Nvidia Agrees to Buy Hugging Face for $12.93 Billion, Promises the AI Hub Will Stay Open
Nvidia has agreed to acquire Hugging Face for $12.93 billion, extending the chipmaker deeper into the software and open-model ecosystem while pledging that the platform will remain open, multi-cloud and compute-agnostic.
Why it matters: The deal could reshape how models, datasets and AI applications are distributed across competing clouds and accelerators, making platform neutrality and vendor dependency material infrastructure questions.
Nvidia has agreed to acquire Hugging Face for $12.93 billion, a deal that would bring one of the AI industry’s largest model, dataset and application-sharing platforms under the ownership of the leading supplier of AI accelerators.
Nvidia announced the agreement on September 3. The company said Hugging Face will remain an open platform and that developers will continue to be able to choose their models, frameworks, cloud providers, inference services and computing platforms without being required to use Nvidia hardware.
That promise matters because Hugging Face has become a central distribution layer for open and open-weight AI. Nvidia says more than 18 million developers, researchers and creators use the platform, which hosts more than 3 million models, 500,000 datasets and 1 million applications. The company also says more than 200,000 businesses use Hugging Face to discover, evaluate, customize or deploy AI.
Independent reporting confirmed the acquisition and its price. The Associated Press reported that Nvidia is buying the platform for $12.93 billion and that the company intends to preserve Hugging Face as an open ecosystem. Reuters reported that the transaction includes roughly $11.9 billion for Hugging Face shareholders plus an equity-retention program of up to about $1 billion for employees joining Nvidia.
The transaction expands Nvidia’s role well beyond chips. The company already sells GPUs, CPUs, networking, systems and AI software, and it has increasingly invested in cloud infrastructure, open models and developer tooling. Hugging Face adds a community and distribution layer where models are shared, tested and deployed across many different hardware and cloud environments.
For developers, the near-term question is whether Hugging Face’s existing neutrality can be preserved in practice as ownership changes. Nvidia’s announcement explicitly says its compute will not be required, and it says the platform will continue to support open-source and open-weight models as well as multi-cloud and multi-accelerator deployment.
For enterprise buyers, the deal also illustrates how the AI stack is consolidating. Hardware, networking, model distribution, evaluation, inference software and deployment tooling are becoming more tightly connected. That can lower integration friction, but it can also increase strategic dependency on a smaller set of vendors.
The most consequential test will therefore be governance rather than branding: whether model discovery, ranking, hosting, evaluation and deployment remain equally accessible across competing chips, clouds and frameworks after the deal closes. Nvidia’s public commitment sets a clear expectation. Developers and enterprise customers will now watch whether platform behavior matches it.