Ant Group’s AI Revolution: Slashing Costs with Chinese Chips

By replacing Nvidia with homegrown semiconductors, Ant Group is cutting AI costs, boosting efficiency, and driving China’s push for tech independence.

Ant Group, backed by Jack Ma, has made a major AI breakthrough. By using Chinese-made chips from Alibaba and Huawei, it has cut AI training costs by 20%. This marks a shift away from dependence on Nvidia, which has long dominated the AI chip market.

Global chip shortages and U.S. export restrictions pushed Ant Group to explore alternatives. The company adopted the Mixture of Experts (MoE) technique, a machine learning approach that activates only the most relevant models.

This reduces computing power needs while improving efficiency. Leading tech firms like Google also use this method.

The performance of Chinese chips now rivals Nvidia’s H800, a version designed for China due to trade restrictions. While Ant Group still incorporates Nvidia hardware, its success with domestic chips signals a larger trend in China’s AI sector.

Companies are moving toward self-sufficiency to lower costs and ensure long-term stability. This development is part of China’s broader effort to reduce reliance on foreign technology.

Other companies, including DeepSeek, are also making strides in AI using homegrown solutions.

Ant Group’s success highlights China’s growing ability to innovate without Western technology. As more firms follow this path, the AI landscape is shifting. Lower costs and greater self-reliance could reshape China’s AI dominance in the years ahead.