On April 21, 2026, SpaceX disclosed it had secured an option to acquire Cursor — the AI-native code editor used by more than half the Fortune 500 — for $60 billion. The alternative: pay $10 billion for a continuing joint development partnership using xAI’s Colossus 2 supercomputer.
The expected timeline: SpaceX lists on Nasdaq under SPCX on June 12. The Cursor acquisition closes roughly 30 days later — around mid-July 2026.
If that timeline holds, the IDE that currently routes your team’s coding through Claude and GPT will, within six weeks, be owned by xAI.
The Deal Structure
SpaceX paid $10 billion upfront for a purchase option and a joint development agreement with Cursor. Under the agreement:
- Cursor gains access to Colossus 2, xAI’s 1-million-H100-equivalent training supercomputer in Memphis
- Cursor is training a net-new model from scratch on Colossus 2, separate from its current third-party API mix
- SpaceX can exercise the option to acquire Cursor outright for $60B following its IPO, or it can let the partnership continue under the $10B collaboration structure
Neither SpaceX nor Cursor has publicly confirmed model exclusivity terms — that detail stays confidential until the deal closes.
Why xAI Needed This
The strategic motivation is not subtle. After SpaceX merged with xAI in February 2026, the combined entity had the world’s largest private AI training supercomputer and Grok, a model that was not winning the coding tools market.
The problem: SpaceX and xAI engineers were using Anthropic’s Claude for technical work instead of Grok. Multiple reports indicate xAI executives were frustrated that their own AI division’s staff preferred a competitor’s model for day-to-day coding. Bloomberg reported that Grok “lagged behind” in coding capability, and that executives had urged staff to prioritize matching Claude’s performance.
Grok Build — xAI’s standalone agentic coding tool — launched in May 2026. It has not achieved Cursor’s distribution. Cursor, by contrast, counts 50–67% of Fortune 500 companies as active users and has become the default IDE for professional software engineers building on AI APIs.
Acquiring Cursor gives xAI the distribution it cannot organically build fast enough.
What Cursor Currently Runs On
Cursor’s product is model-agnostic by design. The IDE exposes Claude, GPT-4o, Gemini, and other third-party models via API alongside its own Composer models. This architecture is what makes the acquisition consequential: Cursor is not just a coding tool, it is the distribution layer for AI model usage among professional developers.
The Cursor Teams pricing restructure announced June 5 reflects this architecture explicitly — effective July 1, every seat now has two separate usage pools:
- First-party Cursor models (Composer and any models Cursor trains natively)
- Third-party API calls (Claude, GPT, Gemini, etc.)
That separation, introduced a month before the acquisition is expected to close, is a structural pricing signal. It creates a billing seam between “Cursor’s own model” and “models from Cursor’s competitors.”
Three Post-Acquisition Scenarios
Scenario A: Full Grok migration (most likely over 12–18 months)
SpaceX routes Cursor’s default model toward Grok and the new Colossus-trained model. Third-party API access remains available but becomes a premium add-on or is repriced to recover margin. Cursor’s competitive edge shifts from “best model selection” to “best xAI model integration.” Builders on Claude Code see Cursor users — many of whom were also Claude API customers — migrate to the xAI ecosystem.
Scenario B: Hybrid indefinitely (possible if antitrust intervention delays integration)
The FTC has opened an antitrust review of the deal. Premature coordination concerns (“gun-jumping”) have already emerged — reports indicate xAI told staff to stop direct contact with Cursor employees after early collaboration raised regulator flags. A consent decree or behavioral remedy could require Cursor to maintain third-party model parity for a defined period.
Scenario C: Deal collapses (low probability, nonzero)
SpaceX pays the $10B exit fee and Cursor remains independent. Under this scenario, the joint development continues but Cursor’s model mix stays competitive. Given the IPO capital, scenario C is unlikely unless antitrust review produces a hard block or SpaceX fundamentally changes strategic direction post-IPO.
The Antitrust Wildcard
The FTC’s review centers on vertical integration: Cursor owns the IDE, SpaceX/xAI owns the training compute and the model, and the combined entity competes directly with both GitHub Copilot and Claude Code — two of the products Cursor currently distributes access to.
The gun-jumping reports add a pre-merger coordination dimension. Antitrust gun-jumping is when merging companies coordinate operations or exchange competitive data before a deal is officially cleared. The fact that xAI issued internal restrictions on Cursor contact suggests awareness of this risk, but also confirms that coordination had already begun.
The deal is large enough ($60B) to require Hart-Scott-Rodino notification and FTC review. Timing into the SPCX IPO creates political pressure on both sides.
What Claude Code Gets If This Deal Closes
Cursor currently acts as a distribution amplifier for Claude. When a Cursor user calls the Claude API through the IDE, that usage flows through Anthropic’s API and contributes to Claude’s developer reach. If xAI acquires Cursor and routes defaults toward Grok, that distribution relationship ends.
The counterpressure: developers who built workflows around Claude’s reasoning and coding quality don’t automatically follow the IDE. Claude Code — Anthropic’s standalone agentic coding environment — is the direct beneficiary of developer friction. The question is whether Claude Code’s onboarding and capability can absorb displaced Cursor users fast enough.
As of early 2026:
- Claude Code: 57% developer awareness, 18% active enterprise workplace usage
- GitHub Copilot: 37% AI coding tools market share, 4.7M paid subscribers, 90% Fortune 100 adoption
- Cursor: 50–67% Fortune 500 company usage
A Grok-exclusive Cursor collapses the market from three-way competition to two-way: Copilot vs Claude Code, with Cursor/Grok as a third entrant building from scratch on distribution it already has but model credibility it doesn’t.
Builder Action Items
Before July 1: Review your Cursor Teams billing tier. The new Standard ($32) vs Premium ($96) split goes live at renewal. Understand how much of your team’s usage is in the third-party API pool, because that’s the pool most exposed to post-acquisition pricing changes.
Before mid-July: Audit your workflow’s model dependency. If your Cursor usage is tightly coupled to specific Claude behaviors — system prompt formatting, tool call reliability, reasoning structure — document those now. Don’t discover the dependency when the default changes.
Model routing flexibility: If you’re building products that run through Cursor’s third-party API, verify that your code can route to the Claude API directly. The architectural change costs an afternoon; the refactor post-acquisition under time pressure costs more.
Claude Code evaluation: If your team hasn’t evaluated Claude Code as a standalone environment, the time to evaluate it is before a forced migration, not during one. Claude Code’s agentic scaffolding, context management, and tool use are different from Cursor’s Composer — but the gap is narrowing.
Watch the antitrust review: If the FTC imposes a behavioral remedy requiring third-party model parity, the urgency drops significantly. This is worth tracking before making major stack changes.
What Isn’t Changing
Cursor’s core IDE experience — the editor, the context window management, the PR review surface, the multi-repo automations from Cursor 3.5 — is not tied to model identity. SpaceX is buying the distribution and the user base, not replacing the product.
Short-term: your Cursor installation works the same way the day after the acquisition as the day before. The model drift happens in default settings, in which models get compute priority on Cursor’s backend, and in how the pricing pools are balanced over subsequent quarters — not in a single cutover.
The builder risk is gradual lock-in through convenience, not a hard switch. That’s exactly the kind of risk worth identifying early.
Deal announced April 21, 2026. SPCX Nasdaq IPO scheduled June 12. Acquisition expected to close ~July 2026. FTC review ongoing.