The company behind Devin is attracting enormous investment as autonomous coding becomes one of the most closely watched areas of enterprise AI.
Software development has become one of the first professional industries where AI agents are moving beyond assistance and toward completing multi-step tasks.
Cognition, the startup behind the Devin AI software-engineering agent, has become one of the most visible companies in this transition.
In September 2026, Cognition announced a $2 billion Series E funding round at a $48 billion valuation, nearly doubling the company’s reported $26 billion valuation from its previous financing in May. (MarketScreener UK)
The funding illustrates how quickly investors are repricing startups operating at the frontier of AI-assisted work.
From Coding Assistant to AI Engineer
Traditional coding assistants generally help developers write individual functions, explain code or autocomplete lines.
Cognition is pursuing a broader model.
Its Devin system is designed to plan software tasks, write code, test changes and work through development projects with limited human intervention. (MarketScreener UK)
That changes the role of AI from a tool that developers operate to a system that can potentially operate inside the development workflow itself.
The distinction matters.
A developer using an autocomplete system remains responsible for directing nearly every step.
An autonomous coding agent can instead be given a larger objective and asked to determine the steps required to complete it.
Revenue Growth Is Becoming as Important as Valuation
Cognition said its run-rate revenue had grown from $492 million to nearly $900 million since its previous funding round. (MarketScreener UK)
That growth is important because the AI startup market is gradually moving beyond demonstrations.
Investors increasingly want evidence that companies can convert AI capabilities into recurring commercial revenue.
The ability to sell autonomous software-engineering systems to enterprises could therefore become one of the most significant tests for the broader agent economy.
Competition Is Intensifying
Cognition is not operating alone.
Factory raised $200 million in September at a $5 billion valuation for its enterprise AI coding-agent platform. Reuters reported that the company allows businesses to use AI agents to build, test and maintain software. (MarketScreener)
Meanwhile, OpenAI, Anthropic, GitHub and other technology companies are also developing increasingly capable coding systems.
The result is an unusually competitive startup market.
Why Coding Is Such an Important AI Market
Software engineering has several characteristics that make it particularly suitable for AI agents.
Code is digital.
It can be tested automatically.
Errors can often be detected through automated systems.
Changes can be reviewed using version-control systems.
And software-development tasks can be divided into identifiable stages.
These characteristics make coding a natural environment for increasingly autonomous AI.
The Enterprise Opportunity
The biggest opportunity may not be replacing developers.
Instead, coding agents could allow existing engineering teams to tackle larger workloads with the same number of people.
An agent might investigate an issue, modify several files, run tests and prepare a pull request while human engineers remain responsible for review and final decisions.
This could change the economics of software development.
The Bigger Picture
Cognition’s $2 billion financing shows that investors are placing enormous value on the idea that AI agents could become active participants in professional work.
Whether that value ultimately translates into long-term business performance will depend on reliability, cost, security and how much autonomy companies are willing to give these systems.
For startups, however, the direction is clear: software development is becoming one of the major laboratories for autonomous AI.