DEAL FACTS — JUNE 16, 2026
● Acquirer: SpaceX (SPCX) via wholly-owned subsidiary X67 Inc.
● Target: Anysphere Inc. (maker of Cursor AI coding agent)
● Deal value: $60 billion implied equity value — all-stock, no cash
● Structure: Cursor merges into X67 Inc., survives as SpaceX subsidiary
● Close expected: Q3 2026 pending regulatory approvals
● Termination fee: $10 billion general / $4 billion antitrust-specific
● Cursor ARR: ~$4 billion total; ~$2.6 billion annualised B2B
● SEC filing: 8-K filed June 16, 2026
What Just Happened
Four days after SpaceX's record $75 billion Nasdaq IPO closed at $160.95 — and on the same day all three major AI lab CEOs were at the G7 summit in France — SpaceX signed a definitive merger agreement to acquire Anysphere, the San Francisco company behind Cursor. The deal is the largest acquisition of an AI developer-tools company ever recorded.
This was not a surprise move internally. SpaceX disclosed in April 2026 that it had secured a formal option to either acquire Anysphere for $60 billion or pay $10 billion to continue a collaborative training arrangement. Two senior Cursor engineers — Andrew Milich and Jason Ginsberg — departed to join xAI in March 2026, weeks before that option was disclosed. The jointly trained model has been in development for several months using xAI's Colossus supercomputing infrastructure and is expected to ship inside both Cursor and Grok Build in the near term.
The acquisition preempted what would have been Cursor's next funding round. TechCrunch reported Cursor was on track to close a $2 billion raise led by Andreessen Horowitz, Thrive, and Nvidia at a $50 billion valuation. SpaceX's $60 billion all-stock offer made that round irrelevant — Cursor shareholders get SPCX shares at a $10 billion premium to the pending VC round instead.
Cursor's Revenue Story — Why SpaceX Paid $60 Billion
| Milestone |
Date |
ARR |
| Early traction |
January 2025 |
$100 million |
| Crossed $1 billion ARR |
November 2025 |
$1 billion |
| Doubled again |
February 2026 |
$2 billion |
| Acquisition date |
June 2026 |
~$4 billion total ARR (~$2.6B B2B) |
| Year-end projection |
December 2026 |
$6 billion (Grok summary projection) |
At $4 billion ARR and the growth trajectory above, $60 billion represents a 15x ARR multiple — aggressive but defensible for a category-leading developer tool growing 100%+ annually. The caveat: Cursor's market share has declined from 41% in June 2025 to approximately 26% in May 2026 — suggesting the growth rate is slowing as Windsurf (ranked #1 LogRocket 2026), GitHub Copilot (83% enterprise adoption), and Claude Code have gained ground. SpaceX is buying a still-fast-growing but decelerating market leader at a premium multiple.
What This Means for Cursor's Model-Agnostic Positioning
Cursor's primary competitive advantage — the reason most developers chose it over GitHub Copilot — is model-agnostic routing. Cursor lets you use Claude Opus 4.8, GPT-5.5, Gemini 2.5 Pro, or its own models depending on task type and cost. You are not locked into one AI provider. This is structurally incompatible with SpaceX's incentives as the owner of xAI (formerly a separate company, merged with SpaceX in February 2026) and Grok Build.
The question every Cursor user now needs to answer: will SpaceX maintain Cursor's Claude and GPT routing, or will it gradually shift Cursor toward xAI's Grok models and the jointly trained Cursor/Colossus model that is already in development? SpaceX has not answered this explicitly. Cursor co-founder Michael Truell said he is "excited to build the most useful AI alongside SpaceX" — which tells you the direction of travel without stating it plainly.
Risk for current Cursor users: Claude routing
Cursor currently routes to Claude Opus 4.8 and Sonnet 4.6. Anthropic and SpaceX are direct competitors — Anthropic has Colossus compute lease agreements with SpaceX that include 90-day termination clauses. Post-acquisition, maintaining Anthropic model routing in Cursor requires SpaceX to keep paying Anthropic API fees to serve Cursor's users. That is a structural incentive conflict. Claude routing may continue in the short term, but expect gradual de-emphasis as the jointly trained Cursor/Grok model ships.
Uncertainty: GPT-5.5 routing
OpenAI and xAI/SpaceX are direct competitors. OpenAI routing in Cursor post-acquisition is even less strategically aligned than Claude routing. However, removing GPT-5.5 from Cursor would be the most disruptive developer experience change possible — many users specifically rely on GPT-5.5 for certain task types. Expect this to be maintained longer than Claude routing, but under review.
What definitely improves: Compute and the jointly trained model
Cursor now has access to Colossus — SpaceX's 100,000+ GPU supercomputing cluster. The jointly trained Cursor/Colossus model that ships soon could meaningfully outperform any model currently available in Cursor on code-specific benchmarks, since it is trained specifically on coding tasks using infrastructure no independent AI startup can match. If SpaceX executes well on this, the new model could compensate for losing Claude routing quality with a better native model.
What This Means for Anthropic and Claude Code
Cursor routed significant token volume to Anthropic's API — every Cursor user selecting Claude Opus or Sonnet was an Anthropic API revenue event. Post-acquisition, that revenue stream is now directly funding a SpaceX subsidiary rather than an independent company aligned with diverse model routing. Anthropic's Colossus compute agreements include 90-day termination clauses — SpaceX now has both a financial incentive (eliminate API payments to a competitor) and a structural mechanism (terminate the compute agreement) to de-prioritise Anthropic's models in Cursor.
For Claude Code specifically, this is a meaningful competitive shift. Cursor was the other major IDE-based coding tool that routed to Claude. Post-acquisition, Cursor's Anthropic routing is in question. That means the primary IDE tool sending users to Claude models is now under the control of a direct competitor — and the alternative for Claude-committed developers is Claude Code itself (terminal), Windsurf (which does not face this conflict), or GitHub Copilot (Microsoft, neutral on Claude routing). See our four-way coding tool comparison for where each stands after this deal.
What This Means for Windsurf
Windsurf — ranked #1 in LogRocket's 2026 AI Dev Tool Power Rankings — is now the only major independent IDE-based AI coding tool. Cursor is about to become a SpaceX product. GitHub Copilot is a Microsoft product. Claude Code is an Anthropic product. Codex is an OpenAI product. Windsurf, owned by Cognition (backed at $25 billion), is the only tool in the top tier not controlled by one of the major AI or tech conglomerates. For developers who chose Cursor specifically for its independence and model-agnostic routing, Windsurf is now the most logical migration target.
The Strategic Picture — Why SpaceX Bought Cursor Now
Three strategic reasons this deal happens four days after the SPCX IPO:
1. Enterprise AI revenue — immediately
Cursor's $4 billion ARR and 4+ million developer users give SpaceX immediate enterprise AI revenue that xAI's Grok never achieved at scale. The deal answers the question SpaceX investors were asking on IPO day: what is the AI monetisation story beyond Starlink compute leases? Answer: developer tools used daily by millions of engineers at enterprise companies.
2. The Colossus compute flywheel
Cursor's massive user base generates coding-specific training data at extraordinary scale. Every edit, refactor, and debugging session in Cursor is a training signal. Running this through Colossus creates a compute-data-model flywheel that could produce a coding model meaningfully better than anything currently available — trained on real developer workflows, not curated benchmark datasets. This is the same flywheel GitHub Copilot built using Microsoft Azure and GitHub's code data.
3. Compete with Anthropic and OpenAI on their strongest ground
Claude Code (Anthropic) and Codex (OpenAI) are the two most used autonomous coding agents in enterprise. SpaceX now has Cursor (IDE), Grok Build (API/CLI), and a jointly trained model on Colossus. That is a complete developer tools stack: IDE for daily coding, API for autonomous agents, proprietary model trained on developer data. SpaceX enters the coding tools market not as a startup but as the owner of the world's third-largest AI compute cluster.
What Cursor Users Should Do Right Now
Nothing immediately. The deal closes Q3 2026 and product changes will not happen overnight. Cursor will continue operating as-is through the integration period. Do not migrate tools based on uncertainty — wait for SpaceX to announce actual product changes.
Watch for model routing announcements. The first concrete signal of where Cursor is headed post-acquisition will be whether SpaceX modifies or removes Claude and GPT-5.5 routing options. That announcement — whenever it comes — is the moment to evaluate alternatives.
Evaluate Windsurf as a hedge. Windsurf is now the only top-tier independent AI coding IDE. Testing it now — before any Cursor changes — means you have a working migration path if you need one. Windsurf Pro is $15/month and its free tier gives 5 Cascade sessions/day. See our full comparison.
If you rely on Claude routing specifically, move to Claude Code or Windsurf for Claude-dependent workflows now rather than later. Claude Code gives you direct Anthropic model access without the SpaceX intermediary layer.
Frequently Asked Questions
Will Cursor still support Claude and GPT-5.5 after the SpaceX acquisition?
SpaceX has not confirmed or denied this. In the short term (through close of the deal in Q3 2026), model routing will almost certainly remain unchanged. Post-close, the strategic incentive is clear: SpaceX owns xAI and Colossus, and has a jointly trained Cursor/xAI model in development. Expect Claude and GPT-5.5 routing to remain available initially but de-emphasised over 12-24 months as the native SpaceX/xAI model becomes the default. Monitor SpaceX's product announcements post-close for official policy.
Is $60 billion too much to pay for Cursor?
At $4 billion ARR and a 15x multiple it is aggressive but not irrational for a category leader in one of the fastest-growing enterprise software categories. The complication: Cursor's market share has declined from 41% to 26% over the past year as Windsurf and GitHub Copilot gained ground. SpaceX is paying a premium for a decelerating market position. The bet is that Colossus + the jointly trained model reverses that deceleration. If the model ships and is meaningfully better than alternatives, the acquisition could look cheap at $60 billion. If the model disappoints and Claude/GPT routing changes drive user exodus to Windsurf, the premium will look expensive.
Why is the deal all-stock and not cash?
SpaceX's 8-K explicitly states the IPO proceeds are not being used to fund the deal. Anysphere shareholders receive SPCX Class A shares — they become SpaceX shareholders, not cash recipients. The all-stock structure aligns Cursor's founders and investors with SpaceX's long-term success and avoids using the $75 billion raised in the IPO for an acquisition that might face antitrust scrutiny. The $60 billion represents 3.4% dilution of SpaceX at IPO valuation.
What happens to Cursor's Windsurf competitor relationship?
Windsurf (owned by Cognition) and Cursor were direct competitors before this deal. Now Cursor is a SpaceX subsidiary and Windsurf remains independent. Windsurf's Devin integration (Cognition's AI software engineer) now competes directly with SpaceX's Grok Build and the new jointly trained Cursor/xAI model. The competitive dynamics of AI coding tools just became significantly more concentrated: two major platform companies (SpaceX and Microsoft/GitHub) now own the two most-used IDE-based tools, while Windsurf is the primary independent alternative.
Sources: CNBC, Reuters via Yahoo Finance, SpaceX 8-K SEC filing June 16, 2026. Tool comparisons: Cursor vs Claude Code vs Codex vs Windsurf · Best AI Coding Agents 2026. All AI news: June 2026 calendar.