Cognition, the company behind Devin — the first AI software engineer designed to work autonomously on real codebases — announced on May 27, 2026 that it has raised more than $1 billion at a $25 billion pre-money valuation.

The round was led by Lux Capital and General Catalyst, with existing investors Founders Fund and 8VC participating alongside new investors including Ribbit Capital, Atreides, and Layer Global.

Eight months ago, Cognition was valued at $10.2 billion after a $400 million funding round. Today it is worth more than double that. In the current venture landscape, that kind of re-rating in under a year is extremely rare.

The number that justifies it is not the brand or the investor list. It is the $492 million annualised revenue run rate, growing at 50% month-over-month for six consecutive months.

Who Is Actually Buying Devin

Cognition counts Mercedes-Benz, NASA, Goldman Sachs, and Santander among its enterprise customers. These are not early adopters testing on side projects. These are institutions with mature engineering teams deploying Devin on production work.

That customer list changes how you read the funding round. This is not speculation on future potential. It reflects revenue already accumulating at an unusual rate, from customers that have options and are choosing to keep spending.

What Devin Actually Does

Devin is an autonomous AI software engineer. Given a task description, it can write code, run tests, fix bugs, search documentation, and ship working software without requiring a human developer to approve every step.

It operates inside a sandboxed environment with access to a terminal, a browser, and code editors. In practice, enterprise teams are using Devin to handle repetitive coding tasks, maintain legacy codebases, write test suites, build internal tools, and accelerate feature development on new products.

The appeal is not that Devin replaces senior engineers — most enterprises using it are deploying it for the work that is clearly defined, time-consuming, and does not require creative judgment. That turns out to be a large share of the total workload.

What This Means for Business

The Cognition raise is significant for three groups in particular.

For business owners evaluating custom software: The cost and time required to build internal tools, automations, and custom applications has dropped dramatically. AI coding agents like Devin handle a meaningful share of the work that previously required a full development team. EDNA’s Omni Apps service sits directly in this space — building AI-powered custom applications that would have cost far more and taken much longer just two years ago.

For data professionals: The definition of technical skill is shifting. Knowing how to direct an AI coding agent — how to write clear specifications, review output, catch edge cases, and integrate the result — is becoming as important as knowing how to write the code yourself. This is exactly the kind of practical, applied AI capability that EDNA Learn trains for. Understanding how to work alongside these tools, not just use them from scratch, is the competitive edge data teams need right now.

For teams evaluating AI investments: Cognition’s revenue growth is one of the most concrete signals available that AI coding agents are no longer in pilot phase. Enterprises with serious engineering needs are deploying them and renewing at scale. The question is no longer “does this work?” The question is “how do we integrate this into our team’s existing workflow?”

At $492M ARR growing 50% month-over-month, the market has answered the first question clearly. The rest is execution.

---

If you want to understand what AI coding agents mean for your team’s capability and your custom software roadmap, the conversation starts here.