi Laboratory. Originally, the Qwen team used a fully integrated architecture with pre-training and post-training stages, managed by Lin Junyang. The new plan will split this into multiple horizontally divided teams operating independently, which conflicts with Lin Junyang’s belief that “large model development requires deep integration across all stages.”

Beyond talent turbulence, Qwen also faces difficulties in market competition. During the 2026 Spring Festival AI red envelope war, Qwen launched subsidy activities within Alibaba’s ecosystem, boosting daily active users (DAU) from 17 million to a peak of 73.5 million—an over fourfold increase. However, after subsidies ended, DAU sharply declined. User retention, engagement depth, and brand recognition still lag behind top competitors, and the core competitive gaps remain unbridged.

Currently, Qwen is under multiple pressures: loss of key talent causing R&D gaps, organizational splits or slowed model iteration, insufficient user retention, weak commercial monetization, and rising costs due to stricter global AI regulations. Alibaba’s large model strategy faces severe challenges.

Despite the crisis, Alibaba has its advantages. Recently, the “Tongyun Ge” AI triangle—comprising Tongyi Laboratory, Alibaba Cloud, and Pingtouge—has emerged as a major force. Alibaba is now one of only two companies worldwide, alongside Google, with top-tier capabilities in large models, cloud computing, and chips. On March 3, Jack Ma and Alibaba executives emphasized that while AI presents huge impacts, it also offers significant opportunities, signaling a firm commitment to advancing AI strategies.

This personnel shakeup at Qwen marks a transition in the large model industry from “single-point technological breakthroughs” to comprehensive competition. Whether Alibaba can quickly fill talent gaps, streamline organizational collaboration, and turn technological advantages into market wins will not only determine Qwen’s future but also influence China’s position in the global large model landscape.