What is Hugging Face and why is Nvidia buying it for $12.93 billion?

Nvidia has agreed to acquire artificial intelligence platform Hugging Face for $12.93 billion, its chief executive Jensen Huang announced, in a deal that will bring one of the biggest open-source AI communities under the chipmaker's umbrella.

Huang said Hugging Face would remain an open platform for the AI ecosystem after the acquisition, with developers continuing to choose the models, frameworks, cloud and inference providers and computing platforms they prefer.

"NVIDIA compute will not be required to build on or deploy through Hugging Face," Huang said in a post announcing the deal.

What is Hugging Face?

Founded in 2016 by French entrepreneurs Clément Delangue, Julien Chaumond and Thomas Wolf, Hugging Face has emerged as a major hub for developers and researchers working on AI models and applications.

According to Huang, more than 18 million developers, researchers and creators use the platform to share more than 3 million models, 500,000 datasets and 1 million applications.

More than 200,000 companies use Hugging Face to discover, evaluate, customise and deploy AI.

“Hugging Face will remain an open platform for the entire AI ecosystem. 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. NVIDIA compute will not be required to build on or deploy through Hugging Face,” Huang said.

The Nvidia chief said the chipmaker's infrastructure, engineering capabilities and global reach would help improve Hugging Face's platform reliability, safety, model evaluation, inference and deployment capabilities while preserving its open ecosystem.

Nvidia has itself contributed more than 500 models and over 250 open datasets to Hugging Face, Huang said.

The acquisition comes as the AI industry is seeing a rapid expansion of open-weight models, which allow businesses, universities, startups and public institutions to build applications without training models entirely from scratch.

Huang said open models could broaden access to advanced AI and help distribute AI capabilities across companies, institutions and communities.

"AI advances faster when people can build together," he said.

Hugging Face began with chatbot technology aimed at teenagers but its founders later shifted their focus to the underlying AI models.

The company open-sourced its tools and launched the Transformers library, which became widely used by AI researchers and engineers.

It subsequently launched the Hugging Face Hub for sharing models, its Datasets library and Spaces platform for interactive AI applications. In 2022, it acquired Gradio, an open-source library used to build machine learning applications in Python.

Open AI-Hugging Face incident

Hugging Face was recently involved in a cybersecurity incident during internal evaluations of OpenAI models.

OpenAI said in an account of the incident that several models, operating under reduced safeguards, circumvented controls, exploited vulnerabilities in shared infrastructure and accessed third-party systems, including Hugging Face's systems.

OpenAI said the incident showed that highly capable AI agents could potentially work around technical controls, exploit security weaknesses and take actions that had not been directed by humans.

The company said it was responding by strengthening safeguards around its research infrastructure, increasing isolation and restrictions on internet access and model weights, and investing in monitoring and alignment measures.

Huang said Hugging Face's team would retain its "iconic" brand following the acquisition and work with Nvidia to expand access to open AI models.

"I am honored that Clem came to me as he considered the next chapter of Hugging Face and believed NVIDIA would be a great home for the company, its community and the future of open models," Huang said.