Nvidia Buys Hugging Face for $13B to Strengthen Its Open Weights Strategy
Set Trending Topics as a preferred source on Google.
Nvidia is acquiring Hugging Face. The chipmaker is paying exactly $12,930,300,000 for the platform that hosts a large share of the world’s open AI models. CEO Jensen Huang confirmed the purchase in a blog post, laying out why Nvidia is putting up that sum and what happens to Hugging Face next. A sale in this range had already been reported earlier.
The Case for the Deal: Reach Plus Infrastructure
Huang opens his argument with the sheer size of the platform. More than 18 million developers, researchers and creators use Hugging Face to share over three million models, 500,000 datasets and one million applications, according to his figures. More than 200,000 companies rely on it to discover, evaluate, customize and deploy AI. The company was founded by Clément Delangue, Julien Chaumond and Thomas Wolf.
On the founders and what they built, Huang writes: “Over the past decade, Clem, Julien, Thomas and the team at Hugging Face have built something remarkable: a vibrant home for the open model developer community.”
His reasoning for the acquisition: what Hugging Face lacks is not a community but the infrastructure to serve that community at a far greater scale. “As the opportunity for open models accelerates, Hugging Face can serve the global AI community at unprecedented scale,” Huang writes. Nvidia’s infrastructure, engineering and global reach, he adds, “can help improve platform reliability, safety, model evaluation, inference and deployment capabilities, while preserving the open ecosystem that made Hugging Face foundational.”
Huang also frames the deal as the seller’s initiative: “I am honored that Clem came to me as he considered the next chapter of Hugging Face and believed NVIDIA would be a great home for the company, its community and the future of open models.”
How It Feeds Nvidia’s Open Weights Strategy
The purchase fits a position Nvidia has been building for months. Huang recently co-signed an open letter on the importance of open weights to the AI economy, alongside representatives of numerous other companies. He sums up its core claim this way: “open weights broaden access to AI and help ensure that AI leadership is distributed across companies, institutions and communities.” Open models, he argues, let startups, businesses, universities and public institutions “build on advanced capabilities without training every model from scratch,” carrying the technology into “factories, hospitals, farms, classrooms and Main Street businesses around the world.” His shortest version of the argument: “AI advances faster when people can build together.” The dispute over that question runs straight through the industry, with some notably contradictory positions among individual players.
In his blog post, Huang lists what Nvidia has contributed so far: more than 500 published models and more than 250 open datasets on Hugging Face, which by the company’s own account makes it the largest contributor of open models and data on the platform. That sits alongside an announced investment of $26 billion over five years in developing open models.
The strategic backdrop: Nvidia’s market position rests not on hardware alone but on CUDA, a software ecosystem built over two decades. Open models that are optimized for Nvidia hardware and shipped through the industry’s largest distribution channel extend that lead into a period when alternatives such as AMD’s ROCm and hardware-agnostic compilers are gaining ground. With Hugging Face, Nvidia is buying that distribution channel outright.
What Happens to Hugging Face Now
Huang makes several concrete commitments in his post. “Hugging Face will remain an open platform for the entire AI ecosystem,” he writes. “Developers will choose the models they want, the frameworks they want, the clouds and inference service providers they want and the computing platforms they want.” And explicitly: “NVIDIA compute will not be required to build on or deploy through Hugging Face.” Support for open source and open weight models from every provider is set to continue, as are multi-cloud and multi-accelerator deployments.
The team stays on and the brand stays intact: the Hugging Face crew will “bring their passion and expertise to a much larger canvas, with their same iconic brand,” Huang writes.
Formally, the deal is not done yet. An acquisition of this size by the currently dominant AI chip company is likely to face antitrust review, and parts of the open source community are already debating how neutral a platform can stay once a hardware maker owns it.
Nvidia was already an investor in Hugging Face before the acquisition. In its last major funding round about three years ago, the company was valued at $4.5 billion, with Salesforce, Google, Amazon, AMD, Intel, Qualcomm and IBM among the backers joining Nvidia.
