Cursor weighs phasing out Chinese AI models amid U.S. security scrutiny
Rising adoption of Chinese open-weight artificial intelligence models by American technology companies is drawing sharper attention in Washington as competition between the U.S. and China intensifies. Cursor is now considering moving away from those models because of geopolitical risk, a shift that follows its use of Moonshot AI technology in its Composer 2 product.
Highlights
- Cursor informed U.S. congressional committees it is weighing phasing out PRC-developed open-weight AI models due to geopolitical risk after SpaceX's $60 billion acquisition.
- House report urges a national American open-weight AI initiative and expanding federal bans on PRC-developed models to all government systems and contracts, escalating supplier compliance pressure.
- The U.S. Commerce Department's Bureau of Industry and Security announced a new interim final rule to block Chinese companies' remote access to Nvidia AI chips via overseas data centers.
Congressional probe spotlights Cursor's model plans
As reported by CNBC, a report from the House Committee on Homeland Security and the House Select Committee on China says Cursor plans to explore phasing out PRC-developed open-weight models because of geopolitical risk. The document says the company informed the committees of those plans during a months-long investigation into the use of Chinese AI models in the American technology stack.Cursor, SpaceX's artificial-intelligence coding platform, built its coding model Composer 2 on top of Chinese lab Moonshot AI's Kimi K2.5. SpaceX acquired Cursor in August at an implied $60 billion equity value, according to the report's account cited by CNBC.
The committees' investigation also included engagement with Airbnb and DoorDash over their use of Chinese AI models. Lawmakers say Chinese systems are gaining traction because they offer lower costs, customization and near-frontier performance, while the U.S. still lacks a competitive open-weight alternative at the price and performance levels the market demands.
Policy response widens pressure on Chinese AI adoption
The House report argues that Beijing is seeking to convert technological gains into global market share and long-term dependence. It recommends launching a national American open-weight initiative that would combine public computing power, shared evaluation infrastructure and lawful access to American frontier models for training.The report also says an order made public in October, which banned DeepSeek from federal civilian agency systems, should be expanded by prohibiting PRC-developed AI models in federal information systems and in the performance of federal contracts. That would increase compliance pressure on suppliers and software companies that serve government customers.
Washington is also focused on Chinese companies gaining remote access to Nvidia chips through overseas data centers and advancing through distillation, a technique in which newer models use older and more established systems to aid training. In a related move, the U.S. Commerce Department's Bureau of Industry and Security announces on Monday an interim final rule aimed at preventing Chinese companies' remote access to AI chips.
We have also discussed the broader context around EU import curbs as Brussels moves toward managed trade to reduce reliance on China.