SpaceX just confirmed its $60 billion all-stock deal for Cursor. Here are the 7 things buried in the fine print that most coverage skipped.
Every outlet ran the same number this week. SpaceX, $60 billion, Cursor, done. Almost none of them ran the part that actually explains why it happened, what it costs if it falls apart, or the one statistic that makes the whole deal look a lot stranger than a press release.
Here’s the full picture.
1. The deal, in one breath
On June 16, SpaceX filed an Agreement and Plan of Merger to acquire Cursor — the AI coding tool built by Anysphere — for an implied $60 billion, paid entirely in SpaceX stock. No cash changes hands. A SpaceX shell entity called X67 merges into Cursor, Cursor survives as a wholly owned subsidiary, and the exact exchange rate gets calculated off SpaceX’s average share price in the seven trading days before closing. Translation: the $60 billion is a moving target until the deal actually closes, which SpaceX is targeting for Q3 2026, pending regulatory approval.
That’s the part every headline had. Here’s what most of them buried or skipped.
2. This wasn’t sudden. It was scheduled.
SpaceX didn’t wake up and decide to buy Cursor this week. Back in April, it disclosed it had already locked in the right to either acquire Cursor for $60 billion later in the year, or pay $10 billion just to keep the existing partnership running if the full buyout fell through. Read that again: SpaceX had a contractual fallback plan worth ten billion dollars in case its sixty-billion-dollar plan didn’t work out. That’s not two companies flirting. That’s an engagement announcement after the venue was already booked.
And if the deal somehow collapses now, even after the formal filing, SpaceX is on the hook for a $1.5 billion termination fee plus $8.5 billion worth of computing resources owed to Cursor. Whatever this is, it isn’t casual.
3. The fine print everyone skipped
A few details buried in the actual merger filing are more interesting than the headline number ever was.
Thrive Capital holds positions in both SpaceX and Cursor. The combined stake is now worth more than $10 billion — meaning one investor is sitting on both sides of this exact transaction, watching both halves of their portfolio move on the same announcement.
Before the merger agreement was even signed, certain Cursor shareholders entered “Revest Agreements” with SpaceX — retention arrangements that tie compensation to staying on after the deal closes. Strip away the legal language and what you’re left with is this: SpaceX wasn’t just buying software. It was locking down the people who built it, before they had the chance to leave with the payout.
There’s also a section of the agreement that spends an unusual amount of space on IPO-related scenarios — lockups, resale restrictions, what happens if shares move on the public market. On its own, boilerplate. In the context of a company that itself just went public days earlier, it’s started a quiet conversation among analysts about whether SpaceX is assembling a stack of AI assets with one eye on a future listing of its own AI division.
4. The number that should bother you
Here’s the twist nobody led with: Cursor’s market share has been shrinking. Per spending data from Ramp, Cursor controlled 41% of the AI coding tools market in June 2025. By this past May, that had fallen to roughly 26% — while Anthropic, in the same period, grew to control about half the category.
So the actual story isn’t “SpaceX bought the market leader.” It’s “SpaceX paid a record price for a company that’s losing ground to a competitor, betting that money and compute can reverse a trend that was already running against it.” That’s a fundamentally different story than the one most coverage told, and it’s the one worth sitting with.
5. Why SpaceX, why now
Context matters here. This deal landed days after SpaceX’s own IPO — the largest in history. Shares priced at $135 and were trading above $200 within days, briefly pushing SpaceX past Amazon and into a near-tie with Microsoft for the title of the fourth most valuable public company in the US. When your stock just minted itself into a trillion-dollar war chest in a week, a $60 billion acquisition stops looking reckless and starts looking like spending money you just printed.
The strategic logic: SpaceX’s AI division, built around xAI (which it absorbed in February), has been trying to close the gap with Anthropic and OpenAI in coding tools. SpaceX and Cursor have reportedly been jointly training a model on SpaceX’s Colossus supercomputer for months, with the output set to ship inside both Cursor and Grok Build. The acquisition formalizes a relationship that was already running.
Here’s the part that should make you raise an eyebrow, though: even as it buys Cursor to compete with Anthropic, SpaceX has separately signed data center deals with Anthropic and Google worth a combined $2.15 billion a month. It’s racing one rival while renting compute from it. That’s not hypocrisy, exactly — it’s just what happens when the AI industry is small enough that everyone needs everyone else’s infrastructure to function.
6. The Karachi twist
One more thing, and it’s not a footnote — it’s the reason this story matters here. One of Cursor’s four co-founders, Sualeh Asif, grew up in Karachi, studied at Nixor College, and represented Pakistan at the International Mathematical Olympiad three years running before any of this happened. He’s Anysphere’s Chief Product Officer now, and his slice of the company’s $29.3 billion valuation reportedly made him a billionaire before he turned 26. We’ve gone deep on his story separately — https://startup.pk/sualeh-asifs-cursor-just-sold-to-elon-musks-spacex-for-60-billion/
7. The bigger question
Step back from the individual facts and a harder question sits underneath all of them: is this deal evidence that AI is consolidating into a handful of vertically integrated empires, or evidence that even the people running those empires aren’t sure their products can win on merit alone so they’re buying market share and locking down engineers instead?
SpaceX didn’t just buy a product. It bought a team it had to contractually bind to stay, a market position that was already eroding, and a relationship it had been quietly building for months before announcing anything. None of that is necessarily wrong. It might be exactly how serious companies are built right now. But it’s a different story than “rocket company buys hot AI startup,” and it’s worth knowing which story you’re actually reading.
And for Pakistan, there’s a second, quieter question worth sitting with: we should be loud and proud about Sualeh Asif. We should also be honest about the fact that he built this in San Francisco, on MIT’s runway, with American venture capital — not here. Celebrating the outcome is easy. Building the conditions that keep the next Sualeh Asif from having to leave is the actual work.
That’s everything. The number was never really the story.
