On June 16, SpaceX announced it was buying Cursor, the AI code editor that millions of developers use every day, for $60 billion, all in stock. It is the largest startup acquisition on record, and it lands just days after SpaceX itself went public in the biggest IPO ever, raising $75 billion.
The number is staggering, but it isn't the story. The story is custody. The place where you write your code no longer belongs to a San Francisco startup; it is now part of one of the most powerful, most vertically integrated empires on the planet. And a code editor is not just any tool: it is the most intimate instrument a developer has.
What was actually bought
Cursor, founded in 2022 by Michael Truell and three MIT classmates, is not a text editor. It is an environment where the AI reads your entire project, suggests code, and, through what are called agents (assistants that carry out coding tasks on their own), handles work from start to finish. It is worth roughly $2.6 billion in annualized revenue. In April it was raising money at around a $50 billion valuation, and that same month it gave SpaceX an option: either buy it for $60 billion, or pay $10 billion to "work together." SpaceX took the first.
"We are excited to share that SpaceX has exercised their option to acquire Cursor in an all-stock transaction with the goal of building the world's most useful AI models," wrote Michael Truell. The deal closes in the third quarter, pending regulatory approval, and Cursor becomes a wholly owned subsidiary.
Why would a rocket company want a code editor? The logic is vertical integration, the same playbook we already saw at Tesla: own the whole chain. SpaceX wants to strengthen its AI business, Grok, on the coding front, where OpenAI and Anthropic already have their own tools. With this purchase, the editor, the model, the data centers, and the connectivity all sit under one owner.
Why this concerns you
An AI editor sees everything. It sees your proprietary code, the keys and passwords left behind in a file, the architecture of your systems, and the way you think through a problem, keystroke by keystroke. To give you suggestions, it sends parts of that context to servers in the cloud. Whoever owns the editor owns that window.
In forensic investigation there is a simple rule: evidence only holds up if you know every hand it passed through, the chain of custody. Your code is your evidence, and that chain just changed hands, to an owner who also owns the model, the stock, a satellite network, and a social network. I have written here before about the risk of depending on a single AI provider. This is the same lesson, one floor up: the dependency now runs through the tool, the model, and the infrastructure at once.
None of this is illegal, or even surprising. This is how power concentrates: not with an attack, with a contract. What changes for those who write code is not the quality of the autocomplete. It is who you trust with your work, and how easily you can leave, if one day you want to.
What you can do now
You don't have to abandon the tool. You have to use it with your eyes open:
- Review the privacy settings: see what gets sent to the cloud, what stays on your machine, and turn off what you don't need.
- Never leave secrets in the code the AI reads: keys, tokens, and passwords belong in a vault, not in a project file.
- Have an exit plan: work with portable formats and configurations so you can switch editors without rewriting everything.
- For anything sensitive, consider models that run locally, without sending your code outside.
- Read the terms when the deal closes: the owner has changed, and the rules for using your data may change with it.
The largest startup acquisition in history did not buy lines of code. It bought the place where they are born. And when someone becomes the owner of the place where you work, it is worth asking who is on the other side of the screen.
Sources: CBS News, Yahoo Finance.
#StaySafe
🙏🖖
