Cognition secures $1 billion with a valuation of $26 billion for its AI coding agent.

      Cognition AI has secured over $1 billion in funding, achieving a valuation of $26 billion for Devin, its AI coding agent. The company's revenue surged from $37 million to $492 million within a year, attracting notable clients like Goldman Sachs, Mercedes-Benz, and the US government.

      Cognition AI has successfully raised more than $1 billion in additional funding, significantly increasing its valuation from $10.2 billion in September to $26 billion. This funding round was co-led by Lux Capital, General Catalyst, and 8VC, with participation from Ribbit Capital, Atreides Management, and Peter Thiel’s Founders Fund. To date, the company has raised over $2.5 billion in total.

      The valuation metrics highlight the company's growth trajectory. Cognition's revenue run rate escalated from $37 million in May 2025 to $492 million today, marking a 13-fold increase in just 12 months. The company aims to exceed $1 billion in annual revenue later this year, serving clients like Goldman Sachs, Mercedes-Benz, NASA, Santander, and various US government sectors.

      Cognition’s main offering is Devin, an AI agent intended to automate the entire coding process. Unlike code completion tools that provide suggestions during coding, Devin functions as a comprehensive coding agent that can autonomously generate functioning software based on task descriptions. It handles planning, writing, debugging, and deployment across complex workflows.

      The most notable claim comes from within Cognition. Co-founder and CEO Scott Wu stated that over 90% of the company's internal code is now created by Devin. If this claim is accurate, it positions Cognition among the most enthusiastic users of its own product in the enterprise software sector, as the company automates its engineering processes using the same tool it markets to others.

      Cognition operates Devin using a combination of its proprietary models alongside those from OpenAI and Anthropic. Wu described this approach as a strategic advantage rather than a reliance on any single provider. He argued that utilizing various models can yield better outcomes as competition intensifies, allowing Cognition to direct clients to the best tools based on their needs and establishing itself as an orchestration layer rather than just a model provider.

      The AI coding sector is experiencing significant growth. Cursor, an AI coding editor developed by Anysphere, rapidly reached $2 billion in annual recurring revenue in about three years and was discussing a $2 billion raise at a $50 billion valuation prior to SpaceX's acquisition deal for $60 billion in April. Both OpenAI and Anthropic are heavily investing in coding functionalities within their foundational models, with Salesforce planning to spend $300 million on Anthropic tokens this year for coding-related applications.

      The competitive landscape is unique. Cognition employs models from OpenAI and Anthropic, the same companies producing competing coding products. Wu addressed this concern, arguing that Cognition's real value lies in the orchestration, planning, and execution logic built on top of foundational models, rather than the models themselves. This represents a belief that while the foundational model layer may become commoditized, the orchestration layer can retain lasting value—an assumption that OpenAI and Anthropic will not develop equally competitive agent products that undermine their own API customers.

      Cognition is also expanding through acquisitions. In July 2025, it acquired the remaining assets of Windsurf, a coding startup that had been in demand during a bidding contest between OpenAI and Google. Google ultimately acquired Windsurf’s top engineering talent and licensing rights for $2.4 billion, while Cognition secured the remaining assets, including technology, clients, and employees who opted not to join Google.

      The Windsurf deal exemplifies the ongoing consolidation trend. Spending on AI-native enterprise solutions soared by 94% year-on-year in early 2026, while traditional SaaS growth slowed to 8%. The influx of capital into AI coding tools is prompting a wave of acquisitions, talent acquisitions, and competitive positioning that resembles the early days of cloud computing, but at an accelerated pace.

      Cognition plans to utilize the new funding to enhance its models, improve customer experiences, and potentially pursue further acquisitions. Wu emphasized on Bloomberg Television that this funding enables Cognition to maintain its independence, a significant statement in light of the SpaceX-Cursor deal and the broader trend of AI startups being absorbed by larger companies.

      The question of Cognition’s independence remains open as the company experiences rapid growth. With a valuation of $26 billion and $492 million in revenue, the valuation stands at approximately 53 times revenue. This level can only be sustained if growth continues at its current speed and the AI coding market does not consolidate around a few dominant players. Major software firms are increasingly developing AI coding capabilities, and foundational model providers are consistently enhancing their coding performance with each release.

      Cognition argues that its advantage lies in execution speed and that the agent layer, rather than the model layer, represents where long-lasting value resides. If Devin can truly automate 90% of a