SpaceX finished what it started in April: on August 14, 2026, it closed its $60 billion all-stock acquisition of Cursor maker Anysphere โ the largest amount ever paid for a startup.
This isn't a new headline number. The deal was announced months ago and covered extensively as a valuation milestone for Cursor's funding history. What's new is that it's actually done: the SEC 8-K confirming the closing dropped this week, the shares have converted, and SpaceX has told the market exactly what it's going to do with the company it just bought. That's the part worth digging into.
Figures from SpaceX's SEC Form 8-K filing and company statements, August 14, 2026.
How the deal actually closed
The transaction was structured as a straight stock-for-stock swap. SpaceX issued approximately 391 million new Class A shares to finalize the purchase, with Cursor's outstanding stock converting into those shares alongside additional restricted stock units and options that SpaceX assumed from Anysphere's cap table. No cash changed hands โ Anysphere's founders, employees, and investors are now paid entirely in SpaceX (Nasdaq: SPCX) equity.
That structure matters. It means Cursor's shareholders are making a direct bet on SpaceX's stock holding its value, not cashing out into something liquid and diversified. Given SPCX has already round-tripped from a $2.1 trillion first-day close in June down to roughly $1.5 trillion by late July before recovering, that's a real variable for anyone who just became a SpaceX shareholder through this deal.

What SpaceXAI actually is
Cursor doesn't stay standalone. It becomes a wholly owned unit inside a newly rebranded division called SpaceXAI, and gets direct access to SpaceX's Colossus supercomputer for training and inference โ the same compute backbone that's been powering xAI's Grok models since SpaceX absorbed that business earlier in the year. On X, Cursor's own account confirmed the team would work on Grok Build, Grok Bot, and the Grok API going forward.
Read that plainly: SpaceX isn't buying Cursor to run it as an independent SaaS business generating its own P&L. It's buying the team, the coding-agent technology, and the $4 billion-plus ARR customer base, and merging all of it into the same AI stack that already includes Grok. This is the second major startup SpaceX has folded into that stack in under a year, after xAI's own ~$1.25 trillion combination with SpaceX.
The exit math for Cursor's cap table
Anysphere's Series D priced the company at roughly $29.3 billion in January 2026. A $60 billion all-stock exit seven months later is about a 2x step-up โ a strong outcome by any measure, but a modest one relative to how fast Cursor's revenue was actually growing. The company reportedly hit $4 billion in annualized revenue by June, up 4x from $1 billion a year earlier. On that growth trajectory, a founder could reasonably have argued for holding out for a bigger multiple in a later round rather than selling into SpaceX stock at 15x trailing ARR.
The counterargument is the one that appears to have won: staying independent means competing against Anthropic's Claude Code, GitHub Copilot, and OpenAI's Codex, all of which are backed by labs that also own the underlying model layer Cursor has always had to license or build on top of. Selling into SpaceX solves that dependency in one move โ Cursor's cofounders now sit inside a company that owns its own frontier models (via xAI) and its own compute (via Colossus), removing the single biggest structural risk an AI coding wrapper carries.
What it means for the AI coding market
This closes the loop on what had been the largest pending M&A deal in the AI coding space, and it changes the competitive map. Cursor was already the category leader by revenue and enterprise adoption; now it's backed by a trillion-dollar-plus parent with its own model lab and its own supercomputer, rather than running as an independently funded startup that has to raise its next round against Anthropic and OpenAI's balance sheets.
For competitors, the calculus shifts. GitHub Copilot has Microsoft's distribution and Anthropic's model access; Anthropic's own Claude Code has the model advantage in-house; OpenAI's Codex has GPT-5's coding benchmarks behind it. Cursor's answer, as of this week, is that it no longer needs to win on model access or compute cost independently โ it can borrow both from SpaceX and xAI while keeping its product and distribution edge. Whether that combination out-executes labs that control the entire stack top to bottom is the question this acquisition was designed to answer.
Bottom line: The SpaceX-Cursor deal stopped being a rumor or a signed term sheet this week and became a completed, SEC-confirmed transaction โ the largest all-stock acquisition of a startup on record. Cursor's founders and investors are now betting their outcome on SPCX stock rather than cash, Cursor's product roadmap is now tied to Grok rather than running independently, and the AI coding market just consolidated its clearest leader into the same balance sheet that already owns xAI. Whoever wins the AI coding wars from here is competing against a company that no longer needs outside capital, outside models, or outside compute to do it.
Get VC data most people never see
โ 100% free
Weekly benchmarks, valuations, and fund data. Join 5,000+ investors. No spam.