📰 Key Takeaways

Cognition, maker of the AI coding agent Devin, is reportedly in talks with investors for a new funding round that could push its valuation sharply higher. The company raised $1 billion at a $26 billion valuation just this past May, and according to Bloomberg, sources say that if it hits a $1 billion annualized revenue (ARR) threshold, the new round could value it at $40 billion or more. When the previous round was announced three months ago, Cognition founder Scott Wu confirmed to TechCrunch that the company’s ARR had already reached $492 million at the time, with enterprise usage of Devin growing 50% month-over-month over the prior six months. Wu, himself a well-known programming prodigy, told TechCrunch that Devin isn’t being sold as a replacement for human engineers — it’s often assigned the long-tail busywork that programmers generally hate doing, like upgrading legacy software or migrating apps from one platform to another. That’s likely a big reason it’s catching on with enterprise customers. The company says its client list includes Mercedes-Benz, NASA, and Goldman Sachs.

💬 JudyAI Lab Take

Cognition’s Devin is reportedly in talks for a new funding round, and if it hits that $1 billion ARR threshold, its valuation could jump from May’s $26 billion to $40 billion or more — a sign that enterprise AI coding agents are commercializing faster than most expected.

Founder Scott Wu says Devin’s growth isn’t coming from replacing engineers — it’s coming from taking on the long-tail work teams don’t want to touch, like legacy system upgrades and cross-platform migrations, with enterprise usage growing 50% month-over-month over the past six months. That’s a trend worth paying attention to if you’re building AI products: targeting the genuinely stuck, nobody-wants-to-do-it tasks inside a team turns out to be an easier way into enterprise accounts like Mercedes-Benz, NASA, and Goldman Sachs than chasing flashy demos.

If you’re building an AI product, it’s worth asking yourself: does your tool solve for the flashy use case, or the grunt work your team actually avoids day to day?

📅 Source Info

🔗 Further Reading