Alibaba and China Telecom on Tuesday unveiled a new data center in Shaoguan, in China’s Guangdong province, that runs on Alibaba’s in-house Zhenwu AI semiconductors. The facility’s initial computing cluster uses 10,000 Zhenwu chips crafted for both AI model training and inferencing, and the partners indicated the installation can handle AI models containing hundreds of billions of parameters.
According to the announcement, the site is designed with growth in mind - the computing footprint is expected to scale up to 100,000 Zhenwu chips. China Telecom and Alibaba highlighted that the cluster’s computing power will be available to multiple industry use cases, explicitly citing sectors such as healthcare and advanced materials as potential beneficiaries.
Alibaba develops its processor technology through its T-head unit. The company, headquartered in Hangzhou, is also among China’s major cloud computing providers. Its cloud division markets AI models, and that business has been identified by the company as one of its faster-growing segments in recent reporting.
The Shaoguan deployment comes as Chinese companies accelerate development of domestically produced semiconductor options. The announcement noted this shift in context of restrictions from the United States that have limited China’s access to certain semiconductor technologies, including AI chips supplied by some foreign vendors.
Industry observers will note this launch follows other recent domestic deployments: last month, a computing cluster built with Huawei’s Ascend 910C AI chips was brought online in China. Both developments underscore continuing investment by Chinese technology firms and carriers in homegrown AI hardware.
For now, the Shaoguan facility’s public description focuses on its technical capacity, the planned scale-up path, and the range of commercial applications the computing cluster is intended to support. Details beyond those points were not provided in the announcement.