Cognition AI, the company behind autonomous coding agents, closed a $2 billion Series E led by Andreessen Horowitz and Accel, nearly doubling its valuation to $48 billion in a matter of months. Existing backers including Founders Fund, General Catalyst, and Avenir participated.
The round is a clean read on where private markets still write the largest checks after foundation-model labs: software that can take a ticket and ship code with limited supervision. Coding agents have moved from demo to budget line in engineering orgs, and Cognition is one of the few specialists priced like a platform company rather than a feature.
The valuation also shows how compressed the cycle has become. Multi-billion-dollar steps that once took years now happen between product releases. That creates fragility if enterprise seat growth slows or if frontier labs give away “good enough” coding agents inside existing subscriptions.
Paired with Mistral’s European mega-round, Tuesday’s financings underline that investors are funding both the model layer and the agent applications that sit on top of it — at prices that assume coding work keeps migrating from humans to systems.
Key takeaway Autonomous coding is no longer a side bet. A $48 billion mark for Cognition says markets think the agent that writes the software may be as valuable as the model that powers the chat window.
Photo: Unsplash (code editor). Sources: Economic Times, Reuters-style roundup coverage of Cognition’s $2 billion Series E at a $48 billion valuation.
