Could Nvidia control how AI models reach developers? Why DOJ is worried about the $12.93 billion Hugging Face deal

The US Justice Department is examining whether Nvidia’s proposed $12.93 billion acquisition of Hugging Face could give the chipmaker greater control over AI model distribution and harm competitors. 

The US Department of Justice (DOJ) is examining whether Nvidia’s proposed $12.93 billion acquisition of open-source artificial intelligence platform Hugging Face could hurt competition by giving the chipmaker greater control over how AI models are distributed, Reuters reported, citing two people familiar with the matter.

Justice Department officials have asked whether the acquisition would allow Nvidia to “capture a fundamental distribution point for AI,” the people told Reuters. Officials are also examining how Nvidia could use the platform to harm rivals in the AI market and whether the deal could give the chipmaker greater leverage over competitors.

The scrutiny comes as the DOJ continues its broader investigation into Nvidia’s business conduct, including its acquisitions, investments, sales practices and whether the company has exerted pressure on customers, Reuters reported.

Nvidia announced the acquisition in September, with the deal expected to close in the first half of 2027. The company has said Hugging Face would continue to support work with a range of chip, cloud and AI providers.

Why Nvidia’s Hugging Face deal matters

Nvidia has become a cornerstone of the technology economy and the artificial intelligence boom, backing major players such as Anthropic and OpenAI while driving data centre development globally to meet AI’s growing computing needs.

The company holds the biggest share of the market for AI chips, which power tools such as ChatGPT and Anthropic’s Claude. Its position has made it an industry force, even as startups and competing chip designers seek to challenge its dominance.

Officials have also looked at Nvidia’s acquisition of Run, which develops software to help AI chips run more efficiently.

Sanchari Ghosh