In April 2026, Michael Truell posted a short reaction on X after SpaceX secured the option to buy his company for $60 billion. “Excited to partner with the SpaceX team to scale up Composer,” he wrote. “A meaningful step on our path to build the best place to code with AI.” Composer is Cursor’s own AI coding model. Two months later, the option turned into a signed deal — the largest acquisition of a venture-backed startup ever recorded.
Four years earlier, Truell and his three co-founders had nothing to show for a year of work.
The year that went nowhere
Truell co-founded Anysphere in 2022 with fellow MIT students Sualeh Asif, Arvid Lunnemark, and Aman Sanger. Their first idea wasn’t a code editor. It was AI-assisted autocomplete for CAD software — tooling meant to speed up the way mechanical engineers design physical parts.
It didn’t work. None of the four founders were mechanical engineers, none were especially passionate about the space, and there wasn’t enough CAD data on the internet to train a model that mattered. Truell later described that stretch as “wandering in the desert.”
They pivoted to a market they actually understood: their own. Coding had what CAD didn’t — founders who were the target customer, and an internet full of code to train on.
Building the whole editor, not a plugin
The obvious move in 2022 was to build an AI plugin for an existing tool like GitHub Copilot. Truell and his co-founders decided against it. Speaking at Y Combinator’s AI Startup School in a fireside chat with YC general partner Diana Hu, Truell explained the reasoning: “We were obsessed with AI’s potential to change software development. But existing tools like GitHub Copilot weren’t pushing the limits.”
Instead of adding AI to a code editor built for humans typing by hand, Anysphere built an editor from the ground up with AI as the default way of working — what would later be called Cursor. The bet was that AI-assisted coding wasn’t a feature. It was a different way to build software, and it deserved its own environment rather than a sidebar in someone else’s.
The revenue ladder
Cursor’s growth, once it found its footing, was compressed into a timeframe that outran most enterprise software comparisons entirely.
The company crossed $100 million in annualized revenue in January 2025. By April, it had tripled to $300 million, according to Dealroom. By June 2025, it had passed $500 million ARR — a 60% jump in two months — alongside a $9.9 billion valuation, per TechCrunch. Cursor crossed $1 billion in ARR by November 2025, the same month Forbes reported that all four cofounders had become billionaires. It reached $2 billion by February 2026, and by June 2026 was generating roughly $4 billion in annualized revenue.
The deal
The relationship with SpaceX started as a partnership, not an acquisition. On April 21, 2026, SpaceX secured an option: buy Anysphere outright for $60 billion later in the year, or pay $10 billion to deepen the two companies’ existing collaboration. SpaceXAI and Cursor were already working together, the companies said at the time, “to create the world’s best coding and knowledge work AI.”
The bigger number arrived on June 16, 2026 — four days after SpaceX’s own Nasdaq debut, a $135-per-share IPO that raised $75 billion and was, by multiple accounts, the largest public offering in history. SpaceX exercised its option: a $60 billion, all-stock acquisition of Anysphere, expected to close in the third quarter of 2026 pending regulatory approval. Reporting at the time put the anticipated net worth of each of the four cofounders at roughly $2.7 billion once the deal closes. Part of the stated rationale was integration — folding Cursor’s coding engine into the broader AI push connected to xAI’s Grok models.
Truell was 25 years old when the deal was announced, an MIT dropout and former Google intern whose four-year-old company had just become the biggest acquisition of a venture-backed startup on record.
What the desert bought them
The CAD detour wasn’t wasted time, in retrospect — it was the filter. It ruled out a market where the founders had no unfair advantage and no personal stake, and it pointed them toward the one where they had both: a problem they lived with every day, in a domain where the training data already existed in the form of every public codebase on the internet.
Cursor didn’t win by being the first AI coding assistant. GitHub Copilot predated it by years. It won by refusing to treat AI as an add-on to a tool built for a pre-AI way of working, and by building the whole environment around a different premise instead. That decision — plus roughly four years of compounding revenue growth most software companies never see in a decade — is what SpaceX paid $60 billion for.
Related founders
- Pieter Levels Built 70 Startups. Four of Them Actually Worked. — an early Cursor investor through his personal fund, levels.vc, who marked the SpaceX-Anysphere deal as his first real investment exit.
- Six Months, $80 Million: How Maor Shlomo Built Base44 Alone — another AI coding platform, built solo and sold to Wix for $80 million, from the same “vibe coding” wave Cursor helped start.
- Anton Osika Built Lovable to a $13.3B Valuation From Sweden — a European AI coding startup that crossed $100 million ARR faster than any software company on record.
The CAD pivot and Truell’s “wandering in the desert” characterization are drawn from a16z’s Speedrun newsletter (“How to Pivot Well,” speedrun.substack.com) and The Wantrepreneur Show’s profile of Truell’s journey with Cursor. The “existing tools like GitHub Copilot weren’t pushing the limits” quote is from Truell’s fireside chat with YC general partner Diana Hu at Y Combinator’s AI Startup School, June 2025. Revenue milestones are per Dealroom.co ($300M ARR, April 2025), TechCrunch ($500M ARR and $9.9B valuation, June 5, 2025), and multiple 2026 reports citing Anysphere’s $1B (November 2025), $2B (February 2026), and roughly $4B (June 2026) ARR figures. The cofounders-become-billionaires milestone is per Forbes (Rashi Shrivastava, November 13, 2025). The April 21, 2026 SpaceX option and Truell’s “excited to partner with the SpaceX team to scale up Composer” quote are per CNBC’s coverage of the option agreement and Truell’s post on X (@mntruell). The June 16, 2026 $60 billion all-stock acquisition, its timing relative to SpaceX’s own Nasdaq IPO, and the “largest acquisition of a venture-backed startup” characterization are per CNBC and Forbes (Sandy Carter, June 16, 2026). Anticipated post-close net worth figures are per Quartz’s coverage of the deal. Truell’s age, MIT dropout status, and Google internship background are per Yahoo Finance and Moneywise’s profiles of Cursor’s CEO published around the deal announcement.
