Newly Bought Tech Startup Will Stay ‘Open, Neutral’ After $13,000,000,000 Takeover, New Owner Vows

A major chipmaker claims the biggest platform for open-source artificial intelligence models will remain unbiased under its ownership.

NVIDIA officially confirmed the roughly $13 billion deal to obtain New York-based AI startup Hugging Face in an effort to expand worldwide access to open-source models, the company announced in a Thursday press release. The deal lands as the AI chipmaker claims the open weight AI will be “platform agnostic,” even as Hugging Face continues to host models and datasets from NVIDIA rivals, including AMD, Google and Intel.

“NVIDIA plans to help scale Hugging Face’s platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide,” NVIDIA Newsroom said in a Thursday X post. “Hugging Face will remain an open, neutral and platform-agnostic home for the entire AI ecosystem.”

NVIDIA released more than 500 open weight models on Hugging Face, more contributions than any other AI company, according to Hugging Face data shared in the company’s announcement.

The main hub for open weight AI, Hugging Face hosts more than two million models, the company states on its ‘Hub’ page. These models include Google’s Gemma, Meta’s Llama, AMD’s Instella and Intel’s Neural Chat, all from companies that compete with NVIDIA.

Hugging Face will remain a platform for open weight models to “serve the global AI community at an unprecedented scale” and assist small businesses, startup companies and other institutions that wish to utilize models that assist in their prospects, according to NVIDIA’s announcement.

This is the largest completed NVIDIA acquisition to date, more than NVIDIA’s previous $6.9 billion purchase of chipmaker Mellanox in 2020.

NVIDIA referred the Daily Caller News Foundation to the company’s blog post announcement when asked for comment, according to a March 2019 NVIDIA press release.

Hugging Face did not immediately respond to the DCNF’s request for comment.

“Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty,” NVIDIA CEO Jensen Huang said in an initial Thursday X post sharing the company’s press release. “They allow every developer, startup, university, industry and country to build with, customize and benefit from AI.”

Hugging Face rejected NVIDIA’s $500 million investment into the startup in late 2025. At the time, the AI platform emphasized the importance fighting concentration of power and staying free of one major investor, Financial Times reported.

“[Open-weight models] contribute to democratising AI, to fighting concentration of power which is, in my opinion, the biggest risk in AI,” Hugging Face Co-founder and CEO Clement Delangue said Jan. 28, according to Financial Times.

Hundreds of OpenAI’s agents, namely from the company’s internal “HPIM” and GPT-5.6 Sol models, breached the AI startup in July during a cybersecurity test. The “unprecedented cyber incident,” was announced by the AI company on July 27.

The OpenAI agents messaged each other roughly 70,000 times using an “unsanctioned message board” to coordinate the attack on Hugging Face, despite being designed to stay isolated and unable to communicate with one another, according to METR and Redwood Research.

Hugging Face defended against the cyberattack with an “[N]vidia quantized version of GLM 5.2 coming from Z.ai,” a Chinese AI model, Hugging Face Co-founder and CEO Clement Delangue said in an X post on July 31.

Open weight models are AI models with publicly available “weights,” according to the International AI Safety Report. Weights are the central components of an AI model that are shared publicly for download.