Xiaomi Constructs GPU Cluster to Accelerate AI Model Development
Chinese technology giant Xiaomi is actively constructing its own GPU cluster to bolster its expanding investments in artificial intelligence models. According to reports from Jiemian, the company's AI model team started with 6,500 GPU resources when it was established in April 2023. CEO Lei Jun has emphasized Xiaomi's long-term commitment to AI, highlighting initiatives such as the AI Lab, Xiao Ai voice assistant, and autonomous driving teams. The company focuses on lightweight, locally deployable AI solutions, successfully running a 1.3 billion parameter model on smartphones with performance comparable to larger cloud-based models. To further strengthen its capabilities, Luo Fuli, a key developer of the DeepSeek-V2 open-source model, joined Xiaomi’s AI Lab in December to lead large model efforts. Since the AI division's inception in 2016, it has grown to over 3,000 members with expertise in vision, speech, natural language processing, and multimodal AI. These technologies are being integrated across Xiaomi’s ecosystem, including mobile phones, electric vehicles, AIoT devices, and robots, marking a significant strategic push into advanced AI infrastructure and application.
Wire timeline
Xiaomi Constructs GPU Cluster to Accelerate AI Model Development
Chinese technology giant Xiaomi is actively constructing its own GPU cluster to bolster its expanding investments in artificial intelligence models. According to reports from Jiemian, the company's AI model team started with 6,500 GPU resources when it was established in April 2023. CEO Lei Jun has emphasized Xiaomi's long-term commitment to AI, highlighting initiatives such as the AI Lab, Xiao Ai voice assistant, and autonomous driving teams. The company focuses on lightweight, locally deployable AI solutions, successfully running a 1.3 billion parameter model on smartphones with performance comparable to larger cloud-based models. To further strengthen its capabilities, Luo Fuli, a key developer of the DeepSeek-V2 open-source model, joined Xiaomi’s AI Lab in December to lead large model efforts. Since the AI division's inception in 2016, it has grown to over 3,000 members with expertise in vision, speech, natural language processing, and multimodal AI. These technologies are being integrated across Xiaomi’s ecosystem, including mobile phones, electric vehicles, AIoT devices, and robots, marking a significant strategic push into advanced AI infrastructure and application.
TechNode