OpenAI has announced that it will stop supplying its models to Cursor, the AI coding tool that SpaceX acquired in a $60 billion deal just two weeks earlier. The shutdown will begin on November 12.

The separation comes amid intensifying competition in the fast-growing generative AI market. According to Fortune Business Insights, the global generative AI market is projected to grow from $103.58 billion in 2025 to $161 billion in 2026. Among developers and businesses, the OpenAI–Cursor split has raised a broader concern: how much choice can developers expect when the leading model makers restrict access to rival platforms?

Buyers have long valued the ability to switch between models based on price, performance, or the task at hand. OpenAI's move suggests that accessibility itself can become a competitive lever. Cursor, a prominent AI coding tool, is set to lose one of its providers as spending on AI platforms and models climbs 63.4% this year.

Why OpenAI says it is walking away

OpenAI tied the termination of the partnership to Elon Musk. According to the company's statement, it lost confidence in SpaceX's willingness to honor the terms of service, citing an alleged pattern of Musk's companies breaching contracts. It pointed to Twitter, now part of SpaceX, and to xAI, which OpenAI said Musk admitted under oath had violated its terms this year. The dispute is the latest flashpoint in the long-running rivalry between OpenAI and Musk, who co-founded the company in 2015 before departing and later launching xAI as a competitor.

Control is also a factor. OpenAI's tailored agreement with Cursor allowed it to terminate the deal within a specified time period following a change of ownership. The company also said its future Astra model heightens the stakes over who gets access to its systems.

OpenAI said it selected the latest stop date permitted under the agreement and described the decision as "incredibly tough." It added that no future models would be made available to Cursor.

How much Cursor actually loses

Cursor's management downplayed the immediate consequences of the termination. Michael Truell, co-founder of Cursor and now a SpaceX executive, said on X on August 29 that OpenAI models account for roughly 5% of Cursor's traffic. He added that the two companies had been negotiating the exit and that Cursor was among OpenAI's earliest customers.

That 5% figure makes the immediate operational impact of the split appear manageable. The broader issue, however, goes beyond the loss of a single provider.

Cursor's core selling point was giving customers the ability to choose among models depending on the task. While OpenAI's absence is unlikely to cripple Cursor, it removes the assumption that developers can access every popular model whenever they need it. The loss is likely to cause more difficulty for companies entering long-term agreements than for Cursor's end users. How the transition plays out after the November 12 cutoff — and whether other model providers take similar steps with rival-owned platforms — will be a key signal for the rest of the developer ecosystem.

The deal that triggered the breakup

According to an 8-K filing with the Securities and Exchange Commission, SpaceX's purchase of Anysphere, Cursor's parent company, was finalized on August 14. The filing values Cursor at $60 billion, with its common and preferred shares convertible into approximately 389.3 million shares of SpaceX Class A stock. According to Cryptopolitan, the all-stock deal ranks among the largest startup acquisitions ever completed.

Cursor is now part of SpaceX's larger AI initiative. In its acquisition announcement, it said it would gain access to "the largest fleet of GPUs in the world," enabling greater computing power to develop more advanced models at lower cost.

That context helps explain why OpenAI's decision carries strategic weight: the company will not supply future models to a platform that is increasingly capable of building competing systems.

A market splitting into camps

The break comes as spending on AI models accelerates. Gartner projected on July 20 that worldwide end-user spending on AI models and platforms will rise from $39.311 billion in 2025 to $64.252 billion in 2026, a 63.4% increase. Spending on generative-AI models alone is forecast to grow 117%.

For programmers, the most immediate effect will be fewer options within one of the most popular coding tools. For the broader AI industry, the implications run deeper: the era when any model could be freely used by anyone is drawing to a close. As AI firms consolidate control over their models, infrastructure, and distribution channels, they have growing incentives to protect those advantages for themselves.