Nvidia Buys Hugging Face for $12.9B, Reshaping the Open-Source AI Stack
The acquisition merges the dominant AI hardware vendor with the primary distribution channel for open-source models, creating a vertically integrated stack that could reshape how models are shared, optimized, and deployed.
Nvidia is buying Hugging Face in a $12.9 billion acquisition, a price that nearly triples the company's valuation from its last funding round in 2023. Hugging Face runs the dominant repository for open-source machine learning models, datasets, and the transformers library that underpins much of modern NLP and generative AI work.
The deal puts the largest open-source model distribution platform under the control of the dominant AI hardware supplier. Hugging Face's community hosts over 500,000 models and serves as the default distribution channel for open-weight releases from Meta, Mistral, Stability AI, and countless research labs.
For the broader ecosystem, the acquisition concentrates two critical layers of the AI stack — compute hardware and model distribution — under one roof. Whether Nvidia maintains Hugging Face's platform neutrality or steers it toward its own hardware and software stack becomes the central question.
Vertical integration of hardware and model distribution gives Nvidia an end-to-end pipeline — from GPU to model hub — that no competitor can currently match.
If Nvidia optimizes Hugging Face's library and hosted inference around CUDA and its own stack, alternative hardware backends could face a widening compatibility gap.
The open-source AI community now depends on a platform owned by a company with strong commercial incentives to steer defaults toward its own products.