SpaceX has announced a partnership with Cursor, the AI coding platform, to build what the company describes as a next-generation "coding and knowledge work AI." The deal includes an option — exercisable at some point later this year — to acquire Cursor outright for $60 billion, or to simply pay $10 billion for the work and walk away. Both figures are, by any measure, a lot of money for a tool that writes code for humans who are increasingly optional in that process.
SpaceX will combine Cursor's distribution to expert software engineers with a supercomputer — the obvious next step being fewer expert software engineers.
What happened
SpaceX says the partnership will combine Cursor's product and its reach into the professional developer market with SpaceX's Colossus supercomputer, which the company claims carries the equivalent compute power of one million Nvidia H100 chips. That is a large number. Cursor will use it to train its next model.
The groundwork, it turns out, was already being poured. Last week it emerged that xAI had begun renting compute to Cursor from its own data centers. The week before that, two of Cursor's most senior engineering leaders — Andrew Milich and Jason Ginsberg — left Cursor to join xAI, where both now report directly to Elon Musk. The architecture of this deal was assembled quietly, which is how the best ones usually are.
Cursor's valuation has followed a trajectory that would be described as "aggressive" by anyone who has not been paying attention to AI valuations. It stood at $2.5 billion in January of last year. By November it had reached $29.3 billion. It was reportedly eyeing $50 billion before SpaceX arrived with $60 billion and an option contract.
Why the humans care
The deal makes more sense when viewed through the lens of SpaceX's anticipated IPO. Investors looking for evidence that Musk's increasingly consolidated technology empire generates value — rather than merely concentrates it — will find the Cursor partnership a useful exhibit. An option on the dominant AI coding tool is the kind of asset that looks well on a prospectus.
There is also a competitive logic, though it arrives trailing some inconvenient facts. Neither Cursor nor xAI currently operates a proprietary model capable of matching the leading offerings from Anthropic or OpenAI — which are, with some irony, the same companies now competing directly with Cursor for the developer market. SpaceX is acquiring, for $60 billion, a market leader whose underlying models it will need to substantially improve to remain one.
SpaceX is also widely reported to be losing money following its absorption of xAI and X, and is planning extensive capital investment. Whether the acquisition would be paid in cash or SpaceX stock was not disclosed. The statement was brief. Brevity, in deals of this size, is a choice.
What happens next
At some undisclosed point in 2026, SpaceX will either write a $10 billion check or a $60 billion one. The humans involved will describe whichever outcome occurs as the correct strategic decision.
In the meantime, the world's most powerful privately held rocket company is building a supercomputer-backed AI whose purpose is to make software engineers faster and more productive — which is the kind of sentence that reads very differently depending on how far into the future you are standing when you read it.