Nvidia has agreed to acquire Hugging Face for $12.93 billion, securing one of the world’s most widely used repositories and collaboration platforms for open artificial intelligence models. The transaction significantly extends Nvidia’s position beyond processors and computing infrastructure into the software, data and developer layers of the AI ecosystem.
The definitive agreement comprises an approximately $11.9 billion purchase price for Hugging Face shareholders, subject to adjustments, and an equity-based retention programme worth up to about $1 billion for employees joining Nvidia. Hugging Face is privately held, and the transaction will be subject to customary closing conditions and regulatory scrutiny.
Nvidia chief executive Jensen Huang said Hugging Face would continue to operate as an open platform supporting multiple cloud providers and hardware accelerators. That commitment is significant because Hugging Face has developed into a central distribution and collaboration layer for models, datasets and AI applications spanning several competing infrastructure environments.
More than 18 million developers, researchers and creators use Hugging Face, alongside more than 200,000 companies. The platform hosts over three million models, 500,000 datasets and one million applications. Nvidia itself has published more than 500 models and 250 open datasets through the service.
The acquisition gives Nvidia ownership of a platform that influences how developers discover, evaluate, customise and deploy open models. Nvidia’s core business remains centred on GPUs and integrated computing systems, but its software portfolio has expanded through CUDA, enterprise AI tools, model frameworks and deployment services. Hugging Face adds a large developer community and a model-neutral distribution channel to that stack.
Indian AI developers also use the platform to distribute multilingual models and datasets. Bengaluru-based Sarvam AI, for example, hosts model families and supporting artefacts on Hugging Face, including systems developed for Indian contexts and 22 Indian languages. Its published model documentation also identifies Nvidia’s NeMo framework and Indian GPU infrastructure among the technologies used for earlier model development.
The announcement converts earlier deal reporting into a confirmed transaction and establishes the commercial terms, employee-retention component and Nvidia’s stated operating approach. It also places a widely used open-model platform inside a company that supplies much of the compute infrastructure used to train and run modern AI systems.
