Nous Research, maker of the open-source Hermes AI assistant, has secured $90 million, according to the Wall Street Journal.
Hermes began as Hermes 3, a family of open-weight large language models built by fine-tuning Meta's Llama 3.1, spanning 8 billion, 70 billion and 405 billion parameter versions and aimed at reasoning, roleplay, agentic tasks and function-calling.
The model was trained using Lambda's cloud infrastructure, including an eight-node 1-Click Cluster, and is distributed through Lambda's chat interface and API, a dependency that underscores how closely the open-source project has grown alongside its compute partner.
The new funding is intended to move Hermes beyond its developer and hobbyist base into enterprise deployment, targeting customers that want tighter control over their data and less reliance on a single AI vendor.
The raise values Nous Research at $1.5 billion and arrives amid a broader wave of AI infrastructure funding: the same report notes Lambda itself is raising up to $4 billion at a $14.5 billion pre-money valuation ahead of a planned 2027 initial public offering, while venture firm Disruptive is assembling a $10 billion megafund, having already committed roughly $7.5 billion across about 10 late-stage investments over two years.
