Biren Technology, a Chinese AI chip start-up, has significantly enhanced the performance of its hardware for training large language models (LLMs) through a collaboration with Tencent-backed Infinigence AI. This partnership boosted the training capacity of Biren’s GPUs by nearly 100%, addressing challenges faced by Chinese AI start-ups in accessing sufficient compute resources, particularly from US-based Nvidia. Infinigence provides on-demand computing through its cloud platform, Infini AI, which integrates chips from multiple suppliers, including Nvidia, AMD, Huawei, and others. Biren’s GPU cluster was added to Infini AI, further advancing AI model development in China.
