FTC Disclosure: TechSifted uses affiliate links. We may earn a commission if you click and buy — at no extra cost to you. Our editorial opinions are our own.
Thirteen days after we covered Cursor 3 and the xAI compute deal, things escalated significantly.
On April 21, SpaceX announced it has the right to acquire Cursor for $60 billion before the end of 2026 — or pay $10 billion to walk away and call it a "collaboration." That $10 billion isn't a venture investment. It's not equity. It's closer to a very expensive option premium. SpaceX paid $10B for the right to maybe pay $60B later.
That's a strange deal structure. But once you understand why it's built this way, it starts to make sense — and the implications for developers are worth thinking through.
Wait — Didn't We Already Cover a Cursor/xAI Deal?
Yes. And this is different.
That deal involved xAI renting GPU compute to Cursor to train Composer 2.5. A backend infrastructure arrangement. We covered it on April 16.
This is a SpaceX deal. Different company — or it was, until February, when SpaceX absorbed xAI outright. So technically these are the same Musk entity now, but the org chart matters here. SpaceX is the parent. It's the one planning the IPO. It's the one that signed this agreement. And it's the one that would eventually own Cursor if the acquisition option gets exercised.
The xAI deal was about training compute. This is about ownership.
The Deal Structure, Actually Explained
Two outcomes are possible by end of 2026.
Outcome one: SpaceX acquires Cursor for $60 billion. The $10B paid now likely counts toward the purchase price, with the remainder financed using SpaceX's publicly traded stock after the IPO.
Outcome two: SpaceX pays $10B and walks away. Cursor keeps the money. No acquisition. SpaceX frames the $10B as compensation for "our work together" — essentially a one-year exclusive compute and collaboration partnership.
SpaceX almost certainly wants to do the acquisition. But it's targeting an IPO this summer at a reported $1.75 trillion valuation, and completing a $60B acquisition before that listing would require updating its confidential financial filings. Messier. So the option structure lets them get working together now, lock in the deal in principle, and close after the listing when they can use public shares to finance it.
It also let Cursor stop raising money. And this is the part that's easy to miss.
Cursor was reportedly hours from closing a $2 billion funding round led by Andreessen Horowitz, Thrive Capital, and Nvidia — at a $50 billion valuation — when SpaceX came in with a $10B collaboration offer and a path to a $60B exit. Cursor halted the fundraise. That choice tells you something about how the Cursor team is reading the situation.
At 25, Cursor CEO Michael Truell just watched his company go from a $50B private round to a $60B acquisition target in the span of one call. MIT dropout, for what it's worth.
The Compute Angle
Here's the part that makes the SpaceX side of this make sense: xAI's Colossus supercomputer in Memphis, Tennessee.
SpaceX describes Colossus as equivalent to 1 million H100 GPUs. That's substantial compute sitting around that needs to generate returns. And Cursor's problem — a good problem to have — is that it identified continued gains from pretraining and reinforcement learning after Composer 2 shipped, but training large models requires infrastructure that no pure-software startup can spin up fast enough.
Colossus solves that immediately. In exchange, SpaceX gets access to Cursor's developer data, IDE distribution, and the most deeply embedded consumer AI product on developers' machines. That's an enormous amount of high-quality behavioral signal for training future models — not scraped from GitHub, but collected during the actual act of programming.
Composer 2.5 — the next model in Cursor's lineup — is being trained on this infrastructure. Cursor's own blog post about the deal led with the compute story rather than the acquisition angle. That sequencing wasn't accidental.
What This Means for the Product Roadmap
Honest answer: the deal probably accelerates the Composer line without immediately changing the IDE.
Cursor 3 just shipped with agent-first workflows, Design Mode, and best-of-N model comparison. You can read our full review. That product isn't getting ripped apart because SpaceX signed an option agreement. The near-term roadmap is already in motion.
What changes is trajectory. Three companies are now assembling vertically integrated AI developer stacks:
- Google: Gemini + TPU + Cloud
- Anthropic: Claude Code + MCP + partner cloud
- SpaceX/xAI: Colossus + Cursor + (potentially) Starlink
If SpaceX acquires Cursor outright, you'd expect deeper xAI model integration, preferential Colossus pricing for Cursor Pro and Business tiers, and eventually a future where Cursor's best models are exclusives you can't get anywhere else.
Pricing is the open question. Cursor Pro is $20/month today. Acquisition targets don't stay at startup pricing forever. Whether Musk pushes for enterprise bundles, Grok API tiers, or higher base pricing after the acquisition closes — nobody knows yet. But watch it closely.
What About Microsoft?
Before SpaceX signed, Microsoft was reportedly looking at Cursor too. That went nowhere — likely a combination of price and the obvious awkwardness of GitHub Copilot being in the same product category.
Worth noting: Microsoft passing on Cursor doesn't make Copilot's position weaker, but it does clarify things. GitHub Copilot adjusted its pricing structure this month as competition in the AI coding space has intensified. If Cursor ends up with Colossus behind it, compute advantage at scale is a real differentiator — not just a marketing claim.
Developer Reactions: The Musk Discount
This is where things get complicated.
A lot of Cursor's user base chose Cursor partly because it was independent. Not OpenAI's stack, not Microsoft's GitHub property, not Google. For a certain kind of developer, that independence mattered.
Now it's potentially owned by Elon Musk. Via SpaceX. Which absorbed xAI. Which is also Grok. Which is also X.
Musk's brand has gotten genuinely complicated among the developer community — to put it diplomatically. Developer forums have been vocal. There's a real concern that the product that earned trust as an independent tool could shift in ways that reflect the parent company's priorities rather than the engineering community's.
The counterargument: Cursor still works. Composer 2 is legitimately excellent. And if Colossus compute makes Composer 2.5 noticeably better, the pragmatic argument wins for most users. Developers, more than most user groups, tend to stay where the tools are good.
But if you're evaluating alternatives, the market has options. Windsurf has been developing quickly — though if you've run into setup or connectivity issues with Windsurf, you're not alone. Claude Code from Anthropic is the other competitor with real momentum right now, particularly for teams already deep in the Anthropic ecosystem.
The Bigger Picture
The $60B Cursor option is partly a SpaceX story and partly a story about where AI coding is heading.
A year ago, this product category barely existed as a serious market. Now there's a quiet bidding war — SpaceX, Microsoft, and implicitly every major tech platform jockeying for control of the software development interface. Because whoever owns the IDE is closest to where code gets written. And whoever's closest to where code gets written has the best dataset for training the next generation of models.
Cursor's value isn't just $60B worth of users. It's $60B worth of proprietary developer context — the queries, the codebases, the corrections made during active programming sessions. That's not something you can reconstruct from public repositories. It's behavioral data collected in the loop of real work.
SpaceX understands this. The option may not get exercised. The IPO timeline could slip. But either way, the collaboration has started and Colossus compute is already flowing to Cursor's training runs. Something has changed, regardless of what happens in December.
For developers using Cursor today: nothing changes immediately. Watch Composer 2.5 — if it's meaningfully better than Composer 2, that's the compute deal delivering. Watch whether Cursor's editorial and product independence holds through the IPO window. And watch pricing. That's usually the first signal that an acquisition is becoming a product reality rather than just a press release.
Nate Calloway covers developer tools and AI infrastructure for TechSifted.
Top comments (0)