Nvidia confirmed today that it is now the owner of Hugging Face for $12.93 billion.

Hugging Face is home to millions of AI tools. It hosts three million models, has one million applications, and more than 18 million developers use the platform. 

There are half a million datasets on it too.

Nvidia’s Plans

Nvidia CEO Jensen Huang shared the news in a company blog post and tried to calm any fears right away.

Huang said Hugging Face will stay open. Developers can still pick any model they want and can use any framework. 

They can choose any cloud or inference provider, and Nvidia hardware won’t be a requirement. He also pointed out that Nvidia has already put over 500 models on Hugging Face. 

It has also shared 250 open datasets. Huang argues this proves Nvidia’s commitment to open tools, not just closed products.

Acquisition

Nvidia builds the chips that power most AI work today. That gives it huge influence already, but owning Hugging Face takes things further.

An open platform under Nvidia’s control could help sell more of its hardware. It also opens a new revenue path. 

Nvidia can bundle its spare computing power with Hugging Face’s tools. That means selling capacity directly to businesses, all wrapped into one package.

Hugging Face

Hugging Face launched back in 2016. And over the years, it raised more than $395 million in funding.

Its biggest round came in 2023, bringing in $235 million. It was led by Salesforce Ventures, and Google, Amazon, IBM, and Nvidia all chipped in.

Nvidia had tried to buy Hugging Face before. Last year, the company turned down a $500 million offer from Nvidia, according to the Financial Times.

But because of Hugging Face’s remarkable growth ($150 million in annualized revenue), the offer increased to today’s $12.93 billion acquisition. 

Back in July, CEO Clem Delangue spoke with TechCrunch. He said the platform’s growth was moving it closer to profitability. 

Delangue also posted about the deal on X. He credited the community for proving open-source AI could rival closed, proprietary systems. 

But he said that scaling further needed more resources. Reaching a larger scale needed more compute, more support, collaboration, and visibility. That’s why he turned to Huang.

Open Models

Jensen Huang has spent months pushing the idea that open-weight AI models are important. He even co-signed a letter urging the US to back open models to compete against China.

Nvidia has been backing that vision. Last month, the Wall Street Journal reported a $6 billion deal between Nvidia and coding startup Poolside to build a strong American alternative to Chinese AI tools.

During a recent earnings call, the company revealed it has poured over $50 billion into AI research labs. 

Cybersecurity 

During an earnings call, he explained that open models are becoming central to digital defense.

He described a new wave of security companies that build fast, automated systems that can watch for attacks around the clock. These tools rely heavily on open models to function properly.

That connects directly to Hugging Face’s own experience. Back in July, Delangue said an open Nvidia model helped protect Hugging Face from an attack. 

Also read: OpenAI Explains the Hugging Face Breach

He said closed, proprietary models failed to catch the threat in time.