April 21, 2026
SpaceX (which merged with xAI in February this year) announced a deal today involving Cursor (an AI code editor; parent company Anysphere). The deal gives SpaceX two paths: the first is to acquire Cursor outright for $60 billion before the end of 2026; the second is to skip the acquisition and instead pay Cursor $10 billion for a technology partnership. SpaceX can choose which path to take within the year.
The $60 billion figure is roughly 20% above Cursor’s current independent fundraising valuation. But SpaceX hasn’t disclosed whether the $60 billion would be paid in cash or stock. Musk’s previous merger of Twitter into xAI was an all-stock deal, and SpaceX’s cash reserves at the end of 2025 fall far short of covering $60 billion. If the payment ultimately comes in SpaceX shares, the actual value will depend on post-IPO stock performance, and the gap between the $60 billion face value and what sellers actually receive could be substantial.
The substance of the $10 billion partnership is already fairly clear. Business Insider reported on April 15 that Cursor has been using xAI’s Colossus supercomputer to train its latest model, Composer 2.5. The core of the partnership is xAI providing compute while Cursor contributes its product and developer user base. The specific form of the payment (lump sum, equity investment, or revenue share) has not been disclosed.
Cursor CEO Michael Truell’s response on X mentioned only the partnership. He said nothing about the $60 billion acquisition option. Meanwhile, Cursor is still pursuing its independent fundraising round (with investors including a16z, Thrive Capital, and Nvidia). Running both tracks simultaneously suggests the Cursor team is in no rush to be acquired.
Talent movement has already begun ahead of any deal. Two senior engineering leads at Cursor, Andrew Milich and Jason Ginsberg, left in March to join xAI, reporting directly to Musk.
SpaceX is planning to go public by the end of June, targeting a $1.75 trillion valuation and raising $75 billion, which would make it the largest IPO in history. This week, SpaceX is conducting closed-door roadshows for institutional investors at its Starbase facility in Texas and its Memphis data center.
TechCrunch’s assessment is that announcing the ability to acquire AI’s hottest product company for $60 billion during an IPO roadshow primarily serves to convince institutional investors that SpaceX has integration capabilities in the AI space. Whether the deal actually closes is secondary. From a funding perspective, the more likely outcome is the $10 billion partnership path: SpaceX trades compute for technology integration and brand association without taking on the financial burden of a full acquisition.
CNBC also noted another dimension in the timeline: this announcement comes less than a week before Musk’s trial against OpenAI CEO Sam Altman in a Northern California court, and OpenAI happens to be an early investor in Cursor.
Over the past two years, major tech companies have developed a new operational playbook: instead of buying companies, they buy only the people and the technology. The industry calls this a reverse acqui-hire.
The mechanism involves signing two separate agreements simultaneously. One is an employment offer for key employees, recruiting founders and critical engineers directly into the acquiring company. The other is a technology license, granting the acquirer rights to the target company’s IP. The company itself does not change hands; remaining employees and operations stay in place. Viewed separately, neither agreement is large enough to trigger antitrust review. That is precisely the design intent.
In March 2024, Microsoft paid $650 million and hired nearly all 70 employees of Inflection AI along with its two co-founders. Mustafa Suleyman became CEO of Microsoft’s AI division. Inflection continued operating under a new CEO, pivoting to B2B API licensing. The FTC opened an antitrust inquiry in June but took no enforcement action.
In July 2025, Windsurf (formerly Codeium, an AI coding tool, roughly 250 employees) went through three ownership changes.
Round one: OpenAI agreed to acquire Windsurf for $3 billion, but the deal fell through after the exclusivity period expired.
Round two: Within hours of the exclusivity period ending, Google signed a $2.4 billion deal to hire CEO Varun Mohan, co-founder Douglas Chen, and roughly 40 senior R&D staff, placing them in DeepMind to work on Gemini’s agentic coding capabilities. The remaining 210 or so employees and company assets stayed behind.
Round three: 72 hours later, Cognition (the company behind Devin) acquired the remaining parts for an estimated $250-300 million in Cognition stock. Cognition CEO Scott Wu offered all Windsurf employees full accelerated vesting (meaning all unvested equity vested immediately). But three weeks later, Cognition laid off 30 people and offered buyouts to the remaining nearly 200.
In December 2025, NVIDIA paid a total of $20 billion to take roughly 90% of Groq’s employees and CEO Jonathan Ross, securing a technology license for Groq’s LPU inference chip architecture. Jensen Huang’s exact words in an internal email were: we did not acquire Groq as a company. Just three months earlier, Groq had raised $750 million at a $6.9 billion valuation. By the GTC conference in March 2026, NVIDIA had already integrated Groq’s chip architecture into its own Vera-Rubin platform.
Amazon executed a similar operation on Adept in June 2024, hiring away roughly 80% of its employees. Google hired away roughly 21% of Character AI’s core team in August 2024. In the first half of 2025, the six major tech companies collectively executed 67 acqui-hires, a 60% increase over the 42 completed in all of 2024.
In a traditional acquisition, the entire company changes hands and all shareholders participate proportionally in the exit. Acqui-hires work differently: the acquirer selectively picks individuals. Those selected receive generous sign-on packages. Those not selected stay behind, holding equity in a company whose core talent and IP have already departed, with the equity’s value significantly diminished.
The Windsurf case is the clearest reference point. Google’s $2.4 billion covered only the roughly 40 people who were hired away. Employees who had joined Windsurf within the past year received nothing from the Google deal. People cried during the all-hands meeting. Cognition’s acquisition 72 hours later brought the price down from OpenAI’s earlier $3 billion offer to $250-300 million in Cognition stock, a 90% reduction. And that stock came with layoffs that began three weeks later.
Deedy Das of Menlo Ventures put it this way: when you join a company, the default social contract is that everyone shares in the big outcomes together. Under this model, employees who joined later bear a disproportionate share of the risk.
Based on public information, SpaceX’s $60 billion / $10 billion structure for Cursor differs from the acqui-hire cases described above. It at least formally includes an option to acquire the entire company.
But several signals suggest the actual trajectory may be closer to an acqui-hire. Key engineering talent has already been flowing to xAI since March. The payment method for the $60 billion remains unclear and may involve SpaceX stock rather than cash. Cursor is simultaneously pursuing independent fundraising, and its leadership appears to prefer remaining independent. If the $10 billion partnership path prevails, SpaceX gains compute-based integration and technology alignment while Cursor remains independent but continues to see its core talent gravitate toward xAI. That outcome would not be far from the practical effect of an acqui-hire.
For Cursor’s rank-and-file employees, the two paths carry very different risks. The $60 billion full acquisition would mean all shareholders exit proportionally, the best possible outcome. The $10 billion partnership path means the company continues operating independently, but if the attrition of key talent persists, the equity held by remaining employees will be progressively diluted. Windsurf’s lesson is precisely this: what determines employees’ fates is not just the final deal structure, but which direction talent was already moving before the deal was struck.