NVIDIA Says $12.93 Billion Hugging Face Deal Should Avoid Antitrust Concerns
NVIDIA believes its proposed 12.93 billion$ acquisition of Hugging Face should be viewed as a force against concentration in artificial intelligence rather than another expansion of its already enormous influence over the industry. During a Q&A discussing the transaction, NVIDIA argued that Hugging Face's open model ecosystem gives developers an alternative to increasingly concentrated proprietary AI platforms, even as the acquisition would place one of the industry's largest model distribution hubs under the ownership of the dominant AI accelerator supplier.
In NVIDIA's acquisition announcement, CEO Jensen Huang confirmed that Hugging Face will remain open to different models, frameworks, cloud providers, inference services, and computing platforms. NVIDIA specifically says its hardware will not be required to build or deploy through Hugging Face. The platform currently serves more than 18 million developers, researchers, and creators, hosting over 3 million models, 500,000 datasets, and 1 million applications, while more than 200,000 companies use its infrastructure.
The regulatory argument was made more directly by NVIDIA Vice President and General Manager of Enterprise Computing Justin Boitano, who said open AI platforms can counterbalance the growing influence of proprietary API providers.
"Hugging Face is almost structurally by definition kind of like a deconcentration platform."
— Quote by: Justin Boitano
NVIDIA's position is that expanding Hugging Face should distribute AI development more broadly across companies, universities, industries, and countries rather than concentrating access around a small number of closed model providers. That argument does not necessarily eliminate regulatory questions. NVIDIA already occupies a central position in AI accelerator hardware, CUDA software, networking, data center systems, inference frameworks, and increasingly open AI models. Owning Hugging Face would add one of the most important discovery, distribution, development, and deployment layers for open weight AI to that stack.
Exciting day for NVIDIA and @huggingface.
— Jensen Huang (@JensenHuang) September 3, 2026
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. They allow every developer, startup, university, industry and country to build with, customize and benefit from AI.
Thank you…
NVIDIA has already been moving heavily into this area. Its growing open model ecosystem across RTX and DGX platforms includes Nemotron models, datasets, inference software, and optimization tools, while NVIDIA says it has contributed more than 500 models and over 250 datasets to Hugging Face. The companies have also worked together for years on infrastructure, including integrating NVIDIA DGX Cloud access into Hugging Face services.
The acquisition also arrives as NVIDIA faces a different competitive challenge from some of its largest customers. OpenAI, Meta, Microsoft, and other major AI companies are investing in custom accelerators that could reduce their dependence on NVIDIA hardware. We recently examined how OpenAI's Jalapeño inference accelerator is challenging NVIDIA Blackwell on efficiency, demonstrating why NVIDIA has strong incentives to remain deeply embedded in the software and developer layers even as alternative silicon expands.
Reuters reports that approximately 11.9 billion$ of the transaction will go to Hugging Face shareholders, while another 1 billion$ is reserved for stock based employee retention. Hugging Face was valued at approximately 4.5 billion$ during its 2023 funding round, making NVIDIA's purchase price a substantial premium for a company whose strategic value extends far beyond its immediate revenue.
Regulatory attention would also arrive during a period when NVIDIA is already becoming increasingly careful around competition concerns. The company recently paused portions of its AI cloud credit program following internal antitrust concerns involving relationships where NVIDIA could simultaneously provide hardware, financing, capacity guarantees, and commercial support to AI cloud providers. The Hugging Face transaction presents a different structure, but it highlights the same broader question surrounding how many layers of the AI economy can become connected to one company.
The central issue may therefore be whether Hugging Face remains genuinely neutral after the acquisition. NVIDIA says developers will continue to choose AMD, Intel, custom accelerators, different clouds, and competing inference platforms if they prefer. If that interoperability remains intact, NVIDIA can make a credible argument that greater investment in Hugging Face strengthens competition between AI models and deployment providers. If NVIDIA hardware or software begins receiving preferential treatment, however, the same platform could become another powerful mechanism for extending CUDA's influence.
Calling Hugging Face a deconcentration platform is an interesting argument because both sides of the debate can be true at the same time. Open models genuinely create competition against closed AI providers, and Hugging Face has become one of the most important pieces of infrastructure supporting that ecosystem. More investment could make those models easier to develop, evaluate, and deploy.
At the same time, NVIDIA is not a neutral infrastructure company entering the market from the outside. It already controls a massive portion of the compute stack used to train and run those models. Adding the platform where millions of developers discover and distribute them gives NVIDIA another strategically valuable position in the AI pipeline.
The real antitrust question may not be whether NVIDIA owns Hugging Face, but whether Hugging Face still behaves like Hugging Face after NVIDIA owns it.
Do you think NVIDIA ownership will give Hugging Face the resources to strengthen open AI, or does combining the largest open model platform with the dominant AI GPU company create too much influence under one company?
