Nvidia is acquiring Hugging Face for $12.93 billion. Both companies confirmed the deal on September 3, 2026, after initial reports of negotiations had surfaced weeks earlier. Nvidia CEO Jensen Huang promises that Hugging Face will remain an open platform and that no one will need Nvidia hardware in the future to build or deploy models there.
Key Points at a Glance
- Nvidia is buying Hugging Face for $12.93 billion, announced on September 3, 2026
- The platform is set to remain open, including for AMD, Intel, and AWS hardware
- Hugging Face has over 18 million users, more than 3 million models, and around 500,000 datasets
- Critics fear that Nvidia will favor its own tools, such as NIM, in the long term
- The deal still needs approval from antitrust authorities
Nvidia and Hugging Face: An Overview of the Deal
The key figures of the acquisition are best presented in a table, as quite a few numbers have accumulated here in a short time:
| Key Figure | Value |
|---|---|
| Buyer | Nvidia |
| Target Company | Hugging Face |
| Purchase Price | $12.93 billion |
| Announcement | September 3, 2026 |
| Hugging Face founded | 2016 |
| Users/Developers | Over 18 million |
| Hosted models | over 3 million |
| Applications (Spaces) | Over 1 million |
| Data records | Over 500,000 |
| Deal statuses | Not yet approved under antitrust law |
By comparison, Hugging Face had raised a total of approximately $395 million in investor funding since its founding, most recently around $235 million in 2023 from investors including Salesforce Ventures. According to a report by The Information, annual revenue was most recently around $150 million. By that measure, the purchase price is enormous, which shows just how much Nvidia values access to the platform itself—not just its current revenue.
Why Nvidia Is Acquiring Hugging Face of All Companies
Hugging Face is regarded in the industry as the “GitHub of AI”: a central hub where developers and companies share and test open-source models, datasets, and small AI applications. It is precisely this role as a neutral meeting place that makes the platform interesting to Nvidia, because whoever sets the tone here occupies a key position at the intersection of the entire AI development ecosystem—regardless of which chip ultimately performs the computations.
Exciting day for NVIDIA and @huggingface.
Open models strengthen safety and cybersecurity, accelerate innovation and adoption, and enable sovereignty. They allow every developer, startup, university, industry, and country to build with, customize, and benefit from AI.
Thank you…
— Jensen Huang (@JensenHuang) September 3, 2026
The acquisition didn’t come out of nowhere. Nvidia had been an investor in Hugging Face since 2023 and had made the company a $500 million offer in 2024, which was rejected at the time. This summer, Hugging Face CEO Clem Delangue reportedly approached Huang himself to discuss a takeover. Delangue explained the move to the community by saying that his company needed more computing power, more support, and more collaboration to continue growing. Huang, for his part, said he felt honored that Delangue had chosen him for this next chapter of Hugging Face.
Strategically, the acquisition fits into a larger pattern: Nvidia has been investing for years in open models and open-source tools to counterbalance closed APIs like those from OpenAI, while simultaneously strengthening the U.S.’s technological position in the global AI competition.
Will Hugging Face really remain open to AMD and other manufacturers?
The community’s biggest concern is the platform’s neutrality. Hugging Face currently supports not only Nvidia chips but also other hardware via extensions like Optimum AMD and Optimum Intel, as well as multi-cloud setups such as those on AWS. Nvidia has publicly assured users that this will not change: It will not require Nvidia hardware, and developers will still be able to work with any accelerators and cloud providers.
Criticism: A Built-in Conflict of Interest
Despite these assurances, a fundamental problem remains: Hugging Face’s value lay precisely in the fact that the platform wasn’t owned by anyone in the hardware industry. The trade magazine *The New Stack* sums it up: If Nvidia hardware is given even slight preferential treatment in the future—for example, because Nvidia’s own NIM microservices and CUDA-optimized containers require fewer clicks during deployment than competing solutions—the platform will lose the very advantage that made it attractive to everyone. Although Nvidia emphasizes that it does not currently intend to make NIM the default setting, observers fear that it could still gain traction through small, barely noticeable advantages in everyday use. Added to this is a practical question: Do AMD, Intel, or AWS want to continue investing engineering resources to maintain their integrations on a platform that now belongs to a direct competitor?
What’s still missing: approval from antitrust authorities
A precise timeline for the completion of the acquisition has not yet been set. The only thing that is clear is that the deal must still be reviewed by competition authorities before it can be finalized. Given the scale of the deal and Nvidia’s already dominant market position in AI hardware, this review is unlikely to be merely a formality.
Conclusion: Nvidia is buying influence in the open AI world
With the $12.93 billion acquisition of Hugging Face, Nvidia secures a massive user base and significant influence over how open AI models will be developed, shared, and deployed in the future. For developers in Germany and Europe—who already must comply with requirements such as the rules described in the article on the 2026 AI labeling mandate —the issue of transparency and independence of the platforms they use becomes even more relevant as a result. Whether Nvidia will truly uphold its promise of neutrality will only become clear in the coming months, once the deal has passed regulatory review and the first product decisions become apparent.
Sources: Nvidia Blog
