Nvidia Acquires Hugging Face to Strengthen AI Open Platform Layout
Nvidia has officially announced the acquisition of open-source AI platform Hugging Face for $12.9 billion, stating that the deal will help expand global developers and organizations’ access to AI.
Nvidia CEO Huang Renxun stated during the announcement of the transaction that “AI advances faster when people can build together. Open models enable startups, enterprises, universities, and public institutions to build on advanced capabilities without the need to train each model from scratch… This is how AI can safely advance, enhance network security and sovereignty, accelerate innovation, and reach factories, hospitals, farms, classrooms, and Main Street enterprises worldwide.”
Hugging Face CEO Clem Delangue stated at a press conference that the goal is to expand the platform’s user base from 18 million to billions in the future, and the partnership with Nvidia will “provide more opportunities to achieve this goal faster”. He added that with the support of Nvidia, Hugging Face can “think about the next decade and truly optimize its impact on the community and the field”.
When asked if he was concerned about antitrust scrutiny, Delangue believed that the acquisition would redistribute power in the AI ecosystem rather than centralize it. He stated that a significant amount of power and control is currently concentrated in the hands of proprietary APIs, placing Nvidia’s strength behind the Hugging Face open-source ecosystem and “allowing for healthy competition”. “We believe that this operation and announcement will enable AI to be more widely distributed globally, avoiding excessive concentration of power.”
Analysts are concerned
Last week’s trading rumors have surfaced, and analysts have different opinions on whether this deal is beneficial to the open-source AI community. Some analysts point out that the deal gives Nvidia control over a key AI distribution channel, which helps provide more resources and support for the open source community, but David Linthicum, founder of Linthicum Research, is concerned about the significant cultural and genetic differences between the two companies. Jason Andersen, Vice President of Moor Insights and Strategy, also questioned how Nvidia establishes business models for third-party AI models and datasets.
Huang Renxun pointed out in the announcement that Nvidia has become the “biggest contributor to open models and data on Hugging Face”, releasing over 500 models and more than 250 open datasets. He said, “Hugging Face will remain an open platform for the entire AI ecosystem. Nvidia’s infrastructure, engineering, and global influence help enhance platform reliability, security, model evaluation, inference, and deployment capabilities, while preserving the open ecosystem that makes Hugging Face the foundation.”
Regarding the question of whether NVIDIA will continue to maintain Hugging Face hardware neutrality, Justin Boitano, Vice President of Enterprise AI at NVIDIA, stated that the path to expanding the Hugging Face developer community relies on trust and providing developers with the freedom to run models wherever they want. “Therefore, we are motivated to remain open and neutral, ensuring that developers can work wherever they want.”