Jack Ma-backed Ant Group Co. has achieved a significant breakthrough in AI development using Chinese-made semiconductors, reducing costs by 20%. The company utilized domestic chips from affiliates Alibaba Group Holding Ltd. and Huawei Technologies Co. to train AI models using the Mixture of Experts (MoE) machine learning approach, achieving results comparable to those from Nvidia Corp.’s H800 chips. Ant Group is now primarily relying on alternatives from Advanced Micro Devices Inc. and Chinese chips for its latest models, though it continues to use Nvidia for AI development. This move underscores Chinese companies’ efforts to use local alternatives to advanced Nvidia semiconductors, which are currently barred from China. Ant Group published a research paper claiming its models outperformed Meta Platforms Inc. in certain benchmarks, though this has not been independently verified. The company aims to leverage its breakthrough in large language models, Ling-Plus and Ling-Lite, for industrial AI solutions in healthcare and finance. Ant Group has made these models open source, with Ling-Lite containing 16.8 billion parameters and Ling-Plus having 290 billion parameters. The company faces challenges in training stability but has deployed healthcare-focused large model machines in several hospitals and healthcare providers across China.