OpenAI Plans to End Direct Model Supply to Cursor After SpaceX Acquisition

Summary

OpenAI plans to wind down its contract that supplies OpenAI models directly to Cursor following Cursor's acquisition by SpaceX. OpenAI says it has proposed November 12, 2026 as the shutoff date and is using the maximum notice period available under its agreement.

The change does not mean OpenAI models become unusable inside the Cursor application. OpenAI has published several alternative access paths, including bringing your own OpenAI API key, using the Codex IDE extension, or connecting through a compatible AI gateway. The main change is to Cursor's own direct model-supply relationship with OpenAI.

Quick Take

• OpenAI intends to end its direct model-supply contract with Cursor after the SpaceX acquisition.
• November 12, 2026 is the proposed transition date, not yet the confirmed final termination date.
• Cursor can keep using the OpenAI models covered by the current agreement during the transition unless access ends sooner.
• OpenAI says it will not provide future models to Cursor under the existing contract.
• Developers can still use OpenAI models in supported Cursor workflows through their own API access, Codex, or compatible gateways.

What changed

On August 28, OpenAI said it had notified SpaceX that it intends to wind down the contract through which OpenAI provides models to Cursor.

OpenAI says the agreement contains a limited cancellation window following a change of control. Cursor announced on August 14 that it had officially been acquired by SpaceX, completing a process that began with an April model-training partnership.

OpenAI's proposed shutoff date is November 12, which the company says gives developers the maximum transition period allowed under the contract. OpenAI's Help Center makes clear that the date still needs to be confirmed between the companies and that Cursor could choose to end access earlier.

Why OpenAI says it is ending the agreement

OpenAI says the decision is tied to the ownership change and its concerns about whether SpaceX will use OpenAI technology within OpenAI's terms.

The company specifically points to past contractual disputes involving other Elon Musk-controlled companies and says that, as AI capabilities advance, it has greater responsibility to ensure future models are used according to its terms.

OpenAI also references its upcoming model, Astra, saying that this increased accountability is part of the reason it is exercising the change-of-control provision now.

This is OpenAI's stated rationale. It should be treated as the company's position rather than as an independently adjudicated finding about SpaceX or Cursor.

What Cursor users can still do

OpenAI has already published migration guidance for developers who want to keep using OpenAI models inside Cursor.

The distinction matters because OpenAI is ending a commercial supply arrangement, not blocking every possible use of OpenAI technology inside the Cursor editor.

Bring your own OpenAI API key

Cursor supports a user's own OpenAI API key for supported local Chat and Agent requests. Those requests are billed through the user's OpenAI API account rather than through a Cursor or ChatGPT subscription.

OpenAI notes that this does not apply to every Cursor feature. Cursor Tab and autocomplete, Auto routing, Cloud or Background Agents, Automations, the Cursor CLI, and Cursor's API and SDK rely on models supplied or routed by Cursor.

Use Codex inside Cursor

OpenAI says the Codex IDE extension can run directly inside Cursor with an eligible ChatGPT subscription or an OpenAI API key.

Because Codex operates as its own extension, it does not depend on OpenAI models appearing in Cursor's built-in model picker.

Route through a compatible gateway

Organizations can also use supported provider integrations or OpenAI-compatible gateways. OpenAI lists examples including Amazon Bedrock and Azure and notes that gateway compatibility and feature support can vary.

Why this matters beyond Cursor

The larger issue is model-supplier dependency.

AI coding products increasingly combine their own interface, agents and orchestration with models supplied by several outside companies. That architecture gives users choice, but it also creates a business dependency: pricing changes, licensing terms, acquisitions or contractual disputes can alter which models a product can bundle.

For developers and enterprise buyers, the Cursor situation makes three questions more important:

• Can users bring their own model credentials?
• Can workflows move between providers without major rework?
• Which product features depend on models supplied directly by the platform?

Products with multiple access paths are generally easier to adapt when provider relationships change.

What to watch next

The most important near-term signal is the final termination date. OpenAI currently describes November 12 as a proposed transition date, not a finalized cutoff.

It is also worth watching whether Cursor and OpenAI reach a different arrangement, how Cursor adjusts its bundled model lineup, and how heavily Cursor leans on models developed within the SpaceX ecosystem after the acquisition.

AI World Scope take

The headline is a supplier split, but the durable lesson is about portability. AI development tools are becoming platforms built on top of other AI platforms, and that makes model access itself a dependency that teams should plan around.

For Cursor users who specifically need OpenAI models, the immediate task is not to abandon Cursor. It is to identify which workflows depend on Cursor-supplied OpenAI access and which can move cleanly to a personal API key, Codex, or another supported route.

Sources & Documentation

• OpenAI — Our decision on Cursor following its acquisition by SpaceX (August 28, 2026)
• OpenAI Help Center — Using OpenAI models in Cursor
• Cursor — Cursor is now a part of SpaceX (August 14, 2026)