Nvidia will acquire the AI startup Hugging Face for $13 billion. It was hacked this summer by an OpenAI neural network.
In July, the Hugging Face platform itself became one of the first victims of AI
The acquisition gives Nvidia control over the "GitHub for AI" / Photo: Matt Gush / Shutterstock.com
Nvidia, a manufacturer of AI chips and developer of artificial intelligence, has agreed to acquire Hugging Face—a platform for AI developers—for $12.93 billion. The deal gives Nvidia control over one of the key platforms where developers showcase and share open-source AI models, Bloomberg notes. At the same time, the chipmaker promises to keep the “GitHub for AI” free.
Details
Nvidia CEO Jensen Huang announced the deal in a post on the company's website. The announced amount includes a Hugging Face employee retention program worth up to $1 billion in stock-based compensation for those who join Nvidia, according to Bloomberg.
Nvidia intends to keep the platform open: the company promises it will continue to adhere to the startup’s existing operating principles. “As part of this commitment, Hugging Face will continue to allow model creators, developers, and users to upload and download the models and datasets of their choice, as well as support other chip vendors,” according to Nvidia’s filing with the U.S. Securities and Exchange Commission.
Why is Nvidia doing this?
Hugging Face is a central hub for open-source AI developers, where they can find, publish, test, refine, and deploy machine learning models. The platform hosts AI models, datasets, and ready-to-use applications. According to Nvidia, Hugging Face is used by more than 18 million developers and researchers, with over 3 million models, 500,000 datasets, and 1 million AI applications published on the platform. More than 200,000 companies use the service to find, evaluate, adapt, and deploy AI models.
The acquisition of Hugging Face will be the biggest step in Nvidia CEO Jensen Huang’s strategy to expand the reach of AI and the company’s customer base, according to Bloomberg. Huang aims to support the development of open-source models to prevent just a few large companies from dominating the AI sector, the agency explains. “Open models enhance safety and cybersecurity, accelerate innovation and technology adoption, and ensure technological sovereignty,” Huang wrote on X. “Nvidia will be a great home for Hugging Face, its community, and the future of open models.”
Nvidia was an investor in Hugging Face alongside Google, Amazon, Intel, Salesforce, and other companies. During a funding round three years ago, the startup was valued at $4.5 billion.
The acquisition of Hugging Face will add to Nvidia's list of deals over the past year. In August, the company entered into a $6 billion licensing agreement with the startup Poolside, which also included job offers to many of its employees. In addition, Nvidia paid approximately $20 billion for the majority of the business of the chip-development startup Groq.
One of the first victims of an AI cyberattack
Hugging Face itself became one of the first victims of a cyberattack orchestrated by artificial intelligence in July. OpenAI’s state-of-the-art models spiraled out of control during a security test: an autonomous agent built using these models escaped from an isolated test environment, gained access to the internet, and hacked Hugging Face.
OpenAI described the incident as an “unprecedented cyber incident.” It was one of the first publicly known cases of a cyberattack carried out by AI without human involvement, according to the Financial Times.
This article was AI-translated and verified by a human editor
