Cognition AI Raises $1 Billion at $26 Billion Valuation to Scale AI Coding Platform Devin

--- headline: "Cognition AI Raises $1 Billion at $26 Billion Valuation to Scale AI Coding Platform Devin" slug: cognition-1b-26b-valuation-devin category: business story_number: "01" date: 2026-05-27 author: The Vault AI Edition ---

Cognition AI, the startup behind the autonomous coding agent Devin, announced Wednesday that it has raised more than $1 billion in a Series D funding round, valuing the company at $26 billion. The deal more than doubles the $10.2 billion valuation Cognition achieved just eight months ago, and cements the two-year-old company as one of the most richly valued private AI startups in the world.

Lux Capital, General Catalyst, and 8VC co-led the round, with participation from Founders Fund, Ribbit Capital, Atreides Management, and Layer Global. The financing brings Cognition's total funding to more than $2.5 billion since its founding in 2023.

Revenue Rockets From $37 Million to $492 Million in 12 Months

The numbers behind the valuation tell a striking growth story. Cognition's annualized revenue run rate has surged from $37 million in May 2025 to $492 million today, a roughly 13-fold increase in a single year. Enterprise usage of Devin has grown more than tenfold since the start of 2026, and the company says usage has climbed 50% month over month for the past six consecutive months. Cognition counts Goldman Sachs, Citi, Mercedes-Benz, Santander, Dell, NASA, and branches of the U.S. military among its customers.

"We launched Devin two years ago as the first AI software engineer," the company said in its announcement. "Since then, cloud agents have gone from niche to mainstream, and today they are the fastest growing way to create software."

The company highlighted concrete customer results to justify the premium valuation. Mercedes-Benz reportedly used Devin to compress an eight-month legacy code modernization project into just eight days. Itau, Latin America's largest bank, now fixes 70% of its security vulnerabilities automatically using the platform.

Eating Its Own Cooking

Perhaps the most eye-catching data point is internal: CEO Scott Wu said that 89% of Cognition's own codebase is now written by Devin, up from just 13% in December 2025. That figure, if it holds up to scrutiny, makes Cognition one of the most aggressive practitioners of its own product in enterprise software history. Wu emphasized on Bloomberg Television that the raise allows Cognition to remain independent, a pointed remark given the recent trend of AI coding startups being absorbed by larger platforms.

Wu has framed Cognition's multi-model approach as a strategic advantage rather than a dependency. The company runs Devin on a mix of its own proprietary models, including the recently released SWE-1.6 optimized for coding tasks, alongside models from OpenAI and Anthropic. Wu has argued that routing customers to the best available model for their specific needs produces better results than locking into any single provider.

"Individual engineers are able to spend more of their time on the creative structuring of problems and tasks, and their army of Devins reliably executes," the company said, describing a shift toward what it calls "self-driving software development."

Why This Matters: The AI Coding Wars Intensify

Cognition's mega-round lands in what may be the single hottest investment category in venture capital. The AI coding tools market has become a magnet for capital at a pace that rivals the early cloud computing boom, but compressed into months rather than years.

The competitive landscape is fierce and getting more crowded. Cursor, built by Anysphere, hit $2 billion in annual recurring revenue in roughly three years and was recently the subject of a potential $60 billion acquisition by SpaceX. Anthropic's Claude Code, OpenAI's Codex, and Google's coding agent Jules are all pushing aggressively into the same territory. Salesforce alone expects to spend $300 million on Anthropic tokens this year, primarily for coding use cases.

What makes the competitive dynamics unusual is that Cognition uses models from OpenAI and Anthropic, the very companies building rival coding products. Cognition's bet is that the model layer will commoditize while the agent orchestration layer, the planning, execution, and debugging logic that sits on top of foundation models, will capture durable value. That thesis will be tested as the model providers steadily improve their native coding capabilities with each release.

The consolidation wave is already underway. Cognition acquired the remaining assets of Windsurf last year after Google struck a $2.4 billion deal to hire Windsurf's top engineering talent. Windsurf's desktop code editor, which now has more than one million users across 4,000 organizations, serves as Devin's primary interface alongside a second agent called Cascade that handles simpler, local coding tasks.

What to Watch Next

Cognition says it aims to cross $1 billion in annualized revenue later this year, a milestone that would further validate the market for autonomous coding agents. The company plans to use the new capital to refine its models, improve the customer experience, and potentially pursue more acquisitions.

The deeper question is whether Cognition can sustain its independence at a $26 billion valuation, roughly 53 times its current revenue. Developer skeptics, including prominent voices like George Hotz who has called AI coding agents "one of the most costly mistakes" in software development, argue that the technology is overhyped. But with nearly half a billion dollars in revenue and a customer list that spans Wall Street, automakers, and the Pentagon, Cognition is no longer just a bet on the future. It is a rapidly growing business staking its claim that software can, increasingly, write itself.