Nvidia Buys the 'GitHub of AI': Inside the $12.93 Billion Hugging Face Deal

Rumored in August, confirmed on September 3: Nvidia has agreed to acquire Hugging Face for $12.93 billion, bringing one of the AI industry’s most important open-model platforms under the auspices of the world’s dominant AI chipmaker — and, with a market capitalization well over $5 trillion, the most valuable company on Earth.

For a company that made its fortune selling the picks and shovels of the AI gold rush, this is a different kind of bet. Nvidia isn’t buying compute or a model lab. It is buying the place where the open-source AI community actually lives — the repository often called the “GitHub of AI models,” home to millions of models, datasets, and applications that developers around the world download, fine-tune, and deploy every day.

The deal, in numbers

The headline figure gets rounded to $12.9 billion, but the actual price is an oddly specific $12,930,300,000. According to Hugging Face co-founder Thomas Wolf, that’s an Easter egg: 129303 is the decimal Unicode value of the 🤗 hugging-face emoji, and #129303 is a green color code nodding to Nvidia’s branding.

Per Reuters, roughly $11.9 billion of the total goes to Hugging Face stockholders, with up to $1 billion earmarked for equity-based retention awards to keep employees from walking out the door. Nvidia’s SEC filing says the deal is expected to close in the first half of 2027, subject to customary closing conditions and regulatory approvals — and given Nvidia’s lofty position in AI infrastructure, those approvals are unlikely to be a mere formality. The filing also flags the possibility that future regulation around open-source AI could affect Hugging Face’s operations or increase compliance costs.

Founded in 2016 by Clément Delangue, Julien Chaumond, and Thomas Wolf as a consumer chatbot app, Hugging Face pivoted to become the de facto home of open models. Its Hub hosts a large share of the world’s open-weight checkpoints — from Meta’s Llama family to Mistral, Qwen, and Nvidia’s own Nemotron models — alongside datasets, Spaces demos, and the Transformers library that arguably did more than any other codebase to standardize how the industry works with open models.

Nineteen mentions of “open”

When reports of the bid first surfaced in August, the immediate concern was obvious: Hugging Face’s value lies partly in being a neutral place to find and deploy open models across Nvidia GPUs, AMD Instinct, Intel Gaudi, AWS Trainium and Inferentia, Google TPUs, and other accelerators. What happens when the platform is owned by just one of those vendors?

Nvidia founder and CEO Jensen Huang is going to great lengths to answer that question before it’s asked. In the official announcement, he makes a series of explicit promises: Hugging Face will “remain an open platform for the entire AI ecosystem,” it “will continue to support open source and open weight models from across the ecosystem, from every model builder,” and — most pointedly — “Nvidia compute will not be required to build on or deploy through Hugging Face.”

The word “open” appears no fewer than 19 times in Huang’s relatively short announcement, underlining just how central that reassurance is to Nvidia’s pitch. Nvidia also uses its SEC filing to reaffirm the commitments in writing: Hugging Face would continue to permit model makers, developers, and users to upload and download models and datasets of their choosing, and to support other silicon vendors.

Huang also points to a recent open letter on open weights that he co-signed alongside executives and researchers from across the industry, including people from Hugging Face. The letter argued that open-weight models are critical to broadening access to AI, strengthening competition, and giving developers more control over deployment and adaptation.

Why Hugging Face said yes

Delangue, for his part, frames the sale as a matter of scale. “10 years after starting Hugging Face, open-source AI is at an inflection point,” he wrote on LinkedIn. “Thanks to the community, we’ve shown that it can be a complement, and even an alternative, to closed-source APIs. But for it to happen at larger scale, it needs more compute, more support, more collaboration and more visibility.”

A few months before the deal, Delangue had gone so far as to call Nvidia the “King of American open-source AI,” pointing to its growing collection of public models, datasets, and Spaces. He says Nvidia has committed to backing Hugging Face while keeping the platform open, independent, and compute-agnostic, with the founders and the existing team staying on. “Together, we think we can make open source the default way to build AI,” he writes — with the stated goal of helping 100 million AI builders “own their intelligence rather than rent it.”

That last phrase is a not-so-subtle jab at the closed-API business models of OpenAI, Anthropic, and Google. It also explains the strategic logic for Nvidia: the more AI gets built on open weights running on infrastructure people control, the more of that infrastructure tends to be Nvidia’s.

The GitHub precedent — and why it understates the problem

Skeptics reach for the obvious historical comparison: Microsoft bought GitHub for $7.5 billion in 2018, promised independence and openness, and largely delivered. But GitHub’s later use of public code to train the proprietary, paid Copilot service still sparked a backlash in parts of the open-source community.

As analyst Janakiram MSV noted in Forbes, the comparison may actually understate Nvidia’s challenge. Neutrality at Hugging Face has a hardware dimension GitHub never had. Hugging Face’s integrations span rival accelerators from AMD, Intel, AWS, and Google. Under Nvidia ownership, continued first-class support for those competing chips becomes a real, ongoing test of just how neutral the platform remains — one that will be measured in engineering investment and release cadence, not press releases.

There’s a data dimension, too. A platform that sees which models are downloaded most, which datasets are used for fine-tuning, and where deployments bottleneck is a strategic intelligence asset. Nvidia insists the platform stays open, but the community will be watching what gets measured — and who gets to see it.

What it means

For developers, the near-term practical answer is: probably nothing. Nvidia has every incentive to keep Hugging Face boringly reliable — the deal isn’t expected to close until sometime in the first half of 2027, and the commitments are explicit and written into regulatory filings. The long-term question is whether “open, independent, and compute-agnostic” survives contact with a parent whose core business is selling compute.

For the industry, the acquisition is the clearest signal yet that the open-weights ecosystem has become infrastructure — valuable enough for the most valuable company on Earth to pay $12.93 billion for its town square. Whether that infrastructure stays genuinely public will depend on the choices Nvidia makes long after the headlines fade, one accelerator SDK and one model card at a time.

The 🤗 emoji, at least, now has the most expensive Unicode codepoint in history.