Nous Research, maker of the open-weight Hermes family of agent models, is in talks for new funding at a $1.5 billion valuation, according to TechCrunch reporting published July 13. The company has built its reputation on an open, decentralized-training philosophy that stands in deliberate contrast to closed frontier labs -- Hermes models are released with inspectable, self-hostable weights, appealing to developers and enterprises wary of routing sensitive workloads through a closed API they don't control.
The backdrop makes the valuation notable: closed frontier labs are commanding valuations in the hundreds of billions -- Anthropic at roughly $965 billion, OpenAI in a comparable range -- while open-weight labs have historically traded at a significant discount given the apparent tension between giving away your core product and building a defensible business around it. A $1.5 billion mark for Nous suggests investors are increasingly comfortable with that tension, betting monetization comes from enterprise tooling, fine-tuning services and hosted infrastructure layered on top of freely available weights rather than from the models themselves.
“For investors, the round is a useful benchmark for how open-weight AI companies are being priced relative to their closed counterparts at this stage of the cycle.”
The competitive set here includes Mistral and Black Forest Labs, both of which have also attracted substantial venture valuations despite an open-weight-first approach, suggesting a genuine and growing investor thesis that open models can build durable, monetizable businesses -- a meaningfully different bet than backing a closed lab racing purely on capability.
For developers and enterprises, a well-capitalized Nous Research means continued investment in an open alternative to closed agent models, a real strategic hedge for any company wary of full dependency on a single closed-model vendor. For investors, the round is a useful benchmark for how open-weight AI companies are being priced relative to their closed counterparts at this stage of the cycle.
The bear case: open-weight companies face a structurally harder monetization path than closed labs, since anyone can self-host the core product without ever paying Nous directly, and a $1.5 billion valuation prices in tooling and services revenue that hasn't yet been demonstrated at scale. What to watch next: whether the round closes at the reported figure, and what specific enterprise or hosted-infrastructure revenue Nous discloses to justify the mark.