Artificial intelligence startup Cognition AI has skyrocketed to a valuation of $10.2 billion following a fresh $400 million investment round, underscoring the growing confidence in AI-powered software engineering. The round was led by Peter Thiel’s Founders Fund with participation from Lux Capital, 8VC, Elad Gil, Definition Capital, and Swish Ventures.
Just months ago, Cognition was valued at $4 billion. Its sharp rise has been fueled by the success of Devin, the company’s flagship AI coding tool, billed as the world’s first AI software engineer. Since its launch, Devin’s growth has been nothing short of explosive: annual recurring revenue jumped from $1 million in September 2024 to $73 million by June 2025.
The company’s momentum accelerated with the acquisition of coding startup Windsurf earlier this year. That move doubled Cognition’s revenue and attracted attention after Google licensed Windsurf’s technology for $2.4 billion, hiring much of its talent before Cognition acquired the rest. Since then, Cognition says its enterprise recurring revenue has climbed more than 30%, boosted by cross-selling between Devin and Windsurf’s products.
Cognition’s client list now includes Goldman Sachs, Citigroup, Palantir, and Dell, reflecting the growing adoption of AI-driven development tools in both finance and technology.
CEO Scott Wu emphasized that the latest funding will help scale engineering efforts, expand research, and accelerate product innovation. Despite the rapid expansion, Wu noted that Cognition has kept its net cash burn under $20 millionsince its founding just two years ago.
The company faces stiff competition from tech giants like Google, Microsoft, and Amazon, all investing heavily in AI-driven software solutions. But with fresh capital, a strong product lineup, and strategic acquisitions, Cognition is betting it can carve out a leadership role in the space.
As the race to define the future of software development intensifies, Cognition AI’s rise signals that investor enthusiasm for artificial intelligence remains strong—and the boundary between human and machine-driven coding is becoming increasingly blurred.
