SpaceX Buys Cursor for $60 Billion, Four Days After Going Public
The largest acquisition of a venture-backed startup ever puts an AI coding tool inside the Musk complex.
On 16 June 2026 SpaceX said it had agreed to buy Cursor, the AI coding editor, for $60 billion in stock. The announcement landed four days after SpaceX itself went public. The company listed on Friday 12 June at $135 a share and was trading above $200 in the pre-market on Tuesday morning, adding close to $1 trillion to its valuation in a few days. The two events are connected: a share price that had just risen that far makes a $60 billion offer, paid in your own stock, much easier to write.
Cursor is a code editor. A developer installs it, it reads the project, and an AI model inside it writes, edits and explains code. The company behind it was founded in 2022 as Anysphere, took part in OpenAI's startup accelerator in 2024, and carried a price tag of roughly $29 billion before the SpaceX deal was announced. It was on course to raise another $2 billion from Andreessen Horowitz, Thrive and Nvidia at a $50 billion valuation. It agreed to be sold instead.
The deal was less a surprise than a conclusion. In April 2026, ahead of the IPO, SpaceX had already announced an unusual arrangement: it would either buy Cursor for $60 billion in stock or pay a $10 billion break-up fee if the deal collapsed. SpaceX expects the acquisition to close in the third quarter of 2026.
Why an editor was priced like infrastructure
Two things made the number possible. The first is that coding tools turned out to be the part of the AI business where money actually changes hands. Cursor rose alongside AI-assisted coding over the two years to mid-2026, raising $900 million in a Series C in June 2025 and a further $2.3 billion later that year — sums normally associated with building factories, not shipping a desktop application. When SpaceX pitched its IPO investors on a total addressable market of around $28 trillion, some $26 trillion of it sat in the company's AI efforts, and the larger share of that was enterprise applications rather than AI infrastructure.
The second is that the economics are still unsettled. Running these tools costs real money: the models behind them consume data-centre capacity continuously. One of the first links between the two companies came in April 2026, when Business Insider reported that xAI had decided to rent out some of its data-centre capacity to Cursor. TechCrunch cited a source who said the $2 billion round Cursor had planned would not have been enough to carry it to break-even. A company can grow quickly and still be expensive to operate, and both were true here.
The buyer's motive is worth stating plainly, because it is not about editors. SpaceX merged with Elon Musk's xAI earlier in 2026, and the deal is intended to help that AI division catch up with the major AI labs. The division spent the first half of the year being rebuilt: all eleven of Musk's xAI co-founders had left by March 2026, and Musk said publicly that the company was not built right the first time and that he was rebuilding it from the foundations up. Buying Cursor brought in an established product and the engineering team behind it, rather than another rebuild.
What this means if you are building something
Ownership changed; the tool did not, yet. Nothing about Cursor is different this week. But an editor that many development teams open every morning now belongs to a listed company with shareholders, a space business and an AI division to justify. Roadmaps, pricing and support commitments follow ownership eventually, not immediately.
Keep the work portable. The practical defence is simple and it costs nothing: your code, tests, deployment scripts and documentation should be ordinary files in an ordinary repository, readable without the tool that helped produce them. If a project can only be maintained by a team using one specific vendor's assistant, that is a dependency nobody wrote into the contract.
Be careful with multi-year commitments. Standardising a whole team on one vendor's agent is reasonable if it makes people faster. Signing a long contract for it in the same quarter that the vendor changes hands is less reasonable. Annual terms and an exit that does not require re-learning everything are the sensible shape right now.
Expect prices to move. A business at this scale that reportedly had not yet reached break-even is a business whose prices are not final. Budget for developer tooling as a subscription that can be repriced, not as a fixed cost.
If you are commissioning work rather than doing it, the question to ask your agency is short: what happens to our project if this tool disappears tomorrow? The acceptable answer is that the team works more slowly for a while. Any other answer is a risk you are carrying without knowing it.
Sources
