Cognition seeks $25B valuation in new Devin funding round
AI

Cognition seeks $25B valuation in new Devin funding round

April 24, 20263 min read
TL;DR

Cognition's Devin AI coding agent could anchor a $25B valuation as the startup races Anthropic and OpenAI for dominance in autonomous software engineering.

Cognition AI is in talks to raise hundreds of millions of dollars at a $25 billion valuation, more than doubling its previous mark, SiliconAngle reported Thursday, citing a Bloomberg story based on people familiar with the matter. No investors were named. Discussions are ongoing and terms could still change.

The company makes Devin, an artificial intelligence coding agent it launched in March 2024 as the world's first "fully autonomous AI software engineer." That claim drew scrutiny immediately, but the product's design was structurally different from what came before.

The bet on Devin

Rather than suggesting completions, Devin generates a step-by-step plan before writing any code, paired with a confidence score indicating how likely the output is to actually work. It handles the full lifecycle: writing, debugging, and deployment, running multiple automated tests before handing results back to the user. A second scan after generation is meant to catch anything the first pass missed. Individual developers access a standard tier; enterprise customers pay more for expanded capacity.

Cognition is pitching the round as a play on momentum. Per SiliconAngle, the company is trying to capture growing investor enthusiasm around AI-generated code, enthusiasm driven primarily by Anthropic and OpenAI.

The competitive context

Both rivals are scaling hard. OpenAI, now valued at $840 billion, plans to nearly double its workforce to 8,000 by end-2026, directing most new hires toward product, engineering, and research. Anthropic is simultaneously testing pricing changes to its own coding tools. As of April 21, Anthropic's pricing page showed Claude Code restricted to Max-tier plans starting at $100 per month, removing it from the $20 Pro tier, The Droid Guy reported. Anthropic called it a test affecting roughly 2% of new signups, but the absence of any notice on the main pricing page caused widespread user confusion before the company clarified.

The Anthropic episode reflects a broader pattern. As the Digital Watch Observatory noted in January, AI companies face mounting pressure to convert infrastructure spending into predictable revenue, and the shift from broad access to tiered gating is becoming an industry reflex. Cognition, still pre-profitability by any reasonable assumption, is raising into that same pressure.

What the valuation requires

A $25 billion figure for a two-year-old product demands a compelling story about future revenue, not just current traction. Cognition's core argument is that autonomous software engineering is a durable category that larger platforms will not simply absorb as a feature. That argument faces real stress. GitHub Copilot normalized AI autocompletion. Cursor moved it to the IDE layer. Each new release from OpenAI and Anthropic narrows the distance between "fully autonomous agent" and "very capable assistant."

Devin's early benchmark results were revised after its initial launch, and its confidence score feature, while genuinely useful, is an implicit acknowledgment that outputs are not always reliable. Enterprise buyers weigh that gap carefully. The European Union's Artificial Intelligence Act, now partially in force, adds compliance complexity for autonomous systems operating in regulated sectors, though exactly how it applies to coding agents remains unsettled law.

None of this forecloses the round. Capital markets for artificial intelligence have consistently priced future scenarios rather than present multiples. Cognition is asking investors to believe that Devin holds a platform-level position that a larger player cannot replicate by bundling one more tool into an existing suite.

Whether that position is defensible at $25 billion is what Cognition's next few quarters will have to show.

---

Frequently asked questions

What is Cognition AI's Devin?
Devin is an AI coding agent that plans, writes, debugs, and deploys code autonomously. Cognition launched it in March 2024, billing it as the first fully autonomous AI software engineer.

What valuation is Cognition currently seeking?
Cognition is in talks for a round that would value it at $25 billion, more than double its prior valuation. The exact amount being raised has not been disclosed beyond "hundreds of millions of dollars or more."

How does Devin differ from tools like GitHub Copilot or Claude Code?
Copilot and similar tools suggest code inline. Devin operates end-to-end: it produces a plan, executes it, tests the output, and handles deployment, positioning itself as an agent rather than an assistant.

Who are Cognition's main competitors?
Anthropic and OpenAI are the most direct rivals in AI-assisted coding, both expanding their toolsets aggressively in 2026. Google and GitHub are also active in the space.