Tencent recently made a significant upgrade to its HPC network, Xingmai, which has boosted its AI capabilities by up to 60% in network communications and 20% in LLM training. This upgrade is part of a broader effort by Chinese tech giants to enhance their technological self-reliance, especially with restricted access to advanced processors like Nvidia’s H100 due to strict US export controls.
The South China Morning Post reported that the Xingmai 2.0 network has greatly improved the efficiency of communication between computing clusters, leading to better utilization of GPU capacities and cost savings for Tencent. This upgrade allows the network to support over 100,000 GPUs in a single cluster, reducing problem-solving time from days to minutes.
Unlike some American tech companies that focus on acquiring cutting-edge technology, Tencent has achieved its performance gains through optimizing its existing infrastructure. The company is also actively promoting its in-house developed LLMs for enterprise applications and offering services to help other companies develop their own AI models.
In the competitive Chinese AI industry, major players like Tencent, ByteDance, Baidu, and Alibaba are engaging in a price war to undercut Western competitors. Tencent recently made its Hunyuan LLM lite version free and reduced prices for standard versions, following similar moves by its rivals.
This upgrade reflects Tencent’s commitment to innovation and self-sufficiency in the face of technological challenges. By enhancing its network capabilities and promoting its AI solutions, Tencent is positioning itself as a key player in the rapidly evolving AI landscape.
