Cognition AI, the company behind the autonomous coding agent Devin, has raised over $1 billion in new funding at a $26 billion post-money valuation — more than doubling its $10.2 billion valuation from September 2025. The round was co-led by Lux Capital, General Catalyst, and 8VC, with participation from Ribbit Capital, Atreides Management, and Peter Thiel’s Founders Fund, bringing total funding to over $2.5 billion.
The numbers behind the valuation are staggering: Cognition’s annualized revenue run rate surged from $37 million in May 2025 to $492 million in May 2026 — a 13x increase in 12 months. The company says it aims to cross $1 billion in annualized revenue later this year. Its customer list now includes Goldman Sachs, Mercedes-Benz, NASA, Citibank, Dell, Santander, and several U.S. government agencies including the Army and Navy.
Perhaps the most striking claim: CEO Scott Wu says over 90% of Cognition’s internal code is now written by Devin. The company runs Devin on a mix of proprietary models and models from OpenAI and Anthropic, positioning itself as an orchestration layer rather than a model company. Enterprise usage has grown more than tenfold since early 2025, and the company expanded through its 2025 acquisition of Windsurf’s remaining assets after Google acquired Windsurf’s talent and licensing rights for $2.4 billion.
Source
Cognition Blog — Series D Announcement · The Next Web · AI Business
Commentary
A company where the product writes 90% of the company’s code is either the most compelling proof-of-product in enterprise software history or the beginning of a very interesting debugging story. The revenue trajectory — $37M to $492M in 12 months — is genuinely remarkable and suggests Devin is solving real problems for real engineering teams, not just generating hype.
The strategic tension is fascinating: Cognition builds on top of OpenAI and Anthropic models while those same companies are building competing coding products. Wu’s bet is that the agent layer — orchestration, planning, execution logic — captures more durable value than the underlying models. That’s a bold position when your model providers are also your biggest potential competitors. With Cursor reportedly at $2B ARR and OpenAI and Anthropic investing heavily in code generation, the AI coding market is shaping up to be one of the most intense competitive battles in tech. The question isn’t whether AI coding tools work anymore — it’s who ends up owning the workflow.