Nvidia is reportedly advancing a deal to acquire Hugging Face for EUR 11 billion, a move that could deepen the hardware giant’s integration across the broader AI industry. The acquisition has not been finalized and could still hit obstacles, but both parties appear to be actively pursuing it, according to reporting citing a person familiar with the matter. A separate source confirmed to CNBC that talks are underway.
Founded in 2016, Hugging Face functions as a cloud repository for AI models, playing a role comparable to what GitHub provides for conventional software. Developers and researchers use the platform to search for models that meet specific criteria, download and run them, fine-tune them into new variants, and upload those versions back to the service.
Why Hugging Face Matters to Nvidia
Hugging Face reportedly drew interest from other potential buyers, including Salesforce. Nvidia had previously invested in the company, though it was not alone: Google and Microsoft also backed the startup. Although Hugging Face is reportedly not yet profitable, it has become the de facto destination for storing, downloading, and working on open-weight models, giving it substantial strategic value regardless of its current finances.
Nvidia and several other tech firms have publicly championed open-weight models as alternatives to proprietary closed-frontier systems trained by companies such as OpenAI and Anthropic. Nvidia has released its own family of open models under the Nemotron name. Acquiring Hugging Face would grant Nvidia significant influence over that ecosystem. The company had also previously struggled to launch its own cloud AI business, and this deal could help restart those efforts.
Strategic Stakes in AI Hardware
The acquisition is especially relevant as OpenAI, Anthropic, and other frontier labs have started investing in their own specialized hardware to train and run models, pursuing vertical integration and reducing their reliance on Nvidia. By adding Hugging Face to its portfolio, Nvidia may be positioned to encourage that segment of the market to keep depending on its chips.
While Hugging Face is best known for hosting large language models, it has increasingly invested in models used for robotics and other physical AI applications. Nvidia is already among the largest players in that field, so the deal could open additional opportunities in physical AI down the line. The transaction remains unfinalized as both companies continue negotiations.
Source
Image: arstechnica.com
