Hermes Agent artwork from Nous Research showing a many-armed classical figure over a bright blue background — Hermes Agent artwork. Image: Nous Research.

Hermes Agent, the open-source AI agent from Nous Research, is moving from hobby projects into the office. On October 7, 2026, Nous announced a $90 million raise to build a business version of the tool. The Wall Street Journal put the company's value at $1.5 billion.

That is a big moment for open-source AI agents. Many companies that want an agent for their teams still pick a closed product from a big vendor.

I'm Nitin from the Debugged Pro team. We build AI agents, chatbots and web apps for clients in India and the USA. Clients keep asking us which agent to trust with their work. So here is a plain look at what Nous announced, what Hermes for Businesses offers, and what to check before you try it.

Hermes Agent Funding: What Nous Research Announced

CEO Dillon Rolnick shared the news in Nous Research's note on its fundraise. He named Nvidia, Microsoft's M12, Samsung, Robot Ventures, Union Square Ventures, Y Combinator and Menlo Ventures among the investors.

TechCrunch's report on the Nous Research round added more detail. Here are the main facts:

Round size: $90 million in a Series B.
Valuation: $1.5 billion, first reported by the Wall Street Journal and confirmed by Nous, according to TechCrunch.
Lead investor: Robot Ventures, per TechCrunch. 1789 Capital, where Donald Trump Jr. is a partner, also joined.
Total funding: $158 million for the three-year-old startup.
Revenue: About $36 million in annualized revenue by mid-September 2026, with a target of over $100 million before the year ends, as the Journal reported.

Rolnick also wrote that "$90 million is modest in modern fundraising." Next to the multi-billion rounds we see each month, he has a point.

What Is Hermes Agent and Why Is It So Popular?

Nous released Hermes Agent under the MIT license in February 2026. That license lets anyone use, change and sell software built on it, with very few limits.

The pitch is simple. The agent learns as you use it. Nous says it remembers your projects, writes reusable "skills" while it works, and carries that knowledge into later sessions.

According to the Hermes website, the agent can:

Live in your chat apps: Telegram, Discord, Slack, WhatsApp, Signal, email and the command line, all sharing one memory.
Run tasks on a schedule: Reports, backups and briefings can run unattended.
Hand work to subagents: Each helper gets its own conversation and terminal.
Browse and create: Web search, browser automation, vision, image generation and text-to-speech are built in.
Stay sandboxed: It supports five backends, including local, Docker and SSH, with container hardening.

Adoption has been fast. Nous says Hermes has been cloned more than 24 million times. By its own internal estimate, it drives about 2.5% of all global AI token usage. The Journal counted more than 22 million downloads since February.

Keep in mind these are the company's own numbers. Clones and downloads also do not equal daily active business users.

How Hermes Agent for Business Works

The new money will fund what Nous calls Hermes for Businesses. Its product page splits the offer into two tiers: Business and Enterprise.

Hermes Business banner from Nous Research with three illustrated figures in space suits on a blue background — The Hermes Business product banner. Image: Nous Research.

Feature	Hermes Business	Hermes Enterprise
Where it runs	Nous infrastructure	Your own on-prem or private cloud
Data storage	Isolated tenant per team	Stays on your systems
Setup	Self-serve	Tailored rollout and onboarding
Login and support	Email invites, admin roles	SSO and SLAs
Billing	Central balance with per-member caps	Contact sales

A few features stand out for teams:

Shared skills library: When one person's agent learns a useful skill, it gets published to a team library. Everyone's agent can then use it.
Spend control: Admins set spending caps for each member and can see every dollar in the portal.
Model choice: Teams can use any model on the Nous inference API or plug in their own provider keys.
Team Gateway: Agents connect to Discord, Telegram, Slack and WhatsApp, plus desktop and terminal apps.

What Hermes Agent Costs Today

The core agent is free. You pay for the models and any hosted services it uses. Nous Portal sells monthly plans with credits for Hermes Agent and access to over 200 models.

Bar chart comparing Nous Portal plan prices for Hermes Agent with monthly credits for Free, Plus, Super and Ultra tiers — Nous Portal plans for Hermes Agent, monthly price against monthly credits. Data: Hermes Agent website. Chart: Debugged Pro.

The paid tiers are Plus at $20, Super at $100 and Ultra at $200 a month. Each adds a 10% bonus in credits. The business page does not list a per-seat price. Instead, a central team balance funds usage.

Why Hermes Agent Matters for Agencies and IT Teams

Personal AI agents have pulled in huge sums this year. The Next Web pointed to Meta's Muse agent and Nvidia's enterprise platform built on the open-source OpenClaw agent. We covered that second one in our OpenClaw enterprise guide for agency teams.

Hermes adds a strong open-source option to that list. Here is why that matters for service businesses.

1. Less Lock-In for Clients

Many clients worry about tying their whole workflow to one AI vendor. With Hermes, the agent code is open and the model is your choice. If prices change, you can switch models without rebuilding everything.

2. Team Knowledge That Builds Over Time

Agencies repeat a lot of work. Think client onboarding, QA checklists, deployment steps and weekly reports. A shared skills library means the best way to do a task can spread across the team, instead of living in one person's head.

3. Clearer AI Budgets

Uncapped AI spending is a real worry for small firms. Per-member caps and one central balance make costs easier to plan. That helps Indian startups watching every rupee and US teams with larger headcounts alike.

4. A New Service to Offer

Open-source tools still need someone to set them up well. Clients will want help with hosting, security, custom skills and connecting the agent to their own systems. That is the kind of work our AI agent development services already cover.

Risks to Check Before You Deploy Hermes Agent

Open source does not mean risk-free. Before you give any agent access to real work, ask these questions:

Where does the data live? Business runs on Nous servers. If your client has strict data rules, Enterprise or a self-hosted setup may be the only fit.
What can the agent touch? Hermes can use a terminal, browse the web and run scheduled jobs. Limit its access to what each task needs.
Who reviews shared skills? Auto-sharing is useful, but a bad or unsafe skill can spread just as fast as a good one.
Are the numbers proven? Adoption figures are self-reported, and the business product is brand new. Expect rough edges.

For a solid starting point on rules and controls, the NIST AI Risk Management Framework is free and practical. Our post on AI agent governance for growing teams also covers simple guardrails.

How to Pilot Hermes Agent in Your Team

Do not roll an agent out to the whole company on day one. Here is the short plan we would follow:

Pick two or three internal tasks. Good starters are meeting summaries, daily stand-up notes and simple research.
Start small. Invite three to five people and set low spending caps for each.
Use a sandbox. Run the agent in Docker or another isolated backend, not on a laptop with full access.
Compare models. Try a cheap model and a stronger one on the same task. Track cost, speed and accuracy.
Review the skills library weekly. Remove anything that looks wrong or touches data it should not.
Write down the rules. Decide what the agent may never do, such as sending client emails without a human check.
Expand slowly. Move to client-facing work only after a few weeks of clean results.

Our Take on Hermes Agent for Business

This raise shows that investors think open-source agents can win inside companies, not just with hobbyists. Nous has real usage and a clear message about control and choice.

The business product is still early, though. Teams should treat it as a pilot, not a full rollout. If it holds up, an open agent that your team can shape and own is a strong option to keep on the table.

Key Takeaways

Nous Research raised $90 million at a reported $1.5 billion valuation on October 7, 2026.
Hermes Agent is free and open source under the MIT license, and Nous says it has been cloned over 24 million times.
Hermes for Businesses adds shared skills, per-member spend caps and team-wide model choice.
Business runs on Nous infrastructure, while Enterprise runs on servers you control with SSO and SLAs.
Adoption numbers are self-reported, so test it yourself before relying on it.
Agencies should pilot it on internal tasks first, with sandboxing and clear rules.

Frequently Asked Questions

What is Hermes Agent?

Hermes Agent is a free, open-source AI agent built by Nous Research and released under the MIT license in February 2026. It remembers your projects, writes reusable skills as it works and lets you choose which AI models it runs on. It works on macOS, Windows and Linux, and through apps like Slack, Telegram and WhatsApp.

How much money did Nous Research raise?

Nous Research raised $90 million in a Series B round. The Wall Street Journal reported a $1.5 billion valuation, and TechCrunch says Nous confirmed it. TechCrunch also reports the round was led by Robot Ventures and brings total funding to $158 million.

What is Hermes for Businesses?

It is a shared Hermes deployment for a whole team. Each member gets a hosted agent, skills spread across the company through a team library, and one central balance pays for usage with spending caps per person. It runs on Nous infrastructure in an isolated tenant for each team.

Is Hermes Agent free to use?

The agent itself is free and open source. You still pay for the AI models it uses. You can bring your own provider keys or buy credits through Nous Portal, where paid plans start at $20 a month. Business accounts are funded from a central team balance.

Should agencies use Hermes Agent for client work?

It is worth a pilot, but start with internal tasks first. Check where data is stored, limit what the agent can access, and test it on a few repeatable jobs. Move to client data only after you have clear rules, logs and spending limits in place.

About the Author

Nitin Kumar is the Lead Project Coordinator at Debugged Pro. He works with clients in India and the USA on web, mobile and AI projects, and helps teams decide which AI tools fit their work and budget. He follows new agent platforms closely so clients can adopt them safely.