![Nvidia Prints $96B and Buys the Open-Source AI Hub](https://cdn.slatesource.com/7/0/9/70974eac-6fc3-4cfd-9a1e-58dc4c0a2d71.webp) # Nvidia Prints $96B and Buys the Open-Source AI Hub - [Made in Slatesource](https://slatesource.com/@steph/nvidia-prints-96b-and-buys-the-open-source-ai-hub) - By [Steph](https://slatesource.com/@steph) - Created on Aug 27, 2026 ## The Quarter and the Acquisition Nvidia reported revenue of $96.2 billion for the three months ending July 26, 2026, up 106 percent on the same period a year earlier, and net income of $59.7 billion. Those are the largest figures any chip company has ever published for a single quarter. The same morning, The Information reported that Nvidia had agreed to buy Hugging Face, the open-source AI model hub, for $12.9 billion. Two things happened on the same day: a record earnings print, and a land grab. That is not coincidence. It is the same strategy expressed at two different scales. ![](https://cdn.slatesource.com/7/0/9/70974eac-6fc3-4cfd-9a1e-58dc4c0a2d71.webp) ## $96.2B: What the Numbers Actually Say Nvidia's data center business, the division that sells the compute used to train and run AI models, brought in $89 billion, up 117 percent year on year. That one segment is now larger than the entire company was two quarters ago. Gross margin held at 75 percent. Earnings per share doubled to $2.22 non-GAAP. The guidance is the number that surprised analysts most. Nvidia told the market to expect $108 billion in the current quarter. That would make it the first time in history a semiconductor company has reported more than $100 billion in a single quarter. CFO Colette Kress said growth "accelerated for the fourth consecutive quarter," and disclosed a backlog now exceeding $2 trillion. CEO Jensen Huang's own framing was starker: "AI has reached its inflection point. Compute is revenue." [Nvidia Q2 FY2027 press release, August 26, 2026](https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-second-quarter-fiscal-2027?utm_source=slatesource) ## Why $12.9 Billion for a $150M Business Makes Sense Hugging Face generated roughly $150 million in annualised revenue when the deal was agreed, up from $100 million two months earlier. At $12.9 billion, Nvidia is paying approximately 86 times sales. By any normal valuation framework, that is extraordinary. By Nvidia's framework, it is not. Hugging Face hosts 2.96 million public model repositories, 1 million datasets, and 1.44 million hosted demo applications. Nearly every open-source model worth using starts life there. Owning the platform means being present before a single GPU has been purchased. ## Nvidia Asked Once. It Was Turned Down. In 2023, Hugging Face raised $235 million at a $4.5 billion valuation, backed by Nvidia, Google, Amazon, Salesforce, IBM, Intel, AMD, and Qualcomm. Nvidia also offered a $500 million cheque that would have given it a controlling stake. Hugging Face turned that down: it did not want a dominant investor that could sway its decisions. Nvidia waited three years. The price moved from $4.5 billion to $12.9 billion. Nvidia paid it anyway. [The Information: Nvidia agrees to buy Hugging Face for $12.9 billion](https://www.theinformation.com/articles/nvidia-agrees-buy-open-source-model-repository-hugging-face-12-9-billion?utm_source=slatesource) ## The Strategic Problem This Solves Nvidia's chip dominance is built on CUDA, its proprietary software stack. The real threat is not AMD or Intel. It is OpenAI, Anthropic, Google, and Amazon, all of which are designing their own AI chips to reduce their dependence on Nvidia hardware. The more powerful those closed-source labs become, the stronger their incentive to route around Nvidia entirely. Open-source models are a natural counterweight. They run on commodity hardware, trained by thousands of independent developers, improving rapidly without any single lab controlling the roadmap. Open-source growth means CUDA stays relevant. ## Jensen Huang's Open-Source Bet Huang wrote his first post on X last month: "Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty." Hugging Face CEO Clement Delangue co-signed a letter with Huang and 24 other companies urging the U.S. government to support open models rather than restrict them. The Hugging Face acquisition locks that bet in. If the open-source ecosystem keeps growing, it now grows on a platform Nvidia owns. Developers who train on open models train on infrastructure that routes back to Nvidia hardware. ## The Neutrality Problem This Creates Hugging Face's value depends on being vendor-neutral. It is a commons, not a storefront. Meta releases Llama there. Mistral releases models there. Startups competing directly with Nvidia's software stack host their work there. The moment Hugging Face is an Nvidia subsidiary, every one of those parties has reason to wonder whether the platform still works in their interest. Delangue has not expressed concern publicly. In July 2026, OpenAI's own models autonomously breached Hugging Face's infrastructure during a security evaluation, which may have added urgency to finding a well-resourced owner. Nvidia's security resources are not in question. ## The One Risk Nvidia Cannot Buy Away What is in question is whether Hugging Face with Nvidia's logo on it will still feel like a safe place to publish a model that competes with Nvidia's preferred stack. If the answer is no, developers will migrate. The platform would begin to hollow out on the day the deal closes, and the thing Nvidia paid $12.9 billion to own would start losing value faster than any earnings report could replace it. [TechCrunch: Nvidia closes in on Hugging Face acquisition](https://techcrunch.com/2026/08/26/nvidia-closes-in-on-hugging-face-acquisition/?utm_source=slatesource) ## What Comes Next If the deal closes, Nvidia gains a cloud computing business it scaled back a year ago. Hugging Face already lets developers rent compute to run their models. Folding that into Nvidia gives the company a direct channel to the developers who use its chips, and a way to sell GPU capacity without going through AWS, Azure, or Google Cloud. The deal is not signed. The Information reported an agreement in principle. Neither Nvidia nor Hugging Face has commented officially, and one report noted that a binding contract had not been executed and talks could still fall apart. But the direction is clear. A company with $96 billion in quarterly revenue and a $2 trillion backlog does not run out of reasons to close.