Summary: On August 24, TechCrunch, citing Business Insider, said Hugging Face had been approached about a sale at a valuation starting at $13 billion or more and had brought in banks to help weigh bids, leaving the buyer unnamed. On August 27, Business Insider supplied the missing piece: the buyer is Nvidia, and the two sides have been talking for several weeks.
On August 27, 2026, Business Insider reported that Nvidia is negotiating to acquire Hugging Face, at a valuation north of $13 billion. Hold on to two caveats: there is no deal yet, and “the talks could still fall apart.” Both companies declined to comment, and the reporting traces back to a single person familiar with the matter. Once the story landed on Hacker News it drew 1,988 points and 926 comments — the biggest potential purchase of open-source infrastructure since Microsoft paid $7.5 billion for GitHub.
What Is Known So Far
Laid out as a timeline, the reporting runs like this. On August 24, TechCrunch, citing Business Insider, said Hugging Face had been approached about a sale at a valuation starting at $13 billion or more and had brought in banks to help weigh bids, leaving the buyer unnamed. On August 27, Business Insider supplied the missing piece: the buyer is Nvidia, and the two sides have been talking for several weeks. That same report also noted that Microsoft had met with Hugging Face too, but those discussions are no longer ongoing.
Why Nvidia Wants the GitHub of Models
Hugging Face is the de facto hub of the open-source AI world, hosting millions of models and datasets. From Nvidia’s side, buying it means owning the front door to the “AI app store”: the more open models that get deployed, the more GPUs it sells. The deal also carries a built-in contradiction that Business Insider names directly — Hugging Face’s value rests on hardware neutrality, and the platform supports rival hardware from AMD, Intel, and others just the same. Under Nvidia’s ownership that neutrality could be undermined; then again, ownership could steer still more AI workloads onto Nvidia chips. Money is not the obstacle: per Business Insider, Nvidia has committed $18 billion to equity investments this fiscal year and already holds $47.9 billion in private companies. The financial logic draws skepticism anyway — as commenters were quick to calculate, recouping $13 billion at Hugging Face’s current revenue would take decades. This would be a strategic purchase, not a spreadsheet one.
From Rejecting an Investor to Weighing a Sale
The history makes this deal more interesting. Hugging Face was founded in New York in 2016 by Clément Delangue, Julien Chaumond, and Thomas Wolf. In 2023 it raised $235 million at a $4.5 billion post-money valuation — with Nvidia among the investors. Earlier this year, Nvidia offered a $500 million investment at a $7 billion valuation, and Hugging Face turned it down, saying it did not want a single dominant investor able to sway decisions. CEO Clem Delangue also stressed at the time that the company was “close to profitability” and that he had a long-term duty to its open-source community — fueling doubt about whether a sale would happen at all. Going from rejecting $500 million to weighing a full sale within a year is this AI infrastructure boom in miniature; over the same stretch, Stripe reportedly acquired OpenRouter for $7 billion.
What the Open-Source Community Is Watching
Those 926 Hacker News comments lay the concerns out. Data comes first: Hugging Face can see which models get downloaded and which datasets are trending — industry-wide insight that would now sit with a chip monopolist; some called the prospect “borderline anti-trust.” Governance is second: Nvidia has a track record of abandoning communities, and if licensing or platform policy shifted, forks would almost certainly follow. Geography is third: a platform founded by French entrepreneurs and treated by Europe as a symbol of “AI sovereignty” would land in the hands of an American chip company — the EU threads got the most heated.
What It Means for Developers and Enterprises
Nothing is signed, so nobody needs to move their models today. Two precautions are worth taking anyway. First, make sure your critical models have an off-platform backup; keeping heavyweight weights in local storage or your own object store costs very little. Second, read the deal as a signal: when “model hosting” itself is worth $13 billion, the infrastructure of the open-source ecosystem is being fenced off by a handful of giants fast, and single-platform dependency belongs in your architecture risk list. And if the talks collapse — both companies’ silence says that is possible — the episode still marks the moment open-source hosting became explicitly strategic real estate.