DOJ probes Nvidia's $13 billion Hugging Face acquisition over AI competition concerns

The U.S. Department of Justice has reportedly opened inquiries into Nvidia's proposed $13 billion acquisition of Hugging Face, the widely used open-source platform for hosting, sharing, and testing AI models. According to Reuters, regulators are concerned that the deal could allow Nvidia to lock in a key distribution channel for its own AI hardware and software products, potentially disadvantaging competitors in the rapidly expanding AI market. The investigation adds to a growing pattern of antitrust scrutiny directed at Nvidia, which has included examinations of its $20 billion Groq acquisition, its purchase of SchedMD, and allegations that the company tied AI chip allocations to purchases of its networking hardware.

Hugging Face serves as a central hub for the open-source AI ecosystem, hosting hundreds of thousands of models that developers and researchers use to build applications. The platform also sells AI infrastructure services. Nvidia has stated that it will continue supporting multiple chip and AI providers after the acquisition closes, which is expected in the first half of 2027. The DOJ investigation reflects broader governmental concern about concentration in AI infrastructure, spanning both the Biden and Trump administrations.

Antitrust concerns and open-source platform control

• Hugging Face platform role: The platform functions as a primary open-source repository for AI model hosting, testing, and collaboration, with hundreds of thousands of models available to developers and researchers.
• DOJ competition concerns: Regulators are examining whether Nvidia could use ownership of Hugging Face to preference its own CUDA ecosystem, GPUs, and software tools over competing hardware from AMD, Intel, and AI chip startups.
• Vertical integration pattern: The investigation follows scrutiny of Nvidia's $20 billion Groq acquisition and its SchedMD purchase, with regulators examining whether Nvidia bundles or ties chip allocations to networking hardware purchases.
• Export control context: The probe occurs alongside existing U.S. restrictions on Nvidia AI hardware sales to China, Iran, North Korea, and Russia, reflecting multifaceted government attention to Nvidia's market position.
• Timeline and commitments: Nvidia expects the Hugging Face deal to close in the first half of 2027 and has committed to continuing support for diverse chip and AI providers post-acquisition.

Why it matters

The outcome of this investigation will significantly affect how open-source AI model distribution remains accessible to developers using non-Nvidia hardware, particularly AMD ROCm, Intel Gaudi, and emerging RISC-V AI accelerators. If regulators impose conditions or block the deal, it could preserve Hugging Face's neutrality as a model hub; conversely, approval without safeguards risks consolidating another layer of AI infrastructure under a single dominant vendor, potentially steering model optimization and deployment tooling toward CUDA-locked implementations.