Nous Research, the startup behind the open-source Hermes artificial-intelligence agent, confirmed on October 7, 2026, that it has raised $90 million in a Series B at a $1.5 billion valuation, according to TechCrunch, which had earlier reported the deal. The Wall Street Journal also covered the round in detail. Robot Ventures led the financing; participants included Nvidia, Union Square Ventures, Menlo Ventures, Samsung, 1789 Capital — where Donald Trump Jr. is a partner — and others, including Y Combinator according to SiliconANGLE’s account of the syndicate.
The new financing brings the three-year-old company’s total funding to about $158 million, according to TechCrunch. Nous will use the capital to accelerate its push into the enterprise sector with “Hermes for Businesses,” which lets companies deploy customized agents that can handle multi-step workflows while keeping data private and secure.
Hermes, released in February under an open-source license, has become a hit among developers and individual users. The company estimates the agent has been cloned or downloaded more than 24 million times and accounts for roughly 2.5% of global AI token usage, according to TechCrunch and SiliconANGLE. Thousands of open-source contributors work on the project, SiliconANGLE reports.
The product ships as a desktop app that can interact with the user’s files, organize documents, use a browser, run code in Docker containers and chain tasks together. Under the hood, Hermes uses popular large language models such as Claude as well as a lineup of models fine-tuned by Nous from open-source bases. The latest collection is called Hermes 4; the company says it customized the models with a dataset of about 5 million synthetically generated records created with an in-house tool called DataForge, of which roughly two-thirds were aimed at improving reasoning.
A router module sends each prompt to the model best suited by inference cost, speed and output quality. Another module personalizes replies based on user preferences; optionally, Hermes can connect to external databases such as RetainDB to retain preferences and project data across sessions, according to SiliconANGLE. That agent architecture with routing and external memory is the foundation on which Nous is now building business editions.
Monetization is structured in three paid editions, according to SiliconANGLE: Hermes Cloud, for consumers who want to launch agents in seconds without standing up infrastructure; Hermes Business, with team accounts, shared balances, spend caps and shared skills; and Hermes Enterprise, aimed at on-premises or private-cloud deployments with single sign-on (SSO) and service-level agreements. TechCrunch summarizes the enterprise push under the “Hermes for Businesses” brand.
On the financial side, The Wall Street Journal reported that Nous was at roughly $36 million in annualized revenue by mid-September 2026 and expects to pass $100 million before year-end. Those projections come from the Journal’s reporting and the company itself; they are not audited results published in a regulatory filing. Moving from $36 million to more than $100 million in a few months would imply a sharp acceleration tied to enterprise adoption.
Nvidia’s and Samsung’s presence in the round fits chipmakers’ and electronics firms’ interest in agent ecosystems that consume inference at scale. Union Square Ventures and Menlo Ventures bring track records in developer and enterprise software. 1789 Capital adds a visible political-investor profile. Robot Ventures, as lead, reinforces the thesis that autonomous agents — not only single-turn chatbots — are the next product layer.
Hermes’s open-source success poses a classic dilemma: how to turn mass free adoption into recurring revenue without alienating the community. Nous’s strategy is to keep the core open while selling managed cloud, team controls and private deployments. If it works, it could serve as a template for other labs that publish agents or models and seek to monetize the operations, compliance and support layer.
Competition is intense. Desktop and enterprise agents proliferate, from assistants embedded in office suites to development environments with autonomous agents and vertical workflow bots. Hermes differentiates itself through its open-source origin, its estimated token share and its own model family. But the token-share figure is a company estimate, not a third-party audited metric, and enterprise retention will depend on security, governance, auditability and total cost of ownership, not only hobbyist popularity or GitHub stars.
Product timing also matters. Hermes 4 and DataForge show that Nous is investing in synthetic data for reasoning, a trend shared by several labs. The router that balances cost and quality can be decisive when companies pay by the token and want to cap bills. Enterprise with SSO and SLAs answers requirements individual users do not demand but IT departments do.
Loose ends remain. There is no published breakdown of how much each investor contributed or of liquidation-preference terms. The $1.5 billion valuation includes, as is usual in such coverage, the new capital in the post-money formulation common in press accounts, though valuation should not be confused with cash on hand or with revenue quality. It is also unclear what share of the $36 million annualized revenue already comes from business customers versus Hermes Cloud consumers, a split that will matter for how investors underwrite the path to $100 million.
Even so, the October 7 confirmation consolidates Nous Research as one of the most visible players in open-source agents with enterprise ambitions. With $90 million fresh, a $1.5 billion valuation and a product that already drives, by the company’s account, a non-trivial share of global token traffic, the test of the coming months will be whether Hermes for Businesses converts developer popularity into durable corporate contracts with measurable retention, security assurances and predictable unit economics.
