Alibaba is using chips developed by its semiconductor subsidiary T-Head to train its next Qwen model, CEO Eddie Wu said during the company’s quarterly earnings call on Thursday. Alibaba’s Bailian model platform is also providing inference services using Chinese-made chips.

The disclosure extends T-Head’s role across both major stages of AI computing: training models and running them to answer user requests. Bailian, known internationally as Model Studio, provides access to models and tools for building AI applications.

Wu said coordination between Alibaba’s own chips and models would improve computing costs and application development efficiency. He also argued that the company’s full-stack AI infrastructure would give it an advantage within China’s domestic computing supply chain. The training work connects two core parts of Alibaba’s AI investment: its Qwen model family and the hardware supporting its cloud services. [CLS, in Chinese]