Alibaba Group and China Telecom are advancing a new data centre initiative in southern China, powered entirely by the e-commerce giant’s proprietary AI chips, as Beijing intensifies efforts to strengthen domestic computing infrastructure.
Summary
Alibaba and China Telecom have introduced a 10,000-chip AI data centre in Guangdong using Zhenwu processors to power large-scale models.
The initiative underscores China’s drive to build self-reliant AI infrastructure amid U.S. semiconductor restrictions and growing demand for compute capacity.
Alibaba aims to scale the cluster to 100,000 chips as adoption expands across industries like healthcare and manufacturing.
The facility, unveiled on Tuesday, is equipped with 10,000 of Alibaba’s Zhenwu chips, built to handle both AI training and inference. The system can support models with hundreds of billions of parameters, placing it among the most advanced computing platforms currently in operation.
Located in Shaoguan, Guangdong province, the cluster is described as a fully domestic project and marks the first deployment of a Zhenwu-powered system of this scale in the Greater Bay Area. Alibaba noted that the chips operate as a unified system, allowing the cluster to function like a single supercomputer with ultra-low latency of around four microseconds.
The rollout highlights how China’s leading tech firms are accelerating the development of in-house AI chips and infrastructure as the country pushes toward technological self-sufficiency.
Push for domestic alternatives accelerates
In recent years, Washington has tightened export restrictions on advanced semiconductor technologies, including AI chips made by Nvidia. These curbs have prompted Chinese firms to fast-track the development of local alternatives across both hardware and infrastructure.
Alibaba has been strengthening its semiconductor ambitions through its T-Head unit, while expanding its footprint in cloud computing. The company now spans the full AI ecosystem—from chip design to data centre deployment and model development—delivered through its cloud division.
Cloud computing remains one of Alibaba’s fastest-growing segments, driven by increasing demand for AI workloads. Across China, investment in large-scale, domestically powered data centres has continued to accelerate.
A comparable project recently went live in Shenzhen, where a 10,000-card cluster powered by Huawei’s Ascend 910C chips began operations, reflecting a broader push among Chinese firms to scale local computing capabilities.
Unlike U.S. giants such as Meta and Microsoft—which are expected to collectively invest hundreds of billions into AI infrastructure—Chinese companies have taken a more focused approach. Investments are being directed toward sectors likely to generate tangible returns, including industrial applications and enterprise services.
Scaling AI for real-world use
The Zhenwu-powered cluster signals China’s shift from experimentation to large-scale AI deployment. Demand for computing power continues to rise as industries increasingly integrate AI into operations.
According to Charlie Zheng, chief economist at Samoyed Cloud Technology Group Holdings, the rollout of domestic clusters marks a transition from “hardware replacement” to “software collaboration” within China’s AI ecosystem.
He noted that adoption has been particularly rapid in government services and urban management, where data sovereignty and security requirements are especially stringent.
Alibaba stated that the cluster improves efficiency in training and inference by around 30%, while boosting single-chip performance nearly tenfold. It is already being applied in sectors such as healthcare and advanced manufacturing.
Access to the system is being offered to small and medium-sized enterprises via China Telecom’s platform, with pricing based on per-card or hourly usage.
Looking ahead, Alibaba plans to expand the cluster to 100,000 chips, a move expected to further lower costs and enhance resource efficiency as demand for large-scale AI computing continues to grow.
