Funding

Nous Research Lands $90M Series B Led by Robot Ventures, Hermes Agent Hits 24M Downloads

The AI agent startup behind the popular Hermes tool has secured $90 million in funding from Robot Ventures, Nvidia, Samsung and Y Combinator, pushing its valuation to $1.5 billion.

·2 min read
Nvidia, Samsung back $90M round for AI agent startup Nous Research
Nvidia, Samsung back $90M round for AI agent startup Nous Research

Nous Research Inc., creator of the widely-used Hermes artificial intelligence agent, announced a $90 million Series B funding round today. Robot Ventures served as the lead investor, with participation from Nvidia Corp., Samsung Electronics Co., Y Combinator and additional backers. The round values the company at $1.5 billion.

Since launching Hermes in February under an open-source license, the agent has accumulated more than 24 million downloads and attracted several thousand open-source contributors. The tool currently represents approximately 2.5% of global token consumption, a measure reflecting the computational resources AI models require.

Hermes operates as a desktop application capable of interacting with user files and programs. Users can instruct the agent to organize documents by category, access browser-based information, execute code within Docker containers and handle similar tasks.

Technical Architecture

The agent leverages established large language models including Claude to execute user requests, while also supporting a suite of LLMs developed internally by Nous. These models represent fine-tuned adaptations of open-source language models.

Nous' newest model collection, the Hermes 4 series, was customized using a training dataset containing 5 million synthetically generated records. The company created roughly two-thirds of these records specifically to strengthen Hermes 4's reasoning performance, utilizing an internal tool called DataForge for data generation.

A routing mechanism directs each user query to the most appropriate LLM based on considerations including inference expenses, processing speed and response quality. An additional personalization module tailors LLM outputs according to individual user preferences.

Users can optionally enhance personalization capabilities by connecting Hermes to external databases, such as the agent-optimized RetainDB relational store. Nous indicates that integrating an external database allows Hermes to preserve user preferences and project information across multiple sessions.

Monetization and Growth

Nous offers three commercial versions of its agent. Hermes Cloud targets individual consumers, allowing them to deploy agents within seconds without infrastructure setup. Hermes Business and Hermes Enterprise serve organizational customers, providing administrators with team-level token consumption controls and enabling end-users to distribute customized agents among colleagues.

According to the Wall Street Journal, Nous generated $36 million in annualized revenue as of mid-September. The company projects this figure will exceed $100 million by year's end. Nous plans to deploy the new capital toward accelerating adoption of Hermes' business-focused editions.