Wu Yongming: Qwen Has Already Trained Its Next-Generation Model on T-Head Chips
8 hour ago / Read about 0 minute
Author:小编   

On November 19, during the earnings analyst call for the second quarter of fiscal year 2027, Alibaba Group CEO Wu Yongming stated that T-Head's three-generation GPU product roadmap is accelerating: the Zhenwu M890 will be fully launched this year, while the V900 and J900 will respectively advance their production supply chains and enter large-scale mass production in 2027 and 2028. Infrastructure built on self-developed domestic computing power is becoming the supporting foundation of Alibaba's full-stack AI technology. It will gradually carry the workloads of the group and Alibaba Cloud customers across model training, model inference, and agentic tasks, and will drive the group's AI revenue growth. Wu Yongming emphasized that Qwen has already trained its next-generation model on T-Head chips, and that Bailian has been providing model inference services based on domestic chips. Alibaba's full-stack AI infrastructure will forge a significant advantage in the domestic computing power supply chain, and the deep synergy between self-developed chips and self-developed models will continue to translate into better per-unit computing power costs and higher R&D efficiency for AI applications.