Tencent to Have 'Very Sufficient' GPU Reserves by Year-End, Nvidia Effectively Exits: China's AI Computing Market Sees Power Shift

According to the latest financial report from Tencent Cloud in China, the company is continuing to advance its independent computing infrastructure, with GPU resource reserves expected to reach a 'very sufficient' state by the end of this year and early next year. This will support the expansion of cloud services and the development of large-scale models. Tencent plans to centrally manage computing resources across its entire platform, balancing internal usage with external rental services to maximize resource utilization.

Due to U.S. export controls, Nvidia's market share in China's advanced AI accelerator market has plummeted from nearly 95% to close to zero. This effectively means the company has exited the high-end AI computing race in China. This significant gap is being rapidly filled by domestic chip manufacturers such as Huawei and Moore Threads, ushering in an unprecedented period of growth.

In this shift of power, Chinese domestic AI chips are accelerating their iteration and large-scale delivery, steadily increasing their market share. Market observers note that this is not merely a supply chain substitution, but a critical turning point in China's journey toward autonomous AI infrastructure—a transition from 'Nvidia-dominated monopoly' to a new landscape of 'multi-vendor coexistence'.

Although some Chinese domestic chips still lag behind Nvidia's top-tier products in performance, major Chinese large-model developers have effectively bridged the parameter gap through deep collaborative optimization in real-world scenarios. Additionally, domestic solutions generally offer more stable power supply and superior localized service advantages. Huawei's recently launched super-node computing solution is a prime example of this trend.