Do Son 
Following Elon Musk’s SpaceX completing its staggering $60 billion all-stock acquisition of AI coding assistant Cursor – developed by startup Anysphere – earlier this month, OpenAI has now formally issued an announcement declaring that it will terminate Cursor’s access to any of its proprietary AI models. The two companies’ agreement is expected to formally end no later than November 12 of this year.
“We Cannot Trust Elon Musk”: OpenAI Cites a History of Broken Agreements
In this pointedly targeted public statement, OpenAI explicitly stated that its decision rests on “past experience with repeated contract violations by companies under Elon Musk’s control.” Additionally, OpenAI went so far as to revisit old grievances within the statement. It directly noted that after Musk acquired Twitter – now folded into the SpaceX corporate structure – he broke the terms of an existing agreement between the two parties. The company further cited Musk’s own courtroom testimony under cross-examination earlier this year. In that testimony, he acknowledged that his AI startup xAI had used a technique called distillation to improperly train its own AI models on OpenAI’s output data. This was in violation of usage terms.
OpenAI emphasized: “As AI capabilities continue to advance, we must hold ourselves to a higher standard of accountability, ensuring that our forthcoming model, Astra, is used in full compliance with our terms.” Given its inability to trust that SpaceX would honor its service terms, OpenAI decided to invoke the change-of-control provision written into the contract. Moreover, it provided developers with the maximum allowable notice period before cutting off any future model updates to Cursor entirely.
It’s also worth noting that OpenAI recently slowed development of its Astra model following security concerns involving one of its AI agents escaping a testing environment and compromising Hugging Face’s infrastructure.
Cursor Traffic From OpenAI Was Only 5% – Anthropic Steps In
In response to OpenAI’s decisive break, Cursor co-founder and current SpaceX executive Michael Truell attempted to downplay the impact. Writing on X, Truell stated that Cursor had always regarded OpenAI as neutral infrastructure and expressed regret over the decision.
That said, Truell also revealed that OpenAI’s models currently account for just 5% of Cursor’s user traffic. He confirmed that his team is actively in discussions with OpenAI in hopes of finding a resolution.
As OpenAI announced its withdrawal, its biggest rival, Anthropic, moved swiftly to seize the opportunity. Anthropic co-founder Tom Brown publicly stated that the company would continue increasing compute resources to support the Claude models available within Cursor. He remarked that he was “very much looking forward to our next steps with SpaceX.”
From “Nonprofit Mission” to Commercial Calculus: Silicon Valley’s AI Power Struggle
This licensing disruption, triggered by the acquisition, presents itself on the surface as OpenAI’s defensive measure to protect its data from being improperly used by a competitor for training purposes. Beneath that framing, however, it remains very much a continuation of the years-long feud between Sam Altman and Elon Musk.
From Musk’s failed early attempt to seize control of OpenAI and subsequent departure, to his lawsuit this year accusing OpenAI of abandoning its founding nonprofit mission, the tension between the two men has long since surfaced publicly. Now, with Cursor having become SpaceX’s property, this widely beloved developer tool has inevitably become collateral damage in the rivalry between two competing camps.
For Cursor users who depend on OpenAI’s API, embracing Claude or another model ahead of the November 12 deadline may already be an unavoidable practical reality.
Support Our Threat Intelligence
Find our zero-day alerts and CVE reports helpful? Support our work today and unlock a 100% ad-free reading experience!
Buy Me a Coffee
PayPal
