Nous Research, the company behind the wildly popular open-source AI agent Hermes, is finalizing a new funding round that would value the startup at $1.5 billion, according to a TechCrunch report published July 13. The round, led by Robot Ventures with significant participation from Union Square Ventures (USV), aims to raise at least $75 million — more than doubling the company’s total raised to date.
Founded in 2023 by Jeffrey Quesnelle, Karan Malhotra, Ryan Teknium, and Shivani Mitra, Nous Research has taken an unorthodox path to a unicorn valuation: give the core product away for free. Hermes Agent, an MIT-licensed autonomous assistant that runs on desktops, VPS instances, or the cloud, has amassed over 214,000 GitHub stars and nearly 40,000 forks since its launch. Users can deploy it on their own hardware starting at $20/month for hosted tiers — a pricing model that turns adoption into recurring revenue without paywalling the source.
From open source to $1.5B
The new round comes less than three months after Nous announced a $50 million Series A led by Paradigm. Prior investors include North Island Ventures, OSS Capital, and former Coinbase CTO Balaji Srinivasan. The rapid valuation jump — from Series A to unicorn territory in a single quarter — reflects the explosive growth of the AI agent market and Hermes’s position as the leading open-source contender against proprietary offerings from OpenAI, Anthropic, and Google.
Unlike those competitors, Hermes runs locally. It can search the web, write and execute code, interpret images, and autonomously acquire new skills through its built-in /learn command — no retraining, no plugin marketplace, no API key required. It integrates natively with Telegram and Discord, the platforms where its developer community already lives.
The Judgment Release and beyond
The funding talks follow the v0.18.0 “Judgment Release” on July 1, a milestone that saw the team close every P0 and P1 issue across the entire repository — 692 priority items resolved in twelve days. The release brought Mixture-of-Agents as a first-class model selector, background subagent fan-out, self-verification with completion contracts, a /journey memory timeline, and production-grade gateway features including scale-to-zero and drain coordination. In total: 1,720 commits, 998 merged PRs, and 370+ community contributors in a single release window.
Two patch releases followed within a week (v0.18.1 and v0.18.2 on July 7), adding another ~660 merged PRs of bug fixes, Windows installer self-healing, and WhatsApp bridge stabilization.
What the money buys
According to sources, the new capital will fund team expansion, enterprise features, and product acceleration. Nous also operates Psyche, a decentralized compute network for training, and has released specialized models for coding and mathematics. The funding round signals that investors are betting not just on Hermes Agent, but on Nous Research becoming a platform company — one where the model is free and the infrastructure, tools, and enterprise services generate the revenue.
The deal has not yet closed, and terms could shift. But at $1.5 billion, the bet is clear: open-source AI agents are not a hobby. They’re a business.
Sources: TechCrunch, GitHub Releases
FAQ
Is the $1.5B funding round closed?
Not yet. According to TechCrunch’s sources, Nous Research is “finalizing” the round and has fielded high investor interest, but the deal has not officially closed and terms could still shift.
What does this mean for Hermes remaining open-source?
Nous Research has consistently shipped Hermes Agent under the MIT license, and the core agent remains free to run on your own hardware. The funding is earmarked for team expansion, enterprise features, and product acceleration — not for paywalling the existing open-source agent. The company’s business model relies on hosted cloud tiers ($20–$200/month) and likely future enterprise services, while keeping the core agent open.
How does a $1.5B valuation make sense for a company giving away its main product?
The math is in the adoption funnel. With 214,000+ GitHub stars, nearly 40,000 forks, and a rapidly growing user base, Hermes has the kind of developer reach that converts into paying customers at the top of the hosted funnel. Investors are betting on the platform play: the agent itself is free, but the infrastructure, enterprise tooling, and cloud hosting generate recurring revenue. It’s the Red Hat model applied to AI agents.
How does Hermes compare to proprietary alternatives like ChatGPT or Claude?
The fundamental difference is where the agent runs. Hermes operates locally on your own hardware (desktop or VPS), meaning no data leaves your machine unless you configure external providers. ChatGPT and Claude are cloud-only. Hermes also features a self-improving skills loop (/learn) that lets it acquire new capabilities from usage without retraining — something proprietary agents don’t offer. The tradeoff: you manage your own infrastructure, but you keep your data and your autonomy.