Lawmakers probe growing use of Chinese AI models in U.S. companies

What happened

The House Committee on Homeland Security and the House Select Committee on China launched a joint investigation into the growing adoption of Chinese AI models by U.S. companies. Initial inquiry letters were sent to Cursor (pending SpaceX acquisition) and Airbnb.

Context and impact

Chinese AI models — including DeepSeek, Kimi, and Alibaba Qwen — accounted for 30-46% of weekly AI token usage at U.S. companies since February 2026 (up from ~11% the prior year). The driver is cost: Chinese models are 60-90% cheaper than OpenAI and Anthropic. The U.S. government cites data security, censorship, and CCP ideological alignment as key concerns.

Details

• Committees: House Homeland Security + House Select Committee on China
• Initial targets: Cursor, Airbnb
• Chinese model share of US enterprise tokens: 30-46% (peak since Feb 2026)
• Measures under consideration: federal procurement bans for government contractors
• Challenge: open-source weights (DeepSeek) cannot be technically banned
• State Department: Chinese AI models promote CCP narratives and censor dissent