This report is from this week’s The Tech Download newsletter. Like what you see? You can subscribe here.
Yesterday, Nvidia confirmed what we all already knew.
The company is planning to acquire Hugging Face, a New York-based startup which has built a repository for open source AI models and recently shot to global recognition after it came under a cyberattack by rogue OpenAI models.
The price tag was a cool $12.9 billion, making Hugging Face Nvidia’s second-biggest purchase, after it paid $20 billion for chipmaker Groq’s assets.
That nearly $13 billion valuation was what it took to fend off other bidders, Nvidia CEO Jensen Huang told CNBC on Thursday.
So why did the world’s most valuable company splash the cash for Hugging Face?
Jensen Huang, chief executive officer of Nvidia Corp., speaks to members of the media following the company’s “Japan AI Ecosystem” reception in Tokyo, Japan, on Thursday, July 16, 2026.
Kiyoshi Ota | Bloomberg | Getty Images
‘Defensive move’
To understand that, first we need to break down what Hugging Face actually does.
The startup has become perhaps the leading platform for building, sharing and running AI models, particularly in the open source and weight ecosystem — which means developers can edit and self-host a model.
Half of Nvidia’s business, Huang said, is “really largely driven by open models.” The chip giant itself is a big developer of open models — with Huang telling CNBC on Thursday that the company was the largest contributor of open models in the world “by far.”
Hugging Face also sells infrastructure and collaboration tools to organisations working with closed, proprietary AI — as well as robot ducks.
The platform’s reach is extensive. More than 18 million people use the platform to share more than 3 million models and 500,000 datasets, and it is used by more than 200,000 companies, Nvidia said in a press release.
The startup has become “one of the most important parcels of real estate in the AI market,” Gil Luria, head of technology research at D.A. Davidson said in a note.
The closest comparable to Hugging Face’s business was GitHub, a software repository that was acquired by Microsoft in 2018 for $7.5 billion, Luria said.
One of the learnings from that transaction was that allowing a tech giant to acquire an “important repository puts everybody else at a disadvantage,” Luria added, making Nvidia’s acquisition of Hugging Face a “defensive move”.
“If one of the big labs, or worse yet Google, owned Hugging Face, it would be in a position to slow down open source AI in the U.S., which would allow the big labs to increase their power,” he added.
Control
In short, Hugging Face is where AI developers hang out — and that’s exactly the kind of space Nvidia would like to control.
“Nvidia gains visibility into customer’s preferences and the AI models they use,” Naveen Chhabra, principal analyst at Forrester, told CNBC. “They can see which models are trending, what datasets customers are downloading, and the architectures that are gaining traction weeks before they hit mainstream tech news.”
The acquisition is an extension of the company’s “strategy to encourage a wide range of AI models to prosper, including both proprietary frontier models and open weight models,” Ian Fogg, research director at CCS Insight, told CNBC.
The chip giant has become one of the biggest strategic backers of tech companies in the world as it’s built up equity stakes across the AI stack, as well as helping to bankroll GPU purchases with increasingly complex financial arrangements.
Now it’s looked to secure one of the world’s leading AI hubs outright — and it beat out other suitors in the process.
Who were they, I hear you ask? Huang wasn’t forthcoming on CNBC’s “Squawk Box.”
“It doesn’t matter who the other bidders were,” he told CNBC. “It only matters who wins.”
News edit
One more thing
As the U.S.-China AI race continues to gather steam — there’s growing scrutiny on American reliance on China for key components used to power data centers, raising the prospect of higher costs and worsening supply chain shortages for the AI buildout
