The deal, announced on Thursday (September 3), values Hugging Face at almost three times the USD 4.5 billion valuation it achieved in 2023.
Nvidia will pay about USD 11.9 billion in cash and other consideration to Hugging Face shareholders, while a further USD 1 billion equity-based retention programme is planned for employees joining Nvidia.
The acquisition is subject to regulatory approvals.
For Nvidia, the acquisition is about much more than adding another software business. Hugging Face has become one of the world’s central meeting places for developers building, sharing, testing, and deploying open-source and open-weight AI models.
Hugging Face says well over 18 million developers, researchers, and creators use its platform. Its repository contains more than three million models, 500,000 datasets, and one million AI applications, while more than 200,000 companies use the service to discover, evaluate, customise and deploy AI.
The importance of that ecosystem is rising rapidly as companies seek alternatives to expensive, proprietary AI systems. Open-weight models can be adapted for specific applications, run on private infrastructure, and deployed across different computing environments.
Nvidia Chief Executive Jensen Huang said Hugging Face would remain an open platform for the wider AI ecosystem. Developers will continue to be able to choose their models, frameworks, cloud and inference providers, and computing platforms, while Nvidia’s hardware will not be mandatory for building or deploying through the service.
Observers will closely monitor that pledge. Hugging Face’s value has been built partly on its neutrality: its platform supports models and tools that can run across hardware from Nvidia and its rivals. Any perception that Nvidia could favour its own chips, software, or services could ultimately undermine the independence that helped make Hugging Face a critical hub for open AI.
The acquisition also reflects a changing competitive landscape for Nvidia. The company remains the dominant supplier of high-end AI accelerators, but some of its largest customers are developing their own chips. Microsoft, Google, Amazon, and other technology groups are investing heavily in custom silicon, while AI developers such as OpenAI and Anthropic are building increasingly proprietary technology stacks.
Nvidia’s ownership of Hugging Face provides it with another crucial avenue to shape the development of AI workloads and determine their deployment locations.
As developers experiment with models on the platform, Nvidia can potentially make its computing infrastructure easier to use across training, fine-tuning, and inference.
That helps explain the price. Hugging Face was reportedly generating annualised revenue of about USD 150 million before the acquisition, meaning Nvidia is paying a very high multiple of sales. But the chipmaker is buying an established global ecosystem at a time when the economics of AI are shifting from simply training giant models to deploying millions of specialised applications.
The deal also marks a remarkable change in Hugging Face’s trajectory. Founded in 2016 by Clément Delangue, Julien Chaumond, and Thomas Wolf, the company initially became known for its open-source machine-learning tools before expanding rapidly into models, datasets, applications, and cloud services.
Nvidia participated in its USD 235 million funding round in 2023, when Hugging Face was valued at USD 4.5 billion.
The two companies have had a complicated courtship. Hugging Face previously rejected a USD 500 million Nvidia investment that would have valued it at about USD 7 billion, amid concerns that taking such a large investment from a dominant chipmaker could compromise its independence.
For Nvidia investors, the transaction is a bet that controlling an important layer of AI software will reinforce demand for its hardware and reduce dependence on a handful of hyperscale customers.
Regulatory scrutiny is likely to be another consideration. Nvidia is already facing intense attention over its position in AI computing, and adding a major developer platform could raise questions about competition and access. Nvidia’s commitment to keep Hugging Face open and support other silicon vendors is therefore likely to be central to the regulatory case.
The acquisition signals that the AI arms race is moving beyond chips and foundation models. Control over the communities, tools, and distribution channels through which developers build AI is becoming increasingly strategically valuable.
With Hugging Face, Nvidia is buying a powerful global gateway into that world—and betting that the company that supplies the computing will also have a say in how the next generation of AI software is created today.