What’s behind Nvidia buying Hugging Face? Look to Microsoft
Nvidia appears to be taking a page from Microsoft’s playbook — a move we happen to like. Nvidia on Thursday said it is acquiring open-source AI platform Hugging Face for $12.9 billion, aiming to defend and expand the ecosystem around its chips as hyperscalers move quickly to develop their own custom chips. Nvidia CEO Jensen Huang wants to drive AI innovation as quickly as possible, with a standardized approach that relies on his company’s hardware. The deal gives Nvidia direct access to Hugging Face’s more than 18 million AI developers and is Nvidia’s second-largest on record, following the $20 billion purchase of assets from chipmaker Groq in December. Hugging Face CEO Clément Delangue and his executive team, including co-founders Julien Chaumond and Thomas Wolf, will join Nvidia, meaning Nvidia also acquires the talent behind the platform. The move reminds us of a similar deal by Microsoft, when in 2018 it purchased GitHub, an open-source platform for traditional software code, for $7.5 billion. Then as now, the acquirer gains access to and control over leading open-source development, collaboration, discovery, and distribution platforms that, when properly leveraged, drive more demand for their core businesses. For Microsoft, that’s its Azure cloud business; for Nvidia, it’s accelerated computing hardware and its CUDA software stack. To be sure, open-source, by its very nature, means the code available on its platform is free to use, which means the deal is not your typical one that immediately adds to the top and bottom lines. But Microsoft has still benefited immensely. Its commitment to keep GitHub true to its open-source roots drives loyalty among developers. Developers have massive influence over which technology a company adopts. So, winning over developers is a fantastic way to ensure those building the world’s greatest companies have an affinity for working with Nvidia and, as a result, push the companies deeper into the Nvidia ecosystem. In other words, appeal to the developers, and they’ll do the selling for you. Moreover, by owning the platform, Microsoft was able to easily integrate its Azure cloud service as a go-to option for developers looking to host software they may have built using resources found on GitHub. From a defensive perspective, the deal also kept GitHub out of the hands of competitors like Amazon and Alphabet . Likewise, Nvidia’s purchase of Hugging Face allows the AI chip giant to go on the offensive by further aligning open-source LLM development with the Nvidia ecosystem, while also playing defense by keeping the platform out of competitors’ hands. Nvidia has also committed to keeping the platform open to all silicon vendors, clouds, and model providers, and will not require Hugging Face users to run models on Nvidia chips. But by more deeply integrating Nvidia’s technology into the Hugging Face platform, Nvidia can incentivize the use of paid products simply by reducing friction for developers and making its offerings a natural next step as projects move from development to deployment. As Huang wrote in a blog post announcing the deal, “We build our own models, libraries and tools in the open so developers everywhere can use them, modify them and build on top of them.” Ultimately, Nvidia wants to tap into the rising demand for compute, whether it comes from open-source or proprietary AI models. By supporting the development across all areas of AI, Nvidia can have greater influence over where that innovation occurs and drive further revenue growth over time. Asked on CNBC about Nvidia’s support for both sides of the AI coin, Huang said that even closed-model labs support open-source models. However, he added: “I recommend that people use closed models as much as they can, you know, because it’s off the shelf, it’s incredibly good, it’s advancing very quickly.” That is the advantage open-source models can provide and why many look to them over proprietary alternatives like Claude or ChatGPT. Notably, in a tweet congratulating Jensen on the deal, Palo Alto Networks CEO Nikesh Arora echoed the importance of balancing open source and open weights with frontier LLMs. (Open-source models provide everything, from the code to training details and the blueprint needed to rebuild from scratch; open-weight models allow one to download the model and tweak it but don’t provide details on the data or training that went into it.) “So thank you for supporting the ‘open’ alternative,” he added. “Of course we have use cases where the bleeding edge abilities will be needed from frontier LLMs, but as we evolve my view on ‘different horses for different courses’ continues to be reaffirmed.” This is nothing new to the world of tech. We’ve seen open-source software like Linux play an important role in the world of computing, right alongside closed ecosystems like Microsoft’s Windows and Apple’s iOS. We’ve also seen Microsoft Office continue to grow despite Alphabet’s free Google Workspace alternatives. The bottom line: Both models can advance AI capabilities and adoption, driving greater demand for compute, the ultimate driver of Nvidia’s sales and earnings growth. (See here for a full list of the stocks in Jim Cramer’s Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
