If AI companies were football teams, Anthropic just signed a star lineup from the transfer market. The Claude-maker has officially acqui-hired the co-founders and most of the team behind Humanloop — a London-born startup best known for its enterprise-friendly AI tools like prompt management, model evaluation, and observability dashboards.
Details of the deal? Top secret. But here’s the vibe: Anthropic didn’t buy Humanloop’s platform or IP — it bought the brains that built it. And in AI, that’s often where the real gold lives.
Humanloop’s co-founders — Raza Habib (CEO), Peter Hayes (CTO), and Jordan Burgess (CPO) — along with about a dozen engineers and researchers, are now officially on Anthropic’s roster. These are the folks who’ve helped big-name clients like Duolingo, Gusto, and Vanta run safe, scalable, and compliant AI systems. Exactly the type of talent Anthropic wants as it doubles down on being the “safety-first” AI company for serious enterprise and government buyers.
Why now? Because in today’s AI race, having a great model isn’t enough. You need the infrastructure — the monitoring, the bias checks, the performance dashboards — to win trust in boardrooms and government offices. Anthropic knows it’s competing with OpenAI and Google DeepMind not just on raw intelligence, but on reliability and compliance. Humanloop’s expertise gives Anthropic that extra edge.
The move comes just as Anthropic flexes new enterprise perks like longer context windows and lands a government deal offering its AI services for just $1 per agency in year one — an aggressive, OpenAI-undercutting strategy.
For Humanloop, it’s a full-circle moment. Born in 2020 out of University College London, raised on Y Combinator and Index Ventures funding, and now absorbed into one of AI’s fastest-growing companies.
Translation: Anthropic isn’t just building better AI brains — it’s building the whole nervous system. And Humanloop’s team? They just became the neurons.
About the Author
