Nvidia Buys Hugging Face for $13B to Strengthen Its Open Weights Strategy

Nvidia is buying Hugging Face for exactly $12,930,300,000. Jensen Huang confirmed the deal in a blog post, and his pitch is blunt: the platform already has the community, but Nvidia says it can add the muscle behind it.

Hugging Face sits at the center of a lot of open AI work. Huang says more than 18 million developers, researchers and creators use it, with over three million models, 500,000 datasets and one million applications hosted there. He also says more than 200,000 companies depend on the platform to find, test, adapt and ship AI.

The message from Nvidia is that this is not about replacing Hugging Face’s identity. Huang says the company will stay an open platform for the broader AI ecosystem, with developers free to pick the models, frameworks, clouds, inference providers and compute platforms they want. He also says Nvidia compute won’t be required to build on or deploy through Hugging Face, and that support for open source and open weight models from every provider will continue.

That promise matters because Nvidia is not just buying a popular website. It is buying a major distribution channel for open models at a moment when its own software stack, CUDA, still sits at the heart of its business. Huang points to Nvidia’s own contribution to Hugging Face — more than 500 published models and more than 250 open datasets — and to a separate $26 billion, five-year investment in open models. Put together, this is Nvidia trying to make sure open AI runs through infrastructure it understands, on terms it likes.

The deal is not finished. It will need antitrust review, and the open source crowd is already arguing about how neutral a platform can stay once a chip giant owns it. Still, the strategic logic is obvious enough: if open models are where the industry keeps moving, Nvidia would rather own the road than keep renting the toll booth.