Nous Research Raises $90 Million for Open-Source Enterprise AI

Nous Research has raised $90 million to build an open-source AI assistant for enterprise customers, the Wall Street Journal reports.

The round, disclosed in a WSJ exclusive, positions the startup as a challenger to closed-model providers such as OpenAI and Anthropic by betting that corporate buyers want AI systems whose underlying models they can inspect, host and customize themselves.

The $90 million figure is the hardest number in the deal. The WSJ did not report an updated company valuation in its headline disclosure, and the names of the participating investors were not included in the syndicated summary of the report.

Why does the open-source angle matter to enterprises?

Nous Research has built its reputation on releasing open-weight models — AI systems whose parameters are published and can be downloaded, modified and run on a customer's own infrastructure. That stands in contrast to the API-only approach of the largest commercial AI vendors.

For enterprise buyers, the pitch is control. Open-source assistants can be:

• deployed inside private cloud environments;
• audited for safety and compliance;
• fine-tuned on proprietary company data without that data leaving the perimeter.

The new funding is earmarked, according to the WSJ report, for bringing an AI assistant built on this open approach to enterprises — moving the company from research and model releases into commercial product territory.

Who is Nous Research?

Nous Research is a US-based AI startup known in the developer community for its open-weight model releases. Its work has circulated widely among engineers who want capable models without the licensing constraints of closed systems.

The company now joins a cohort of open-source AI players — including Mistral AI in Europe and Meta with its Llama model family — that are competing for enterprise workloads against closed-model leaders.

What does a $90 million round signal?

The size of the round signals investor appetite for the open-source enterprise AI thesis. Assistants — products that perform tasks for workers rather than simply answering questions — have become the commercial battleground in enterprise AI, with Microsoft, Google and a wave of startups all shipping copilots and agents.

An open-source assistant would differentiate on procurement terms, data governance and customization. Those are the three criteria most often cited by large organizations that restrict their AI deployments to approved vendors.

What comes next?

The WSJ report did not specify a launch timeline for the enterprise assistant, the pricing model, or the identity of pilot customers. The open questions for Nous Research are whether an open-weight assistant can match the performance of frontier closed models on agentic tasks, and whether the company can build the enterprise sales and support infrastructure that the $90 million is presumably meant to fund.

The round gives Nous Research the capital to make that transition from research lab to enterprise vendor. Whether corporate buyers reward openness with contracts is the test the funding now sets up.