Nous Research Raises $90M Series B for Open-Source Hermes Agent

Nous Research has closed a $90 million Series B financing that values the company at $1.5 billion, according to reporting by The Wall Street Journal, making the open-source AI lab one of the most richly valued independent agent startups. The company announced the round on October 7, 2026, saying the capital will fund an enterprise push for its open-source Hermes agent and the development of a mobile app.

The round was led by Robot Ventures, according to Axios reporting cited by RuntimeWire. The company's announcement named NVIDIA, Microsoft's M12, Samsung Next, Union Square Ventures, Y Combinator and Menlo Ventures among the participating investors. The company did not disclose individual check sizes.

The closing brings Nous Research's cumulative capital to approximately $160 million. TechCrunch reported in July that Nous was finalizing a round of at least $75 million at the same $1.5 billion valuation, but the financing had not yet closed at that point. The announced $90 million figure is the completed financing rather than a recycled version of the earlier fundraising discussion.

Hermes, the Open Agent at the Center of the Bet

Hermes is the product the new capital is meant to scale. According to the company's documentation, Hermes is an open-source AI agent that can browse the web, edit files, write code and learn additional skills from user activity. The Wall Street Journal reported that Hermes has been downloaded more than 22.7 million times since February.

The agent ships under an MIT license, and its documented capabilities include persistent memory, reusable skills, web search, code execution, scheduled tasks and connections to messaging platforms and other tools. Users can run Hermes on their own computers or servers. Nous Research also offers hosted model access, tools and cloud hosting through a product called Nous Portal, which is where the company sees its paid business growing.

The distinction matters for how the funding will be judged. GitHub stars and download counts measure developer curiosity, not customer demand. The Series B funds the unglamorous work of turning an open tool into something enterprises will pay for: deployment tooling, privacy controls, security reviews and administration features.

The Monetization Question Every Open-Source Lab Faces

Nous Research's pitch is that openness is distribution. The company provides the software openly while charging companies for deployment, security, management and enterprise features around it, according to reporting by The Wall Street Journal. The management team expects annualized revenue to exceed $100 million by year-end, up from $36 million in mid-September.

That projection is a company forecast, not realized revenue, and it deserves the usual caution. Moving from $36 million in annualized revenue in mid-September to more than $100 million by December would require a sharp acceleration in enterprise deal-making. The enterprise push the financing backs will test whether developer enthusiasm for an open agent translates into durable corporate spending.

The company was founded in 2023 by Jeffrey Quesnelle, Karan Malhotra, Ryan "Teknium" and Shivani Mitra, according to TechCrunch's July reporting. The founders built their early reputation through open work on AI models and software. In a note accompanying the financing, chief executive Dillon Rolnick described Nous as a team that began with few resources and built in public, and said the company was formed to resist the concentration of AI behind closed products. Its early work included fine-tuning Meta's Llama models to better reflect individual users' preferences.

Why the Agent Ecosystem Is Watching

For the AI agent ecosystem, the round is a meaningful data point on a question that has hung over the space all year: can an open-source agent support a real enterprise business? Nous Research is now capitalized to answer it at scale, with backing from both specialist crypto-adjacent funds like Robot Ventures and corporate venture arms like NVIDIA's and Microsoft's M12.

The investor mix itself tells a story. Chipmakers and cloud platforms have a direct interest in agents that run on their infrastructure, while enterprise-focused funds are betting that the hard part of the agent market is not the model but the plumbing — permissions, memory, audit trails and the tooling that lets an agent operate safely inside a company.

That is also where the risk sits. An agent that can edit files and execute code on a user's behalf needs strict boundaries, reliable audit trails and human oversight, and enterprise buyers will price that operational risk into every deal. If Nous Research can sell managed Hermes deployments that clear those bars, the $100 million forecast may look conservative. If developer adoption stays hobbyist, the $1.5 billion valuation will be harder to defend. Either outcome will be instructive for the next open agent lab raising at a fraction of that price.

For now, the company's trajectory is one of the clearest experiments in the agent economy. Coverage of the financing by Tech Startups' funding roundup and RuntimeWire's report underscores how closely the ecosystem is tracking it, and the AI Agents Directory news brief placed the raise alongside a busy day of agent funding and tooling news.