OpenAI announced it is winding down its contract providing OpenAI models to Cursor, "with a proposed shutoff date of November 12, 2026" — which OpenAI states is the maximum notice period its contract allows. This follows directly from SpaceX's acquisition of Cursor's parent company Anysphere for $60 billion, which closed on August 13.
OpenAI's stated reason is contract-compliance history, not a product dispute
OpenAI's own post states the decision plainly: "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 cites two specific precedents directly. First, after Musk acquired Twitter — now part of SpaceX — the company "broke the terms of our contract (alongside many others)," per OpenAI, linking to what it says is documentation of that violation. Second, OpenAI states that "under oath earlier this year, Musk admitted that xAI, now also part of SpaceX, had violated OpenAI's terms of service (terms which are similar to xAI's own)." Worth being precise about what OpenAI is and isn't claiming: it isn't alleging Cursor itself has done anything wrong — the stated concern is about what SpaceX, as Cursor's new owner, might do with continued access, based on the contract history of other Musk-controlled entities.
The Musk admission behind OpenAI's second precedent, sourced directly
OpenAI's post cites that under-oath admission without quoting it. Forbes reports on the testimony directly: during cross-examination in Musk's own lawsuit against OpenAI, Musk admitted that xAI "partly" trained its Grok models by distilling OpenAI's outputs — training on another company's model outputs rather than solely on independently licensed or collected data, a practice OpenAI's terms of service prohibit. The irony is structural, not incidental: Musk sued OpenAI over its conduct, and in defending that case he was questioned closely enough to concede a practice that is itself the terms-of-service violation OpenAI now points to. Worth being precise about what the admission does and doesn't establish: "partly" doesn't specify a share of Grok's training data or which generation was involved, and this environment could not independently fetch forbes.com to quote the testimony verbatim — but the setting matters. Testimony damaging to Musk's own side, extracted through cross-examination in a case he initiated himself, carries stronger provenance than a leak or a rival's accusation.
This also lands squarely inside an already-live industry fight over distillation. The White House accused Moonshot AI of distilling Anthropic's Fable to build Kimi K3 in July, and Anthropic CEO Dario Amodei has called for "stopping industrial-scale distillation" without naming a target. Musk's sworn admission gives that generic framing a specific, first-party American example — xAI, conceded by its own founder, against a fellow American lab — that undercuts any version of the distillation debate that treats the practice as something only Chinese labs do to American ones.
The mechanism and the timing
OpenAI states its custom agreement with Cursor "gives us a limited time window to cancel it after a change of control" — meaning this notice had to be delivered within a specific window following the SpaceX acquisition's close, not at OpenAI's own discretion on any timeline. OpenAI also connects this directly to its own upcoming model: "As AI capabilities advance, we also have a new level of accountability to ensure our upcoming model, Astra, is being used in accordance with our terms." OpenAI states it chose to hold the cancellation to the latest permitted date specifically to maximize developer access time, rather than cutting access immediately.
What OpenAI says about the relationship and the affected developers
OpenAI's post states it has "worked with Cursor for nearly four years and have enormous respect for their team, their product, and what they've built for the developer community," and separately: "We know that the people most affected by this decision are the developers who rely on OpenAI models in Cursor. We care about their experience in this transition and we're ready to go above and beyond to support them." OpenAI doesn't specify in this announcement what that support will concretely involve.
Cursor's CEO responds, and Anthropic commits to the opposite path
Cursor CEO Michael Truell answered directly on X: "We're sorry to see that OpenAI put out a note saying they plan to block Cursor users from accessing OpenAI models in three months." He adds a figure OpenAI's own announcement didn't include — "OpenAI models serve about 5% of Cursor user traffic" — and states Cursor is "speaking with the OpenAI team to resolve this," leaving open whether the cutoff is final or still negotiable. The sharper line is the close: "Cursor was one of the very first users of OpenAI, we've worked closely with their team for years, and we've trusted their platform to be neutral infrastructure for our business." That's a direct rebuttal, in substance if not in name, of the "neutral infrastructure" framing OpenAI's own post relied on to justify the cutoff — Truell is arguing that trust in exactly that neutrality is what's being broken here.
Separately, and worth reading against the low 5% traffic figure, Anthropic co-founder Tom Brown posted that "Cursor has been a trusted partner of Anthropic since Sonnet 3.5. We'll continue to increase compute to support Claude models in Cursor and are excited for what comes next with them at SpaceX" — amplified approvingly by Thariq (@trq212), who called Cursor's team best-in-class for bringing AI coding to the world. Anthropic explicitly naming SpaceX in a commitment to increase, not restrict, access is a notable contrast with OpenAI's decision: the same change-of-control event both companies are reacting to produced opposite responses from the two labs, and Anthropic's own recent history with Musk's companies — the xAI distillation admission covered above — makes the difference more pointed, not less.
What to expect next
- Watch whether "speaking with the OpenAI team to resolve this" produces an actual change to the November 12 date. Truell's post stops short of disputing OpenAI's contract-violation rationale and stops short of confirming the shutoff will happen on schedule — it's a request to negotiate, not a rebuttal of the facts OpenAI cited.
- Watch how Cursor replaces OpenAI-model access by November 12 for the roughly 5% of traffic Truell says runs on OpenAI models. Cursor already has deep SpaceXAI/Grok integration following the acquisition, and now an explicit Anthropic compute commitment — worth watching which absorbs that share.
- Watch whether this becomes a template for other frontier labs' contracts with SpaceX-owned products, given OpenAI explicitly frames the decision around a documented pattern across multiple Musk-controlled companies rather than a Cursor-specific concern — and given Anthropic has now publicly staked out the opposite position.
References: OpenAI — Our decision on Cursor following its acquisition by SpaceX, read directly in full · OpenAI's announcement on X · Forbes reporting on cross-examination testimony in Musk v. OpenAI, as described — this environment could not independently fetch forbes.com to quote the testimony verbatim · @mntruell on X — Michael Truell's response · @NotTomBrown on X — Tom Brown's response · @trq212 on X · related coverage: The Cursor Acquisition Closes: SpaceX Owns the $60B Deal It Signed in June · The Open-Weight Fight · Frontier Arcade: trends & predictions