Last Monday, SpaceX announced it's acquiring Cursor — the AI coding tool used by over a million developers — for $60 billion in stock. The deal is expected to close in mid-July.
If your team uses Cursor, you should read this before that happens.
Why People Chose Cursor in the First Place
Cursor wasn't the first AI coding tool. But it became the dominant one because of one thing most of its competitors didn't offer: it didn't care which AI model you used.
You could run Claude for nuanced refactoring, switch to GPT for quick completions, use Gemini when you had a big context window problem, and fall back to Cursor's own Composer model for everyday edits. All of this inside one IDE. No re-learning, no switching tools, no negotiating with procurement about which vendor relationship the company approved.
For a 6-person engineering team at a nonprofit, or a 12-person dev shop building government software, that flexibility mattered. You didn't have to bet your entire dev workflow on a single AI vendor. You could use what worked for the task.
That era is now ending.
What SpaceX Buying Cursor Actually Means
SpaceX merged with xAI — Elon Musk's AI company — earlier in 2026. xAI makes Grok. Grok just released its V9-Medium model in the same week the Cursor acquisition was announced. That timing is not a coincidence.
The strategic logic is obvious: SpaceX wants Cursor's 1M+ developer user base running on Grok. The code that Cursor users generate every day is extraordinary training data. And right now, xAI's Grok is behind OpenAI and Google in the agentic coding tools race. Cursor fixes that — fast.
Here's what the developer community is asking, loudly, on Hacker News (207 points, 147 comments at time of writing):
Will Cursor stay model-neutral after the acquisition closes?
Nobody at SpaceX or xAI has answered that question with a public commitment. The official line so far is that Cursor will continue to operate independently. Acquisitions always say that. They almost never mean it for long.
The uncomfortable math: SpaceX paid $60 billion for a company doing $4 billion in ARR. That's a 15x revenue multiple. At those economics, you need to extract substantial additional value. Routing users toward Grok, reducing the margin paid to Anthropic and OpenAI for model API access, and folding Cursor's developer data into xAI's training pipeline are all obvious ways to do that. None of them are good for people who chose Cursor specifically for its neutrality.
There's also the data question. Cursor's Composer model trains on user code sessions. Once that training pipeline is inside xAI's Colossus infrastructure, your code — if your contract allows training data use — may be feeding Grok. If you're building anything sensitive (government systems, healthcare apps, client-confidential code), that's a conversation you need to have with your legal team before July.
The Bigger Pattern Nobody's Talking About Enough
The Cursor acquisition completes a consolidation that's been happening all year. Look at where the major AI coding tools sit now:
- Cursor → owned by SpaceX/xAI (Grok)
- Windsurf → backed by OpenAI (GPT) after 2025 acquisition
- GitHub Copilot → Microsoft/OpenAI (GPT)
- Claude Code → Anthropic (Claude)
Every major AI coding tool is now either owned by or deeply dependent on a specific AI lab. The era of model-neutral coding assistants is effectively over.
What this means for small orgs: the tool choice is now also a vendor bet. When you adopt a dev tool, you're implicitly choosing an AI lab. If that lab raises prices, restricts access, or gets blocked by export controls (see: Fable 5 earlier this month), your whole dev workflow takes the hit.
What to Do Before the Deal Closes
You have roughly three to four weeks before the Cursor acquisition closes. That's not a lot of time, but it's enough to not be caught flat-footed.
1. Audit your Cursor usage and contract.
Figure out what version of Cursor your team is on, whether you have an enterprise contract, and what the data training terms are. If you're on the default individual or team plan, your code sessions may be used for training. If you've never checked, now is the time.
2. Test an alternative, seriously, before you need one.
Don't wait until xAI announces changes to start evaluating options. Pick one alternative and run a real project through it over the next two weeks. The alternatives that don't lock you into a specific lab:
- Windsurf is OpenAI-backed now, but still has broad model support and a free tier that covers most daily workflows. It's been a genuine competitor to Cursor on complex multi-file edits.
- VS Code with model-agnostic plugins (Continue.dev, Cline, Aider) gives you full control over which model you send prompts to, at the cost of more setup. For teams running local models via Ollama, this is the most sovereign option.
- Self-hosted coding assistants via open models (Qwen3, GLM, DeepSeek) are now legitimately capable for most day-to-day coding tasks and put you completely outside the Big Tech lab ecosystem. Setup is real work but the operational independence is real too.
3. Separate your sensitive code work from cloud-model tools now.
If your team works on anything regulated or client-confidential, establish a clear policy before July about what goes into a cloud-backed coding tool and what doesn't. This should have been your policy already, but the Cursor acquisition is a good forcing function to actually write it down and enforce it.
4. Watch what SpaceX says — and what they don't say.
If SpaceX wants to keep Cursor's developer base, they need to make a clear, public commitment to model neutrality soon. If that commitment doesn't come in the next two weeks, treat it as a signal that the commitment won't come at all.
What This Looks Like for a Real Team
We've been helping a 9-person public-sector contractor work through exactly this transition over the last week. Their setup: three devs heavy on Cursor, all of them using Claude via Cursor's model picker for most tasks, working on government contract code that they don't want in anyone's training data.
We helped them stand up Continue.dev against a local Qwen3 instance for the day-to-day completions work, with a clear escalation path to call out to a commercial API for complex problems. Total setup time: one sprint. Total additional monthly cost: the electricity. Their sensitive work stays on-premise; their commercial API calls go through a contract that explicitly opts out of training.
That's not the right setup for every team. But the decision-making process is: understand what you're relying on, understand what the new owner's incentives are, and make a conscious choice before the choice gets made for you.
The Bottom Line
Cursor built a great product by staying neutral. That neutrality is what made a million developers trust it with their daily work. SpaceX paid $60 billion for that trust — and the users, data, and market share that came with it.
Whether that trust survives the acquisition is SpaceX's choice to make. But whether your team is prepared for the case where it doesn't — that's yours.
If your team is working through an AI tool transition and you want help figuring out the right setup for your context, we're pretty good at this. We set up dev workflows for real teams, not slide decks.