OpenAI announced it will terminate access to its AI models via Cursor, following SpaceX's acquisition of the AI coding startup earlier this month. The company cited concerns over SpaceX's compliance with OpenAI's terms of service, referencing prior contract breaches involving Elon Musk's companies. The planned cutoff for model access is set for November 12, 2026, and OpenAI confirmed it will not provide future model support to Cursor as it winds down their agreement.

Cursor CEO Michael Truell acknowledged that about 5% of Cursor's user traffic currently relies on OpenAI’s models and stated that discussions with OpenAI are ongoing to resolve the situation. Truell highlighted Cursor’s longstanding relationship with OpenAI, emphasizing their trust in its technology as a neutral infrastructure. However, OpenAI’s decision reflects growing tension after SpaceX finalized a $60 billion deal to acquire Cursor’s parent company, Anysphere, on August 14.

This development is part of an extended feud between Elon Musk and OpenAI leadership, including CEO Sam Altman and President Greg Brockman. Musk, who co-founded OpenAI in 2015 but left the board in 2018, has been highly critical of its for-profit transition and filed a lawsuit in 2024 contesting the company’s restructuring. Although Musk lost the lawsuit earlier this year, he has vowed to appeal and continues to attack OpenAI leadership publicly.

The move also fits a broader industry pattern where AI model providers limit access to partners linked with competitors. For instance, Anthropic recently restricted its Claude AI model access for Windsurf, a competitor, even as it continues working with SpaceX. As the AI-assisted coding market heats up, OpenAI prepares for a planned public offering next year, while rivals like Anthropic are also exploring IPO possibilities at multibillion-dollar valuations.