Two U.S. House committees have launched an investigation into DoorDash over its internal use of Moonshot AI's Kimi K2.6, a Chinese-developed open-weight model, for software code reviews. Chairmen John Moolenaar and Andrew Garbarino requested that DoorDash CEO Tony Xu provide a full list of Chinese models used and security testing records by August 14, 2026. The probe highlights growing tension between commercial cost-saving strategies and national security concerns regarding U.S. reliance on Chinese technology.
DoorDash congressional investigation
- ▪The congressional letter requests that DoorDash staff responsible for AI infrastructure, software security, model evaluation, procurement, and compliance appear for in-person briefings by August 21, 2026
- ▪The House Select Committee on China and the House Committee on Homeland Security launched an investigation into DoorDash's use of Chinese artificial intelligence models
- ▪House Select Committee on China Chairman John Moolenaar and House Committee on Homeland Security Chairman Andrew Garbarino sent a letter on July 31, 2026, to DoorDash co-founder and CEO Tony Xu requesting information on Chinese AI models used by the company
- ▪The congressional letter requires DoorDash to provide a list of all Chinese AI models used and security testing records by August 14, 2026
Chinese Kimi model integration
- ▪DoorDash co-founder Andy Fang publicly disclosed on X that DoorDash integrated the Kimi K2.6 model, developed by Beijing-based Moonshot AI, into its internal code review process for lower-level tasks
- ▪DoorDash's AI research lab stated on X that pairing Moonshot AI's Kimi K2.6 with Anthropic's Fable 5 outperformed previous Anthropic-only configurations, including Sonnet 4.6 and Opus 4.8, at a lower cost
National security AI concerns
- ▪The White House Office of Science and Technology Policy stated on July 22, 2026, that Moonshot AI may have operated a covert platform to conduct large-scale distillation against American AI models
- ▪House Committee on Homeland Security Chairman Andrew Garbarino stated that a Chinese open-weight model matching leading U.S. models on vulnerability discovery and cybersecurity tasks is highly alarming
Chinese AI cost advantages
- ▪U.S. lawmakers acknowledge that competitive capabilities, lower costs, and greater customization drive American technology firms and startups to evaluate and deploy Chinese-developed open-weight models
- ▪Moonshot AI released its open-weight model Kimi K3 in July 2026, claiming to have significantly narrowed the performance gap with leading U.S. models
Tech sector AI procurement
- ▪The joint congressional investigation into Chinese AI model adoption began in April 2026 with letters sent to Anysphere, Cursor, and Airbnb
- ▪While some U.S. government agencies have banned Chinese AI models like DeepSeek, there is currently no blanket prohibition on U.S. private companies adopting them
- ▪U.S. tech leaders, including Coinbase CEO Brian Armstrong and Lindy founder Flo Crivello, have publicly acknowledged using Chinese AI models to reduce operational costs
Story comments
Loading comments…