Nous Research, the startup behind the open-source Hermes agent, is reportedly close to securing a new financing round that would value the company at $1.5 billion. TechCrunch, citing three sources with knowledge of the deal, reported that the round is expected to bring in at least $75 million, led by Robot Ventures with significant participation from USV and other investors.

If completed on those terms, the deal would mark a sharp step-up for a young company founded in 2023 and would underline how quickly investor appetite has shifted toward AI agents, especially products that blend open-source distribution with paid hosted services. Nous Research declined to comment to TechCrunch, and the reported investor participants did not respond to that outlet, so the financing remains unconfirmed by the company.

A fast valuation jump around Hermes

According to TechCrunch, Nous Research had previously raised a total of $70 million from investors including Paradigm, Robot Ventures, North Island Ventures, OSS Capital, and Balaji Srinivasan, citing Crunchbase data. A new round of at least $75 million at a $1.5 billion valuation would therefore represent not just fresh capital, but a strong market signal that investors believe Hermes can become more than a popular open-source project.

That matters because AI infrastructure funding has often flowed first to model builders and hyperscale platforms. The reported Nous Research round suggests that agent-layer startups are also commanding large private-market valuations when they can show developer traction and a credible path to recurring revenue.

The company’s pitch appears to sit at the intersection of two current investor themes: open-source AI and autonomous or semi-autonomous software agents. Hermes is positioned as both a developer-friendly system that can run locally and a commercial product that can be consumed as a hosted service. That dual model gives investors multiple stories to underwrite: community adoption, enterprise deployment flexibility, and subscription revenue.

What Hermes is and why it stands out

TechCrunch describes Hermes as a competitor to OpenClaw, another agent product that drew attention for running locally on a PC and performing tasks on a user’s behalf. In Nous Research’s case, one differentiator cited by TechCrunch is that Hermes launched with built-in skills including web search, coding, and image understanding.

TechCrunch also reported that Hermes was designed to learn from usage and expand its skill set without manual intervention. That is an important claim because it points to a more adaptive agent framework rather than a fixed collection of prompts and workflows. However, the available reporting does not include independent technical validation of how reliably that learning process works in production, under what safeguards, or with what performance tradeoffs.

What is clearer from the coverage is the product direction. Hermes can be run on a desktop or a virtual private server, according to TechCrunch, and users can interact with agents directly or receive messages from them through apps such as Telegram and Discord. That architecture fits a growing segment of agent tooling aimed at persistent, always-on automation rather than one-off chatbot interactions.

Nous Research has also released language models focused on coding and math, according to TechCrunch. That broader model work helps explain why investors may see the company as more than a single agent app. It suggests Nous Research is trying to build both the underlying model expertise and the user-facing orchestration layer needed to support autonomous software.

Open source distribution, paid hosting, and the business model question

One reason Hermes appears to be drawing attention is the scale of its open-source footprint. TechCrunch reported that the project has roughly 214,000 GitHub stars and nearly 40,000 forks. Those are substantial numbers for any developer product and especially notable for a company founded only a few years ago.

Still, GitHub stars and forks are imperfect proxies for durable usage. They can signal enthusiasm, experimentation, and community visibility, but they do not necessarily translate into active deployments, enterprise contracts, or long-term monetization. In this case, the reported traction is useful as an adoption signal, but not as proof of commercial performance.

TechCrunch also noted that Nous Research offers a cloud-hosted version of Hermes, with paid tiers ranging from $20 to $200 per month. That detail is central to the funding story. Open-source agent projects often struggle to convert popularity into revenue, especially when self-hosting is part of the value proposition. By offering a hosted product, Nous Research can target users who want the capabilities of Hermes without managing infrastructure themselves.

That hosted-versus-self-hosted split is increasingly important in enterprise AI. Some teams want local control for security, cost visibility, or customization. Others want convenience, vendor support, and faster deployment. A company like Nous Research can potentially serve both ends of that spectrum if it can keep the open-source community engaged while building enough product value on top to justify paid plans.

Evidence, claims, and what remains unverified

The strongest factual reporting in this story comes from TechCrunch, which attributes the financing details to three unnamed sources familiar with the deal. AI Insider and Crypto Briefing appear to echo the same core claim through syndicated or derivative coverage, but the provided evidence from those sources does not add new reported facts.

Several points should therefore be treated carefully.

First, the financing is still reported as being in talks or nearing completion, not formally announced. The exact round size, valuation, and investor composition could still change before closing.

Second, the reported product differentiation around Hermes’ built-in skills and automatic learning comes from media description of the product rather than an independently published benchmark or audit. Without additional technical documentation in the provided evidence, it is not possible to verify how those capabilities perform across real workloads.

Third, the GitHub metrics cited by TechCrunch are meaningful community indicators, but they are not the same as customer counts or enterprise penetration. Likewise, the existence of pricing tiers for the hosted version shows a monetization path, but the available reporting does not disclose revenue, conversion rates, or retention.

Finally, the investor interest itself is also reported through sources rather than public statements. TechCrunch says Nous Research fielded a high level of interest from investors. That may well be accurate, but it remains sourced reporting, not a company-confirmed metric.

Why this matters for AI builders and enterprise buyers

For AI builders, the reported Nous Research round is another sign that the market is rewarding teams that can package AI agents as usable systems, not just model demos. The feature set described for Hermes — local execution, hosted deployment, built-in skills, integrations with Telegram and Discord, and support for coding workflows — reflects what product teams increasingly care about: orchestration, persistence, usability, and control.

For founders, the Nous Research story highlights a playbook that is becoming more common in enterprise AI and open-source AI: use GitHub distribution to build mindshare, then layer on a managed service for teams that do not want to handle deployment. That can work particularly well in categories like AI agents, where runtime reliability, background execution, and tool access matter as much as raw model quality.

For enterprise buyers, the appeal of Hermes is not simply that it is open source. It is that open-source AI can offer optionality. Organizations that need tighter data control may prefer local or private deployment on a virtual private server. Others may choose the hosted option for speed. In both cases, the challenge will be operational maturity: agent reliability, permissioning, observability, and failure handling tend to matter more in production than community popularity.

The financing also adds competitive pressure across the agent market. OpenClaw is mentioned in the reporting as an early comparison point, but the broader implication is that startups offering AI agents now have to prove both adoption and business discipline. Investors may be willing to pay up for growth, yet they will also look for evidence that agent workflows can be sticky enough to support sustainable subscription revenue.

What to watch next

The first signal to watch is whether Nous Research formally announces the round and confirms the valuation, amount raised, and investor list. Until then, the headline numbers remain reported rather than official.

The second is whether the company provides harder commercial data around Hermes, such as active hosted users, enterprise deployments, or usage-based growth. That would help separate broad developer interest from actual business traction.

Third, watch the product roadmap. If the reported new funding is meant to expand Hermes’ product line and business model, the next moves could include stronger enterprise controls, more integrations, better hosted deployment options, or tighter links between Hermes and Nous Research’s coding and math models.

Finally, pay attention to how the company handles the tension between open-source AI and monetization. The most durable companies in this category tend to make self-hosting attractive enough to win trust while keeping the managed product compelling enough to drive revenue.

Creati.ai perspective

The reported Nous Research financing is notable less for the headline valuation than for what it says about where AI investors think value is shifting. Capital is not only chasing frontier models; it is also flowing to teams that can turn model capabilities into agent products developers actually adopt. Hermes appears to have reached that threshold of relevance, at least in community terms.

But this is also where the next phase of scrutiny begins. In AI agents, popularity is easier to demonstrate than dependable automation. For Nous Research, the key question is whether Hermes can convert open-source momentum into repeatable, trusted workflows for paying users. If it can, this round will look like an early bet on a durable platform. If not, it will be another reminder that agent hype and agent businesses are not the same thing.