DOJ Questions Nvidia's $12.93B Hugging Face Acquisition

A question about the shelf, not the chip

Most antitrust attention on Nvidia starts with its share of the AI chip market. The Hugging Face line of questioning starts somewhere else: with distribution. Hugging Face is where a large share of the open-model world is published, searched, compared and downloaded. A developer choosing a model, a dataset or a deployment path will often start there, whatever hardware the work ends up running on.

That is the distribution role the DOJ is reportedly examining: whether owning a platform used to discover, compare and obtain open models could give Nvidia new ways to influence competition. In its September announcement, Nvidia described Hugging Face as a platform with more than 18 million developers, more than 200,000 companies, more than 3 million models, 500,000 datasets and 1 million applications. Owning it would not, by itself, force anyone onto Nvidia hardware, and Nvidia has promised the opposite in plain terms. The question officials are reportedly asking is subtler: whether ownership could be used to disadvantage rivals or to gain leverage in negotiations, in ways that would not require an explicit ban on anyone. Potential mechanisms could include ranking, defaults, featured placement or integration choices. These are examples of how platform ownership might influence discovery and adoption, not specific practices Reuters reported the DOJ had identified in this inquiry.

None of this has been established. The reporting describes questions, sourced to two people familiar with the matter, and the department declined to comment on matters before it. Treat any account that upgrades this to a case against Nvidia as having added a step the reporting does not contain.

The deal this inquiry now sits on top of

The transaction under examination is large enough that its structure deserves a plain reading. The $12.93 billion headline divides into about $11.9 billion payable to Hugging Face stockholders and up to about $1 billion set aside as equity for employees who stay. The agreement was signed on September 2, 2026 and announced the next day. Nvidia has said it expects to close in the first half of 2027, and the closing conditions include the required regulatory approvals. Until then this is an agreed acquisition, not a completed one. The inquiry comes while the acquisition remains proposed and before its expected first-half 2027 closing. Regulators can continue gathering information, seek remedies or, if they believe the legal standard is met, challenge a transaction before it closes. Reuters reported no formal challenge to this deal, and the current inquiry does not establish that one will follow.

Nvidia addressed the neutrality question at announcement, before the DOJ reporting. Jensen Huang said Hugging Face "will remain an open platform for the entire AI ecosystem", with developers free to choose their models, frameworks, clouds, inference providers and computing platforms, and that "NVIDIA compute will not be required to build on or deploy through Hugging Face." The department's questions, as reported, are effectively a test of how much weight that assurance should carry, and whether it needs to be backed by conditions rather than taken as a statement of intent.

Groq, SchedMD and Run:ai

The Hugging Face questions have emerged alongside a years-long DOJ examination of Nvidia's acquisitions, investments and sales practices. Reuters reported that officials have also scrutinized the Groq licensing arrangement and the SchedMD and Run:ai acquisitions. The reporting does not establish that every inquiry belongs to one formal proceeding.

Reuters reported that officials have also scrutinised Nvidia's conduct after its 2025 Groq arrangement, in which it licensed the chip startup's technology and hired its engineers. Officials explored whether Nvidia unfairly offered larger allocations of AI chips if customers also purchased its networking equipment and Groq processors. Read carefully, that is an allegation about tying AI chips, where Nvidia has a leading market position, to entry into a newer market, and it is reported as something officials explored, not something they concluded.

Two software acquisitions have also been examined. SchedMD, bought in December 2025, makes open-source software that helps run supercomputers, including systems relevant to national security. Its purchase prompted concern among AI specialists about continued access, and Nvidia has vowed to continue open access and support. Run:ai, which makes software for running AI chips more efficiently, has also been looked at. The pattern across these lines of questioning is consistent: each one asks whether a position Nvidia already holds can be extended into an adjacent layer, distribution, scheduling, orchestration or a rival's processor line, rather than whether any single price was too high.

What to watch

Three markers will show whether this stays at the level of questions. The first marker is whether the inquiry advances beyond the reported questions: for example, through a publicly reported formal challenge or an announced remedy. Reuters did not report a challenge or finding of wrongdoing. The absence of a public announcement does not establish that no additional information has been requested privately. The second is the closing timetable. The deal is due in the first half of 2027; a delay attributed to regulatory review would say more than any statement. The third is Hugging Face itself. If the platform's treatment of rival hardware, clouds and model providers visibly changes before closing, that would be evidence in a debate that, for now, consists entirely of a promise and a set of questions about whether the promise is enough.

Sources

[1] Max A. Cherney, Jeffrey Dastin and Jody Godoy, "Exclusive — US DOJ is inquiring about Nvidia's purchase of Hugging Face, sources say," Reuters, October 28, 2026 (primary source; read in full via syndication, and the syndication copy is the version read) — the inquiries and the asking-only stage, the distribution-point question, the questions about harming rivals, leverage and Hugging Face's competitors, the wider examination's scope and its Biden-to-Trump continuity, no indication of Trump involvement, the Groq allocation line as explored rather than concluded, SchedMD and Run:ai, the parties' comment status, and the cautions that the inquiry may result in no action and that investigations are not evidence of wrongdoingRead source
[2] Nvidia, Hugging Face acquisition announcement, September 3, 2026, and Form 8-K terms — total value about $12.9303 billion (about $11.9 billion to stockholders plus up to about $1 billion in retention equity), agreement signed September 2, 2026, expected close in the first half of 2027 subject to regulatory approvals; platform scale figures are Nvidia's announcement figures; Jensen Huang's pledge that the platform "will remain an open platform for the entire AI ecosystem" and that "NVIDIA compute will not be required to build on or deploy through Hugging Face"
[3] PYMNTS, "DOJ Probes Nvidia's Hugging Face Deal Over AI Competition Concerns," October 28, 2026 — CNBC's same-day confirmation of the inquiry with a separate unnamed source; Bloomberg's September reporting on whether the Groq licensing structure avoided antitrust reviewRead source
[4] Seoul Economic Daily, "U.S. Justice Department Probes Nvidia's Hugging Face Deal Over AI Competition," October 29, 2026 — Reuters summary corroborating the distribution-hub framing, the Groq allocation question and the SchedMD and Run:ai scopeRead source

THE ALFW VERDICT

Read the stage, not the headline. What Reuters reported is questions, inside an examination that predates this deal, from a department that declined to comment on the matter itself. That is a long way from a challenge. But the question being asked is the right one to track, because it is not about price or market share in chips. It is about whether the place where open models are found, compared and downloaded can be owned by the largest seller of the hardware those models run on, on the strength of a promise to keep it neutral. Nvidia's commitment to keep the platform open is a stated assurance, not evidence by itself of how the platform will operate after closing. The important question is whether the deal's structure, the company's commitments and any applicable safeguards would preserve meaningful access for rival models, chips and cloud providers. The inquiry has raised that question; it has not answered it.