SpaceX buys Cursor for $60 billion: when your code editor joins an empire

On June 16, 2026, SpaceX signed a merger agreement to acquire Anysphere, the company behind Cursor, the AI-powered code editor most used by professional developers, at an implied equity value of $60 billion, entirely in stock. It was exactly four days after SpaceX's record Nasdaq IPO. A first of its kind: the largest acquisition ever in developer tooling.

For regular readers, this is act three of one story. After Kickbacks, which built a business on Anthropic's land with no lease, after the US government forcing Fable 5 offline, here is the next floor up: the tool itself, your editor, the software you write every line in, absorbed into an industrialist's mega-company. The question is no longer "what am I building on?" but "who owns what I build on?"

We dug into this from the regulatory filing (SpaceX's SEC Form 8-K), CNBC, Bloomberg and TechCrunch coverage, and SemiAnalysis's infrastructure analysis. Here is what was signed, how we got here, and, this is the heart of it, why the deal exposes a dependency most teams never saw coming.

What was signed on June 16

The deal is an all-stock merger, structured as a reverse triangular merger: a SpaceX subsidiary called X67 Inc. merges into Cursor, which survives as a wholly owned SpaceX subsidiary. Each Cursor share (common and preferred) converts into SpaceX Class A stock. Here are the exact terms, as they appear in the 8-K.

Two nuances most headlines flatten, but that matter. First, the deal is signed, not closed: it remains pending regulatory approvals, targeted for the third quarter of 2026. Second, because the ratio depends on the average share price over the seven trading days before a still-future closing, the exact number of shares Cursor's founders will receive is not yet locked. And since it is a stock swap, the capital raised at the IPO does not fund the deal (per CNBC's analysis): SpaceX pays in paper, not cash.

This signing is not a total surprise: it exercises an option secured on April 21, 2026. On that date, SpaceX (via xAI) gave itself the right to either buy Cursor for $60B later in the year, or pay $10B for their "work together." The exact nature of that $10 billion is reported inconsistently across outlets, a payment for the training partnership per CNBC, Bloomberg and Wikipedia; a "break-up fee" per TechCrunch. We flag it as contested rather than pick a side.

How we got here: the consolidation of Musk's empire

To understand how a space company buys a code editor, you have to trace two mergers, which must not be confused.

In the February 2026 merger, xAI was dissolved as a separate entity. Grok, the X platform and the Colossus supercomputer in Memphis were folded into SpaceX, and the AI division renamed "SpaceXAI." Elon Musk's own phrasing sums it up: "xAI will be dissolved as a separate company, so it will just be SpaceXAI, the AI products from SpaceX." The name "SpaceXAI" you see in the official tweets is therefore not a company in the legal sense: it is the brand of the AI division of a single entity, SpaceX.

The financial context is worth recalling, because it explains the logic: per the IPO filing, xAI burned cash heavily, an operating loss on the order of $6.35 billion in 2025 against roughly $3.2 billion in revenue. Absorbing Cursor buys, in one move, more than $4 billion in annualized revenue, massive distribution among expert developers, and an ideal training ground for its code models.

Cursor: from $100M to $4B in revenue in eighteen months

Cursor was founded in 2022 by four MIT alumni, Michael Truell (CEO), Sualeh Asif, Arvid Lunnemark and Aman Sanger. The product is a code editor (a VS Code fork) where AI is the center, not the margin. Its annualized recurring revenue (ARR) trajectory is one of the fastest in software history.

At the time of the deal, Cursor claimed more than one million paying users (over 2 million total), roughly 50,000 enterprise teams, and deployment across 64% of the Fortune 500, the latter two being company-stated figures, not independently audited. In April 2026, Cursor was even in talks to raise at around $50B; the $60B SpaceX option short-circuited that round.

The technical detail that suddenly turns strategic: Cursor has always been multi-model. The editor lets the user pick between Claude (Anthropic), GPT (OpenAI), and its in-house "Composer" models (including Composer 2.5, launched May 18, 2026, built on the Kimi K2.5 base). In other words, the most popular product for AI coding was, until now, one of the largest customers of its future direct competitors' APIs. Hold onto that point: it is central to everything that follows.

Colossus: the firepower behind the deal

If SpaceX can train a frontier code model, it is thanks to Colossus, the supercomputer in Memphis, Tennessee. The official April 21 tweet referred to a "million H100 equivalent Colossus training supercomputer", a marketing phrase describing compute capacity in equivalent terms, to be taken as such. The infrastructure numbers documented by SemiAnalysis give the real scale:

At that scale, Colossus 2 would be the first gigawatt-scale AI data center. This is the infrastructure Cursor was already using jointly, since the April partnership, to co-train a model meant to ship "soon" in Cursor and in Grok Build, SpaceXAI's coding agent. The name, performance and exact release date of that co-trained model have not, at this stage, been confirmed by a reliable source: we stay cautious on that point.

The real stake: Cursor depended on its future competitors

Here is why this acquisition is far more than another funding-news line. Cursor, now owned by SpaceXAI (which makes Grok), built its success on access to Anthropic's (Claude) and OpenAI's (GPT) models, that is, Grok's two most direct competitors in coding. An editor owned by one AI lab relies, for its flagship product, on rival labs' models.

This is not a textbook hypothetical. The precedent exists, dated and documented. In 2025, when rival Windsurf was about to be acquired by OpenAI, Anthropic cut off most of Windsurf's access to Claude. An Anthropic co-founder put it bluntly: "it would be odd for us to sell Claude to OpenAI." Transpose it: it would be at least as odd for Anthropic and OpenAI to keep feeding, at full tilt and wholesale prices, an editor now controlled by xAI/Grok. The risk that third-party models get progressively restricted inside Cursor, by contract, by price, or by rate limit, is real, and it is exactly what would push users toward the co-trained in-house model. The acquisition isn't just a stake: it's a captured distribution channel.

Reactions, antitrust and concentration

The deal immediately fueled debate. On Hacker News and in developer communities, two concerns dominate: concentration (one group now owning the editor, the model, the training cloud, and a social network for distribution) and conflict of interest, with Elon Musk controlling SpaceXAI while holding outspoken public positions elsewhere. On the competition-law side, specialist analyses (notably at IPWatchdog as early as May 2026, on the partnership) already flagged the antitrust implications of a SpaceX–Cursor tie-up: vertical integration from compute down to the end tool, and privileged access to the coding data of millions of developers.

On that last point, data, vigilance is warranted for any company whose teams code in Cursor: your repo context, your prompts, your code snippets pass through a tool whose parent also has every interest in training models. Nothing indicates misuse to date; but data governance is changing hands, and that is a parameter to reassess.

Our take: the dependency trilogy

What follows is our analysis.

Three stories in two weeks, one lesson. Kickbacks showed you can build a business on someone else's surface, with no lease. Fable 5's suspension showed the model itself can be cut off by a decision beyond you. The Cursor deal closes the loop: the tool you work in can swing, overnight, into the empire of a player who also owns the model and the cloud. When the IDE, the model and the infrastructure belong to the same mega-company, the question is no longer the tool's quality, Cursor remains excellent, but what stays under your control.

The answer, for a business, isn't to flee good tools. It's to never mistake a tool for a foundation:

• Abstract the model and the tool. Your critical workflows shouldn't depend on one specific editor or one hard-coded model name. Switching IDE or model provider should stay a choice, not a rebuild.
• Keep your assets portable and yours. Code, repositories, reusable prompts, documentation, tests: these are your assets. As long as they're standard and exportable, the tool on top stays a replaceable component.
• Assess concentration risk, not just price. Before standardizing a whole team on a tool, ask: what happens if it's acquired, restricted, or has its access to third-party models cut? A slightly less flashy but neutral and open tool is often the more resilient bet in production.
• Watch data governance. Knowing where your code context and prompts go, and under what terms, is now part of a dev tool's due diligence.

This is exactly how we build custom software for our clients (see our work): AI and tools as interchangeable components behind your own abstractions, on assets you own, so that a headline like today's stays a piece of industry news, not a risk to your business. Want to know where your stack would break if a key tool changed hands or policy tomorrow? Tell us about your situation (or contact us) and we'll come back within 48 hours with a concrete read on your dependency.

Timeline (as of June 16, 2026)

This is a developing story; the deal is signed but not yet closed.

• 2022, Anysphere/Cursor founded by four MIT alumni.
• March 2025, xAI absorbs X (Twitter): ~$113B combined.
• Early 2025 → June 2026, Cursor's ARR goes from $100M to over $4B.
• February 2, 2026, SpaceX absorbs xAI (~$1,250B combined); the "SpaceXAI" division is born.
• April 21, 2026, SpaceXAI and Cursor announce their partnership and the acquisition option ($60B, or $10B for the joint work).
• May 18, 2026, Cursor launches its in-house model Composer 2.5.
• June 12, 2026, SpaceX's record Nasdaq IPO (SPCX), ~$1,750B valuation.
• June 16, 2026, SpaceX signs the Cursor acquisition for $60B in stock; closing targeted for Q3 2026.