# CodeRabbit Raises $143M to Govern AI-Generated Code > Source: > Published: 2026-09-03 02:08:43+00:00 Your AI coding agent just opened 17 pull requests. Your team has capacity to seriously review maybe three of them today. Welcome to the problem CodeRabbit just raised $143 million to solve. The AI code review company, now valued at $1.5 billion after its Series C, launched **Agentic Change Management** this month — a governance layer built for the uncomfortable reality that AI generates code faster than engineering teams can evaluate it. ## The Numbers Behind the Problem This is not a hypothetical crisis. GitHub is processing 275 million AI agent commits per week in 2026. Merged pull requests have grown from 25 million per month in January 2023 to 90 million today. Agent-opened PRs specifically went from 4 million per month in September 2025 to 17 million by March 2026 — a 4x jump in six months. The average individual developer is opening [59% more PRs than they were eight months ago](https://www.coderabbit.ai/blog/the-pull-request-lives-on-ai-gave-it-a-bigger-job). The industry framing has crystallized around a single tension: “Code is abundant. Judgment is scarce.” Ashish Chaturvedi of HFS Research puts it plainly: “The problem is increasingly not the ability to generate code but the limited amount of human attention available to review it.” Your CI/CD pipeline got faster. Your review capacity did not. ## What Agentic Change Management Actually Does CodeRabbit’s ACM platform has three functional components, each targeting a different layer of the PR overload problem. ### Triage The intake system scores incoming PRs across value, urgency, risk, dependencies, readiness, and reviewer fit — then routes them accordingly. Low-risk, low-complexity changes get automated handling. High-consequence changes get flagged for human attention. The goal is to stop treating every PR as equally deserving of a senior engineer’s focus. ### Change Stack The comprehension layer. Raw diffs are notoriously bad at communicating intent, especially when an AI agent generated the code and cannot explain its reasoning. [Change Stack transforms those diffs into semantic explanations](https://www.coderabbit.ai/blog/introducing-agentic-change-management) — showing blast radius, downstream dependencies, and architectural impact alongside the actual changes. A reviewer can understand a change in minutes rather than hours of line-by-line archaeology. ### Security Rather than scanning only the changed lines, this component applies codebase-wide analysis — examining relationships across files, services, and data flows to surface complex vulnerabilities that line-level review misses. When it finds something, it routes a fix back through the PR workflow rather than creating a parallel ticketing mess. The Security tier runs at $40 per seat per month on top of the base plan. ## Open Source Gets the Whole Thing Free All public GitHub repositories get the entire ACM suite for free — automatically, with no application required. Triage, Change Stack, Security, Review — everything. CodeRabbit is backing this with a [$10 million direct-cost pledge](https://www.coderabbit.ai/blog/coderabbit-expands-its-commitment-to-open-source) covering model inference, compute, security analysis, and infrastructure — not inflated list-price accounting. Their previous pledge was $1 million; they delivered $1.2 million. This one is 10x larger. Major open source projects already on CodeRabbit include Vue.js, Bun, pnpm, Mermaid, Nuxt, and Ant Design. That Datadog participated as a strategic investor alongside being a customer is a signal — product-market fit tends to look like your customers writing checks. ## The Governance Question Nobody Wants to Ask ACM is genuinely useful. But there is a meta-problem worth naming: if an AI agent writes the code and another AI system decides which PRs deserve human attention, the humans in the loop are reviewing what two AI systems agreed they should review. That is a specific kind of accountability gap. Stephanie Walter of HyperFRAME Research frames the concern clearly: teams need to understand “who defines scoring criteria, what evidence supports each decision, how model drift is detected, and who remains accountable when it misses a risky change.” For a 50-person engineering team, the Security tier alone costs $2,000 per month. [Governance is becoming a line item](https://www.infoworld.com/article/4208611/coderabbit-targets-ai-generated-code-overload-with-agentic-change-management.html) — and someone has to own the budget for it. ## What Teams Should Do Now If you maintain a public repository, enable CodeRabbit today — there is no reason not to, and the free tier is the full product. For teams already paying for GitHub Copilot, ACM addresses a gap that Copilot’s code review feature does not: prioritization and organizational impact rather than just correctness. The deeper question is strategic. Teams adopting AI code review tools are making a decision about where human judgment sits in the software delivery process. ACM makes that decision explicit, which is more honest than the alternative — pretending that reviewing 90 million PRs per month at human speed was ever going to work.