Chinese AI company Moonshot AI's open-source large language model "Kimi K3" achieved performance similar to major AI research institutions like OpenAI and Anthropic at a far lower cost, sparking U.S. suspicions of "distillation attacks." U.S. Treasury Secretary Scott Bessent claimed in a Fox Business interview on the 21st (local time), "We have discovered watermarks of American LLMs in many Chinese AI models," and stated, "we will closely investigate this issue in the coming days or weeks," adding that "if foreign models steal technology from our great companies, we will impose sanctions on them." Last month, Anthropic suggested to the U.S. Senate Committee on Banking that Alibaba in China may have engaged in "distillation" of Anthropic’s model, calling it "the largest distillation attack ever." Distillation is a model replication and compression technique used by AI researchers for decades, where a pre-trained large model acts as the "teacher" and a smaller "student" model is created by imitating the teacher through learning from its outputs.