CodeRabbit raises $143M to tame the flood of AI-generated code
The money in AI coding is moving from generating code to governing it: CodeRabbit raised $143 million at a $1.5 billion valuation to help teams keep up with the flood of machine-written pull requests. Elsewhere, Ahrefs shipped an agent workspace for marketers, and some of AI's biggest backers started buying old-school service firms to rewire them.
CodeRabbit raised $143 million in Series C funding at a $1.5 billion valuation to help engineering teams manage the explosion of AI-generated code. The round was co-led by Atomico and Smash Capital, with BMW i Ventures, Datadog, Hirtle Callaghan, SineWave Ventures and Scenic Management joining, alongside existing backers CRV, Scale Venture Partners, Flex Capital, Pelion Venture Partners, Harmony Partners and Engineering Capital. It lands less than a year after the company's $60 million Series B: CodeRabbit says revenue grew more than 5x year over year, its platform runs more than 2 million code reviews a week, and customers include Adyen, BMW, Indeed, JFrog, NVIDIA and Trivago.
The company's argument is that plain review no longer scales. Code now arrives from developers, nontechnical staff and autonomous coding agents, so CodeRabbit is shipping what it calls Agentic Change Management — a control layer with three pieces: Triage routes low-risk pull requests to automation and flags high-stakes ones for human reviewers, Change Stack explains what each proposed change does, and Security hunts for vulnerabilities in production code. CEO Harjot Gill sums up the problem: "Every change creates a decision for the team."
That framing is the real story. The first phase of the AI coding boom produced a glut of generated code; the next phase is about governing that output, and the money is following the bottleneck. Datadog's participation hints that code review is becoming an observability problem as much as a quality gate, and CodeRabbit's pledge to spend more than $10 million over the next year keeping its agents free for open-source maintainers shows it still needs the community that made it.
Ahrefs launched Letaido, an agent-powered marketing workspace, with its new Agent A marketing agent running on top. Teams hand agents multistep jobs — keyword clustering, content-gap analysis, technical site audits — schedule recurring workflows, and set up continuous monitoring of their own sites and competitors, with a shared visual interface so several people can refine the same output. Automations keep running on built-in hosting after a conversation ends, and the platform connects to Notion, Slack, HubSpot, Google Ads and WordPress with native access to Ahrefs' link and keyword data.
Ahrefs is positioning this as the answer to a shift it tracks closely: discovery is moving off search engines and into AI tools, and the SEO giant wants to own the agent layer that monitors it. Early customer Foundation Marketing says keyword research and bottom-of-funnel content audits, once a 40-hour week, now take about an hour. The bet is that data moats, not model access, decide who owns marketing automation.
Thrive Holdings, the year-old holding company from OpenAI backer Thrive Capital, closed a $2 billion raise at a $12 billion valuation to buy traditional service businesses and infuse them with AI. The New York Times reported the round, backed by SoftBank, D1 Capital and others. Thrive Capital's Josh Kushner has been one of the most aggressive investors in the model layer; Thrive Holdings is the vehicle for the next stage of the thesis — acquire unglamorous services firms and rewire them with AI agents.
It is a telling sign of where AI value capture is heading. The investors who financed the model buildout are now convinced the outsized returns come from applying it to industries that haven't been touched, and a wave of AI-flavored rollups is likely to chase the same playbook.
What to watch: whether CodeRabbit's open-source pledge survives its enterprise expansion — and whether marketing teams actually trust agents with their dashboards.
Your team's review queue — is it drowning in AI-generated pull requests? Tell us in the comments.
