Nvidia reportedly reaches a 12.9 billion dollar deal for Hugging Face
Nvidia and Hugging Face reportedly reached a 12.9 billion dollar agreement, although neither party has announced it. The reported price for the open-model repository is nearly triple its 2023 valuation. The same day, WRTN raised 72.2 million dollars, while Arga Labs secured 10 million dollars to build software twins for enterprise agents.
Artificial Intelligence··Morning
The buyer search turns into a reported Nvidia deal
Nvidia and Hugging Face have reportedly reached a 12.9 billion dollar acquisition agreement, although neither company has announced it publicly. That price is nearly triple the open-model repository's 4.5 billion dollar valuation in 2023. The reported deal follows Hugging Face's rejection two years ago of a 500 million dollar Nvidia investment that would have valued the company at 7 billion dollars. Days earlier, the company was said to be sounding out buyers at 13 billion dollars or more. Nvidia was already an investor alongside Salesforce and Google in Hugging Face's 2023 funding round.[1]
WRTN's overseas revenue supports a larger round
Seoul-based WRTN Technologies raised about 72.2 million dollars in a Series C led by Coreline Ventures and Eugene Asset Management. The round valued the consumer AI company above 1 trillion won, or about 722 million dollars, and brought total funding to roughly 166 million dollars. WRTN runs the entertainment products Crack and OOC alongside its AI portal. The company says OOC passed 10 billion won, or about 7 million dollars, in monthly revenue within three months of launching in North America. It also said overseas sales have overtaken domestic revenue.[2]
Arga builds enterprise software twins for agents
Arga Labs raised a 10 million dollar seed round to build training environments for enterprise AI agents. General Catalyst led the round, with Box Group, Emergence, Gradient and SV Angel participating. The company creates digital twins of business software such as Salesforce, Workday and email clients while preserving their permission systems and webhooks. Those environments can be changed and replayed as often as a training run requires. Co-founder Phillip Li gives the example of a lead created in Salesforce while a colleague reaches the same business through HubSpot, leaving the agent to determine that both entries refer to one company.[3]
