SpaceX bought Cursor. The coding stack has a new owner.

SpaceX announced on June 16 that it would acquire Anysphere, the company behind the Cursor AI coding tool, for $60 billion in an all-stock transaction. The deal closes in the third quarter pending regulatory approval. Cursor goes into SpaceX, which already owns xAI, which is now competing directly with Anthropic and OpenAI in the AI coding market.

A meaningful fraction of the professional software development community uses Cursor every day. The product reached $2 billion in annual recurring revenue earlier this year, which puts Anysphere among the fastest-growing business software companies on record. Most teams that adopted Cursor adopted it for technical reasons. As of this month, the technical and the political are no longer separable.

The shape of the deal

The acquisition is all stock. SpaceX is the acquirer, and SpaceX completed its own IPO on June 13. SpaceX bought xAI in February. xAI now absorbs Cursor. The combined entity has a foundation model (Grok), a coding product (Cursor), a chip strategy through Nvidia partnerships, and a launch vehicle attached to the same shareholder.

There is no recent corporate analogue. The closest comparison is Microsoft in the late 1990s, which integrated Windows, Office, and Internet Explorer into one stack. The structural difference is that Microsoft was a public company with a board, a CEO who was not a celebrity, and a long history of operating in regulated markets. SpaceX has none of those features at the same scale. The decision-making is concentrated in a way Microsoft's never was, even at its most monopolistic.

Risks under the new owner

Cursor itself has not changed since Monday. The product still works. The keyboard shortcuts are the same. The autocomplete still calls the same model providers behind the scenes. Teams that adopted Cursor for technical reasons can keep using it for those reasons.

What has changed is the long-term risk profile, and that risk has three real components that working teams should think about separately.

Product direction is one piece of it. A new owner can change what the tool prioritises. Acquisition promises of no change tend not to survive the first product cycle. Cursor's competitor list (Windsurf, Zed, the built-in VS Code chat with Anthropic, GitHub Copilot for teams that left and might return) is short enough that switching costs are real but not prohibitive.

Dependency on the owner is the second piece. Some engineers will object to xAI's parent company on grounds that are not technical. That is a legitimate reason to leave a tool. Other engineers will object to nothing. That is also legitimate. Both positions need to be respected inside the same organisation, and the second-order question is whether your tool choice becomes a recurring internal debate. For many teams it now will, and the cost of that recurring debate is real even when no one ends up switching.

Data is the third piece. Cursor processes source code from the companies that use it. The terms of use as of today say what they say. The terms of use as of next year may say something different. A new owner has the legal latitude to change terms with notice. The parent company in this case has demonstrated, across other portfolio companies, a willingness to change terms quickly when the strategy demands it. Legal review of an existing Cursor contract is a reasonable thing to do this quarter, not next year.

Precedent and limits

This is a real consolidation in developer tooling. It joins a short list of recent acquisitions where a tool that mattered to working developers became part of a larger empire: GitHub into Microsoft, npm into GitHub into Microsoft, Atlassian's acquisitions across the SaaS layer. Each of those changed the surrounding politics slightly. This one changes it more, because the surrounding company is among the most polarising in technology and because the founder is a public political figure in his own right.

It is not, on its own, the end of Cursor as a product. Post-acquisition teams typically retain focus for a while after close. The tool will probably stay competitive in feature parity with its peers. Every team using Cursor should be able to name a replacement and have it written down.

The shorter version

If you use Cursor, keep using it. Also evaluate Windsurf, the built-in VS Code chat, and the current version of Copilot, and form a real opinion on each, so that the question of what you would switch to is not theoretical.

For everyone else, the developer tooling layer is becoming as interesting as the foundation model layer became in 2024. The cast of characters now includes one founder whose other companies cover launch vehicles, satellites, neural interfaces, and social media.