Full story
Nvidia buys Hugging Face
Nvidia agreed to acquire the open-source AI platform Hugging Face for $12.9 billion, a deal the company announced as it moves beyond hardware and further up the AI stack.
“Hugging Face will "remain an open platform for the entire AI ecosystem,"”
In a blog post, Nvidia CEO Jensen Huang wrote that Hugging Face would “remain an open platform for the entire AI ecosystem,” while adding that “Together, we will scale Hugging Face's platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide.”
Hugging Face CEO Clément Delangue told CNBC that his company approached Huang over the summer and that “a few weeks later, here we are,” describing the moment as a turning point for open-source AI.
Delangue said he decided Nvidia was “a perfect home” for Hugging Face, and he said discussions “went quite fast to get a deal done.”
Open models, users, and scale
The acquisition price is framed as a major step up from earlier valuation levels, with the deal described as $12.93 billion and with Hugging Face’s prior valuation cited as $4.5 billion in 2023.
The platform’s footprint is detailed in the reporting as “more than 18 million developers, researchers and creators” using Hugging Face to share “more than 3 million AI models, 500,000 datasets and 1 million applications.”
Nvidia also said “More than 200,000 companies use the platform to discover, evaluate, customize and deploy artificial intelligence applications,” positioning Hugging Face as a collaboration and deployment hub rather than only a model repository.
Huang emphasized that Nvidia would not require Hugging Face users to build or deploy through Nvidia compute, saying “Developers will choose the models they want, the frameworks they want, the clouds and inference service providers they want and the computing platforms they want.”
Cybersecurity and regulatory risk
The deal comes after Hugging Face was “recently at the center of a hacking incident,” which Delangue said raised concerns about the rapid evolution of powerful AI and cybersecurity tools.
“Delangue, a proponent of open-source models, blamed engineering mistakes”
Delangue told CNBC that he blamed “engineering mistakes for the recent attack on Hugging Face,” and he said Nvidia used “an Nvidia version of a Chinese open model to resolve it.”
Nvidia’s SEC filing flagged regulatory risk as a potential obstacle, noting that “other parties are lobbying governments to restrict open-source AI models,” and warning that restrictions on widely used open-source models could affect Hugging Face’s platform and Nvidia’s business.
In the same reporting, Nvidia committed to keeping Hugging Face’s platform open so that model makers, developers, and users can “freely share and access models and datasets,” while continuing to work with chip vendors beyond Nvidia.