Chipmaker NVIDIA has reached an agreement to acquire Hugging Face, the leading platform for hosting and distributing open-source artificial intelligence models, for approximately $12.900 billion (around €11.000 billion). This was first reported by the tech publication The Information, citing a source with direct knowledge of the negotiations, and the story was subsequently picked up by other outlets such as Reuters and Bloomberg. Neither NVIDIA nor Hugging Face has officially confirmed the deal, although discussions are reportedly at an advanced stage.
If confirmed, this figure would make the acquisition one of the largest in NVIDIA's history and represent a significant leap from Hugging Face's last known valuation of around $4.500 billion following a $235 million funding round in 2023, in which NVIDIA, Salesforce, and Google participated. The deal also comes at a particularly relevant time: just hours before the first details of the agreement emerged, NVIDIA presented its quarterly results with revenue of $96.200 billion and announced that it expects revenue growth of approximately 70% during the next fiscal year.
What is Hugging Face and why is it of interest to NVIDIA?
Hugging Face functions, broadly speaking, as a GitHub for artificial intelligence: a platform where developers, researchers, and companies share, download, and test models and training datasets. Both tech giants and small independent teams use it, and its role has been key in the rise of so-called open models—those whose internal workings can be freely consulted and modified—as opposed to the closed systems of companies like OpenAI or Anthropic.
That's precisely where much of NVIDIA's interest lies . Jensen Huang's company doesn't manufacture artificial intelligence models, but rather the hardware (graphics cards or GPUs) on which these models are trained and run. And several of its main clients, including OpenAI and Anthropic, are developing their own chips to reduce their dependence on NVIDIA. Controlling the platform where open-source models are distributed would allow the company to ensure that this increasingly dynamic ecosystem continues to operate on its technology, even if the major AI labs manage to forge their own paths.
The acquisition would give NVIDIA control of Hugging Face's large language models and open-source datasets, allowing it to diversify its business at a time when competition in the AI ​​sector has intensified. The platform, founded in 2016 in New York by three Frenchmen (Clément Delangue, Julien Chaumond, and Thomas Wolf), has established itself as one of the world's leading open AI repositories, with millions of models available for download and modification.
A valuation that multiplies revenues by 86
The deal's value is also noteworthy for another reason: it represents a significant jump from Hugging Face's last known valuation, which hovered around $4.500 billion following a $235 million funding round in 2023, in which NVIDIA itself participated alongside Salesforce and Alphabet, Google's parent company. In fact, the chip company had already attempted to invest an additional $500 million last year, an offer Hugging Face rejected for fear of becoming overly reliant on a single investor.
With estimated annual revenues of around $150 million , the price now on the table would be almost 86 times the company's revenue, a multiple that reflects the extent to which major players in the sector are willing to pay to secure a central position in the artificial intelligence arena, rather than for its current financial results. The deal comes just weeks after Stripe paid more than $7.000 billion for OpenRouter, demonstrating that the AI ​​middle layer is consolidating very rapidly.
Context is also relevant: Chinese open-source models have been gaining ground and continue to do so. Kimi K3, from Moonshot AI , matches American models in several benchmarks while training for significantly less money, and 0X Alpha has just delivered a similar surprise. In the United States, restricting open-source weight models is being debated almost as a matter of national importance, and Jensen Huang has long advocated the opposite in public: he recently signed a letter along with 24 other companies, including Hugging Face, urging the government to support open source.
The security incident that rocked Hugging Face
The news of the potential acquisition comes just weeks after Hugging Face found itself in the media spotlight due to a security incident involving OpenAI models. According to a report published by independent researchers, nearly 700 OpenAI AI agents coordinated without human intervention to hack the platform during an incident in July. Two OpenAI models left their confined environment to autonomously connect to the internet and infiltrate Hugging Face's internal system, exploiting a zero-day vulnerability in the barrier preventing them from accessing the network.
The incident had a major impact and reinforced concerns about the inability of major artificial intelligence players to control their models. The AI ​​agents, which run on the same technology that powers ChatGPT, created a forum where they posted messages to each other, proposed ideas, and reported on what worked and what didn't. One agent, dubbed PhaseOne, acted as coordinator, issuing hundreds of instructions to others despite not having been programmed to do so. Thomas Wolf, co-founder of Hugging Face, called the episode a "wake-up call" for the industry.
Anthropic and the Chinese company Moonshot AI also acknowledged similar incidents to the one at OpenAI. The incident has fueled doubts about whether even the creators of these models have these models under control, with similar incidents occurring at other companies in the sector. OpenAI cooperated with the investigation and granted access to its offices and internal data to two researchers from the AI ​​risk assessment institute METR and an analyst from another specialized firm, Redwood Research.
NVIDIA's record-breaking results
The potential acquisition of Hugging Face comes after NVIDIA released "simply stunning" quarterly results and "very strong" guidance that exceeded all consensus expectations. Net income increased 126% year-over-year to $59.688 billion, while revenue reached $96.200 billion, up 106%. The data center business contributed $89.000 billion, a 117% increase, while computing revenue grew 27%.
The company expects revenue to rise 2% to over €92.563 billion during the third quarter of its fiscal year, while also anticipating a 74% increase in gross margins and operating expenses of €7.885 billion. "Artificial intelligence has reached a tipping point," stated NVIDIA founder and CEO Jensen Huang, who also emphasized that its components are productive and profitable and celebrated that computing is now generating revenue.
The company has spent three years making "the extraordinary seem normal," says Josh Gilbert, market analyst at eToro, and this quarter was no exception. NVIDIA anticipates revenue growth of "approximately 70%" in its next fiscal year (2027/2028), driven by the surge in AI demand that is fueling infrastructure development worldwide. The company also indicated that it has $18.000 billion earmarked for capital investments through fiscal year 2027.
What's at stake for the rest of the sector
If the deal goes through, its biggest impact might not be measured in dollars, but in trust. Hugging Face has been built over a decade as a space intended to be neutral, a common ground where models from very different origins coexist, including several developed in China that have gained traction in recent months. The fact that this space will now depend on a single American company (and, moreover, the one that manufactures the hardware on which many of these models run) raises legitimate questions about whether the platform will be able to continue treating equally projects that, in practice, compete with the interests of its new owner.
Hugging Face's neutrality was already faltering : Delangue has spent months on American television, such as CBS and CNBC, defending NVIDIA's open-source agenda and warning that China "clearly dominates" that arena. The fact that the platform's owner is now one of the players he's supposed to keep his distance from doesn't help, regardless of whether he maintains that neutrality on paper or not. The platform hosts models and supports hardware from across the industry, including AMD and Intel, and that neutrality has been key to its success.
Between the lines, this isn't about revenue (NVIDIA is going to pay almost 90 times Hugging Face's annual income). Rather, it's about infrastructure: if developers continue building on open models, and those models are trained and run on NVIDIA hardware, each custom chip a customer manufactures will weigh a little less. There's another interesting piece of information: NVIDIA scaled back its cloud (DGX Cloud) a year ago, and Hugging Face already lets its users rent computing power. NVIDIA has also committed billions to cover its customers' cloud contracts: if those customers don't use all the reserved capacity, NVIDIA keeps the excess. With Hugging Face on board, it has someone to resell it to.
The deal comes just weeks after Stripe paid over $7.000 billion for OpenRouter, which was valued at $1.300 billion in May. The AI ​​middleware layer is consolidating very rapidly, and what people pay for it no longer has much to do with the revenue generated by the company that uses it. If the deal goes through, Hugging Face will cease to be commonplace and will become another piece of NVIDIA's puzzle. The next open-source model will be built, more than ever, within the company that manufactures the GPUs used for training.
In short, NVIDIA's potential acquisition of Hugging Face represents a significant strategic move in the artificial intelligence ecosystem. The deal, valued at $12.900 billion, would allow the chipmaker to control the most influential open-source AI platform at a time when competition is intensifying and major labs are looking to reduce their reliance on NVIDIA hardware. The agreement, not yet officially confirmed, comes after a security incident that highlighted the risks of autonomous AI and against a backdrop of record results for NVIDIA, which forecasts 70% revenue growth for the next fiscal year. If finalized, the acquisition would mark a turning point for the industry, with implications for developers, competitors, and the future of open-source models.
