Nvidia has been in talks to acquire Hugging Face for more than $13 billion

Nvidia may be looking to significantly expand its influence over the artificial intelligence landscape through a potential acquisition of Hugging Face. According to sources familiar with the matter, the chip giant has been engaged in discussions recently that could see the open source AI platform valued at more than 13 billion dollars. While Microsoft reportedly held meetings with the company as well, those talks are no longer active, leaving Nvidia as a primary contender in what would be one of its most substantial deals to date.

The move comes at a time when Nvidia is aggressively deploying its massive cash reserves into strategic investments. With billions already committed to private firms and an additional equity budget set aside for the remainder of its fiscal year, the company is well positioned to make such a leap. This isn’t the first time Nvidia has eyed Hugging Face; the chipmaker was part of a funding round in 2023 and previously offered a half billion dollar investment that was declined because Hugging Face wanted to avoid having a single dominant investor steer its direction.

For Nvidia, owning Hugging Face would mean controlling the central hub where millions of developers share and build AI models and datasets. By integrating this community directly into its orbit, Nvidia could create a powerful incentive for developers to optimize their workloads specifically for Nvidia hardware. It effectively turns a neutral repository into a gateway that could funnel more business toward its high end GPUs.

However, such a merger brings significant risks regarding the perceived independence of the platform. Founded by French entrepreneurs who built it as an inclusive space for all creators, Hugging Face currently supports hardware from various competitors including Intel and AMD. If acquired by the world’s leading AI chip manufacturer, there are concerns that this spirit of neutrality could vanish, potentially alienating users who rely on the platform precisely because it remains agnostic toward whose silicon is powering the model.