Nvidia CEO Jensen Huang used his first post on X to support wider access to open AI models, sharing a letter that argued against premature restrictions on the technology. The post appeared on July 24 and pointed to support from Nvidia, Google, OpenAI, Meta, AMD, Cloudflare, GitHub, IBM, Hugging Face, and LM Studio.
Open models give users access to their weights, and sometimes their source code, allowing people to run them locally and modify how they work. Closed models usually stay under the control of a company and are accessed through cloud services or APIs, with their weights kept private.
That distinction has become more important as locally downloadable models from outside the United States, including DeepSeek, have offered alternatives to services from companies such as OpenAI. Open models can be run without relying on a remote service, which gives users more control but also makes the technology harder for its original developers to monitor.
Why the companies support open models
The letter presented open weights as a way to broaden access to AI, support competition, and help more businesses take part in the growing industry. It also argued that open models can improve safety and cybersecurity by allowing more people to inspect and test them.
Huang shared the letter’s central argument in his post, writing that AI will affect every industry and be developed by every country. He also said the world needs both leading closed models and leading open models.
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.
AI will transform every industry, power every company, and be built by every country.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.… pic.twitter.com/t02bi51N4C
— Jensen Huang (@JensenHuang) July 24, 2026
The document acknowledged that open weights carry risks. Once released, they can be copied and changed without the original developer’s control, while modified versions may be difficult to track or reverse. Its position was that those risks should be addressed without banning open weights altogether.
The debate also has an international dimension. Open models can spread access to AI beyond the companies building the largest cloud-based systems, while the United States continues to seek an advantage in the field. Restricting models too early, the letter argued, could limit competition or push development to other countries.
For PC users, the discussion matters because open models can run locally on hardware with enough processing power and memory. That approach can offer more control over data and customization, though it also places the costs of running the model on the user’s own system.
Huang’s first post did not announce a new Nvidia product. Instead, it placed the company behind a broader policy position shared by several of the biggest names in AI.
Tell us what you think about the push for open AI models in the comments, and follow us on X, Bluesky, YouTube, and Instagram.
