SpaceX's $60 Billion Cursor Bet: The Deal That Could Reshape AI Coding

When SpaceX announced on April 21 that it had secured an agreement giving it the right to acquire AI coding startup Cursor for $60 billion — with a fallback $10 billion partnership option — the announcement landed like a thunderclap across Silicon Valley. By some measures, it is the largest option-to-buy arrangement ever struck in the technology industry, and it signals Elon Musk’s most ambitious pivot yet: transforming SpaceX from a rocket company into an AI powerhouse before its summer IPO.

The Deal Structure

The transaction is built around two tiers. SpaceX has the right to either acquire Cursor outright for $60 billion — which would rank among the largest technology acquisitions in history — or pay $10 billion for what the company described as compensation for “our work together.” The dual-track structure reflects SpaceX’s calculated caution: the company is delaying the final acquisition decision until after its IPO, partly to avoid updating confidential financial filings and partly to finance the purchase with freshly issued public stock.

Cursor is built by Anysphere, a San Francisco startup founded in 2022 that has raised more than $3 billion from investors including Andreessen Horowitz and Benchmark. The company’s AI-powered code editor has achieved remarkable market penetration among professional developers, offering real-time code generation, inline AI chat, and multi-file editing powered by frontier language models including OpenAI’s GPT-5.5 and Anthropic’s Claude Sonnet 4.6.

The $60 billion figure is a staggering multiple for a company that has yet to go public or disclose revenue at scale. For context, Microsoft’s acquisition of GitHub — which now houses Copilot — cost $7.5 billion in 2018.

Why Colossus Changes Everything

What makes this potential union particularly compelling is the compute matchup. SpaceX’s Colossus supercomputer in Memphis, Tennessee — the same facility that provides training infrastructure for Anthropic’s Claude models under a deal announced last week — runs the equivalent of roughly one million H100 GPUs. Pairing that infrastructure with Cursor’s distribution network and developer mindshare would allow SpaceX to build and deploy coding models that no independent AI lab could match on throughput alone.

“The combination of Cursor’s leading product and distribution to expert software engineers with SpaceX’s Colossus training supercomputer will allow us to build the world’s most useful models,” SpaceX said in a statement accompanying the announcement.

That claim deserves scrutiny. Training a frontier coding model requires sustained access to compute at a scale that even well-funded AI labs find difficult to guarantee. Colossus gives SpaceX that guarantee in-house, without dependence on cloud provider pricing or availability. A proprietary coding model trained on SpaceX’s own infrastructure and deployed through Cursor’s interface would be architecturally unconstrained in ways that models built on rented cloud compute are not.

Microsoft Passed

Adding an ironic dimension to the deal: Microsoft, which invested $13 billion in OpenAI over several rounds and positions GitHub Copilot as its primary AI coding offering, reportedly had the opportunity to acquire Cursor and declined. Internal documents surfaced during the ongoing Musk v. Altman trial in Oakland suggest Microsoft has consistently prioritized its OpenAI partnership over alternative AI investments — a strategy that may now look costly given Cursor’s trajectory and valuation.

The dynamic is particularly striking given that Microsoft’s 2018 investment in OpenAI, now being scrutinized in court, was reportedly motivated at least partly by the fear of losing OpenAI to Amazon rather than genuine belief in the company’s near-term capabilities. Whether Microsoft’s pass on Cursor reflects the same kind of organizational paralysis, or a genuine strategic bet on Copilot, remains an open question.

The Competitive Landscape

The move reshapes competitive dynamics in AI coding at a moment when the sector is already in flux. Cursor, GitHub Copilot, Replit, and agent-first coding platforms like Cognition’s Devin are all competing for the developer workflow. A SpaceX-owned Cursor with access to Colossus-scale compute would become the first coding AI platform with direct, in-house ownership of frontier model training — a structural advantage none of its current competitors possess.

It also positions SpaceX — and by extension Musk — as a direct competitor to OpenAI in the developer tools market, even as the Musk v. Altman trial plays out. Musk left OpenAI’s board in 2018 and launched his own AI lab, xAI, in 2023. The Cursor deal represents a third AI coding vector: separate from xAI’s Grok and from OpenAI, built on SpaceX’s aerospace-grade infrastructure.

For Anthropic, which currently has a compute deal to train on Colossus, the question becomes what happens to that arrangement if SpaceX begins developing its own competing models. Compute sharing is a precarious relationship when the host becomes a direct competitor.

IPO Dynamics

SpaceX’s choice to delay the final transaction until after its IPO is strategically precise. The company filed confidential IPO documents earlier this year and is widely expected to debut on public markets in the second half of 2026. At its most recent private valuation of approximately $350 billion, SpaceX would need to issue roughly 17% of its post-IPO share count to finance a $60 billion all-stock acquisition — a dilution that would likely be far easier to absorb with public market liquidity than in the private markets.

The strategic logic is clear: owning the dominant AI coding environment would generate recurring software revenue and developer data that could power future model improvements, creating a flywheel that extends SpaceX’s AI ambitions well beyond aerospace.

How SpaceX’s public market investors will react to a $60 billion software acquisition is a different question. Rocket companies trading on public markets are not typically valued like AI software companies, and the move would represent a fundamental redefinition of SpaceX’s business model.

What Comes Next

The deal leaves several important questions open. Cursor’s existing investors — who participated in funding rounds that implied a far lower valuation — will need to approve any acquisition. The company has also been in discussions for a fresh $2 billion funding round, negotiations that SpaceX effectively preempted with its acquisition offer. Those potential investors are now in a holding pattern.

On the regulatory side, a $60 billion acquisition by a company controlled by Elon Musk — who is simultaneously running a competing AI lab (xAI) and a social media platform (X) with strong developer communities — would almost certainly draw antitrust scrutiny. The DoJ and FTC have both signaled heightened interest in AI market consolidation.

Whether the $60 billion option is ultimately exercised or SpaceX settles for the $10 billion partnership route, this agreement represents a fundamental shift in how the rocket company thinks about itself. Musk’s vision of an AI-first aerospace company, armed with developer tools and frontier training compute, may be the most consequential — and controversial — bet he has placed since signing the original OpenAI charter in 2015.