> ## Content Index > Fetch the complete content index at: https://www.tao.media/llms.txt > Use this file to discover other available public pages before exploring further. # Nvidia Agrees to Acquire Hugging Face for $12.9 Billion - URL: https://www.tao.media/nvidia-agrees-to-acquire-hugging-face-for-12-9-billion/ - Published: 2026-08-27T13:47:06.000Z - Updated: 2026-08-27T20:52:16.000Z - Description: The deal gives Nvidia control of one of AI’s most important open-model platforms, extending its reach from the compute layer into model discovery, deployment and an increasingly important robotics ecosystem. - Author: Bart Hillerich - Tags: Hugging Face, Nvidia, News, AI Nvidia has agreed to acquire Hugging Face for $12.9 billion, in one of the largest acquisitions in the chipmaker’s history and a major expansion of its position across the AI development stack. The deal was [first reported by The Information](https://www.theinformation.com/articles/nvidia-agrees-buy-open-source-model-repository-hugging-face-12-9-billion?ref=tao.media), citing a person with knowledge of the agreement, and subsequently [reported by Reuters](https://www.reuters.com/technology/nvidia-talks-acquire-hugging-face-13-billion-deal-business-insider-reports-2026-08-27/?ref=tao.media). Hugging Face has become one of the central platforms for developers to discover, publish, fine-tune and deploy open AI models, many of which ultimately run on Nvidia hardware. The teams have also worked directly together across cloud infrastructure, models and robotics. Most recently, the companies expanded their robotics partnership in July by bringing Nvidia Isaac GR00T 1.7, Isaac Teleop, datasets and robotics workflows into Hugging Face’s LeRobot ecosystem, with Nvidia Cosmos 3 integration also planned. The purchase price represents a significant premium to Hugging Face’s existing business. The company recently reached roughly $150 million in annualized revenue, up from around $100 million only two months earlier. At $12.9 billion, Nvidia is paying more than 80 times Hugging Face’s current annualized revenue. Hugging Face was valued at just $4.5 billion in its previous major funding round in 2023, when Nvidia participated alongside investors including Google, Amazon, AMD, Intel, Salesforce and Qualcomm. The $235 million round cemented Hugging Face’s position as one of the most valuable infrastructure companies emerging from the generative AI boom. The strategic value, however, goes well beyond revenue. ## Nvidia Moves Up The Stack Hugging Face acts as a distribution layer for open AI. Its platform grew to 13 million users, more than 2 million public models and over 500,000 public datasets by the end of 2025, while more than 30% of Fortune 500 companies now maintain verified accounts on the platform. Hugging Face has also said Nvidia has become the strongest contributor among major technology companies publishing repositories on its platform. That makes the acquisition particularly important as Nvidia looks to protect its dominant position in AI computing, something teams like OpenAI are starting to build against. [OpenAI’s First Custom AI Chip Beats Nvidia Blackwell in Early Inference TestsOpenAI’s first custom inference chip is already outperforming Nvidia, AMD and Google hardware across several AI workloads, despite being a first-generation design.![](https://storage.ghost.io/c/78/0b/780ba906-b1a7-4bf0-873c-bdd5c32e5331/content/images/icon/Group-1321319358-33388ff8-1537-45f4-8127-48478a8ef043.png)Intelligence | Bittensor News, Insights, StoriesBart Hillerich![](https://storage.ghost.io/c/78/0b/780ba906-b1a7-4bf0-873c-bdd5c32e5331/content/images/thumbnail/ChatGPT-Image-Aug-26--2026--11_36_00-AM-357b5748-fc0e-4f41-b099-968b67405419.png)](https://www.tao.media/openais-first-custom-ai-chip-beats-nvidia-blackwell-in-early-inference-tests/) The Information reported that Nvidia views the growing open-model ecosystem as an important counterweight to closed AI developers such as OpenAI and Anthropic. Those companies, along with Google, Amazon and other hyperscalers, are investing heavily in custom AI chips that could reduce their reliance on Nvidia hardware over time. Open models, meanwhile, are generally designed to run across broadly available infrastructure, where Nvidia GPUs remain dominant. The closest comparison may be Microsoft’s acquisition of GitHub. Microsoft agreed to acquire GitHub for $7.5 billion in 2018, gaining ownership of the platform that had become the central repository and collaboration layer for software developers. Microsoft ultimately kept GitHub operating as an independent and open platform. > GitHub sits at the center of how developers discover, store and collaborate on software, while Hugging Face plays the equivalent role for AI models. It's also important to note where the acquisition fits into Nvidia’s existing stack: CUDA already owns much of the execution layer, while Hugging Face sits around model discovery, weights, libraries and deployment workflows. Owning Hugging Face therefore moves Nvidia one layer higher. Nvidia already benefits when developers train and run workloads on its GPUs. Hugging Face gives the company exposure much earlier in the process, starting when a developer searches for a model and continuing through downloading weights, accessing datasets, fine-tuning, inference and deployment. ## Surprise Hugging Face "Microduck" Launch Hours after news of the acquisition emerged, Hugging Face provided a tangible example of where that ecosystem is heading. Hugging Face-owned Pollen Robotics unveiled Microduck, a $399 open-source biped robot designed to let users train new behaviors themselves using reinforcement learning. The 25 cm robot weighs around 800 grams and packs 15 motors, a camera, LiDAR and two IMUs, alongside an articulated beak capable of picking up objects. Seven trained movements are included out of the box, including walking, sitting and standing, grabbing objects, roller skating and recovering after falling over. [Hugging Face Launches $399 Open-Source Microduck Robot for Reinforcement LearningThe robot brings sim-to-real reinforcement learning, customizable behaviors, and physical AI experimentation to developers for under $400.![](https://storage.ghost.io/c/78/0b/780ba906-b1a7-4bf0-873c-bdd5c32e5331/content/images/icon/Group-1321319358-31bdfa06-7830-45e4-a0a6-d118d9be77ea.png)Intelligence | Bittensor News, Insights, StoriesBart Hillerich![](https://storage.ghost.io/c/78/0b/780ba906-b1a7-4bf0-873c-bdd5c32e5331/content/images/thumbnail/kickabout-c166453f-e678-47f2-9382-d53572906518.webp)](https://www.tao.media/hugging-face-launches-399-open-source-microduck-robot-for-reinforcement-learning/) More importantly, those behaviors are not simply hard-coded animations. Microduck is built around a simulation-to-real workflow in which users can train behaviors inside a physics simulator, deploy the resulting policy onto the physical robot, refine the simulation and then publish the trained behavior for other users. Hugging Face says the SDK, simulation environment and full reinforcement learning training stack are open source under an Apache 2.0 license. Pre-orders opened on August 27 at $399, with shipments expected before Christmas 2026. ## Hugging Face's Growth & Nvidia Deal Implications Microduck is a small product relative to a $12.9 billion acquisition, but it illustrates how far Hugging Face has expanded beyond its original role as a model repository. The platform is increasingly bringing models, datasets, software libraries, simulation environments, training infrastructure and physical hardware into the same open ecosystem. Instead of simply downloading a language model from Hugging Face, developers can now train a robotics policy in simulation, deploy it to a real machine and share the resulting behavior back with the community. That expansion also aligns closely with Nvidia’s own ambitions around physical AI. Nvidia has been building a robotics stack around Jetson computers, Isaac simulation and robotics software, GR00T foundation models and Cosmos world models. Hugging Face adds a distribution and developer layer on top of that infrastructure, potentially giving Nvidia a direct channel to the communities building and sharing the models that eventually run on robots. The deal does introduce a major question around Hugging Face’s neutrality. Part of the platform’s appeal has been its ability to support models and hardware across competing ecosystems. Nvidia itself invested in Hugging Face alongside several of its biggest rivals in 2023, including AMD, Intel, Google and Amazon. Nvidia owning the platform outright creates an incentive to make CUDA, Nvidia GPUs and its broader software stack the easiest path from a model on Hugging Face to a running application. Microsoft faced a similar concern when it acquired GitHub and responded by maintaining GitHub as an independent, open platform.