Anthropic's head of threat intelligence, Jacob Klein, stated that foreign adversaries are illegally accessing their Claude models to train competing technology and sell copycat versions at lower prices. The company specifically named Chinese AI labs Moonshot AI (developer of Kimi K3), DeepSeek, MiniMax, and Alibaba (developer of Qwen models) in its allegations. Klein highlighted a growing illicit ecosystem, including the dark web, used to evade controls and create fraudulent accounts for large-scale model access. The Trump administration had previously deemed such distillation that undermines American proprietary information as "unacceptable," vowing to explore measures to hold foreign actors accountable. Anthropic, valued at nearly $1 trillion, is expected to go public as early as October.