SpaceX’s artificial intelligence coding platform Cursor is considering moving away from using Chinese AI models due to “geopolitical risks,” CNBC has learned.
Cursor, which SpaceX acquired in August for a stock value of $60 billion, had built its Composer 2 coding model based on Kimi K2.5 from Moonshot AI, a Chinese lab.
The company is one of several U.S. companies adopting the Chinese model, whose use has proliferated around the world this year, worrying lawmakers in Washington. With improved Chinese models, the battle for AI supremacy between the US and China will reach new heights in 2026.
But a report prepared by the House Committee on Homeland Security and the House Select Committee on China, shared exclusively with CNBC ahead of Friday’s announcement, says Cursor now plans to “consider phasing out the China-developed Indiscriminate Weight Model due to geopolitical risks.”
The report follows a months-long investigation into the introduction of Chinese AI models into the U.S. technology stack, and Cursor said he “informed the committee” of those plans. Airbnb and DoorDash also engaged in investigations into China’s use of AI models.

SpaceX announced in June that it had entered into a definitive agreement to acquire Cursor. The transaction was completed in August. SpaceX is developing the Grok family of models, and Cursor is involved in training them.
Cursor, SpaceX and DoorDash did not respond to CNBC’s requests for comment. Airbnb declined to comment. The U.S. Department of Commerce’s Bureau of Industry and Security did not respond to requests for comment.
Introduction of AI in China
The report found that the cost, customization, and near-state-of-the-art performance of China’s AI models are driving adoption, and that the U.S. lacks competitive open-weight alternatives at “the price and performance the market demands.”
“Beijing is now seeking to translate these technological advances into global market share and long-term dependence,” the report said. The Chinese embassy in the UK did not respond to requests for comment. China’s Foreign Ministry did not immediately respond.
It also provides recommendations to address identified issues.
These include the launch of the “America’s National Nondiscrimination Initiative,” which combines public computing power, shared assessment infrastructure, and “legal access” to the American Frontier Model for training.
The report also says the use of “Chinese-developed AI models” in federal information systems and the performance of federal contracts should be prohibited, building on an October order banning DeepSeek from federal civilian agency systems.
Separately, the US government is also concerned that Chinese companies are gaining ground by accessing NVIDIA chips remotely through overseas data centers and using “distillation,” where newer models use older, more established models for training.
The U.S. Department of Commerce’s Bureau of Industry and Security (BIS) on Monday announced an interim final rule to prevent Chinese companies from remotely accessing AI chips.
