• DOJ Antitrust Division issues detailed subpoenas probing Nvidia's $12.93B takeover of open-source platform Hugging Face.
• Federal enforcers evaluate whether owning the developer hub creates an unfair distribution chokepoint favoring Nvidia's CUDA architecture.
• European Commission competition regulators coordinate with Washington on potential cross-border antitrust enforcement action.

WASHINGTON & SAN FRANCISCO — The Department of Justice's Antitrust Division has significantly escalated its formal inquiry into Nvidia Corporation's proposed $12.93 billion takeover of Hugging Face, issuing sweeping Civil Investigative Demands to artificial intelligence startups, cloud hyperscalers, and rival semiconductor manufacturers. According to individuals familiar with the regulatory proceedings who spoke on the condition of anonymity, government investigators are scrutinizing whether allowing the world’s preeminent AI chipmaker to absorb the planet’s central open-source model repository will construct an anti-competitive software bottleneck designed to permanently entrench its hardware dominance.

Hugging Face, which hosts more than 1.2 million open-source AI models, datasets, and collaborative coding repositories utilized by millions of software engineers worldwide, has long served as the de facto Switzerland of the machine learning developer ecosystem. The platform enables computer scientists to publish, evaluate, and fine-tune machine learning architectures across a diverse array of competing hardware substrates, ranging from AMD's Instinct accelerators and Intel's Gaudi chips to Google's proprietary Tensor Processing Units (TPUs) and Amazon’s Trainium processors. The announcement of Nvidia's acquisition prompted immediate alarm among antitrust enforcers that Nvidia could subtly optimize platform architectures, default inference kernels, and benchmarking pipelines to favor its proprietary CUDA software stack.

Antitrust investigators are focusing heavily on allegations of ecosystem lock-in and tie-in mechanisms. Depositions conducted by DOJ counsels have questioned rival chipmakers and venture-backed AI founders regarding Nvidia’s hardware allocation policies, bundled software licensing terms, and historical acquisitions of algorithmic optimization startups such as Groq’s assets and SchedMD. Regulators are examining whether acquiring Hugging Face would allow Nvidia to monitor early-stage algorithmic innovation before it reaches the commercial market, giving the company asymmetric intelligence to preserve its roughly 88 percent market share in AI training silicon.

Nvidia leadership and corporate spokespersons have vigorously defended the transaction, arguing that the acquisition will inject hundreds of millions of dollars in sustained capital into maintaining open-source infrastructure that has historically struggled to achieve standalone profitability. In formal regulatory filings submitted to the Federal Trade Commission and Department of Justice, Nvidia pledged to preserve Hugging Face’s open, hardware-neutral platform architecture, maintaining public APIs and continuing support for non-Nvidia hardware backends. Executive leadership stressed that the partnership aims to democratize advanced model quantization and distributed inference tools for independent developers globally.

Despite these assurances, regulatory sentiment in Washington and Brussels has hardened against mega-mergers across the generative AI supply chain. European Commission competition authorities in Brussels confirmed they are coordinating closely with American counterparts, warning that vertical integration between foundational software hubs and monopoly hardware providers risks stifling the next wave of open-source algorithmic breakthroughs. The DOJ is expected to conclude its initial second-request review by late winter, with antitrust attorneys weighing whether to seek an emergency preliminary injunction in federal district court to block the acquisition.

Congressional testimony scheduled for later this month is expected to address cross-agency coordination, focusing specifically on supply chain resiliency, sovereign compute clusters, and long-term capital allocation strategies designed to safeguard domestic competitiveness.