CodeRabbit, the AI code review company, has closed a $143m Series C at a $1.5bn valuation — co-led by Atomico and Smash Capital, with Datadog and BMW i Ventures among the new backers. Revenue has grown more than 5x year-on-year. The company now runs over two million automated code reviews a week across more than 17,000 customers, including Adyen, Indeed, BMW and Nvidia. Alongside the raise, CodeRabbit is launching “Agentic Change Management” — a layer built specifically to govern, prioritise, explain and monitor code changes coming from AI agents, not just human developers.
The framing matters more than the funding number. A year ago, “AI code review” meant a linter with better manners — a tool that caught bugs a human reviewer might miss. CodeRabbit’s pitch has shifted: the problem it now says it solves is that a meaningful share of the code hitting production was written by an agent that doesn’t get tired, doesn’t ask for a second opinion, and can generate a week’s worth of pull requests before lunch. Review isn’t the bottleneck anymore — knowing what to trust is. That’s a different, harder problem, and it’s the one every team using Claude Code, Cursor or Copilot at scale is quietly running into.
So what
This is a governance problem before it’s a tooling problem. Any team that has handed real production work to AI coding agents — including us — needs a review discipline that assumes agent-generated code arrives faster than a human can meaningfully audit it line by line: automated checks that catch what a rushed human reviewer won’t, paired with judgement about which changes actually need a person to slow down and look. Tools like CodeRabbit are one part of that; the other part is architectural discipline and senior oversight that doesn’t get automated away. If you’re weighing how to get the speed of AI-assisted development without losing control of what ships, that’s exactly what our AI-assisted development approach is built around. Get in touch if you want to talk through what that looks like for your team.