On September 2, NUPIAO reported that Nvidia is expected to reach a deal this week to acquire Hugging Face for nearly $14 billion.
As of now, the final terms haven’t been locked down, and both the price and timeline could still shift. Neither side has issued an official announcement, and both declined to comment.
Reportedly, Nvidia isin advanced talks to acquire AI startupHugging Face, with a total deal value of about $14 billion. People familiar with the matter said the cash consideration is roughly $12.9 billion, and a deal could be reached as early as this week; there’s also a $1 billion employee retention package aimed at keeping Hugging Face’s core R&D staff.
If the deal goes through, this ten-billion-dollar-plus transaction would rank among Nvidia’s largest acquisitions to date. It would also be the biggest AI deal globally so far this year, and it could reshape the competitive landscape of the open-source AI industry.
The talks are drawing close attention across Silicon Valley. Some market watchers say that once a deal is signed, it will likely face tough scrutiny from U.S. antitrust regulators, with the competitive risks of combining hardware and software ecosystems expected to be a key focus.
Hugging Face was founded in 2016 by three French entrepreneurs in New York. The company initially dreamed of building a fun chatbot for teenagers. After open-sourcing its Transformers library in 2018, its path changed dramatically: it gradually became a global AI model-sharing and collaboration platform for developers, often called “the GitHub of AI.”

Public operating data shows that the Hugging Face Hub now hosts more than 3 million public AI models and over 1 million datasets, bringing together more than 13 million developers worldwide. It has over 18 million monthly visitors, about 5 million registered users, and more than 2,000 businesses paying for its Enterprise Hub. A large number of open-source model teams around the world—including Meta, Google, DeepSeek and Zhipu AI—use the platform as their primary channel for distributing models and iterating versions.
Last April, it acquired French open-source robotics company Pollen Robotics and started selling robots. It has also rolled out paid enterprise hosting, model inference deployment and custom development services, gradually building a commercial revenue system, with its business reaching almost every corner of the AI value chain. According to foreign media estimates, Hugging Face’s current annualized revenue is about $150 million, and revenue growth is on a steep upward trajectory.
On the capital side, Nvidia was already an investor in Hugging Face’s Series D round in 2023, alongside tech giants including Google, Amazon, Intel and Salesforce. After that round, Hugging Face was valued at $4.5 billion post-money. Compared with the current offer of nearly $14 billion, that’s nearly a threefold increase.
Earlier this year, Hugging Face turned down a $500 million investment offer from Nvidia. Just months later, the two sides are now in talks over a full acquisition—a sign of how determined Nvidia is to stake its claim in open-source AI.
Looking at Nvidia’s strategic moves in recent years, paying a high price for Hugging Face isn’t a short-term financial play. It’s a key bet aimed at shaping the AI ecosystem over the medium and long term.
The most direct strategic payoff is owning the key gateway for global open-source model distribution, which would complete Nvidia’s industrial loop from underlying compute hardware and the CUDA development framework all the way to the upper-layer model distribution channel. Nvidia’s commercial edge has long been rooted in the GPU compute hardware market, while Hugging Face controls a massive developer base. Once the acquisition closes, Nvidia’s compute products could reach open-source model developers directly, connecting compute supply with model development.
At the same time, the market sees the deal as a defensive move by Nvidia to hedge against industry risks. OpenAI, Anthropic and other closed-source model makers are accelerating work on their own AI server chips, aiming to reduce their reliance on Nvidia GPUs, and the compute market is becoming more diversified. Nvidia management’s core view is that the more vibrant the open-source AI ecosystem becomes, the more dispersed global compute demand will be—making it hard for any single closed-source player’s in-house chips to cut into Nvidia’s base.
Beyond that, owning Hugging Face would give Nvidia a stronger say over how open-source models are adapted and where community tech trends go. Combined with Nvidia’s own Nemotron open-source model project, this could create an ecosystem play that ties together hardware compute, in-house models and the open-source community—cementing Nvidia’s role as infrastructure for the global AI industry.
This year, Nvidia has also closed a string of software-ecosystem acquisitions. In the first half, it completed the purchase of Israeli compute-scheduling company Run:ai, a deal already approved by EU regulators; the products will strengthen Nvidia’s capabilities in GPU cluster resource scheduling and hybrid-cloud compute management. In June, Nvidia also fully acquired U.S. predictive AI startup Kumo AI for at least $400 million.
In Silicon Valley’s tech M&A market, AI middleware infrastructure has become a hot target for the giants. Not long ago, U.S. online payments company Stripe just swallowed AI model routing platform OpenRouter for more than $7 billion. Nvidia’s latest ten-billion-dollar talks could set off an even bigger wave of acquisitions in the open-source platform space.