It’s rare to see an option contract make the front page, but that is what landed on 21 April 2026. SpaceX disclosed that it has signed a deal with Cursor - the AI coding tool made by Anysphere - giving it the right to buy the startup outright for $60 billion later this year, or to walk away with a $10 billion payment for the joint work the two teams are doing in the meantime.
That is not a typical M&A structure. It is closer to a call option on a company.
The Shape of the Deal
The reporting across Bloomberg, CNBC and TechCrunch lines up on the basics:
- Partnership now. Cursor and SpaceX are jointly building what SpaceX describes as a “next generation coding and knowledge work AI”.
- Two exit ramps later in 2026. SpaceX can either acquire Cursor for $60 billion, or pay Cursor $10 billion for the collaborative work and go its own way.
- Compute is the dowry. The project plugs Cursor’s product and developer reach into Colossus, the xAI supercomputer that SpaceX claims has compute equivalent to around a million Nvidia H100 chips.
- People have already moved. Two senior Cursor engineering leaders, Andrew Milich and Jason Ginsberg, have reportedly left Anysphere to join xAI, reporting directly to Elon Musk.
In effect, SpaceX has pre-paid for optionality. It gets to test the integration before committing $60 billion, and Cursor gets a very large floor under its near-term outcome either way.
From $2.5bn to $60bn in Fifteen Months
The valuation trajectory here is the part that makes your eyebrows lift.
- January 2025: Cursor valued at roughly $2.5 billion.
- May 2025: around $9 billion.
- November 2025: $29.3 billion post-money after a $2.3 billion Series D.
- Early 2026: reportedly targeting a $50 billion private round, with Andreessen Horowitz, Nvidia and Thrive Capital circling.
- April 2026: a $60 billion option from SpaceX.
That is a 24x move in fifteen months. Anysphere is only four years old. Cursor, the product, has ridden the shift from “AI autocomplete” to “AI agent that can reason across your repo” better than almost any other tool on the market, and the financials have followed developer adoption with unusual speed.
Why SpaceX, and Not xAI Directly
This is the part that is worth chewing on. xAI is the AI company in Musk’s orbit. SpaceX is the rocket company. So why is the rocket company holding the option?
The simplest read is capital. In February 2026, SpaceX and xAI merged in a deal Musk valued at around $1.25 trillion. Colossus - the supercomputer sitting at the centre of this - is an xAI asset, but the combined entity has SpaceX’s balance sheet behind it. A $60 billion call option needs a serious counterparty, and SpaceX is that counterparty.
The second read is strategic. xAI has been playing catch-up on coding. OpenAI has Codex and the broader ChatGPT developer funnel, Anthropic has Claude Code, and Google has deeply embedded Gemini into its own developer stack. Cursor is the one independent coding tool with a product loved by engineers and genuine distribution inside developer teams. For xAI, owning the interface is a cheaper way to become relevant to developers than trying to out-market an already crowded field.
What Cursor Gets
The obvious answer is compute and cash. The more interesting answer is distribution into a specific kind of user.
Cursor’s growth so far has been bottom-up: individual developers, then teams, then enterprises. Plugging into Colossus and xAI’s model pipeline gives Anysphere access to a scale of training and inference compute that would be hard to buy on the open market right now, at any price. If the joint product is genuinely “coding and knowledge work AI” rather than just a better autocomplete, that compute is the difference between “fast editor” and “agent that can refactor a codebase overnight”.
The $10 billion walk-away fee is also worth dwelling on. Even if the acquisition never closes, Cursor has effectively locked in a nine-figure, possibly ten-figure, payout from the partnership. That changes the risk profile for the Series E investors reportedly getting involved.
What It Means for the Rest of the Market
A few second-order effects worth watching:
- Consolidation at the top. Developer-facing AI is narrowing fast. If this deal closes, three of the main AI coding interfaces (Codex, Claude Code, Cursor) will sit inside three of the main frontier labs (OpenAI, Anthropic, xAI). Independent coding tools become a much harder place to start a company.
- Pricing pressure. When the interfaces are owned by the model providers, expect aggressive bundling. Free or near-free Cursor tiers backed by Colossus compute would put meaningful pressure on GitHub Copilot and smaller independents.
- Model neutrality becomes a question. Cursor has always let users pick a model, including Claude and GPT. If xAI ends up owning it, the commercial gravity will pull towards Grok and whatever successor xAI is training on Colossus. Whether Cursor stays genuinely multi-model is something developers should watch closely.
- The IPO backdrop. SpaceX is reported to be considering an IPO. Owning a fast-growing AI revenue line like Cursor - rather than just rockets and satellites - materially changes the narrative for public investors.
The Honest Uncertainty
Options are exercised, or they are not. SpaceX can still walk away, pay the $10 billion, and leave Cursor to a more traditional exit or IPO path. That path is not obviously worse for Anysphere. A fully independent Cursor with $10 billion in the bank and a deep technical collaboration already done is still one of the most interesting companies in software.
What this deal does is set a price. For a four-year-old startup making a code editor, the market - or at least Elon Musk’s slice of it - is now willing to sign a contract that says Cursor is worth $60 billion. That number will shape every other conversation in AI developer tools for the rest of the year.