Hermes agent maker Nous Research in talks for new funding at $1.5B valuation - TechCrunch

A billion and a half. That’s the new price tag on a lab best known for Hermes—an agent framework, not just another chatbot weight. TechCrunch reports Nous Research is in talks for a $1.5B valuation. This isn’t a bet on a model; it’s a bet on the end of the nerd-on-call. Why this matters for the compute world: Agentic inference is the true datacenter killer. A prompt asks for a token; an agent burns through a cluster’s worth of latency and accelerating memory. This valuation signals that the next era of silicon isn’t about linear training runs, but about orchestrating multi-step loops—which means a new pricing tier for every TPU and H100 out there. The bigger signal? Open-source pragmatism is beating closed-source mystique. Nous isn’t chasing AGI white paper glory; they’re shipping tools that make machines act. While hyperscalers argue over model genius, Nous is monetizing the plumbing of digital autonomy robustly enough to hit a unicorn-plus milestone. Case closed: Big models think; Hermes agents do. And in this market, "doing" is worth $1.5B. ```json { "key_insight": "Agentic execution, not raw model intelligence, is the new valuation driver for compute and AI labs.", "confidence": 92 } ```