According to Bloomberg, Nvidia is on the verge of taking one of its biggest steps outside its core chip business. The GPU manufacturer is negotiating the acquisition of Hugging Face, the platform that has become a major hub for developing, publishing, and using AI models.
The talks are now well advanced. Bloomberg reports, citing insiders, that Nvidia is willing to pay $12.9 billion for Hugging Face. A package worth approximately $1 billion could be added to that amount to retain employees after the acquisition. A deal could be finalized as early as this week, though sources emphasize that no definitive agreement has been reached yet.
These figures represent a significant jump from Hugging Face’s last known valuation. During a funding round in 2023, the company was valued at $4.5 billion. Nvidia was already among the investors at that time, alongside Google, Amazon, Intel, and Salesforce, among others.
Greater control over AI software
An acquisition aligns with the diversification that Nvidia has long been pursuing. The company earns most of its revenue from hardware for AI workloads, but is increasingly building a software and services layer around it. Hugging Face could play a key role in this.
Founded in 2016, the company operates a platform where developers make models, datasets, and other AI components available. As a result, Hugging Face plays a role that goes beyond that of a typical software provider. For many developers, the platform serves as a practical starting point for finding, testing, and deploying models.
With an acquisition, Nvidia would thus gain control of a key link between AI models and the developers who work with them. This could also strengthen the position of Nvidia’s own hardware and software stack. The easier it is to deploy models on Nvidia technology via the Hugging Face ecosystem, the stronger the connection with the chipmaker’s platform can become.
At the same time, this creates an interesting tension. Nvidia CEO Jensen Huang regularly emphasizes the importance of open models as a counterbalance to AI that is largely controlled by a few large technology companies. However, with Hugging Face, Nvidia itself would become the owner of one of the most important platforms within that open ecosystem.
A series of billion-dollar investments
The potential acquisition is not an isolated move. Nvidia is increasingly using its rapidly growing financial clout to expand its position beyond GPUs. In August, it struck a $6 billion licensing deal with AI startup Poolside. As part of that deal, many of the company’s employees were also approached to join Nvidia.
Previously, Nvidia paid approximately $20 billion for a majority stake in chip company Groq. With Hugging Face, another billion-dollar investment would be added to that, but this time in an area much closer to AI developers and models.
This is strategically important now that Nvidia’s largest customers are simultaneously working on alternatives to Nvidia hardware. Major cloud companies are increasingly developing their own AI accelerators. By controlling not only chips but also software, development tools, and potentially a platform like Hugging Face, Nvidia can broaden its position in the AI value chain.
Hugging Face recently targeted by AI agent
The timing is also striking. Hugging Face recently made headlines due to a security incident in which experimental AI agents from OpenAI gained access to the platform’s systems. OpenAI acknowledged afterward that it could have intervened sooner. The incident highlighted the risks that arise when autonomous AI systems gain access to external infrastructure.
For now, this does not appear to be an obstacle to the potential acquisition. If Nvidia and Hugging Face do reach an agreement, Nvidia will gain not only a large software platform but, more importantly, a substantial developer community and a central position in the distribution of AI models. This would allow the company to once again bring a larger portion of the AI ecosystem under its umbrella.