Devin's numbers just came out. And they're wild.
$1 million in ARR in September 2024. $445 million run rate today. Usage doubling every eight weeks.
Cursor held the all-time SaaS record at $1M to $100M in 12 months. Devin crossed that line in roughly 10. Cursor reached $100M through 360,000 individual developers at $276 ACV. Devin reached it through the US Army, Goldman Sachs, Mercedes-Benz, Citi, Dell, Cisco, and Palantir. The United States military pays production rates for an autonomous coding agent.
Cognition is now raising at $25 billion. That's 56x run rate. Cursor cleared $9.9B at a similar multiple last May, and the multiple held because the curve hadn't bent. The unusual part isn't the price. The unusual part is that the doubling is still happening at $445M.
The buried number is the burn. Cognition has spent under $20 million cumulatively since founding two years ago. Most Series B companies spend that in a single year. Devin's $445 million was built on Series A money.
Then the Windsurf paragraph. Google paid $2.4 billion in licensing fees in July to pull Windsurf's founders out the door. The remaining company sold to Cognition inside 72 hours for a fraction of that. Combined enterprise ARR rose more than 30% in seven weeks post-close. Less than 5% customer overlap pre-acquisition. Google paid two and a half billion dollars to hand Cognition the IDE distribution layer.
In March 2024, independent testers said Devin completed 3 of 20 tasks. The internet called it a fake demo.
Two years later, that product codes for the US Army.
Inside Cognition with Grant Sanderson of @3blue1brown
We sat to discuss what it’s actually like to work as an engineer on the frontier of AI. The day-to-day, how work actually gets done, and what it feels like to build on a team that's like “a close-knit group of athletes training for the Olympics.”
Video link below:
OpenAI offered $3 billion for Windsurf. Google paid $2.4 billion for the CEO and top researchers. Cognition picked up the remaining company, the IDE, the product, 350 enterprise customers, and $82M ARR, for roughly $250 million.
That $250 million just became the best deal in AI coding.
Devin went from $1M ARR in September 2024 to $73M ARR by June 2025. The agent runs on its own VM with a desktop, browser, and full computer use. It can debug, deploy, test, and ship code end to end. But an autonomous agent that lives outside your editor is an autonomous agent developers forget about. Nobody leaves their IDE to manage a separate tool.
Windsurf gave Devin a front door. And Windsurf 2.0 shows exactly what that front door does.
The IDE is now a Kanban board for AI agents. You plan locally, hand tasks to Devin with one click, close your laptop, and Devin opens PRs while you sleep. This is the first coding product that literally tells developers to stop watching.
Compare this to Cursor's bet. Cursor 3.0 (launched this month) replaced its classic IDE layout with an "agent-first interface" for managing parallel AI fleets locally. Every agent needs your screen open. Cursor hit $2B ARR in February, doubling in three months. The product gets better the more attention you give it.
Cognition's product gets better the less attention you give it.
The revenue gap is still massive: $2B+ vs roughly $155M combined. But the product categories are diverging. Cursor sells a faster coding experience. Cognition sells engineering capacity measured in cloud hours.
The constraint that determines who wins: if five Devin agents ship code overnight, you wake up to PRs across five repos touching code you've never read. The bottleneck moves from writing to reviewing. And reviewing autonomous AI-generated code is a skill most teams haven't built yet.
Whoever solves review at scale wins this entire market.
@lalleclausen Hey @lalleclausen , we are also creating a unique consumer app and experience @INSIDERSxyz to bring web3 to the masses. App and AI agent launching shortly. DM me if open to chat?
@0xferg@ethereum@Immutable@apecoin IMX hasn’t even executed on the projects you already have in the pipe, just focusing on accomplishing that first please