Cognition raises $1bn at up to $26bn valuation

Cognition, the startup behind AI coding agent Devin, said it had raised more than $1 billion in a new funding round, valuing the company at roughly $25 billion to $26 billion depending on the measure used. The deal is another large wager on AI coding tools, a market where startups are trying to move beyond autocomplete assistants and into software tasks that would usually sit with engineers.

Cognition closed a $400 million financing at a $10.2 billion post-money valuation only eight months ago, TechCrunch reported. In its own funding post, the company said annualised revenue run-rate had reached $492 million. Those figures help explain why investors are treating coding agents less like a novelty and more like an emerging enterprise software category. They also show how quickly the market has repriced a company built around one flagship product.

TechCrunch reported the new deal as a $25 billion pre-money round, while Bloomberg reported put the valuation at $26 billion. The gap is mostly about framing rather than a disagreement over the size of the raise. Either way, the round places Cognition among the most highly valued private AI startups focused on developer tools rather than foundation models.

In a Bloomberg interview, chief executive Scott Wu said Cognition wants flexibility in how it builds Devin as the underlying model market changes.

“As the model layer continues to heat up and get more competitive, what we see is that working with the combination of models is actually much better than having to rely on any single model.”

— Scott Wu, Bloomberg

Wu also said the capital “allows us to stay independent and continue as an independent business”. That argument matters because Cognition is pitching Devin as a software layer that can sit above competing model providers rather than depend on one of them. Investors appear to be backing that position at infrastructure-style valuations, not the multiples usually attached to a single application.

Cognition used its funding post to repeat its description of Devin, saying it launched the product two years ago “as the first AI software engineer”. For investors, the more important line in the same post was the revenue run-rate. It suggests customers are already spending meaningful sums on software that aims to do more than code completion or chat-based help. That gives the company a stronger case than startups still selling a product vision ahead of measurable demand.

For the developer-tools market, the round suggests capital is still flowing to startups that can turn access to AI models into products teams use inside real workflows. The money also shows some investors are looking beyond the model makers themselves and toward companies that package those models into software engineers will actually use. Cognition still has to show Devin can keep expanding inside engineering groups without being undercut by cheaper assistants or copied by larger platforms. Even so, the size of the raise, coming so soon after the last one, shows investors are still willing to back AI coding companies that can pair a clear product pitch with fast revenue growth.