Why SpaceX bought Cursor
Jan-Erik Asplund
TL;DR: With Anthropic's Claude Code and OpenAI's Codex competing directly against Cursor, SpaceX acquired Cursor for $60B to pair the fastest-growing SaaS company in history with xAI's 700K+ GPU compute clusters, applying the same vertical integration playbook it used to drive space launch costs from $10K/kg to ~$100/kg to AI, from chip fabrication to model training to developer-facing application layer. For more, check out our full reports on Cursor (dataset) and SpaceX (dataset).


We previously covered Cursor at $65M in November 2024, $100M in February 2025 and $200M in April 2025, and SpaceX at $4.6B in September 2023, Starlink at $7.7B in June 2024, September 2025, December 2025 and February 2026. We're following up after SpaceX's $60B all-stock acquisition of Cursor in June 2026.
Key points via Sacra AI:
- After GitHub Copilot ($400M ARR in November 2024) pioneered AI autocomplete in your IDE in 2021, Cursor (founded 2022) launched the first AI-native IDE in 2023 and became the fastest-growing B2B SaaS company in history, growing from $1M to $100M ARR in ~12 months by the end of 2024 and establishing & leading AI coding as the top AI application layer category versus Cognition’s Devin (Cognition at $492M ARR in April 2026, focused on agentic coding) and fast follower Windsurf ($82M in ARR in July 2025, team acquihired by Google in July 2025 then company acquired by Cognition).
- Key Cursor partner & model provider Anthropic began competing directly with Cursor, launching Claude Code in February 2025, which combined with the launch of Opus 4.5 in November 2025: (1) surpassed Cursor in annualized revenue run rate in late 2025 and hitting $2.5B annualized revenue by February 2026 and prompting OpenAI to launch Codex ($1B in January 2026) and (2) changed the game from AI assisted coding in the IDE to end-to-end AI agent coding in the terminal, forcing Cursor ($4B annualized revenue in 2025, up from $1.2B at the end of 2025) to declare a state of emergency and prioritize training its own model.
- In April 2026, SpaceX & Cursor announced a deal that gave SpaceX the right to acquire Cursor outright for $60B, giving Cursor access to large-scale dedicated compute via xAI (700K+ GPUs coming online across its Colossus clusters) to go after its model ambitions and xAI access to the massive dataset of developer edits & completions it needs to build a competitive coding model, with SpaceX ultimately exercising the option and paying ~2% of its now-$2.5T market cap (vs. 5% of its $1.25T market cap as of the deal announcement) for a 15x multiple on Cursor's annualized revenue and providing Cursor (~-20% negative gross margin) a home with a long-term durable player to finance its continued growth.
- Adding Cursor ($4B annualized) to xAI ($3.2B revenue in 2025) makes it into a roughly $7B annualized AI business ($1.8B from ads, ~$500M AI revenue for xAI,’s Grok $900M from X, $4B from Cursor) plus $26B in new contracted compute from Anthropic ($1.25B/month) and Google ($920M/month) gives SpaceX's xAI line of business roughly $33B in forward contracted revenue, potentially making it larger than Starlink ($11.4B in 2025) and SpaceX’s launch business ($4.1B in 2025) combined (with the caveat that the compute deals are terminable contracts representing committed rather than realized revenue).
- After building its own rockets, engines, and Starlink satellites from off-the-shelf components to drive space launch costs from $10,000/kg on Falcon 1 to a target of ~$10-100/kg on Starship, SpaceX is now applying the same vertical integration playbook to AI, from chip fabrication at the joint Tesla-SpaceX-xAI Terafab semiconductor factory to compute through its Colossus compute clusters (700K+ GPUs) to Grok model training & Cursor as the developer-facing application layer, with this vertical integration giving SpaceX its differentiation and upside versus Anthropic and OpenAI which operate principally in the model layer.
For more, check out this other research from our platform:
- SpaceX (dataset)
- Why SpaceX acquired xAI
- CoreWeave of space
- SpaceX vs Verizon vs AT&T
- SOTA model nightclub hype cycle
- OpenAI's side chip
- Starlink at $4.1B/year growing 121%
- Anduril, SpaceX, and the American dynamism GTM playbook
- The biggest mistake defense startups make
- Ross Fubini, Managing Partner at XYZ Capital, on the defense tech opportunity
- Shield AI (dataset)
- Scott Sanders, chief growth officer at RRAI, on the defense tech startup playbook
- Anduril (dataset)