Industry, Business & Companies · 2026-04-22

Mask took Cursor.

English translation of the Chinese original. This version is generated for international readers and may be refined over time.

Interactive reading view ↗

English | Chinese Original

English translation of the Chinese original. This version is generated for international readers and may be refined over time.

Date: 2026-04-22

On March 12th, two core engineering directors, Andrew Mirich and Jason Ginsberg, announced their departure from Cursor to join XAI and SpaceX. Both served as Head of Engineering Project in Cursor and were jointly responsible for the complete operation of the company ' s product works.

Musk relayed a two-person notice of departure, saying, "xAI was not built for the first time and is now rebuilding from scratch.

Then the xAI began to provide Cursor with the algorithm. Cursor plans to train its latest AI programming model Computer 2. 5 on xAI infrastructure, using tens of thousands of GPUs.

Meanwhile, Cursor is seeking a round of $2 billion in new financing, valued at over $50 billion. Less than a year before the previous round of $900 million.

The signal of the deal is clear enough. Now, SpaceX push confirmed.

xAI on the next game of what?

xAI is slowly becoming an AI infrastructure company from an AI model company.

Renting power to Cursor is just the beginning. Amazon, Microsoft and Google are all renting to other companies and making huge profits. XAI has a lot of GPU, and it doesn't have to be for its own use, it can be rented.

This explains why Musk is digging through Cursor's engineers: building the foundations for Grok's programming capacity. Cursor's team is the best IA programmers, dig 'em up, run 'em up with xAI -- it's a closed loop.

If this closed loop runs through, XAI is not just AI Model, it's an integrated platform for AI + AI programming capabilities.

And Cursor? Cursor has a $50 billion valuation, an annualized income of over $6 billion, and a large number of corporate clients. But it has a fatal question: it doesn't have its own calculus.

All its model training depends on someone else - OpenAI, Anthropic, Google, now XAI. That means it's in the hands of others.

The logic of SpaceX/XAI acquisition Cursor is quite straightforward.

xAI is smart and has no programming data. Cursor has programming data, no stable algorithm.

The two merged, xAI became the sub-infrastructure of Cursor, and Cursor became the export of programming capacity for xAI. Musk has X, Grok, SpaceX, XAI, Neuralink-Cursor to complement the AI programming, and the entire ecology is tied together.

And don't forget, SpaceX is making space data centres. If this works, Cursor's AI programming model can run directly into space algorithms. It sounds science fiction, but it's SpaceX moving in the direction.

Cursor is now in the position that it is the strongest company in the AI programming field, but it is subject to human control.

Because Cursor knew that he couldn't do it, and now Exit is fine.

Anthropic has Claude Code, OpenAI has Codex, Google has Gemini Code. Everybody's trying to figure it out. Cursor, to remain competitive, has to have a big back.

XAI now gives it the ability to calculate, which is the first step. What's the second step? It's to turn "competence" into "capitalism."

For Cursor investors, the acquisition by xAI means exit. For Cursor's team, being acquired by xAI means that their product can run directly across Musk's entire ecology.

It's more sexy than being listed independently.

The greatest risk to this transaction is time.

If Cursor ' s income continues to grow at a high rate, the path to independent listing becomes clearer and it does not have to be sold. If XAI ' s space data centre is built more slowly than expected, the synergy of the merger is less obvious.

There is another variable: regulation. FTC is now becoming more and more rigorous in its acquisition of large technology companies. SpaceX acquisition xAI, it's good to say, will SpaceX/XAI acquisition Cursor trigger an antimonopoly review? It's hard to say.

The issue of SpaceX/XAI acquisition of Cursor is not a question of whether or not, but of what terms and when.

Cursor needs power, xAI needs programming data. The interests of the two families are naturally tied up.

The question now is how fast Musk can move.

Sources: