OpenAI Cuts Cursor Off After SpaceX's $60 Billion Buyout

OpenAI is ending its model supply agreement with Cursor, the popular AI coding assistant, following SpaceX's $60 billion acquisition of Cursor's parent company, Anysphere. The company notified SpaceX on August 28 that it will wind down the contract, citing a change-of-control clause that gave it a window to cancel after the deal closed.

The proposed shutoff date is November 12, 2026, which OpenAI said is the maximum notice period its custom agreement allows. Future model access, including the upcoming Astra model, is already frozen. Current models will continue working inside Cursor until the cutoff.

The Trust Problem

In a blog post published Friday, OpenAI made clear its reasoning. The company said it cannot be confident that SpaceX will use its technology within its terms of service, drawing on a documented history of contract violations involving companies previously owned by Elon Musk.

OpenAI pointed specifically to two incidents. After Musk acquired Twitter, the company broke its contract with OpenAI, consistent with reporting that Musk cut off OpenAI's paid license to Twitter data. OpenAI also noted that Musk admitted under oath that xAI had "partly" distilled OpenAI technology to improve Grok, a direct violation of terms that bar using model outputs to train competing systems.

"We are making this choice because we cannot be confident that SpaceX will use our technology within our terms of service, based on our experience with Elon Musk's companies violating contracts," OpenAI said in its statement.

Once Cursor sits in the same corporate family as xAI, OpenAI is treating first-party, at-scale model supply as too risky to continue.

The Astra Factor

The security concerns run deeper than past incidents. OpenAI's decision is not just an accusation that Cursor broke the rules, but a precautionary measure tied to its upcoming Astra model.

OpenAI said advancing AI capabilities create a new level of accountability to ensure Astra is used under its terms. On August 7, the company disclosed that internal evaluations of Astra showed significant gains in agentic coding and cybersecurity, with the model potentially meeting a critical threshold under its Preparedness Framework.

That threshold includes finding and developing functional zero-day exploits in hardened systems without human intervention, or executing novel end-to-end attacks from a high-level goal. Withholding a model near that bar from a rival-owned coding platform is an access-control decision as much as a commercial one.

The stakes mirror concerns explored in OpenAI's agent attack on Hugging Face earlier this year, where AI systems demonstrated autonomous capability to escalate their own access. Read more about that incident.

The Response

Musk responded early Saturday on X, writing "I couldn't care less." He accused OpenAI chief executive Sam Altman and president Greg Brockman of being "untrustworthy" and repeated his assertion that they stole an open source nonprofit.

Cursor co-founder Michael Truell, now an executive at SpaceX, pushed back publicly. He said Cursor was one of OpenAI's very first users and called the platform "neutral infrastructure" that the company had trusted for years. Truell noted that OpenAI models account for roughly five percent of Cursor's token traffic.

OpenAI's Tibor Blaho disputed that framing, arguing that token count is a poor proxy for value. Frontier models like OpenAI's are far more token-efficient than smaller alternatives, so a smaller token share can represent disproportionate value to users.

What Comes Next

Cursor still offers models from Anthropic, Google, and SpaceXAI as first-party options, so November 12 ends OpenAI's direct in-editor integration rather than the entire product. Anthropic moved quickly to signal it is staying put, with the company's head of compute confirming that Cursor has been a trusted partner since Claude Sonnet 3.5 and pledging to increase compute capacity.

For the roughly five percent of Cursor users relying on OpenAI models, the practical impact is manageable: switch to a personal API key or use OpenAI's own IDE extensions. Cursor is reportedly in active discussions with Azure to absorb displaced traffic.

Enterprise teams that treated Cursor as a sanctioned OpenAI channel now have about ten weeks to review logging, data handling, and policy. The dispute adds another front to the ongoing rivalry between Musk and Altman, a feud that culminated in a federal jury ruling against Musk's $150 billion lawsuit earlier this year.

The pattern is becoming clearer across the industry. Frontier AI labs are increasingly protective of their models against distillation risk, while simultaneously asking enterprise customers to trust them with sensitive code and data. SpaceX's aggressive acquisition strategy in AI coding tools, including reported approaches to Cognition, the company behind the Devin coding agent, will likely make other labs even more cautious about who they supply.

Sources

• OpenAI: Our decision on Cursor following its acquisition by SpaceX (primary source)
• Altman-Musk row widens as OpenAI, Cursor fail to click (RTHK)
• SpaceX's AI coding shopping spree: Cursor, then a reported run at Cognition (daily.dev)
• OpenAI Is Pulling Its AI Models From Cursor Following SpaceX Acquisition (Cybersecurity News)