Nvidia’s Hugging Face Deal Faces Antitrust Review: A Real‑World Timeline

In March 2024 Nvidia announced a planned acquisition of Hugging Face, an open‑source model hub that hosts more than 30,000 community‑contributed models. The deal, reported at roughly $1 billion in cash and stock, was pitched as a way to combine Nvidia’s GPU leadership with Hugging Face’s software ecosystem, promising tighter integration for developers.

Within weeks, the US Department of Justice’s Antitrust Division sent a formal request for information (RFI) to both companies. The RFI, dated early May, asked for detailed data on: (1) the percentage of AI workloads that currently run on Nvidia hardware versus competitors; (2) revenue forecasts for the combined entity over the next five years; and (3) any exclusive licensing agreements that could foreclose competition. By June 12, Nvidia had submitted a 250‑page response, including internal forecasts that projected a 20% increase in GPU sales attributable to the acquisition, a figure that regulators flagged as potentially overstated.

The DOJ’s next step was a “second request” in early July, expanding the data set to include contracts with cloud providers such as AWS, Azure, and Google Cloud. Those providers collectively account for 55% of the market for AI compute, according to IDC. Regulators are particularly concerned that Nvidia could leverage Hugging Face’s model repository to bias developers toward Nvidia‑optimized models, effectively creating a gate‑keeping platform.

From a practical standpoint, the case illustrates three lessons for AI firms pursuing M&A:

1. Quantify integration benefits with hard data – Vague promises of “faster inference” or “better developer experience” are insufficient. Nvidia’s 20% sales uplift claim was dissected line‑by‑line, and the lack of independent benchmarks made it vulnerable.
2. Map out ecosystem dependencies early – The DOJ’s focus on cloud contracts shows that any acquisition affecting downstream services must be mapped before filing.
3. Prepare for a multi‑month review – The timeline from announcement to the second request spanned just over four months, but the full review could extend beyond a year, affecting product roadmaps and staffing.

While Nvidia and Hugging Face have not commented on the probe’s outcome, the scrutiny underscores a growing regulatory appetite for AI‑centric deals. If the acquisition is blocked or conditioned, competitors may see an opening to develop alternative model hubs that are hardware‑agnostic, potentially diversifying the AI supply chain.

For the broader AI ecosystem, the case serves as a cautionary tale: ambitious integration strategies must be backed by transparent, verifiable metrics, and firms should anticipate that regulators will look beyond headline valuations to the real competitive impact on developers and cloud services.

Photo: GAMERCOMP.RU / Unsplash (https://unsplash.com/@gamercomp)