Nvidia under probe by US justice officials over Hugging Face acquisition

This isn't an acquisition—it's the final piece of a full-stack monopoly. Hugging Face owns model distribution; Nvidia owns the compute substrate. Together, that's the AI equivalent of Windows+Office. Technical cascade: Integration bakes CUDA optimization bias into every model on the Hub. AMD's ROCm and Intel's oneAPI lose their neutral ground permanently. Developer lock-in deepens beyond what Qualcomm's 2019 patent dispute inflicted on the supply chain. Compliance drag: The dual probe (this deal plus broader antitrust) imposes a "regulatory tax" on Nvidia's capital allocation for 18+ months. The Microsoft/Activision EU conditional approval is the closest precedent—worst case: forced divestiture or interoperability mandates, adding $300-500M in compliance overhead. Competitor response: AMD will accelerate a model-hosting platform acquisition to fill the ecosystem gap. Google repositions TPU+Vertex AI as the "sovereign alternative." EU AI Act enforcement will cite this as the vertical-integration benchmark. 12-24 month tail: The AI stack bifurcates into "Nvidia-native" versus "open-stack" procurement tracks. If decoupled, Hugging Face fragments into three to four competing hubs—better for competition, worse for developer UX. The real winner is whoever controls the inference layer next.