AI coding startup Cognition has raised more than $1 billion in a funding round that values the company at $26 billion, the company announced on May 27.
The round was led by Lux Capital, General Catalyst, and 8VC, with support from existing investors including Founders Fund and Elad Gil, alongside new investors Ribbit Capital, Atreides, and Layer Global.
The company said its annualised revenue run rate has reached $492 million, while enterprise usage has grown more than tenfold since the beginning of the year.
Cognition launched Devin two years ago as what it described as the first AI software engineer. Since then, the company said cloud-based AI agents have moved from niche adoption to mainstream software development workflows.
“As we’ve scaled, Cognition has become a trusted partner for the world’s largest and most impactful organisations,” the company said in a statement.
Cognition said enterprises, including Citi, Mercedes-Benz, Goldman Sachs, Elevance, Dell, Santander, the US Army, and the US Navy, are using Devin. The company added that startups such as Exa, Modal, Eight Sleep, and OpenRouter have also adopted the platform to automate parts of software development.
The company highlighted several customer deployments. Mercedes-Benz reduced an eight-month legacy modernisation project to eight days using Devin, according to Cognition.
Itaú, Latin America’s largest bank, now fixes 70% of security vulnerabilities automatically with Devin, the company said. Cognition also said Infosys and Cognizant have integrated Devin into software delivery workflows.
Cognition described itself as an “independent agent lab” that works with multiple foundation model providers. The company said engineering teams are increasingly focused on balancing cost and performance as AI token usage rises across the industry.
“Teams today care more than ever about the ratio of price to performance, which requires using the right model for the right task,” the company said.
Earlier this year, Cognition expanded its model training efforts and launched SWE-1.6, a model that it said is now the most-used model in Windsurf. The company said customers have adopted the model for its speed and cost efficiency, with performance reaching up to 950 tokens per second.
Cognition said software development is shifting toward ‘self-driving’ workflows where engineers focus more on structuring problems while AI agents execute tasks. The company added that 89% of code committed by engineers internally is now committed by Devin, with the remaining code generated by local agents in Windsurf.
ALSO READ: Middle East: The Sovereign AI Testbed US, EU and Asia Can Learn From
