OpenAI Cuts Off Cursor as Musk Ties Collide With Billion-Dollar Revenue

OpenAI is ending its partnership with Cursor, the AI coding startup now tied to Elon Musk’s empire, even though the arrangement had become worth more than $1 billion in annualized revenue to the ChatGPT maker. The move shows how seriously OpenAI is taking the risk of doing business with Musk-linked companies as it prepares for a public offering next year.

The decision is more than a commercial break-up. It reveals the extent to which OpenAI is willing to sacrifice short-term revenue, developer access and a major customer relationship to avoid possible misuse of its models by SpaceX, which recently acquired Cursor in a reported $60 billion transaction.

OpenAI said it is winding down the relationship because it cannot be confident SpaceX will follow its terms of service, citing what it described as past contract violations involving Musk-controlled companies. The company’s stance lands at a delicate moment: OpenAI is trying to present itself as a larger, more diversified and more dependable business ahead of a likely IPO.

What happened between OpenAI and Cursor?

OpenAI said late last week that it will discontinue its partnership with Cursor, one of the most widely used AI coding editors in the market. Cursor had been a major customer of OpenAI’s models, purchasing access so developers using the product could tap into ChatGPT-maker technology inside the editor.

People familiar with the matter told WIRED that Cursor ranked among OpenAI’s top five customers by revenue at the beginning of 2026. By spring, OpenAI was estimating that the business could generate more than $1 billion in annualized revenue.

That makes the split notable not just because of the money involved, but because OpenAI is voluntarily walking away from one of its most valuable enterprise relationships. The company framed the move as a trust and governance decision, not a pricing dispute or a product misalignment.

OpenAI said it was ending the partnership because it could not be confident SpaceX would use its technology within the company’s terms of service, pointing to what it described as a pattern of contract violations by Musk’s businesses.

OpenAI declined to comment beyond its public statement. Representatives for SpaceX and Cursor did not immediately respond to requests for comment.

Why is Elon Musk at the center of the breakup?

Elon Musk is the reason the partnership became politically and strategically untenable for OpenAI. Cursor’s acquisition by SpaceX brought the startup into Musk’s orbit, and OpenAI appears to have concluded that the risks of keeping its models available outweighed the revenue it was earning.

For OpenAI, the concern is not simply that Musk is a former cofounder turned rival. The deeper issue is that his companies are now direct competitors in the AI race, while also maintaining access to OpenAI technology through commercial relationships. That combination creates a classic conflict-of-interest problem.

OpenAI also referenced statements Musk reportedly made during a deposition in his lawsuit against the company, suggesting xAI had used OpenAI models to train its own systems. Musk sued OpenAI alleging the company had strayed from its original charitable mission; a federal jury dismissed the case earlier this year.

The company’s public reasoning suggests that trust is now a business variable. If OpenAI believes a customer might use its models in ways that undermine its competitive position, it is willing to cut off even a highly lucrative account.

How much revenue is OpenAI giving up?

OpenAI is giving up a customer that, according to people familiar with the situation, was projected to bring in more than $1 billion in annualized revenue. That figure had not previously been reported.

The loss matters, but it lands differently now than it would have a few years ago. OpenAI reportedly now brings in more than $40 billion in annualized revenue across multiple product lines, including subscriptions, advertising inside ChatGPT and sales of access to its AI coding product Codex.

That broader revenue base gives OpenAI room to make strategic cuts that would once have been too costly. The company is signaling to investors that it can depend less on a single customer or a single distribution channel, and that it can absorb a hit in order to protect long-term competitive interests.

Still, the company acknowledged a downside. Ending the relationship may weaken its ties with developers, a group OpenAI has spent years trying to cultivate.

• Cursor had been a major source of model usage and revenue.
• OpenAI now has a larger and more diversified business base.
• The company may lose some goodwill among developers who relied on the integration.
• The decision reduces exposure to a Musk-controlled business.

What does this mean for developers?

For developers, the immediate effect is uncertainty about how one of the most popular coding tools will access frontier models going forward. Cursor built its reputation partly by offering developers convenient access to models from OpenAI and other providers inside a coding environment.

OpenAI’s withdrawal could change the product experience for some users, especially if access to OpenAI models becomes limited or unavailable inside Cursor. It may also force the startup to rely more heavily on other model providers.

At a broader level, the episode shows how dependent many developer tools remain on a small set of AI model makers. A product like Cursor can become a platform for different model families, but those partnerships can disappear quickly if the business or ownership structure changes.

Why does this matter for OpenAI’s developer strategy?

It matters because OpenAI has spent years positioning itself as the default AI platform for builders. Losing a marquee coding customer risks undercutting that message, even if the company can afford the revenue loss.

The company appears to be making a trade-off: preserve model integrity and competitive control, even if it means alienating a valuable audience that expects broad access and flexibility.

How did Cursor respond?

Cursor’s cofounder and chief executive, Michael Truell, responded publicly on X hours after OpenAI’s announcement. He argued that OpenAI’s models accounted for only a small slice of Cursor’s user activity, saying they served roughly 5% of user traffic.

Truell’s point seemed aimed at minimizing the commercial importance of the relationship and suggesting that Cursor would not be badly hurt by the loss of OpenAI’s models.

OpenAI’s head of core products, Thibault Sottiaux, pushed back, saying that token usage should not be treated as a direct measure of revenue or value created. He asked Truell to show the math behind the 5% figure.

Truell said OpenAI models served about 5% of Cursor user traffic, while OpenAI’s product chief responded that usage volume is not the same thing as revenue or value.

The exchange highlights a common tension in the AI market: raw usage data can be easy to cite, but it does not always reveal the actual financial or strategic importance of a partnership.

Why was the partnership so complicated in the first place?

The relationship had always been unusual because OpenAI and Cursor were at once collaborators and competitors. OpenAI’s own coding business, Codex, has grown into a meaningful product line, which means it increasingly competes with third-party tools that rely on OpenAI’s models.

At the same time, Cursor had been one of the clearest examples of a startup that could build a successful product on top of OpenAI’s technology while still offering developers an interface and workflow distinct from ChatGPT.

That balance held for more than a year, and at times the companies appeared to have a workable coexistence. Cursor remained a customer, and OpenAI benefited from usage and revenue while also learning from one of the most popular developer tools in the market.

But Musk’s involvement appears to have changed the calculus entirely.

What changed after SpaceX acquired Cursor?

The acquisition made Cursor part of a Musk-controlled ecosystem, and that raised the stakes for OpenAI. Once the startup was inside SpaceX, OpenAI had to consider whether its models might be exposed to a rival with both strategic motivation and technical ability to exploit them.

OpenAI’s public explanation suggests it concluded that the risk of model leakage, misuse or contractual violation was too high. In that sense, the breakup is about trust in governance as much as trust in a specific customer.

How does this fit into OpenAI’s bigger business strategy?

This fits into OpenAI’s effort to look like a mature, diversified company rather than a research lab dependent on a few partners. As it prepares to go public next year, the company is trying to demonstrate that it can manage risk, control its distribution and avoid overreliance on powerful allies or adversaries.

OpenAI reportedly has more than $40 billion in annualized revenue now, and the business is spread across consumer subscriptions, ad products and developer offerings. That gives the company leverage to cut off a lucrative customer if doing so helps protect strategic assets.

In investor terms, the message is clear: OpenAI wants to be seen as a business that can choose discipline over convenience.

Why would an IPO make this decision more important?

An IPO would put even more pressure on OpenAI to show predictable governance and reduced single-customer risk. Public-market investors often reward recurring revenue, strong controls and lower exposure to controversial counterparties.

By severing ties with a Musk-linked customer despite the revenue hit, OpenAI may be trying to prove that it can act like a company with a long-term risk framework rather than one chasing every dollar of near-term growth.

What role did OpenAI play in Cursor’s rise?

OpenAI was not just a vendor to Cursor; it was also an early backer. The company’s startup fund participated in Cursor’s 2023 seed round, helping support the startup at a stage when many investors still viewed AI coding assistants as speculative.

That early involvement makes the breakup even more striking. OpenAI once helped nurture the company, and at one point reportedly explored acquiring it. Those talks never got far, but they show how closely the two companies had been linked before Cursor became a major product in its own right.

By the time OpenAI launched GPT-5 in 2025, Truell was featured prominently in the company’s marketing. That earlier public relationship now reads like a snapshot from a very different phase of the AI market.

What does Anthropic’s role tell us about the market?

Anthropic’s response shows that AI model providers are making selective decisions about which partners they will support when competitive stakes rise. After OpenAI’s announcement, Anthropic cofounder Tom Brown said Claude would remain available in Cursor.

That could be interpreted as a willingness to stay flexible, but the business reality is more complicated. Anthropic also has its own infrastructure dependence tied to Musk’s orbit, which means the company may have incentives to maintain a workable relationship.

Anthropic has previously shown that it is willing to cut off access when strategic conditions change. Last year, it severed ties with the AI coding startup Windsurf after reports that OpenAI might acquire it. Windsurf later ended up being bought by Cognition.

Anthropic has previously suggested that selling Claude to a company OpenAI might acquire would be awkward, underscoring how quickly AI partnerships can become politically charged.

The fact that Claude remains available in Cursor shows that model access is becoming a negotiation tool across the industry, not just a technical feature.

Is this about competition or principle?

It is about both. Model vendors talk about policy compliance, but they are also competing for position in a market where coding tools are becoming one of the most important gateways to developers.

In practice, the companies that control the best models will increasingly choose where those models can be used, and with whom they are willing to build.

What the OpenAI-Cursor split says about the AI industry

The break-up is a sign that the AI market is maturing from a phase of broad cooperation into one of tighter strategic boundaries. Early on, model providers needed as many distribution channels as possible, and startups needed access to frontier models to build products quickly.

Now, the biggest companies are competing for the same customers, the same developers and the same infrastructure. That overlap creates friction.

For OpenAI, the decision to cut off Cursor suggests that control is becoming more important than reach. For Cursor, it means surviving in a world where access to leading models can change overnight depending on ownership and strategy.

The broader industry implication is that partnerships are no longer just commercial arrangements. They are becoming tests of trust, governance and competitive alignment.

Timeline of the dispute

The sequence below shows how a startup partnership turned into a strategic break.

What comes next?

OpenAI’s decision is unlikely to be the last time a major AI company cuts off a valuable customer over competitive concerns. As frontier model providers become bigger businesses, the strategic risk of selling to rivals will only rise.

For Cursor, the immediate task is to reassure users and developers that its product will keep working with competitive model coverage. For OpenAI, the challenge is to prove that losing a billion-dollar customer was a disciplined business move, not a sign of instability.

The larger question is whether the industry can sustain a model in which the same companies are both suppliers and competitors. In this case, OpenAI answered that question by drawing a hard line.

That line may cost the company money in the short term. But it also says something powerful about where the AI business is headed: the era of easy partnerships is ending, and strategic trust is becoming one of the most valuable assets in the market.