Alibaba Launches AI Data Center Powered by 10,000 Homegrown Chips
China's tech giant bypasses US restrictions with a new, homegrown AI chip cluster.
Alibaba Group, in collaboration with China Telecom, has officially activated a new AI data center in Shaoguan, China. The facility's core infrastructure is built around 10,000 Zhenwu AI accelerator chips, which were designed and manufactured entirely within China by Alibaba's T-Head semiconductor division. This strategic deployment is a direct countermeasure to U.S. export controls that restrict access to advanced chips from companies like Nvidia. The Zhenwu chips are engineered specifically to support the computational demands of large-scale AI models, capable of handling models with hundreds of billions of parameters.
This initiative is more than a reactive move; it represents a calculated shift toward technological self-reliance. By developing its own silicon, Alibaba gains tighter control over performance, cost, and the scaling of its AI systems, mirroring strategies employed by other global tech giants. The data center, operated by China Telecom, will provide AI computing power to critical sectors including healthcare, materials science, and government. Alibaba and China Telecom have outlined an ambitious roadmap to expand the cluster's capacity tenfold, from 10,000 to 100,000 chips. This development intensifies the global AI infrastructure race, as Chinese firms build domestic alternatives to navigate constrained global semiconductor supply chains.
- The data center is powered by 10,000 domestically produced Zhenwu AI chips from Alibaba's T-Head division.
- It's a strategic response to US export restrictions, aiming to scale to 100,000 chips in the near future.
- The facility will provide compute for AI models powering industries like healthcare and materials science.
Why It Matters
It accelerates China's push for semiconductor self-sufficiency and reshapes the global AI compute landscape amid trade tensions.