Nvidia has confirmed the Hugging Face acquisition, a deal worth $12.93 billion. The move ends weeks of speculation about the AI platform’s future. Nvidia CEO Jensen Huang detailed the purchase in a blog post, saying Hugging Face will keep operating as an open platform once the sale closes.
Hugging Face hosts three million models, one million applications, and half a million datasets. More than 18 million developers use the platform. The Hugging Face acquisition gives Nvidia direct ownership of the largest hub for open and open-weight AI models. It is a resource Nvidia had previously relied on only as an outside partner.
What Nvidia says will stay the same after the Hugging Face acquisition
Huang wrote that developers will keep choosing their own models, frameworks, clouds, and inference providers. “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 wrote.
Nvidia has already published more than 500 models and 250 open datasets on Hugging Face. Huang said the company builds open models to get developers around the world using its hardware. That incentive now extends to owning the platform those models run through.
A reversal after an earlier rejection
Hugging Face turned down a $500 million offer from Nvidia last year, according to the Financial Times. Since then, the company has grown its revenue to roughly $150 million on an annualized basis, The Information reported last month. Founder and CEO Clem Delangue discussed the growth in a July interview with TechCrunch. He said the rate is pushing the platform “close to profitability.”
Delangue framed the new deal as more than a sale. In a post on X, he thanked the community for proving an alternative to closed-source APIs could work at scale. He then explained why Hugging Face needed a partner. “But for it to happen at a larger scale, it needs more compute, more support, more collaboration, and more visibility. That’s why we went to talk to Jensen, who offered to do exactly that with us,” Delangue wrote.
Hugging Face was founded in 2016. It had raised more than $395 million before the acquisition, according to Crunchbase. That total includes a $235 million round in 2023, with Salesforce Ventures leading and Google, Amazon, IBM, and Nvidia investing alongside it.
Open models and Nvidia’s broader strategy
Nvidia’s acquisition of Hugging Face fits a pattern. The company struck a $6 billion deal with coding startup Poolside last month to develop open models, according to the Wall Street Journal. On its most recent earnings call, Nvidia said it has put more than $50 billion into AI frontier labs.
Huang has pushed for open-weight models as a matter of national strategy. He co-signed a letter arguing they strengthen the United States’ position against competitors such as China in AI development. On the earnings call, he pointed to cybersecurity as an area where open models matter most. He said they let companies run continuously operating autonomous systems that defend against attacks.
That argument has a direct precedent at Hugging Face itself. Delangue said in July that an open Nvidia model helped the platform defend against a cyberattack, after a proprietary model failed to stop it. Days earlier, OpenAI acknowledged that one of its unreleased models had breached Hugging Face’s systems. That episode, Delangue suggested, is part of why acquiring Hugging Face made sense to Nvidia.