Okay so this one genuinely stopped me in my tracks when I saw it.
Last week, SpaceX announced it had struck a deal with an AI startup called Cursor. The terms: SpaceX gets the right to acquire the company by end of 2026 for $60 billion. Or, if they decide to walk away, they pay $10 billion just for the year of working together. Ten billion dollars. To walk away. That’s not a typo.
And the guy on the other side of this deal? Michael Truell. 25 years old. MIT dropout. Worth an estimated $1.3 billion. Running what many people are calling one of the fastest growing software companies in history.
Let me tell you the whole story because it’s genuinely one of the more interesting ones coming out of the AI boom right now.
Who Is Michael Truell?
Truell grew up in New York City and went to Horace Mann, a private prep school in the Bronx. Smart kid, clearly — he got into MIT for computer science and math. During his first year he landed a summer internship at Google at 18, working on language models. That internship turned out to be more important than he probably realized at the time.
While he was there, he met Ali Partovi — an early investor in Facebook and Airbnb — who was recruiting for something called the Neo Scholars program. It’s an accelerator for young tech talent, very selective, thirty spots a year. Partovi gave Truell a coding test. Truell finished it faster than anyone Partovi had seen. He got one of the thirty spots.
Back at MIT, Truell connected with three other students who were equally obsessed with where AI was heading: Sualeh Asif, Aman Sanger, and Arvid Lunnemark. All of them were in computer science. All of them were watching GitHub Copilot come out and thinking the same thing — this is impressive but it’s barely scratching the surface of what’s possible.
By 2022, as graduation approached, all four made a decision that most of their classmates thought was insane. They turned down job offers from Google, Meta, and other big tech companies — the kind of offers MIT CS graduates dream about — and dropped out to start a company.
The False Start Nobody Talks About
Here’s the part that most coverage skips over.
Cursor wasn’t their first idea. Not even close.
Their first attempt was building AI tools for mechanical engineers. CAD software, design tools, that kind of thing. They spent nearly a year on it. And it went nowhere.
Truell described it later as “wandering in the desert.” The team had identified what they thought was a market opportunity but they had no domain expertise in mechanical engineering, no real passion for it, and the customers could tell. It just didn’t click.
Their second idea — a message encryption project — also failed.
Most people don’t know this because by the time Cursor became famous, the early failures got buried under the success story. But it matters because it tells you something about how these founders operate. They were willing to kill two projects and start over rather than keep forcing something that wasn’t working.
The pivot to AI coding almost didn’t happen. Truell said in an interview that they had actually avoided the space initially because they thought it was too competitive. GitHub Copilot was already out there. It seemed crowded.
Then they looked more carefully and realized: Copilot was a plugin. It lived inside someone else’s environment. You were always working within Visual Studio Code’s constraints, and AI was just kind of bolted on top.
What if you built the whole thing from scratch? Not a plugin. An entirely new code editor where AI wasn’t an add-on but the actual foundation of how the thing worked?
That was the idea that became Cursor.
What Cursor Actually Is
Cursor is a code editor — think of it like Microsoft Word but for writing software. The most popular one before Cursor came along was Visual Studio Code, made by Microsoft. Millions of developers use it every day.
What Cursor did was take Visual Studio Code, fork it (basically copy the open source version and rebuild on top of it), and redesign the whole experience around AI. Instead of AI being a helpful suggestion on the side, AI in Cursor is woven into everything. It can understand your entire codebase. It can make complex edits across multiple files at once. It can explain what code does, find bugs, write new features — and do all of this with a level of context that earlier tools simply couldn’t match.
Cursor launched in March 2023 and developers immediately started talking about it. By late 2023 it had indexed 150,000 codebases. The word of mouth was the kind that companies spend millions trying to manufacture and almost never actually achieve.
Nvidia CEO Jensen Huang called Cursor his “favorite enterprise AI service” and said that almost everyone of Nvidia’s roughly 40,000 engineers uses it. When Jensen Huang is saying that publicly, people pay attention.
The revenue numbers are almost hard to believe. Anysphere crossed $100 million in annual recurring revenue in January 2025, and topped $500 million ARR by June 2025. From $100 million to $500 million in five months. And then they crossed $1 billion ARR by late 2025. The fastest B2B software company to reach that milestone. Faster than Slack. Faster than Salesforce. Faster than basically anything.
The Problem That Nobody Saw Coming
Here’s where it gets complicated.
Cursor built its product on top of AI models made by OpenAI and Anthropic. They used these models to power the AI features that made Cursor so good. That was fine — until OpenAI and Anthropic decided they wanted a piece of the AI coding market themselves.
OpenAI launched Codex. Anthropic launched Claude Code. Suddenly the companies whose models Cursor depended on were also competing directly with Cursor. Both providers spent the last year vertically integrating into coding.
This is the nightmare scenario for any startup. Your suppliers become your competitors. You’re paying them for access to their technology while they use that technology to build something that directly threatens you.
Truell acknowledged this publicly. He didn’t pretend it wasn’t happening. He said the market was “standardizing on a couple solutions” and that Cursor needed to adapt fast.
The question was how.
Why SpaceX? This Is the Part That’s Actually Interesting
When you first hear “SpaceX buys AI coding startup,” it sounds completely random. What does a rocket company want with a code editor?
But it actually makes a lot of sense once you think about it.
SpaceX is preparing for an IPO that could be the largest in history. SpaceX is targeting a $2 trillion IPO valuation. To justify that valuation, SpaceX needs to show it’s more than just a rocket company. It’s been positioning itself as an AI and technology company, especially since it merged with Elon Musk’s xAI earlier this year in a deal that valued the combined company at $1.25 trillion.
xAI has the Colossus supercomputer — 550,000 GPUs scaling toward one million. That’s an enormous amount of computing power. More than most countries. But computing power alone isn’t a product. You need software, you need applications, you need something that developers actually want to use.
Cursor is exactly that. It’s the most popular AI coding tool in the world right now. It’s used by developers at Fortune 500 companies and solo founders alike. It has real revenue, real growth, and real brand recognition in the developer community.
The xAI alliance gives Cursor a foundation model partner that is not also a competitor, plus access to the 550,000 GPUs in Elon Musk’s Colossus cluster. In return, xAI gets Cursor’s enterprise install base, training data, and the IDE itself.
For Cursor, this solves the problem. Instead of depending on OpenAI and Anthropic — who are now competitors — they get a model partner who has every incentive to help them win.
The Deal Structure Is Clever
By the end of 2026, SpaceX will either acquire the AI coding startup for $60 billion or pay it $10 billion to walk away. Former White House AI czar David Sacks confirmed the structure on the All-In podcast, framing the $10 billion as the cost of a one-year option on the hottest application in AI.
Chamath Palihapitiya — a prominent tech investor and SpaceX shareholder — went on the same podcast and said Musk actually got a bargain. His argument: SpaceX is planning to fund any acquisition with stock issued at its IPO valuation of around $2 trillion. But today SpaceX is worth closer to $1 trillion. So effectively, Musk is paying for a $60 billion asset with stock that the market is about to revalue upward — meaning the real cost to SpaceX shareholders is significantly lower than the headline number.
Cursor’s valuation went from $2.5 billion in January 2025 to $29.3 billion by November 2025. Before the SpaceX deal landed, Cursor was reportedly in talks to raise at a $50 billion valuation. SpaceX came in at $60 billion and moved fast enough that other potential investors didn’t have time to react. Multiple people on the All-In podcast noted how quickly it all came together — investors apparently asking “where are the wiring instructions?” before the deal was even fully processed.
Prediction markets are pricing roughly a 74% to 83% chance SpaceX actually goes through with the full acquisition by end of year.
What Happens Next
This deal reshapes a few things.
For Cursor, it means stability. They’re no longer at the mercy of OpenAI and Anthropic for their core technology. They have a compute partner with more GPUs than almost anyone on the planet and a very strong financial incentive to see Cursor succeed.Elon Musk Just Bet $60 Billion on a 25-Year-Old MIT Dropout — The Cursor AI StoryElon Musk Just Bet $60 Billion on a 25-Year-Old MIT Dropout — The Cursor AI Story
For the AI coding market, it means the race is on. OpenAI has Codex. Anthropic has Claude Code. Google has its own coding tools. And now SpaceX/xAI has Cursor. Every major AI company is trying to own the developer workflow.
For developers — the people actually using these tools — more competition is generally good news. It drives faster improvements, better pricing, and more options.
And for Truell personally? Three years ago he was a 22-year-old MIT dropout who’d just failed at his first two startup ideas. Today he’s worth $1.3 billion, running a company that Elon Musk is willing to pay $60 billion for, and still only 25 years old.
Not a bad three years.
The Bigger Picture
What this story really shows is how fast things are moving in AI right now. Cursor went from zero to a billion dollars in annual revenue in about two years. It went from a $400 million valuation to a $60 billion acquisition target in less than two years after that.
No company has grown this fast. And Cursor isn’t even the only one. The whole AI sector is producing these kinds of numbers repeatedly, which is either the most exciting thing happening in technology or the most terrifying sign of a bubble, depending on who you ask.
Either way, it’s worth paying attention to. Because whatever happens with the SpaceX deal, the tool that Truell and his MIT classmates built in 2022 has already changed how millions of developers write code. That part isn’t going away.
Come back tomorrow for the next update — there’s always something happening.