Alibaba said it is developing a next-generation AI model with 5 trillion to 10 trillion parameters and unveiled a new AI chip at its annual Apsara conference in Hangzhou, according to Reuters. The announcements are part of the company’s effort to build more of its AI technology stack in China.
Reuters reported that the planned model could be up to four times the size of Alibaba’s Qwen 3.8 Max. The company’s T-Head unit introduced the Zhenwu V900 chip, which Alibaba said delivers triple the performance of its predecessor and can be used to form large clusters for massive AI models. Mass production is expected by early 2027.
For learners, the story highlights two major parts of modern AI development: model scale and the computing hardware needed to train and run large systems. It also shows how AI competition involves chips, data centers and software together.
Reuters reported that Alibaba is targeting more than 20 gigawatts of data-center capacity for Alibaba Cloud by 2032. The company is developing models, semiconductors and data-center infrastructure as part of the same strategy.