The all stock acquisition of Cursor, the AI powered coding editor, commits a large slice of SpaceX's record IPO cash to an in house AI stack anchored by Colossus, its million chip training supercomputer.
SpaceX has agreed to acquire Anysphere, the company behind the Cursor AI coding editor, in a $60 billion all-stock deal that, if it closes in the third quarter of 2026 as expected, would convert a record-setting IPO windfall into a single bet on an in-house AI stack. The price, structure, and timing were first reported by Gizmodo, citing a regulatory filing.
The transaction is more than an acquisition of a popular developer tool. Read as SpaceX appears to want it read, it is a vertical-integration move that pairs Cursor's product and its grip on a community of expert software engineers with Colossus, SpaceX's in-house training supercomputer, which the same Gizmodo report puts on the scale of a million Nvidia H100 chips. The April 2026 cooperation announcement that produced this acquisition right also gave SpaceX a $10 billion alternative: pay Cursor for joint work rather than buy it. SpaceX chose to buy.
The choice is harder to dismiss than the wire treatment of a typical $60 billion tech deal because it lines up with the way SpaceX has talked to public investors about its post-IPO plans. The IPO itself, which the Gizmodo report describes as the largest in history, raised roughly $75 billion in cash at a $135 per share offer price, with shares trading near $211 as of the report, a roughly 50 percent gain for the initial buyers. Spending a meaningful share of that paper wealth on a single counterparty is a different kind of decision than spending the same amount over several quarters across multiple targets. It commits SpaceX to one thesis.
The thesis has three parts. First, product: Cursor is one of the most widely used AI coding editors, and the company also sells AI coding agents, software that can write and edit code on its own. Second, distribution: Cursor is already installed in the workflows of a large base of professional software engineers, the same audience SpaceX needs to build applications on top of Colossus if its training hardware is to be useful to anyone outside SpaceX's own labs. Third, compute: Colossus, the training supercomputer, is the production backbone for xAI's Grok models, and it has been the clearest physical asset behind the company's claim that it can train frontier models at scale. Buying Cursor lets SpaceX move from selling raw compute and finished models to owning the toolchain that turns that compute and those models into a product a paying developer can use.
There is a risk ledger attached, and a serious read of the deal puts it next to the strategic case rather than after it. According to the Gizmodo report, SpaceX is currently unprofitable, the $28.5 trillion total addressable market the company has used to pitch investors (of which roughly $26.5 trillion is attributed to AI) is SpaceX's own projection rather than a third-party forecast, xAI has recently lost its non-Musk co-founders, and SpaceX has warned public-market investors that some features of Grok have been problematic. The personnel turnover and the product warning are the kind of disclosures a company puts in a prospectus when it wants to acknowledge trouble without owning it.
The April agreement also explains why the deal is structured as all stock rather than cash. The $60 billion price tag matches the acquisition right SpaceX negotiated two months ago, which by itself implies that Cursor's founders and investors were willing to accept SpaceX equity at a value that, given the post-IPO share price, has only risen since the original right was granted. Paying in SpaceX shares also means that the dilution from the deal lands on every existing public shareholder, not on a private buyer. That is the trade-off public investors accepted by buying into the IPO in the first place.
What to watch next is concrete. The deal has a Q3 2026 expected close, which means a regulatory review window and a chance for any competing bid for Anysphere to surface, although none has been reported. SpaceX's next earnings cycle will be the first to report Colossus utilization and revenue contribution from Cursor as a combined unit, and the first test of whether the $28.5 trillion addressable market the company has been pitching to investors survives contact with actual customer numbers. The acquisition right and the $10 billion alternative were a single bet on which way the relationship would go. SpaceX has now made the choice. The next quarterly report will be the first real test of whether the bet pays off.