Alibaba-affiliate Ant uses Chinese, U.S. chips to cut AI costs

Ant Group, an Alibaba-affiliated company, is leveraging a combination of Chinese and U.S.-made semiconductors to develop more efficient artificial intelligence models, as reported by a source familiar with the matter. By utilizing a mix of chips from different suppliers, including Alibaba, Huawei, Advanced Micro Devices, and Nvidia, Ant Group aims to reduce the time and cost of training AI models while also minimizing reliance on a single supplier.

This strategic approach aligns with the industry trend of utilizing a mixture of experts, a technique that allows models to be trained with less computational power. Ant Group recently published a paper showcasing its ability to effectively train its own MoE models using lower-cost hardware, resulting in a 20% reduction in computing costs.

Ant Group, known for operating Alipay, one of China’s leading mobile payment apps, was founded by Jack Ma, who is also affiliated with Alibaba. According to a report by Bloomberg, Ant Group has been incorporating chips from various suppliers, emphasizing Chinese-made chips for training AI models. The company has recently announced significant enhancements to its AI solutions for healthcare, which are currently being utilized by prominent hospitals and healthcare institutions in key Chinese cities.

The healthcare AI model developed by Ant Group incorporates DeepSeek’s R1 and V3 models, Alibaba’s Qwen, and Ant’s proprietary BaiLing technology. This model not only assists in answering medical queries but also enhances patient services, as stated in the company’s press release. Despite efforts by the U.S. to restrict China’s access to advanced semiconductors for AI development, companies like Nvidia are still able to supply lower-end chips to Chinese businesses.

Ant Group’s innovative approach to AI development, utilizing a diverse range of semiconductor suppliers, showcases its commitment to staying at the forefront of technological advancements. By leveraging both Chinese and U.S.-made chips, the company is poised to continue driving innovation in AI models and contributing to the evolution of healthcare solutions in China and beyond.