OpenAI Cuts Off SpaceX-Owned Cursor: Model Access Ends November 12

Two weeks after SpaceX closed its $60 billion all-stock acquisition of Anysphere, the startup behind the wildly popular AI coding editor Cursor, OpenAI has served notice: it intends to wind down the contract that supplies GPT models to Cursor, with a proposed shutoff date of November 12, 2026. The announcement, published late Friday, August 28, turns model supply into an explicit weapon in the long-running and increasingly personal feud between Sam Altman and Elon Musk — and it leaves millions of developers caught in the crossfire.

What OpenAI announced

In a blog post titled “Our decision on Cursor following its acquisition by SpaceX,” OpenAI wrote: “Today, we notified SpaceX that we intend to wind down our contract providing OpenAI models to Cursor, with a proposed shutoff date of November 12, 2026. To maximize the time that developers can retain access to our models through Cursor, we are giving the maximum notice provided by our contract.”

The stated rationale is blunt. OpenAI says it “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.” The company added that it will not provide future models to Cursor as the agreement winds down — meaning the entire GPT family, including GPT-5.6 Sol, Terra, and Luna, will disappear from Cursor’s model picker over the coming months.

Notably, November 12 is the maximum notice OpenAI’s custom contract allows. OpenAI is framing the timing as developer-friendly — giving coders the longest possible runway to migrate — but the subtext is unmistakable: the company no longer trusts its models sitting inside a Musk-controlled product.

The backdrop: a $60 billion deal that changed everything

SpaceX completed its acquisition of Anysphere on August 14, 2026, according to financial filings. The deal, valued at $60 billion in an all-stock transaction, brought one of the fastest-growing developer tools in history under the same corporate umbrella as X and xAI — both of which Musk had folded into SpaceX earlier this year ahead of its June public listing.

That corporate entanglement is precisely what alarmed OpenAI. With xAI’s Grok now a sibling product to Cursor, OpenAI’s models would effectively be running inside a direct competitor’s flagship developer tool. Coverage from multiple outlets has pointed to concerns about model distillation — the practice of using a frontier model’s outputs to train a rival — as the unspoken fear behind the decision, especially given ongoing litigation between OpenAI and xAI over exactly that allegation.

The acquisition also transformed Cursor’s role: once a “neutral infrastructure” platform that offered models from every major lab, Cursor is now part of the Musk ecosystem, and OpenAI has concluded that neutrality can no longer be assumed.

Cursor’s response: a shrug and a talking point

Cursor CEO Michael Truell responded on X Friday night, downplaying the impact: “OpenAI models serve about 5% of Cursor user traffic, and we’re speaking with the OpenAI team to resolve this. 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.”

The 5% figure is the company’s central talking point — a message to developers and investors that losing GPT access, while annoying, is not existential. Claude models from Anthropic have long dominated Cursor usage, and Grok models will presumably gain prominence now that xAI sits inside the same corporate family.

Musk himself responded with characteristic hostility on Saturday, writing: “I couldn’t care less. Scam Altman and Greg Stockman are utterly untrustworthy,” before repeating his longstanding claim that Altman and President Greg Brockman “stole an open source nonprofit.” Musk sued OpenAI, Altman, and Brockman in 2024 over the company’s for-profit conversion, lost the case earlier this year, and has vowed to appeal.

Anthropic pounces

The most strategically interesting response came from Anthropic. Co-founder and Chief Compute Officer Tom Brown posted on X Friday night: “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.”

The irony was not lost on observers. In June 2025, Anthropic itself abruptly blocked Windsurf — then Cursor’s chief rival — from accessing Claude models ahead of that company’s acquisition, citing similar concerns. Replit CEO Amjad Masad reminded Brown of that history directly, and Docker executive Mat Velloso replied: “This was a great opportunity to stay quiet.” Meanwhile, Anthropic is now renting compute capacity from SpaceX, making it a simultaneous partner and rival in Musk’s orbit — a posture that sums up the strange, double-crossing geometry of today’s AI industry.

Why this matters beyond the drama

First, model supply is now a competitive weapon. When frontier models can be switched off based on who owns the customer, “neutral infrastructure” is dead. Every AI application builder now has to price in the risk that their model provider may one day see them as a strategic threat — or be acquired by one.

Second, the consolidation wave has a human cost. Developers who built workflows around GPT-5.6 inside Cursor now face a forced migration, and Cursor’s Tab completion and agent modes will need to rebalance around Claude, Grok, and open-weight alternatives. The “vibe coding” market is heating up just as OpenAI prepares for an expected IPO next year and Anthropic holds preliminary banker meetings for a listing that could value it as high as $2 trillion.

Third, the feud is now structural, not personal. With SpaceX public, xAI embedded alongside Cursor, OpenAI heading to market, and litigation ongoing on multiple fronts, the Altman–Musk conflict has become a fault line that splits vendors, tools, and developer ecosystems. Choosing a coding assistant is increasingly choosing a side.

What happens next

The November 12 date is described as “proposed,” and Truell says the companies are still talking — so a negotiated extension isn’t impossible. But with OpenAI citing contract violations and Musk responding with insults, a graceful de-escalation seems unlikely. Developers depending on GPT models inside Cursor should treat the 74-day clock as real, audit which of their workflows touch OpenAI models, and evaluate Claude, Grok, or direct API access before the deadline forces the decision for them.

One era of AI tooling ended quietly on August 14, when a neutral Cursor became part of the Musk empire. On August 28, OpenAI made that shift impossible to ignore.