US-China AI rivalry heats up as Chinese models gain ground

China's AI capabilities have rapidly advanced, narrowing the gap with the United States in just a few months. While US companies like Anthropic and OpenAI have invested billions in top-tier AI models, Chinese firms have opted for open-source alternatives such as DeepSeek, Qwen, and MiMo. Moonshot AI's recent unveiling of Kimi K3, a model that matches or surpasses Anthropic's Claude Fable 5 in key benchmarks, underscores China's growing prowess.

This move has prompted US President Donald Trump to threaten restrictions on US access to Chinese AI models, with the Treasury Secretary warning of potential sanctions over intellectual property theft. US AI firms, Anthropic and OpenAI, have accused Chinese labs of using their models for distillation, a process where advanced reasoning is harvested through excessive queries.

China rejects these allegations, labeling them as "technological bullying." The US remains committed to maintaining its lead in frontier AI, recognizing its potential impact on military, economic, and technological supremacy. While China's open-source models have provided cost-effective solutions, analysts caution that the US AI market may face a correction in the near future, as Chinese developers continue to build comparable models at a fraction of the cost.