Imagine if the company that made the printing presses also bought the biggest public library, promised to keep letting anyone check out books, and pinky-swore the shelves would still stock rival publishers. That's roughly the pitch Nvidia is making with its latest acquisition.
$12.9 Billion for the Model Store
Nvidia has agreed to acquire Hugging Face, the open-source hub that hosts roughly three million AI models, half a million datasets, and a million applications used by more than 18 million developers, in a deal reported around $12.9 billion. The structure includes about $11.9 billion payable to Hugging Face stockholders plus up to $1 billion in retention equity for employees joining Nvidia.
More than 200,000 companies use the platform to find, evaluate, and deploy AI models, which makes Hugging Face less a single product and more the connective tissue of the entire open AI ecosystem. The deal is expected to close in the first half of 2027, pending regulatory approval, and ranks as Nvidia's second-biggest acquisition on record.
Openness, Now Brought to You By the Company That Sells the Chips
Nvidia says it's committed to keeping Hugging Face open — letting developers upload and download whatever models they want, and continuing to support rival silicon vendors like AMD. That's a meaningful promise to make publicly, which also means it's one regulators and the open-source community will be watching closely.
The real tension: Nvidia already dominates the hardware AI runs on. Owning the place where most of the world discovers and downloads the models that run on that hardware is either a savvy vertical bet or exactly the kind of chokepoint antitrust lawyers get out of bed for. Probably both.
Whether "open" survives an acquisition this size is the only chart that matters here — everything else is just a very large number with a dollar sign in front of it.
If your business is trying to figure out which AI models and tools are actually worth building on top of, that's exactly the kind of untangling we help clients with — let's talk.
Source: TechCrunch