Nvidia Unveils New Open-Source AI Model. CEO Jensen Huang Says Free AI Means More Chip Demand.
Nvidia is deepening its push into open-source artificial intelligence, unveiling a new model that can run on a single graphics processor just weeks after CEO Jensen Huang publicly jumped into the debate over whether the U.S. should embrace open AI development.
The chip giant on introduced Nemotron 3.5 Lightning, a “lightweight” open-source AI model that companies can download, modify and use without paying Nvidia or seeking permission. It is Nvidia’s first open-source model release since Huang called on Washington to avoid restrictions that he argues could undermine U.S. leadership in artificial intelligence.
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 timing is significant for Nvidia as open-source AI models could help broaden access to advanced technology while also driving demand for the company’s core business of graphics processing units, or GPUs, used to train and run AI systems.
“Free AI should be great for hardware,” Huang told Axios last month. “Free AI should be great for chips.” Nemotron 3.5 Lightning is designed to run on a single GPU in a personal computer, potentially making advanced AI capabilities accessible to companies and developers that do not have the enormous computing budgets required for some frontier models.
Nvidia said the model was developed particularly for AI agents, programs capable of performing tasks autonomously in the background. CrowdStrike, CodeRabbit and Harvey are among the companies that have already tested and customized the model, according to Nvidia.
The release comes less than three weeks after Huang made his first post on X. “Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty,” Huang wrote.
Huang also published an open letter arguing that open-weight models give businesses greater control over the technology they use, encourage competition and reduce costs. He joined other major technology executives pushing policymakers to preserve access to open models while “avoiding premature restrictions” that could shift innovation outside the United States.
The debate intensified following the release of Kimi K3 from Chinese startup Moonshot AI. The model narrowed the performance gap between Chinese systems and some of the most powerful U.S. models, raising concerns in Washington about national security and the possibility that Chinese developers could benefit from American technology.
Part of the controversy centers on “distillation,” a technique in which outputs from a more advanced AI system are used to help train a smaller model. Some policymakers have raised concerns that Chinese companies could use the technique to replicate capabilities developed by U.S. AI labs, prompting questions about whether additional restrictions or sanctions are needed.
Nvidia said it also used distillation to give Nemotron 3.5 Lightning capabilities similar to those of its larger Nemotron models, highlighting how the technique itself has become increasingly important across the industry.
The company’s open-source push also comes as other Silicon Valley leaders make similar arguments. Meta CEO Mark Zuckerberg on Monday published a manifesto calling for American open-source AI models to become the world’s best as Meta released Muse Spark, a new coding model. “Our goal should be for American open source models to be the best globally,” Zuckerberg wrote.
Nvidia has also expanded its open AI efforts into cybersecurity. Days after Huang’s letter, the company launched an AI safety consortium with Microsoft and other participants focused on using open models and open-source software to strengthen cybersecurity.
Alongside Nemotron 3.5 Lightning, Nvidia introduced NeMo Switchyard, software designed to determine the most appropriate and cost-effective AI model for a particular task. The new model will be available through Hugging Face and Nvidia.
