The ‘deeply nerdy’ founders behind Nvidia’s $13 billion bet on hugging face

A trio of nerdy Frenchmen came to New York to build a sassy chatbot for teens. A decade later, their startup has grown into an AI platform that’s a pillar of the tech ecosystem, and a linchpin in the strategy of the world’s most valuable company.

Nvidia agreed on Thursday to buy their AI company, called Hugging Face, for roughly $13 billion, doubling down on the chipmaker’s effort to challenge the dominance of proprietary AI models from OpenAI and Anthropic. With the deal, Nvidia is helping foster the ecosystem Hugging Face built for so-called open-weight AI tools that can be downloaded and tweaked by anyone.

But when Clément Delangue and Julien Chaumond first arrived at the Manhattan offices of venture capital fund Betaworks for its “BotCamp” startup program in the sweltering summer of 2016, their goal was more modest. They wanted to build a general-purpose chatbot that wasn’t boring like Apple’s Siri. It was more than seven years before the launch of OpenAI’s ChatGPT.

“It won’t always give you the right answer, but it will always give you a funny one,” Delangue, the chief executive, said during a demonstration that summer, in which the app traded selfies with Chaumond, the chief technology officer. “And of course, it’s always very sassy.”

The name of the app—Hugging Face—was a placeholder they had put on their application for the program, based on the title of their favorite emoji. They intended to change it later but never did.

The following spring, Thomas Wolf, a physicist turned patent lawyer, joined the founders as the chief science officer.

“They’re deeply nerdy. They’re deeply into science,” John Borthwick, managing partner of Betaworks, now an investor, said in an interview.

Raised in northern France, Delangue has long been engaged in digital collaboration and open systems. As a teenager, he sold motorbikes on eBay and later worked for the company, according to a profile on the website of Sequoia Capital, one of Hugging Face’s investors.

As a university student, he developed a platform for students to share notes and insights from their classes with others, according to a TEDx talk he gave in Paris in 2012.

Delangue was already working in product and marketing for startups in New York when he and Chaumond—a mathematician who had worked for startups and the French government—teamed up. They pursued a shared interest in natural-language processing, a type of computer program that they thought could be used to make a chatbot, and Chaumond flew from Paris to New York.

“It was the first time we worked together IRL for more than a day at a time,” Chaumond wrote.

They took an online class at Stanford with Richard Socher, a proponent of natural language processing, and formed a study group on the topic. Wolf was in the class, too, Socher said.

“I was very impressed by them. I thought they asked clever questions,” said Socher, who is now founder and chief executive of AI startup Recursive. He sometimes goes tree-climbing on Bay Area hikes with Delangue.

A major turning point for Hugging Face came in 2018, when Google released a new version of its transformer, the neural network technology at the heart of modern generative AI models.

Wolf told his co-founders he would take the weekend to make an open-source version of it accessible in the most popular language for AI researchers, Delangue said in a 2024 podcast episode. By Monday, the Hugging Face version had gone viral, and six-months later it became a new business model.

“We think we need to pivot,” Delangue recalled telling investors.

As AI models grew in size, developers started to need a new platform to host them. Hugging Face built one. It effectively made a sharing platform for AI models that’s similar to what Microsoft’s GitHub is for code from software developers—a role that has grown as the AI revolution has taken off.

Today, there are millions of open-weight models on Hugging Face.

The company made news in July when it shared that it had been hacked by an autonomous swarm of AI agents, which it later discovered had escaped from an internal cybersecurity test at OpenAI. The company tried to dissect the attack logs using proprietary AI models, but was unable to because of security guardrails, so instead turned to an open-weight model from a Chinese developer.

Since the incident, Delangue has been outspoken about the need to preserve and promote access to high-quality, open-weight AI models so that potential cyberattack targets can effectively defend themselves. Nvidia has been, too.

Borthwick said he thinks the Nvidia investment makes sense for Hugging Face’s goal to democratize agentic AI beyond the 18 million developers currently on the platform. “Nvidia is enabling that growth by this acquisition,” he said.

Delangue said Thursday on CNBC that he approached Nvidia over the summer to pursue a deal. He saw that the world was facing a choice between a path where proprietary models like those from Anthropic or OpenAI dominate the field, and another where open-weight models proliferate.

“There’s a path where open-source AI is available to everyone, and everyone can actually become like an owner, a builder of AI and not just renting it or using it from other people,” Delangue said. “We needed to double down on open-source AI to really distribute the technology as much as we can all over the world.”

Write to Sam Schechner at Sam.Schechner@wsj.com and Robbie Whelan at robbie.whelan@wsj.com