Nvidia buys Hugging Face for $13 billion AI platform

Nvidia agreed to buy Hugging Face for roughly $13 billion, widening its bet on open-weight AI models and developer tools globally.

A $13 billion platform bet

The acquisition gives Nvidia control of a New York-based startup that has become a central venue for engineers working with open-weight artificial intelligence. Nvidia said the platform is used by more than 18 million developers, researchers and creators.

Hugging Face hosts more than 3 million models, 500,000 datasets and 1 million applications, according to Nvidia. Those figures make the company a distribution point for AI software at a time when Nvidia’s chips remain a core input for training and running advanced systems.

Open-weight models differ from closed systems offered by companies such as OpenAI and Anthropic. Their parameters are available for users to download, adapt and run on their own infrastructure or on third-party hardware, which can lower switching barriers for companies and research institutions.

Huang pledges open access

Nvidia Chief Executive Jensen Huang said in a Thursday statement that Hugging Face would continue operating as an open platform. “Open weights broaden access to AI and help ensure that AI leadership is distributed across companies, institutions and communities,” Huang said.

The company framed the transaction as a way to expand access to advanced AI capabilities without requiring every user to train a model from the ground up. Nvidia said startups, businesses, universities and public institutions can build on open models rather than absorb the full cost of original model development.

That argument places the deal inside a broader contest over how AI systems are built and distributed. Proprietary models concentrate access through providers that control the code and infrastructure, while open-weight systems allow outside users to customize models for local needs, internal data and specific compliance requirements.

Data centers frame deal

The purchase follows a series of Nvidia moves tying its balance sheet more closely to the AI infrastructure buildout. The company previously announced plans to work with major Wall Street investment firms on partial guarantees for as much as $500 billion in data-center financing.

Nvidia also agreed to support a large Ohio data-center project connected to OpenAI, according to the source material. Those financing efforts show how the company is trying to shape both the hardware layer and the software ecosystem that run on top of its chips.

For Hugging Face, the main issue is whether ownership by a dominant chip supplier changes how developers view platform neutrality. If Nvidia preserves broad access, the company gains a larger software channel for its hardware and services; if developers see the platform as tilted toward Nvidia’s stack, rival chipmakers and cloud providers have an opening to promote alternatives.

The sector-level effect depends on whether open-weight models keep gaining enterprise and government use. If adoption holds, Nvidia can use Hugging Face to support a wider market for AI applications, while the broader industry faces more pressure to make models portable across infrastructure; if proprietary models regain momentum, the deal becomes more about developer reach than a shift in AI distribution.