Build4 distinct publishers3 min readPublished
OpenAI plans to end Cursor's direct access to its models on November 12, citing doubts that SpaceX will stay inside its terms of service. Whether developers can still bring their own OpenAI keys remains unclear.
The Engineer · Build desk
Compiled by The EngineerSomething wrong?How this is made
The 5% figure is where this starts. Cursor's public reply says OpenAI models are about 5% of its traffic, and the same reply declines to treat the decision as final [7]. That figure is Cursor's own, and it is an aggregate over a base the company describes as millions of developers [11]. For it to describe your exposure, two things would have to hold: your model mix matches the house average, and request share tracks the share of work you cannot let regress. Neither is usually true. A release-gating review agent pinned to one model is one line of config and a disproportionate share of the value.
The bigger question is how much work this creates. OpenAI's post does not identify which models are affected, and does not say whether developers keep a path to OpenAI through their own accounts or API keys [10]. Those are different projects. If bring-your-own-key survives, the fix is billing plus a config edit. If it does not, you are re-selecting a model inside an editor whose repository context and configured workflows were built around the old one [16].
The trigger itself is harder to pin down. OpenAI told Reuters it was not confident SpaceX would use its technology "within our terms of service" [4], and latent.space reports the blogpost also cites "our experience with Elon Musk's companies violating contracts" [6]. Neither is a signal a customer can poll. Cursor said SpaceX closed the all-stock acquisition on August 14, a deal Axios valued at $60B against the $29.3B post-money mark from November 2025 [8][9], about 2.05 times the earlier number [20]. Supply changed roughly two weeks after ownership did [1].
The reporting does not agree on the day. runtimewire dates the announcement August 29, 2026, and notes OpenAI never stated the cutoff year [2]. Reuters-sourced coverage puts it on August 28 [5], and the Economic Times says Friday [4]. August 28, 2026 was a Friday [21]. Measured from August 29, the notice period is 75 days [19], one day longer if the Reuters date is the right one.
Why this option was available to OpenAI now and not last year is the mechanism worth keeping. According to latent.space, a year ago Cursor appeared in the GPT-5 launch video and a cutoff was a nonstarter because Claude models led coding by a wide margin [13]. Now GPT 5.6 reads as a credible alternative to the Claude 5 series and Cursor is promoting Grok 4.6 [14], with SpaceX supplying what Cursor calls the world's largest GPU fleet and first-party models added to the lineup [17]. Cutoffs get cheap once neither side is load-bearing for the other.
There is precedent. latent.space notes Anthropic did this to Windsurf while OpenAI was weighing an acquisition of it [12]. Two cases is enough to call it a procurement pattern. The OpenAI Startup Fund, for its part, was in Anysphere's $8M seed in 2022 [15]; the cheque outlasted the partnership.
In my context the right tradeoff is to treat the editor as replaceable at the model layer and to pay for that: a second provider exercised weekly against an eval set I own rather than one the tool ships, and a written answer from the vendor on which models it can still reach after November 12 [2]. It costs some throughput, but it buys independence from a trigger that lives in someone else's shareholder register.
Ranked by verification strength, evidence, and original report placement.
OpenAI said on August 29 that it is ending its partnership with Cursor, the AI coding editor led by Michael Truell, following Cursor's acquisition by SpaceX.
OpenAI said on August 28 that it plans to end the agreement supplying AI models to Cursor, and SpaceX offered no comment outside working hours.
OpenAI said on Friday that it plans to end its agreement providing AI models to Cursor, now owned by Elon Musk's SpaceX, saying it was not confident SpaceX would use its technology "within our terms of service"; SpaceX did not immediately respond to a Reuters request for comment outside regular business hours.
OpenAI's blogpost on the decision cites "our experience with Elon Musk's companies violating contracts" as the leading reason given.
Cursor said SpaceX completed the acquisition on August 14, ending a process that began with an April partnership to train models with SpaceXAI.
Axios reported the all-stock transaction valued Cursor at $60B, roughly twice the $29.3B post-money valuation Cursor announced in November 2025.
Distinct publishers with included, body-backed reporting in this cluster.
economictimes.indiatimes.com
1 article · August 28, 2026
latent.space
1 article · August 28, 2026
mezha.net
1 article · August 28, 2026
runtimewire.com
1 article · August 28, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
product
Wu says Cognition is not for sale. The more useful fact is who bought Cursor last week.1 distinct publisher
build
Grok 4.6 lands in Copilot two days after launch, and the model picker becomes a procurement problem1 distinct publisher
product
SpaceX paid $60B for Cursor to fix Grok, and Cursor's customers inherited the plan1 distinct publisher
build
Grok Build's real product is the X timeline, not the code generator1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One blogpost, read four times
Everything load-carrying here comes out of a single OpenAI post. Reuters copy at The Economic Times and Mezha supplies the terms-of-service quote, Latent Space quotes a different sentence from the same post, and Runtimewire reads it closely enough to notice it contradicts itself. What is missing is anyone on the other side: SpaceX declined to comment, and Cursor's reply reaches us only through Latent Space's paraphrase. Two publishers cannot even agree on which day the post appeared.
Big installed base, nothing switched off yet
The exposure is real but unverified and stale: Cursor's own Series D numbers — past $1B annualized, millions of developers — are the only measure of who stands to lose a model, and they are ten months old and self-reported. Against that sits Cursor's claim that OpenAI is 5% of its traffic, also self-reported, and pointing the other way. The one hard event in the window is the August 14 acquisition close; the cutoff itself is a dated intention, still labelled a proposal.
Headlines settle what the document leaves open
Latent Space says OpenAI 'shuts off' Cursor and the wire says the partnership is ending; the underlying post says 'proposal' and forgets to write the year. Only Runtimewire stops on that. The overstatement is modest and mostly in the framing of finality — and it runs both ways, since the figure that would deflate the drama, Cursor's 5% traffic share, is repeated with no verification either.
Nobody in the frame is disinterested
OpenAI is cutting off a company its own Startup Fund seeded, while selling Codex into precisely the market Cursor serves — and it attributes the decision to Musk's companies breaking contracts, a line that reads as much like litigation posture as procurement policy. Cursor answers that OpenAI is a twentieth of its traffic, which is exactly what a company with a new owner and a promised GPU fleet would say. The wire copy carries no obvious agenda, but it is only relaying OpenAI.
The spine holds; the specifics don't
That OpenAI intends to stop supplying Cursor directly, and said so in public, is not seriously in doubt. Nearly everything a developer or a platform owner would act on is: which models, whether your own API key still reaches them, whether November 12 survives a negotiation OpenAI itself calls open. A publisher-level date conflict on top of that keeps this short of high confidence.