Prime Highlights:
The company is using a combination of Chinese and U.S.-made semiconductors to enhance the efficiency of AI model training, reducing both costs and time.
Ant’s approach has led to a 20% decrease in computing costs, thanks to the use of lower-cost hardware for training AI models.
Key Background:
Ant Group, an Alibaba affiliate, is optimizing its artificial intelligence (AI) development by incorporating both Chinese and U.S.-made semiconductors, according to a source familiar with the matter. This strategy not only reduces the cost and time required to train AI models but also mitigates the company’s dependence on a single supplier, such as Nvidia. By combining chips from different sources, Ant Group can leverage a more diversified approach in AI development, which aligns with the industry trend of utilizing multiple networks, including the “mixture of experts” (MoE) method. This technique enables training models with less computational power.
Earlier this month, Ant Group revealed that its use of lower-cost hardware had led to a 20% reduction in computing expenses. The company’s AI model development is further supported by a mix of semiconductors from both domestic and international suppliers, including chips from Alibaba, Huawei, and Advanced Micro Devices (AMD). Although Nvidia chips remain a part of Ant’s AI infrastructure, the company is increasingly turning to alternatives from Chinese suppliers and AMD to reduce reliance on any single source.
In addition to these technological advancements, Ant Group also announced significant updates to its AI solutions for the healthcare sector. The new AI capabilities are already in use by seven major hospitals and healthcare institutions across key cities such as Beijing, Shanghai, Hangzhou, and Ningbo. Ant’s AI models, based on DeepSeek’s R1 and V3 models, Alibaba’s Qwen, and Ant’s own BaiLing, are designed to assist with medical inquiries and improve patient services.
The company’s move to diversify its semiconductor supply comes as the U.S. continues to impose restrictions on China’s access to advanced AI chips. While Nvidia can still sell lower-end chips to China, these limitations have accelerated efforts by Chinese companies like Ant Group to rely more on domestic chipmakers for AI development.
