Nvidia built its power in artificial intelligence by selling the chips powering nearly the entire race toward frontier models. Now it also wants to control one of the places where those models are distributed, tested, and turned into products. On September 3, the company announced an agreement to acquire Hugging Face for approximately $12.93 billion.
The financial figure is massive, but it tells only part of the story. According to Nvidia's filing with the Securities and Exchange Commission, the transaction includes approximately $11.9 billion for Hugging Face shareholders and up to roughly $1 billion in an equity retention program for employees joining Nvidia. The deal is expected to close in the first half of 2027, subject to regulatory approvals.
Why Hugging Face is worth so much
Over the years, Hugging Face has become a kind of social and technical infrastructure for open AI. Nvidia states that the platform is used by more than 18 million developers, researchers, and creators, hosting over 3 million models, 500,000 datasets, and around one million applications. More than 200,000 companies use it to discover, evaluate, customize, and deploy artificial intelligence systems.
It is difficult to find a perfect equivalent, but the most intuitive comparison is GitHub: not because Hugging Face does the exact same thing, but because it sits at the intersection where a technical community publishes components, builds reputation, experiments, and makes moving from research to product easier.
For Nvidia, this means entering a layer of the value chain where its dominance had so far been less direct. The company controls an enormous share of AI hardware and the software ecosystem tied to CUDA. With Hugging Face, it can get closer to the moment when a developer chooses a model, a dataset, a framework, and a deployment method.
Nvidia promises the platform will remain open
The most delicate question is obvious: what happens to a platform founded as the home of the open ecosystem when it is acquired by the leading AI hardware maker?
Nvidia has pledged that Hugging Face will continue to support diverse models, frameworks, clouds, and inference providers, and that Nvidia hardware will not be required to build or deploy systems via the platform. In its SEC filing, the company also reaffirmed its commitment to allowing support for other chip manufacturers.
It is a strategically necessary promise. The value of Hugging Face relies precisely on its perceived neutrality: if it were to become merely a showcase for the Nvidia ecosystem, it would lose part of the community that made it important.
The battle is no longer just about models
The acquisition demonstrates how AI competition is broadening. At the beginning of the generative race, the contest seemed primarily about who had the most powerful model. Then it turned into a battle for cloud infrastructure, chips, data, developer tools, agents, and distribution.
Nvidia finds itself in a unique position because it sells infrastructure even to entities that could become its competitors in other segments. OpenAI, Anthropic, Google, Meta, and dozens of independent labs use or have used Nvidia hardware. The more the market grows, the more Nvidia earns, even when its own model isn't the winner.
Hugging Face reinforces this exact strategy: positioning itself at the crossroads of the entire ecosystem. If open AI continues to grow, Nvidia benefits. If companies want proprietary models running on Nvidia hardware, it benefits. If they want open-weight models optimized for local or cloud inference, Hugging Face can become the gateway.
The China question
In the document filed with the SEC, Nvidia explicitly highlights a geopolitical risk: many highly popular open models originate in China and are then downloaded, modified, and used by developers worldwide. Any regulatory restrictions limiting the ability to host or distribute models from certain regions could have a material impact on the platform.
It is an important detail because it demonstrates that Hugging Face's value also depends on its global nature. The platform works because a European researcher can test a Chinese model, an American company can publish a dataset, and a startup can compare dozens of architectures without being locked into a single ecosystem.
An acquisition that will draw regulatory scrutiny
It is difficult to imagine a transaction of this scale going unnoticed. Nvidia is already the dominant player in the AI GPU market. Buying one of the world's leading hubs for the distribution of models and tools could raise questions about competition, non-discriminatory access, and the risk of favouring its own hardware or software.
This is precisely why Nvidia’s conduct after any eventual closing will be crucial. The platform will have to remain credibly open not just in statements, but also in its APIs, integrations, pricing, and the visibility granted to competing models.
The real stakes
Nvidia is not simply buying a website with millions of users. It is buying a privileged position in the place where AI becomes a product. It is a step consistent with a broader transformation: value is no longer concentrated solely in the model or the chip, but across the entire chain connecting compute, software, distribution, and developers.
If the deal is approved and Hugging Face manages to retain the trust of its community, Nvidia will have done something far more significant than adding a new company to the group: it will have set foot inside the town square of open AI.
