Cursor said in a blog post published under its own byline that SpaceX has completed its acquisition of the company, closing a process the two sides first announced in April. The post frames the deal as the conclusion of a partnership built to speed up Cursor’s model training, not a change in how the product works day to day.
This is the outcome of a trade AI Insiders has been tracking since spring. Bloomberg reported in May, citing a person familiar with the matter, that Cursor’s annualized revenue had crossed $3 billion and that SpaceX held a call option to buy the company for $60 billion once SpaceX’s own shares began trading. Bloomberg also reported separately, later that same month, that xAI’s legal team had warned staff to limit contact with Cursor employees while a distinct acquisition track was already underway. xAI has since folded into SpaceX and taken the name SpaceXAI. Cursor’s post does not use the $60 billion figure or confirm whether that was the final price. It states the outcome and skips the arithmetic that got there.
Cursor’s stated reason for the deal is compute. The company says SpaceX gives it “access to the largest fleet of GPUs in the world,” which it expects will let it train stronger models at lower cost. As evidence, the post points to Grok 4.6, the model xAI shipped earlier this month, calling it an early look at what the two companies can build together. That is Cursor’s own characterization of its new owner’s technology, made in an announcement designed to reassure customers, not a claim verified by an outside source.
What the post does not say matters more than what it does. It gives no transaction terms: no price, no equity structure, no mention of whether the $60 billion figure from Bloomberg’s reporting held. It says nothing about what happens to Cursor’s pricing or its enterprise contracts, the exact question AI Insiders flagged when covering xAI’s internal warning in May. And it does not address the question closest to Cursor’s product identity: whether the company will keep letting customers choose among competing labs’ models inside its editor.
That last point is the real tension for builders who use Cursor. Over the past year, the product’s pitch rested partly on being neutral ground: developers could route a coding task to Claude, to a GPT model, or to Gemini depending on which performed best that week. Cursor’s new owner has its own frontier lab in Grok, and the blog post itself treats Grok 4.6 as a preview of the combined roadmap. Nothing in the announcement says exclusive routing is coming. Nothing rules it out either.
The historical parallel is worth naming. When a distribution platform acquires or gets acquired by a model maker, the acquiring party’s incentive to favor its own model tends to grow over time, even when day-one messaging promises no change (Codeium’s Windsurf saga and Google’s periodic Gemini defaults inside Workspace both followed that arc). Cursor’s post promises the product stays “familiar.” It does not promise the model menu stays the same.
For teams running multi-model workflows through Cursor, the next ninety days are the window to watch for signals: any change to which models ship as defaults, any shift in pricing tiers, and any language in renewal contracts about model availability. If Grok becomes the fastest or cheapest option inside Cursor and nothing else changes, that is a product decision. If it becomes the only fully supported option, that is a different company than the one those contracts were signed with.
Cursor announced the completed SpaceX acquisition on its own blog on August 17, 2026.
