The National Security Agency, Cybersecurity and Infrastructure Security Agency and Federal Bureau of Investigation accused six Chinese artificial-intelligence companies of conducting industrial-scale model distillation to extract proprietary capabilities from U.S. systems. The Sept. 8 advisory named DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun and Z.AI, alleging they obtained billions of tokens through millions of requests since at least late 2024 and that Chinese authorities were likely aware. The agencies said the firms targeted Claude, GPT, Gemini and Grok using APIs, cloud services, aggregators, overseas proxy networks, fraudulent accounts and automated systems to evade restrictions and detection. Anthropic, OpenAI, Google and SpaceX were identified as affected U.S. companies, while the extracted capabilities were allegedly used to improve mathematical reasoning, software coding, agentic functions and question answering. China rejected the advisory as unfounded. Ministry spokesperson Mao Ning described China's AI progress as domestic scientific innovation and urged Washington to stop making accusations, while Beijing called for greater cooperation on artificial intelligence. Treasury Secretary Scott Bessent separately warned that a Chinese AI lead could threaten U.S. security and said opposition to large domestic data centers could undermine competitiveness as China expands infrastructure projects.