Chinese technology giant Alibaba has introduced a new artificial intelligence chip called the Zhenwu M890 as the company accelerates efforts to reduce China’s dependence on foreign semiconductor technologies. The announcement comes amid ongoing U.S. export restrictions targeting advanced AI processors supplied by companies such as NVIDIA.
Developed by Alibaba’s semiconductor subsidiary T-Head, the Zhenwu M890 is designed to support AI training and inference workloads, with a particular focus on agentic AI systems capable of executing complex autonomous tasks. According to the company, the chip delivers three times the performance of its previous-generation Zhenwu 810E processor and includes 144GB of GPU memory with bandwidth reaching 800GB per second.
Alongside the processor, Alibaba also introduced a new server platform called Panjiu AL128, which integrates 128 AI accelerators into a single rack infrastructure. The company additionally revealed Qwen 3.7-Max, a large language model optimized for coding and advanced AI workflows capable of maintaining long-duration processing sessions.
Alibaba stated that more than 560,000 units from the Zhenwu chip family have already been shipped across multiple industries in China. The company plans to continue expanding its in-house semiconductor roadmap with future generations of processors expected in 2027 and 2028.
Industry analysts note that while the new processor still trails the most advanced Western AI chips in raw performance, it is increasingly viewed as a viable domestic alternative within the Chinese market. The launch reflects Beijing’s broader strategy to strengthen local AI infrastructure and reduce reliance on imported semiconductor technologies.
Earlier this year, Alibaba launched a large-scale data center in Shaoguan built around 10,000 Zhenwu chips in partnership with China Telecom, with plans to scale the deployment significantly in the coming years.
The company also confirmed plans to invest more than 380 billion yuan over three years into cloud computing and AI infrastructure as it pushes to develop a fully integrated ecosystem spanning semiconductor design, AI models, cloud services and computing infrastructure.
