The US Department of Justice has asked whether Nvidia's $12.93 billion purchase of Hugging Face would hand the chipmaker control of a critical distribution point for artificial intelligence, according to a Reuters report published October 28. Two people familiar with the matter said investigators want to know how Nvidia could use the platform to disadvantage rivals, and whether owning it would give the company leverage over the way open models reach developers.
Nvidia signed the agreement on September 2, announced it the next day, and told investors the deal would close in the first half of 2027. The price is about $11.9 billion payable to Hugging Face stockholders plus up to $1 billion in equity-based retention pay for employees joining Nvidia. Neither the Justice Department nor Nvidia would comment on the report. Hugging Face did not respond.
The inquiries are not a standalone merger review. They sit inside a wider investigation into Nvidia's conduct that opened under the Biden administration and has continued under the Trump administration, the sources said.
What the Justice Department Is Asking
Recent questioning has covered four fronts, according to the report:
- Nvidia's other acquisitions, including its December 2025 purchase of SchedMD, the company behind the Slurm scheduler used across most of the TOP500 supercomputers.
- Its investments and licensing arrangements, including a 2025 Groq licence and engineer-hiring deal that Bloomberg reported was structured to avoid a full antitrust filing.
- Sales practices, including whether buyers of networking gear or Groq processors were offered larger chip allocations.
- Hugging Face's competitors, and whether an owner could tilt ranking, defaults or support toward Nvidia silicon.
An investigation is not evidence of wrongdoing and may end without any action. A Justice Department spokesperson told CTOL that it would be inappropriate for the department to conduct a merger investigation under the guise of a conduct investigation.
Why Hugging Face Is the Distribution Point
Hugging Face is where most of the open-model world actually lives. Nvidia's own announcement put the platform at more than 18 million developers, researchers and creators, sharing over 3 million models, 500,000 datasets and 1 million applications, with more than 200,000 companies using it. Nvidia describes itself as the largest single contributor of open models and data on the site, with more than 500 models and 250 open datasets.
That makes the acquisition unusual. Nvidia is not buying a rival's product line, it is buying the shelf where everyone else's products sit. Broadcom, Google and AMD all sell silicon that runs models developers fetch from that shelf.
Nvidia's commitments are on paper in the 8-K filed for the agreement. The company says Hugging Face will stay open to model makers and users of their choosing, will keep supporting other chip vendors, and that Nvidia compute will not be required to build on or deploy through the platform. Chief executive Jensen Huang framed the deal as scaling the platform rather than capturing it.
What Approval Would Cost Nvidia
The money is not the problem. The $12.93 billion equals roughly 13 percent of Nvidia's revenue in the quarter to July 26, which came in at $96.2 billion, and about 17 percent of first-half operating cash flow of $74.4 billion.
A single platform cannot be split the way a product line can. The realistic remedy is a set of binding conditions on access, default placement and support for competing chips, or abandoning the deal outright. Nvidia has been here before: the FTC sued to block its $40 billion Arm purchase in December 2021, and the deal collapsed in February 2022.
Beijing adds a second layer of supervision. The same 8-K warns that many of the most popular open models originated in China and that restrictions on supporting models from any region could materially hurt the platform. China's antitrust regulators already issued a preliminary finding in September 2025 that Nvidia had violated the conditions attached to its Mellanox approval.
The strongest counterargument is that open weights travel. Files can be mirrored to other hubs, and Nvidia says its hardware is optional. Replacing the place where a developer expects to find three million models is the harder problem, and it is the one the Justice Department is now asking about.
Nvidia reports fiscal third-quarter results on November 17, its first report since the questions became public. Approval, not price, now sets the pace of the deal.
Comments