Nvidia is buying Hugging Face. The chipmaker announced today it has agreed to acquire the open-source AI platform for $12.93 billion, one of the biggest acquisitions in Nvidia’s history and a major move deeper into the software layer of AI.

Hugging Face has spent the past decade building what’s become the default home for the open model developer community. More than 18 million developers, researchers, and creators use the platform to share and work with AI. It currently hosts more than 3 million models, 500,000 datasets, and 1 million applications, and over 200,000 companies use it to discover, evaluate, customize, and deploy AI systems. For a lot of builders, it’s basically the GitHub of AI models.

Nvidia CEO Jensen Huang confirmed the deal in a blog post Thursday, framing it as an extension of the platform rather than a takeover of it. “Together, we will scale Hugging Face’s platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide,” Huang wrote. He was direct on the point that’s likely to matter most to Hugging Face’s existing community. Nvidia compute will not be required to build on or deploy through the platform, and Hugging Face will keep supporting open source and open weight models from every model builder, across multiple clouds and multiple hardware accelerators, not just Nvidia’s own.

The relationship between the two companies goes back further than this week. Nvidia already held a stake in Hugging Face, having joined a 2023 funding round alongside Salesforce and Google that valued the company at $4.5 billion. Earlier this year, Hugging Face reportedly turned down a separate $500 million investment offer from Nvidia at a $7 billion valuation, before ultimately agreeing to this full acquisition at close to double that number.

The price is a steep jump given the numbers involved. Hugging Face’s annualized revenue sits at around $150 million, which makes the $12.9 billion price tag a significant multiple on current earnings. Part of the strategic logic seems tied to where the rest of the AI industry is heading. Closed source labs including Anthropic and OpenAI are actively developing their own proprietary chips, which could eventually reduce how dependent they are on Nvidia’s GPUs. Owning one of the most widely used distribution points for open models gives Nvidia a foothold deeper into the AI stack that doesn’t rely purely on chip sales.

This is Nvidia’s second largest acquisition on record, behind only its roughly $20 billion purchase of Groq’s assets late last year. It also lands in the same week as another notable consolidation move in AI infrastructure, Stripe’s approximately $7 billion acquisition of OpenRouter, the platform developers use to route requests across different AI models. Taken together, the two deals point to a broader trend of larger tech companies buying up the infrastructure layer that sits between AI models and the developers building on top of them.

The announcement also comes about a month after Hugging Face experienced a security breach, which the company’s CEO Clem Delangue has said was caused by engineering mistakes rather than a flaw in open models themselves. Delangue told CNBC the incident actually reinforced his belief in open source AI, saying Hugging Face needs to “double down” on it going forward.

For now, the deal has been agreed but not yet closed, and the usual regulatory steps still apply before it’s finalized.

What is Hugging Face?
Hugging Face is an open-source AI platform where developers share and access machine learning models, datasets, and applications. It’s widely used across the AI industry as a central hub for open source AI work.

Will Hugging Face still be open source after the Nvidia acquisition?
According to Nvidia, yes. The company has said Hugging Face will remain an open, multi-cloud platform, and that Nvidia compute won’t be required to use it.