Cognition raises $1B at $25B pre-money valuation

Cognition raised more than $1 billion at a $25 billion pre-money valuation.

Cognition just pulled in a monster round: more than $1 billion at a $25 billion pre-money valuation, or $26 billion post-money. The company behind Devin says enterprise demand is rising fast, and the numbers attached to this round explain why investors are paying attention.

A valuation jump that says a lot

The headline number is the valuation, but the bigger signal is how quickly Cognition moved from a $10.2 billion post-money valuation in September to $26 billion now. That is a huge jump in eight months, even by AI startup standards.

For a company that sells an autonomous coding agent, that kind of pricing says investors think the market is still early and that the winners may not all be the foundation model labs. It also says enterprise software buyers are willing to pay for agents that can do real work inside large codebases.

• September 2025 round: $400 million
• September 2025 valuation: $10.2 billion post-money
• May 2026 round: more than $1 billion
• May 2026 valuation: $25 billion pre-money, $26 billion post-money

Devin’s enterprise pitch is getting louder

Devin was introduced as an autonomous AI software engineer, and Cognition is now leaning hard into enterprise adoption. The company says customers include Mercedes-Benz, NASA, Goldman Sachs, and Santander.

That customer list matters because it moves the story beyond demos and social media clips. If a coding agent is getting used inside regulated, high-stakes organizations, the product has crossed into a very different category of software buying.

“We’re seeing a lot of enterprise interest in using Devin for real engineering work,” said Scott Wu, Cognition’s co-founder and CEO, in the company’s launch materials for Devin.

Cognition also says enterprise usage of Devin has grown 50% month-over-month for the past six months. That is the kind of growth rate investors want to see when they are writing checks at this scale, especially in a market where every major AI lab is pushing its own coding tools.

The competition is getting crowded

The funding comes at a time when the coding-agent market is getting tighter, not looser. Anthropic’s Claude Code, OpenAI’s Codex, and Google’s Jules are all pushing into the same territory: tools that can read code, make edits, run tasks, and help developers ship faster.

Cognition’s edge is that it is still independent and focused. That matters because the big model makers can bundle coding features into broader platforms, while a startup has to prove it can win on product depth, workflow fit, and trust.

• Claude Code comes from a model company with deep research and infrastructure resources
• Codex sits inside OpenAI’s broader developer stack
• Jules gives Google another way into AI-assisted programming
• Devin is pitched as an autonomous engineer, not just a code helper

There is also a strategic wrinkle here: Cognition acquired the remaining parts of Windsurf last year, which gave it more surface area in the developer tools world. That move suggests the company is not trying to be a narrow agent demo. It wants a bigger role in how teams write and maintain software.

What this round means for the rest of AI coding

When a startup raises more than $1 billion in a single shot, the message is bigger than cash. It tells the market that investors still believe there is room for specialized AI coding companies even as the giants crowd in.

The question now is whether Cognition can turn that belief into durable product adoption. Revenue is already high for a startup at $492 million in annualized run-rate, but the next test is whether Devin becomes a standard tool inside engineering teams or stays a high-profile option for a few large customers.

If Cognition keeps growing enterprise usage at anything close to 50% month-over-month, the company will keep pulling attention away from general-purpose assistants and toward agents built for specific jobs. If that growth slows, the valuation will look a lot less comfortable. For now, the market has clearly chosen to pay first and ask harder questions later.