What happened: Nvidia confirmed it will acquire Hugging Face for $12.93 billion. CEO Jensen Huang said the site stays an open platform and that Nvidia chips will not be required to build or deploy on it. Reuters split the figure as about $11.9 billion to investors plus up to $1 billion in employee retention equity.
Huang’s post: more than 18 million developers, more than 3 million models, 500,000 datasets, and 1 million applications. Nvidia already lists more than 500 models and 250 datasets there. Context for builders: the AI content stack.
Key takeaways
- Price is $12.93 billion, confirmed September 3 on the Nvidia blog.
- Huang pledged multi-cloud, multi-accelerator support and no chip mandate.
- Hugging Face keeps hosting open-weight models from other labs.
What Nvidia actually bought in the Hub
The asset is distribution: the default shelf where open models get discovered. Owning that shelf does not automatically lock inference to one GPU vendor, which is why the “chips not required” line is in the first paragraph of the announcement.
Cofounder Clément Delangue called open-source AI an inflection that now needs more compute and visibility. That is the strategic bet. Whether the Hub stays credibly multi-vendor is the test after close, not the press language.
What to watch next
Watch the close timeline and any regulator review. Watch whether non-Nvidia models keep featured placement. Watch if Hub inference defaults start to prefer Nvidia runtimes. Primary source: Nvidia’s acquisition post.