Cursor Built an AI Agent Called Sand. Then SpaceX Bought Cursor.
Cursor made its name as a tool for people who write code. Now, according to a July 9 report from The Information, it is quietly building one for people who do not. The project is codenamed Sand, and it marks the first time Anysphere — the company behind the AI-native code editor Cursor — has built a product aimed at the office worker rather than the software engineer. The timing is what makes it interesting. Sand is entering development at the exact moment its parent company is being absorbed into Elon Musk’s SpaceX in the largest acquisition of a venture-backed startup on record. Whether Sand ever reaches the public may not be a decision Cursor gets to make alone.
What Sand Is Actually Built to Do
According to sourcing cited by TweakTown and The Next Web, Sand is designed as a general-purpose assistant rather than a coding-specific tool. It is meant to answer emails and texts, organize spreadsheets and documents, and manage the kind of administrative overhead that fills a typical workday — while still being able to reach into engineering tasks when needed. That positions it as a direct competitor to Anthropic’s Claude Cowork and OpenAI’s ChatGPT Work, both of which are chasing the same “agent that runs your whole workday” category. Internal rollout to Cursor employees began in late June 2026, meaning the company is already using its own staff as the first test group before any public release.
A Timeline That Runs Straight Through the SpaceX Deal
Sand’s development reportedly began around April 2026 — the same month SpaceX secured an exclusive option to acquire Anysphere or walk away for a $10 billion breakup fee, and the same month Cursor began leasing compute from SpaceXAI. On June 16, 2026, SpaceX exercised that option, agreeing to buy Anysphere in an all-stock deal valuing the company at $60 billion — disclosed in a Form 8-K filing with the SEC just four days after SpaceX’s own record-setting Nasdaq debut. The price represented roughly double the $29.3 billion valuation Anysphere had commanded in its $2.3 billion Series D round in November 2025, a round backed by Accel, Coatue, Thrive Capital, Andreessen Horowitz, DST Global, NVIDIA, and Google. CNBC reported the deal is expected to close in the third quarter of 2026, subject to regulatory approval. Cursor had reportedly turned down two separate acquisition approaches from OpenAI and been examined — then passed over — by Microsoft before the SpaceX arrangement emerged.
Three-Way Race for the Office Agent
Sand is not entering an empty field. Anthropic’s Claude Cowork and OpenAI’s ChatGPT Work have already staked out the “AI agent for knowledge workers” category, and both have been expanding their enterprise footprints. Sand’s edge, if it has one, is Cursor’s existing developer trust and its multi-agent orchestration technology — first proven in Cursor 2.0’s parallel-agent architecture and refined further in Cursor 3’s agent-first interface. Applying that same “run several agents at once, each in its own sandbox” approach to office work rather than code is a genuine technical differentiator, at least on paper.
Get N G’s stories in your inbox
Join Medium for free to get updates from this writer.
Comparison Snapshot
What Happens If SpaceX Says No
The uncomfortable possibility is that Sand never ships at all — not because the product fails, but because it becomes redundant to a strategy Musk is setting for the combined company. SpaceX is reportedly integrating Cursor’s coding data into Grok’s training pipeline and giving Cursor access to xAI’s Colossus supercompute cluster, which suggests the acquisition is about more than adding a coding tool to SpaceX’s portfolio — it is about acquiring data, compute, and engineering talent after xAI’s own co-founders reportedly departed by the end of March 2026. A consumer-facing productivity agent, built to compete with two of the best-funded AI labs in the world, may or may not fit into that plan. Until the acquisition closes — expected sometime in Q3 2026 — Sand exists in a strange limbo: real enough to be tested internally, uncertain enough that nobody outside Anysphere and SpaceX can say whether it survives the merger.
For now, Sand is a case study in how consolidation in AI is starting to reshape not just who owns which company, but which products even get permission to exist. Cursor built its reputation on staying out of the model wars. Its next act may be decided entirely by the outcome of one.
What do you think — does Sand ship as an independent product, or does it quietly disappear into SpaceX’s broader AI roadmap?
