Cognition’s $40B Target Shows AI Coding Is Still The Hottest Bet

Cognition is reportedly trying to raise more than $1 billion at a valuation above $40 billion, a number that shows how aggressively investors are still pricing the AI coding market.

PYMNTS, citing Bloomberg, says the Devin maker is in early talks with investors for a new funding round that would lift its valuation by more than 50 percent from the $26 billion mark it reached in May. Tech in Asia also reported the same $1 billion raise and $40 billion valuation target.

This is moving at a speed that would have sounded unserious in enterprise software a few years ago. Cognition was valued at about $10.2 billion in late 2025, then raised more than $1 billion at a $26 billion post-money valuation in May 2026. If the new round lands near $40 billion, investors will be saying that AI coding agents are not a side product. They are a new software platform category.

The reason is Devin. Cognition sells Devin as an AI software engineer that can plan tasks, write code, fix bugs and work through software problems with more autonomy than a normal coding assistant. The product sits in the same broad market as Cursor, Claude Code, GitHub Copilot, OpenAI Codex and Replit’s agentic tools, but Cognition has been especially aggressive in framing Devin as a worker rather than a feature.

That framing is exactly why the valuation is so high. If AI coding agents can reduce the time needed to build and maintain software, then the market is much bigger than developer productivity subscriptions. It touches consulting, enterprise IT, internal tools, app maintenance, testing, security fixes and the whole backlog of software work companies never quite get to.

There is still a large unanswered question. Faster coding does not automatically mean better software. The more code agents generate, the more companies need testing, review, security checks and architecture discipline. That is why AI code-testing startups are also gaining attention. The bottleneck may shift from writing code to trusting and maintaining it.

We are already seeing this across the agent market. xAI has launched Grok Bot as an always-on AI teammate, while River AI is trying to build more user-owned AI infrastructure after a huge funding round. The direction is clear: investors want agents that do work, not only chatbots that explain work.

The risk is valuation gravity. A $40 billion private valuation demands enormous revenue growth and enterprise retention. AI coding is hot, but it is also crowded, and the biggest model companies can bundle coding tools into platforms customers already use. Cognition will have to prove that Devin is strong enough to remain a standalone company rather than a feature competitors can imitate.

For now, the market is choosing optimism. AI coding is one of the few generative AI categories where customers can see a direct productivity argument. If Cognition can turn that argument into durable enterprise revenue, the valuation may make sense. If not, this round will become another test of whether AI startup pricing has moved faster than the businesses underneath it.