SpaceX's $60 billion agreement to buy Cursor is the bluntest signal yet that AI coding tools have moved from developer side project to strategic infrastructure.
The deal announced today says more than another funding round could. SpaceX has agreed to buy Anysphere, the San Francisco company behind Cursor, for $60 billion in stock, after an April partnership that gave Elon Musk's newly public company the right to make the acquisition. Business Insider reported that Cursor, founded in 2022 by Michael Truell and three MIT classmates, has pushed past $1 billion in annualized revenue in under a year. You don't pay that price for a clever code editor. You pay it because you think the next software stack will be written with AI sitting inside the workflow.
That is the real point here. Cursor started as a tool developers used to generate, edit, and explain code inside an editor. Now it is being treated like a core asset for SpaceX and its AI ambitions around Grok. Axios reported that the all-stock deal is the largest acquisition of a venture-backed startup to date, excluding Musk's earlier xAI transaction. SpaceX had just come through a record IPO that valued it around $2.5 trillion, so the company is using a public-market moment to grab one of the fastest-growing names in software. Frankly, that is not subtle.
The original draft overstated one important fact: this is an agreement, not a completed acquisition. The New York Post reported that the merger is expected to close in the third quarter of 2026 and that it includes termination fees, including $4 billion if regulators block it. That distinction matters because you should read the deal as a live test, not a finished trophy. Regulators, SpaceX shareholders, Cursor employees, and enterprise customers all still have a stake in what happens next.
The money is following the workflow
Cursor is not alone, and that is why the SpaceX move lands so hard. Cognition, the company behind the Devin programming agent, raised $1 billion at a $26 billion valuation in late May, according to Bloomberg reporting cited across the market. Its valuation had been $10.2 billion in September 2025, and its revenue run rate reportedly climbed from $37 million to $492 million over the same period. If those figures hold, this is not investor excitement drifting around a vague AI category. It is money chasing tools that are already changing how code gets written.
Supabase sits in a different part of the stack, but it belongs in the same story. CNBC reported that the database infrastructure company raised $500 million at a $10.5 billion valuation in early June, roughly doubling its valuation in eight months. That matters because many prompt-built apps still need authentication, storage, databases, and deployment rails. You can ask a model to build the front end. The app still has to run somewhere.
Lovable shows the other side of the market. Business Insider reported that the Swedish app-building startup, launched in 2024, reached $400 million in annual recurring revenue and a $6.6 billion valuation, with investors including CapitalG and Menlo Ventures. Its pitch is simple enough for a non-engineer to understand: describe the application you want, and the system builds it. That has obvious appeal. It also creates obvious risk, because software made quickly is not automatically software you can maintain.
The phrase vibe coding came from Andrej Karpathy in February 2025, and it captured something real. You tell the machine what you want, accept more generated code than you would have a few years ago, run it, complain when it breaks, and repeat. Merriam-Webster listed the term as slang in March 2025, and by now it is not only slang. It is a market map.
Still, don't confuse adoption with maturity. Cursor itself had an embarrassing moment in April 2025 when an AI support bot invented a non-existent login policy and told users they needed separate subscriptions for multiple devices, before the company corrected it. That is a small example, but it is useful because it punctures the clean story. These tools are powerful, and they can still be confidently wrong in front of paying customers.
For founders, the lesson is not to slap AI coding into a pitch deck and wait for a term sheet. The bar is higher now. Investors are looking for revenue, usage, retention, and a real place in the development workflow. Cursor has that. Cognition is trying to prove it with agentic programming. Supabase has it because prompt-built software still needs infrastructure. Lovable has it because non-technical users want working apps, not tutorials about how software is made.
SpaceX buying Cursor does not mean every AI coding startup gets a SpaceX-sized exit. It means the category has crossed into strategic territory, where the buyers are not only software companies and the valuations are no longer built on novelty. If you build in this space, you now have to answer the harder question: are you a feature developers try, or are you part of how software gets shipped?
Also read: ChatGPT's majority is gone and OpenAI's IPO story just got harder to tell • Andreessen Horowitz bets $9 million that AI reliability is a category, not a feature • Plaud reached $250 million in recurring revenue without a single venture dollar and is now targeting $500 million in 2026 sales
