Ant Group’s AI breakthrough: Cuts AI training costs by 20% using Chinese chips over Nvidia’s

Ant Group, the fintech giant backed by Jack Ma, has quietly made a bold move in AI development—one that could shake up the industry. According to a report from Bloomberg, the company has successfully trained AI models using Chinese-made semiconductors from Alibaba and Huawei, cutting training costs by 20% compared to Nvidia’s premium H800 chips—all without sacrificing performance.

“Jack Ma-backed Ant Group Co. used Chinese-made semiconductors to develop techniques for training AI models that would cut costs by 20%,” Bloomberg reported, citing sources familiar with the matter.

“Ant used domestic chips, including from affiliate Alibaba Group Holding Ltd. and Huawei Technologies Co., to train models using the so-called Mixture of Experts machine learning approach, the people said. It got results similar to those from Nvidia Corp. chips like the H800, they said, asking not to be named as the information isn’t public.”

At the core of this shift is a machine learning method called Mixture of Experts (MoE), which splits AI tasks among different model components and only activates the necessary ones. This lets Ant get more out of less powerful chips, reducing its reliance on high-end U.S. hardware. According to internal figures, training 1 trillion tokens now costs about 5.1 million yuan ($706,000), down from 6.35 million yuan ($880,000) when using Nvidia gear.

This move comes at a time when U.S. export controls on advanced chips have intensified, prompting Chinese tech firms to explore alternatives. Ant still uses some Nvidia hardware, but its pivot to local chips underscores a bigger shift: China’s leading companies are actively working to reduce dependence on foreign technology.

The research was published in early March 2025 and picked up more widely on March 23. For Ant, it’s more than a technical win—it’s a signal that Chinese firms can still compete at a high level, even with restricted access to top-tier chips.

While Nvidia remains the gold standard for AI training, Ant’s success could mark a turning point. It’s the kind of result that might encourage other Chinese companies to double down on local solutions—and add more pressure to an already tense global chip race.

Nvidia, the long-reigning GPU titan, is facing a challenge from tech giants like Google, OpenAI, and Meta, who are building their own chips to cut costs and gain independence from Nvidia’s expensive grip.

Ant Group is a subsidiary of Alibaba Group, a company founded by Jack Ma. The company was founded as Alipay in 2004. Over the years, Ant Group has evolved into a major player in the global fintech landscape, gaining significant influence and prominence.