MAIN CONTENT
Inference Efficiency Shields Cloud Revenue Despite Training Headwinds
According to Reuters reporting from Beijing on November 18, 2026, Tencent executives sought to reassure investors after shares of Chinese cloud service providers tumbled on Tuesday following Washington's announcement of rules restricting remote access to advanced United States artificial intelligence chips via overseas data centers. Chief Strategy Officer James Mitchell explained that Tencent's cloud business today primarily runs inference workloads rather than large-scale model training, enabling its optimized models to operate effectively on a surprisingly limited number of domestic chips.
At the same time, Tencent President Martin Lau described the tighter American restrictions as unfortunate, noting that they leave Chinese technology firms with less advanced computing power. Lau cautioned that the measures will directly affect artificial intelligence model training and are likely to slow the rollout of Tencent's next flagship model in the near term.
"We actually don't see an impact, because our cloud business today is primarily running inference." James Mitchell.

















