Alibaba unveils China’s most powerful AI chip and trillion‑parameter model plans

Chinese tech giant Alibaba has used its annual conference in Hangzhou to showcase what it calls China’s most powerful AI chip and to sketch out plans for the next generation of its Qwen foundation models, targeting an enormous 5–10 trillion parameters. Parameters are the learned weights inside an AI model; the more you have, the more capacity the system has to capture patterns in data and handle complex tasks.

The new chip is designed to accelerate inference on Alibaba Cloud and within the Qwen model family, delivering faster responses and lower energy use for generative AI services. The launch comes against the backdrop of ongoing U.S. export controls on advanced semiconductors to China, and clearly serves as a signal that Beijing’s tech champions intend to build more of the AI stack — from chips to models — domestically.

Alibaba says it will use the chip to train upgraded Qwen models with stronger multimodal abilities, such as understanding images and long documents alongside text. But ultra‑large models also raise familiar concerns: soaring power demand, difficulties in testing safety at scale, and potential misuse for surveillance or military applications. The announcement underscores how quickly the U.S.–China race for AI leadership is shifting from software alone to full‑stack AI infrastructure, with both sides investing heavily in cutting‑edge compute.