US intelligence and cybersecurity agencies have formally identified six Chinese artificial intelligence companies as engaging in industrial-scale efforts to replicate the capabilities of leading US frontier models.
What Happened
In a joint release issued Tuesday, the National Security Agency (NSA), the Cybersecurity and Infrastructure Security Agency (CISA), and the Federal Bureau of Investigation (FBI) named DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI as the firms involved. The agencies alleged that these companies have been attacking US models since at least late 2024 to extract capabilities from variants of Claude, GPT, Gemini, and Grok. According to the agencies, the firms "likely" acted with "Chinese government awareness" during these operations.
Why It Matters
The core of the accusation rests on the economic and strategic advantages gained through this process. The agencies stated that China-based AI companies conducting industrial-scale distillation against US models see "significantly shorter AI development timelines and reduced financial expenditures in training a frontier model." By copying capabilities rather than developing them from scratch, these firms may have spared billions in Chinese development costs, potentially altering the competitive landscape for global AI leadership.
The Bottom Line
US agencies have publicly accused six major Chinese AI firms of aggressively distilling capabilities from top US models to save time and money, alleging state-level awareness of these operations.